The Journal of Pediatrics - lib.ajaums.ac.irlib.ajaums.ac.ir/booklist/94457.pdf · Tridapalli, Fabrizio Sandri, Maddalena Menarini, Emanuela Ferretti and Giacomo Faldella . 23. Comparative

The Journal of Pediatrics

Volume 150, Issue 2, Pages ,119-212 (February 2007) THE EDITORS' PERSPECTIVES

1. Depression in parents following the diagnosis of cystic fibrosis • SHORT COMMUNICATION Page A1 Robert W. Wilmott

2. Lipids and diabetes • SHORT COMMUNICATION Page A1 Stephen R. Daniels

3. Evaluating neonates with congenital cytomegalovirus infection • SHORT COMMUNICATION Page A1 Sarah S. Long

4. Outcomes for young adults with fetal alcohol syndrome • SHORT COMMUNICATION Page A2 Alan H. Jobe

5. A “beta” way to treat hypertension • SHORT COMMUNICATION Page A2 Thomas R. Welch

6. Cortisol levels as preterms age • SHORT COMMUNICATION Page A3 Alan H. Jobe

7. Rocky Mountain Spotted Fever: Missed opportunities for early treatment • SHORT COMMUNICATION Page A3 Sarah S. Long

8. Table of contents • MISCELLANEOUS Pages A4-A7

9. Editorial board • EDITORIAL BOARD Page A8

10. Information for Readers • MISCELLANEOUS Page A9

11. Announcements • MISCELLANEOUS Page A10

NOTES FROM THE ASSOCIATION OF MEDICAL SCHOOL PEDIATRIC DEPARTMENT CHAIRS, INC.

12. Factors Influencing Community Pediatrics Training in Residency • EDITORIAL Pages 119-120.e2 Cynthia S. Minkovitz, Anita Chandra, Barry S. Solomon, Lee M. Sanders, Holly A. Grason and Carol Carraccio

EDITORIALS

13. Is There a Role for β-Adrenergic Blockers in Treating Hypertension in Children? • EDITORIAL Pages 121-122 Russell W. Chesney and Deborah P. Jones

14. PET Scanning for Infants with HHI: A Small Step for Affected Infants, A Giant Leap for the Field • EDITORIAL Pages 122-124 Mark A. Sperling

MEDICAL PROGRESS

15. Functional Echocardiography: An Emerging Clinical Tool for the Neonatologist • ARTICLE Pages 125-130 Martin Kluckow, Istvan Seri and Nick Evans

16. 50 Years Ago in The Journal of Pediatrics: Observations on the clinical use of v-cillin in pediatric practice • SHORT COMMUNICATION Page 130 Hugh S. Lam

SPECIAL ARTICLE

17. Pediatric Workforce: A Look at Pediatric Infectious Diseases Data from the American Board of Pediatrics • ARTICLE Pages 131-133.e1 Linda A. Althouse and James A. Stockman

ORIGINAL ARTICLES

18. Efficacy and Safety of Extended Release Metoprolol Succinate in Hypertensive Children 6 to 16 Years of Age: A Clinical Trial Experience • ARTICLE Pages 134-139.e1 Donald L. Batisky, Jonathan M. Sorof, Jennifer Sugg, Michaelene Llewellyn, Michael Klibaner, James W. Hainer, Ronald J. Portman and Bonita Falkner

19. Diagnosis and Localization of Focal Congenital Hyperinsulinism by 18F-Fluorodopa PET Scan • ARTICLE Pages 140-145 Olga T. Hardy, Miguel Hernandez-Pampaloni, Janet R. Saffer, Mariko Suchi, Eduardo Ruchelli, Hongming Zhuang, Arupa Ganguly, Richard Freifelder, N. Scott Adzick, Abass Alavi, et al.

20. Longitudinal Lipid Screening and Use of Lipid-Lowering Medications in Pediatric Type 1 Diabetes • ARTICLE Pages 146-150.e2 David M. Maahs, R. Paul Wadwa, Kim McFann, Kristen Nadeau, Matthew R. Williams, Robert H. Eckel and Georgeanna J. Klingensmith

21. Altered Basal Cortisol Levels at 3, 6, 8 and 18 Months in Infants Born at Extremely Low Gestational Age • ARTICLE Pages 151-156 Ruth E. Grunau, David W. Haley, Michael F. Whitfield, Joanne Weinberg, Wayne Yu and Paul Thiessen

22. Cranial Ultrasound Scanning and Prediction of Outcome in Newborns with Congenital Cytomegalovirus Infection • ARTICLE Pages 157-161 Gina Ancora, Marcello Lanari, Tiziana Lazzarotto, Valentina Venturi, Elisabetta Tridapalli, Fabrizio Sandri, Maddalena Menarini, Emanuela Ferretti and Giacomo Faldella

23. Comparative Efficacy and Cost of Asthma Care in Children with Asthma Treated with Fluticasone Propionate and Montelukast • ARTICLE Pages 162-167 David A. Stempel, Denise T. Kruzikas and Ranjani Manjunath

24. Clinical Spectrum and Histopathologic Features of Chronic Hepatitis C Infection in Children • ARTICLE Pages 168-174.e1 Parvathi Mohan, Camilla Colvin, Chevelle Glymph, Roma R. Chandra, David E. Kleiner, Kantilal M. Patel, Naomi L.C. Luban and Harvey J. Alter

25. Fetal Alcohol Spectrum Disorders in Young Adulthood • ARTICLE Pages 175-179.e1 Hans-Ludwig Spohr, Judith Willms and Hans-Christoph Steinhausen

26. Clinical and Laboratory Features, Hospital Course, and Outcome of Rocky Mountain Spotted Fever in Children • ARTICLE Pages 180-184.e1 Steven C. Buckingham, Gary S. Marshall, Gordon E. Schutze, Charles R. Woods, Mary Anne Jackson, Lori E.R. Patterson and Richard F. Jacobs

27. Parental Depression Following the Early Diagnosis of Cystic Fibrosis: A Matched, Prospective Study • ARTICLE Pages 185-191 Claire Glasscoe, Gillian A. Lancaster, Rosalind L. Smyth and Jonathan Hill

28. 50 Years Ago in The Journal of Pediatrics: Intrathecal hydrocortisone in the treatment of tubercolous meningitis • SHORT COMMUNICATION Page 191 Albert M. Li

29. Prolonged QTc Intervals and Decreased Left Ventricular Contractility in Patients with Propionic Acidemia • ARTICLE Pages 192-197.e1 Daniela Baumgartner, Sabine Scholl-Bürgi, Jörn Oliver Sass, Wolfgang Sperl, Ulrich Schweigmann, Jörg-Ingolf Stein and Daniela Karall

30. 50 Years Ago in The Journal of Pediatrics: Spontaneous pneumothorax in the first ten days of life • SHORT COMMUNICATION Page 197 James M. Greenberg

GRAND ROUNDS

31. A Newborn Infant with Protracted Diarrhea and Metabolic Acidosis • ARTICLE Pages 198-201 Amir Bar, Arieh Riskin, Theodore Iancu, Irena Manov, Ayala Arad and Ron Shaoul

CLINICAL AND LABORATORY OBSERVATIONS

32. Gaucher Disease: Progressive Mesenteric and Mediastinal Lymphadenopathy Despite Enzyme Therapy • ARTICLE Pages 202-206 T. Andrew Burrow, Mitchell B. Cohen, Ronald Bokulic, Gail Deutsch, Arabinda Choudhary, Jr, Richard A. Falcone and Gregory A. Grabowski

33. Recipient Twin Limb Ischemia with Postnatal Onset • ARTICLE Pages 207-209 Roland Spencer Broadbent

34. Congenital Patent Ductus Venosus: An Association with the Hyper IgE Syndrome • ARTICLE Pages 210-212 Keren Sagiv-Friedgut, Michaela Witzling, Ilan Dalal, Chana Vinkler, Eli Someh and Arie Levine

35. Fifty Years Ago in The Journal of Pediatrics: Classification and etiological factors in mental retardation • SHORT COMMUNICATION Page 212 Pasquale J. Accardo

LETTERS TO THE EDITOR

36. Erythromycin for treatment of feeding intolerance in preterm infants • CORRESPONDENCE Page e30 Bai-Horng Su and Hung-Chih Lin

37. Efficacy of oral erythromycin for treatment of feeding intolerance in preterm infants • CORRESPONDENCE Pages e30-e31 Sandra Mascarenhas and Mithilesh Lal

38. Reply • CORRESPONDENCE Page e31 Pracha Nuntnarumit, Pakaphan Kiatchoosakun, Wacharee Tantiprapa and Suppawat Boonkasidecha

39. Cardiovascular support in the preterm: Treatments in search of indications • CORRESPONDENCE Pages e31-e33 Istvan Seri

40. Reply • CORRESPONDENCE Page e33 Keith J. Barrington

THE EDITORS’ PERSPECTIVES

Depression in parents following thediagnosis of cystic fibrosis

Parents of 45 children recently diagnosed with cystic fibrosis(CF) were tested with the Beck Depression Inventory at two timepoints by Glasscoe et al at the University of Liverpool. Overall, therewas no difference in the rates of depression between parents ofchildren affected by CF and matched control couples. However,when the data were stratified by age, the parents of children diag-nosed with CF who were less than 9 months old at the time of theinitial study had an increased rate of mild depression compared tocontrols and to the parents of older children with CF. These symp-toms of depression were still present, although less common, onfollow-up 9 months later. Medical providers caring for young infantsdiagnosed with cystic fibrosis, for example, through newborn screen-ing programs, should carefully assess the mood of the parents as theyare vulnerable to developing depression.

—Robert W. Wilmott, MDpage 185

Lipids and diabetesCardiovascular disease is an important adverse occurrence in

patients with diabetes. Dyslipidemia is an important risk factor forcardiovascular disease. There have been few longitudinal studies oflipid abnormalities in young patients with Type 1 diabetes. In thisissue, Maahs et al report on a study in which they evaluated lipidabnormalities in patients with Type 1 diabetes. They found thatnon-HDL cholesterol was persistently above 130 mg/dL in 28%, �160 mg/dL in 10.6% and � 190 mg/dL in 3.3%. Medications tolower LDL-cholesterol had been started in only 23/360 patients.These results suggest that although lipid abnormalities are common,treatment is not common in children and adolescents with Type 1diabetes.

—Stephen R. Daniels, MD, PhDpage 146

Evaluating neonates with congenitalcytomegalovirus infection

Although options for treating infants with congenital cyto-megalovirus (CMV) expand, they are difficult. Accurate predictionof outcome if untreated would be very helpful. Ancora et al in thisissue of The Journal offer up evidence for a new tool – cranialultrasonography (US). In 57 neonates with congenital CMV infec-tion, investigators performed neonatal and follow-up US at birth,with repeated study up to 3 months of age if normal and to 6 monthsof age if abnormal. They prospectively evaluated neurodevelopmentand hearing through 24 months and school age, respectively (meanfollow-up 42 months). All 10 of 18 infants with symptomatic CMVinfection who had an abnormal neonatal US had at least one sequela,whereas none of the 8 symptomatic infants with normal US had longterm sequelae (P � .001). In 37 asymptomatically infected neonates,3 of 37 with normal US developed sensorineural hearing loss and oneof two neonates with abnormal US developed serious neurologicsequelae.

The investigators have shown that in neonates with symp-tomatic congenital CMV infection, US is a valuable screening tool topredict outcome – comparable with other imaging modalities, lesscostly, available at the bedside, and with no radiation. Their data alsogive reason to believe that further study might prove the value of USin asymptomatically-infected infants as well.

— Sarah S. Long, M.D.page 157

This Month in

THE JOURNAL OFPEDIATRICS

February 2007 • Volume 150 • Number 2

Copyright © 2007 by Mosby, Inc.

The Journal of Pediatrics (ISSN 0022-3476) is published monthly by Elsevier Inc., 360 Park Avenue South, New York, NY 10010. Business and Editorial Offices:1600 John F. Kennedy Blvd., Suite 1800, Philadelphia, PA 19103-2899. Accounting and Circulation Offices: 6277 Sea Harbor Drive, Orlando, FL 32887-4800.Periodicals postage paid at New York, NY, and additional mailing offices. POSTMASTER: Send address changes to The Journal of Pediatrics, Elsevier PeriodicalsCustomer Service, 6277 Sea Harbor Drive, Orlando, FL 32887-4800.

The Journal of Pediatrics February 2007 1A

˜ Outcomes for young adults with fetalalcohol syndrome

Fetal alcohol syndrome (FAS) is diagnosed by maternal his-tory and physical findings in infants and early childhood. Somechildren have the incomplete findings called the fetal alcohol syn-drome spectrum disorder (FASD). Spohr, Willms, and Steinhausenreport the 20-year outcomes of 37 individuals with FAS and FASDthat they followed from early childhood. Although the physicalfindings become less apparent, the intellectual and behavioral prob-lems persist and prevent most of the individuals from successfullyworking or living independently. This report highlights the dismaloutcomes resulting from fetal alcohol embryopathy.

—Alan H. Jobe, MD, PhDpage 175

˜ A “beta” way to treat hypertensionAs we have learned more about the pathophysiology of hy-

pertension, a host of new drug classes have been developed to targetthese various mechanisms. Many of these agents have much less inthe way of side effects than earlier drugs; angiotensin convertingenzyme (ACE) inhibitors and angiotensin 2 receptor antagonists, tocite two examples, are now widely used in children.

These drugs have displaced other categories of agents which,although effective, often had unacceptable side effects in children.Beta blockers such as propranolol, for example, have been lessfrequently employed because of troublesome features such as a se-dating effect. Recently, however, new classes of beta blockers, with amore cardioselective profile and greater tolerability, have been usedwidely in adults. There has been a paucity of data upon which to basepediatric use of these agents.

In the current issue of The Journal, Batisky et al report anexceptionally well-designed trial of an extended release formulationof the beta 1 selective blocker metoprolol. The trial was powerful fora number of reasons, not the least of which was the presence of aplacebo arm. Although some eyebrows might be raised by this, inreality, the risk to children with asymptomatic hypertension of thedegree in this trial posed by a few weeks without treatment is closeto nonexistent. On the other hand, the presence of this group makesthe results of the trial even more meaningful. The drug proved safeand effective, with a very favorable side effect profile. An accompa-nying editorial by Chesney puts this trial into perspective. Currently,beta blockers are chosen as first-line agents by pediatric nephrolo-gists less than 10% of the time; in contrast, the drugs are quitepopular in adult medicine. The study of Batisky et al may set thestage for a major change in this approach.

—Thomas R. Welch, M.D.page 134 (article)

page 121 (editorial)

2A February 2007 The Journal of Pediatrics

Cortisol levels as preterms ageCortisol is a potent developmental regulator that is normally

very low in the fetus until just prior to term. Following preterm birth,the newborn must increase plasma cortisol to support multiple phys-iologic and metabolic adaptations. Cortisol then remains elevatedrelative to fetal levels until term equivalence. In animal models, fetalcorticosteroid exposures can result in altered neuroendocrine regu-lation and behavior as the newborn grows to become an adult.Grunau et al report the developmental trajectory of salivary cortisolin large cohorts of 23-28 weeks, 28-32 weeks, and term infants at 3,6, 8, and 18 months corrected age. The preterms have low cortisollevels at 3 months, but high levels at 8 and 18 months, indicatingdysregulation of cortisol homeostasis relative to term infants. Themultiple effects of cortisol on developing systems suggest that thesealtered cortisol levels as preterms age may contribute to collateralchanges in physiologic and metabolic responses that may be longlasting.

—Alan H. Jobe, MD, PhDpage 151

➙

Rocky Mountain Spotted Fever:Missed opportunities for early treatment

Buckingham et al from six institutions in the southeastern andsouth-central United States pooled 92 hospitalized cases of RockyMountain Spotted Fever (RMSF) to permit re-assessment of thedisease, management, and outcome 30 years after the last large caseseries. Although this retrospective series has limitations of un-doubted low case ascertainment and probable low symptom/signascertainment, the major findings stand alone. The infection andpotential for morbid or fatal outcome of RMSF has not changed,and physicians continue to miss opportunities for early diagnosis.Although 86% of patients in the series had been taken for medicalcare prior to hospitalization, RMSF had not been suspected ortreated in most patients. Treatment of RMSF should be begunempirically on the basis of clinical suspicion. The challenge forphysicians is consideration of the diagnosis. Buckingham et al reportan incomplete classic constellation of symptoms in most patients atfirst outpatient evaluation (median of 2 days of illness), as well as in58% of patients at the time of hospitalization and first anti-rickettsialantimicrobial therapy (median of 7 days of illness). The finding inthis study that a medical outpatient visit early in the course of illnesswas significantly associated with delay in therapy reminds us thatbarring another diagnosis, patients with febrile illnesses have diseasesand diagnoses “in-progress” (rather than diagnoses of viral illnesses)requiring continuous re-thinking as the course evolves.

—Sarah S. Long, MDpage 180

➙


NOTES FROM THE ASSOCIATION OF MEDICAL SCHOOL PEDIATRIC DEPARTMENT CHAIRS, INC.

Factors Influencing Community Pediatrics Training in Residency 119

Cynthia S. Minkovitz, MD, MPP, Anita Chandra, DrPH, Barry S. Solomon, MD, MPH, Lee M. Sanders, MD, MPH, HollyA. Grason, MA, and Carol Carraccio, MD, Baltimore, Maryland, Arlington, Viriginia, and Miami, Florida

EDITORIALS

Is There a Role for �-Adrenergic Blockers in Treating Hypertension in Children? 121

Russell W. Chesney, MD, and Deborah P. Jones, MD, Memphis, Tennessee

PET Scanning for Infants with HHI: A Small Step for Affected Infants, A Giant Leap for theField 122

Mark A. Sperling, MD, Pittsburgh, Pennsylvania

MEDICAL PROGRESS

Functional Echocardiography: An Emerging Clinical Tool for the Neonatologist 125

Martin Kluckow, FRACP, PhD, Istvan Seri, MD, PhD, and Nick Evans, DM, MRCPCH, Sydney, Australia, and LosAngeles, California

50 Years Ago in The Journal of Pediatrics—Observations on the Clinical Use of V-Cillin inPediatric Practice 130

Hugh S. Lam, MRCPCH, Hong Kong

SPECIAL ARTICLE

Pediatric Workforce: A Look at Pediatric Infectious Diseases Data from the American Board ofPediatrics 131

Linda A. Althouse, PhD, and James A. Stockman III, MD, Chapel Hill, North Carolina

ORIGINAL ARTICLES

Efficacy and Safety of Extended Release Metoprolol Succinate in Hypertensive Children 6 to 16Years of Age: A Clinical Trial Experience 134

Donald L. Batisky, MD, Jonathan M. Sorof, MD, Jennifer Sugg, MS, Michaelene Llewellyn, BSN, RN, MAS, MichaelKlibaner, MD, PhD, James W. Hainer, MD, MPH, Ronald J. Portman MD, and Bonita Falkner, MD, for the Toprol-XLPediatric Hypertension Investigators, Columbus, Ohio, Wilmington, Delaware, Houston, Texas, and Philadelphia, Pennsylvania


February 2007 • Volume 150 • Number 2



Diagnosis and Localization of Focal Congenital Hyperinsulinism by 18F-Fluorodopa PET Scan 140

Olga T. Hardy, MD, PhD, Miguel Hernandez-Pampaloni, MD, PhD, Janet R. Saffer, PhD, Mariko Suchi, MD, PhD,Eduardo Ruchelli, MD, Hongming Zhuang, MD, PhD, Arupa Ganguly, PhD, Richard Freifelder, PhD, N. Scott Adzick, MD,Abass Alavi, MD, and Charles A. Stanley, MD, Philadelphia, Pennsylvania

Longitudinal Lipid Screening and Use of Lipid-Lowering Medications in Pediatric Type 1 Diabetes 146

David M. Maahs, MD, MA, R. Paul Wadwa, MD, Kim McFann, PhD, Kristen Nadeau, MD, Matthew R. Williams, BS,Robert H. Eckel, MD, and Georgeanna J. Klingensmith, MD, Aurora, Colorado

Altered Basal Cortisol Levels at 3, 6, 8 and 18 Months in Infants Born at Extremely LowGestational Age 151

Ruth E. Grunau, PhD, David W. Haley, PhD, Michael F. Whitfield, MD, Joanne Weinberg, PhD, Wayne Yu, BSc, and PaulThiessen, MD, Vancouver, British Columbia, Canada

Cranial Ultrasound Scanning and Prediction of Outcome in Newborns with CongenitalCytomegalovirus Infection 157

Gina Ancora, MD, Marcello Lanari, MD, PhD, Tiziana Lazzarotto, MD, Valentina Venturi, MD, Elisabetta Tridapalli, MD,Fabrizio Sandri, MD, Maddalena Menarini, MD, Emanuela Ferretti, MD, and Giacomo Faldella, MD, Bologna, Italy, andQuebec, Canada

Comparative Efficacy and Cost of Asthma Care in Children with Asthma Treated with FluticasonePropionate and Montelukast 162

David A. Stempel, MD, Denise T. Kruzikas, PhD, and Ranjani Manjunath, MSPH, Bellevue and Seattle, Washington,Phoenix, Arizona, and Research Triangle Park, North Carolina

Clinical Spectrum and Histopathologic Features of Chronic Hepatitis C Infection in Children 168

Parvathi Mohan, MD, Camilla Colvin, MPH, BS, Chevelle Glymph, MPH, Roma R. Chandra, MD, David E. Kleiner, MD,Kantilal M. Patel, PhD, Naomi L.C. Luban, MD, and Harvey J. Alter, MD, Washington, DC, and Bethesda, Maryland

Fetal Alcohol Spectrum Disorders in Young Adulthood 175

Hans-Ludwig Spohr, MD, Judith Willms, MD, and Hans-Christoph Steinhausen, MD, PhD, Berlin, Germany, and Zurich,Switzerland

Clinical and Laboratory Features, Hospital Course, and Outcome of Rocky Mountain SpottedFever in Children 180

Steven C. Buckingham, MD, Gary S. Marshall, MD, Gordon E. Schutze, MD, Charles R. Woods, MD, MS, Mary AnneJackson, MD, Lori E. R. Patterson, MD, and Richard F. Jacobs, MD, as the Tick-borne Infections in Children Study Group,Memphis and Knoxville, Tennessee, Louisville, Kentucky, Little Rock, Arkansas, Winston-Salem, North Carolina, and KansasCity, Missouri

Parental Depression Following the Early Diagnosis of Cystic Fibrosis: A Matched, ProspectiveStudy 185

Claire Glasscoe, PhD, Gillian A. Lancaster, PhD, Rosalind L. Smyth, MD, and Jonathan Hill, FRCPsych, Liverpool, UnitedKingdom

50 Years Ago in The Journal of Pediatrics—Intrathecal Hydrocortisone in the Treatment ofTubercolous Meningitis 191

Albert M. Li, MB, Shatin, Hong Kong

continued on page 6AThe Journal of Pediatrics February 2007 5A

Prolonged QTc Intervals and Decreased Left Ventricular Contractility in Patients with PropionicAcidemia 192

Daniela Baumgartner, MD, Sabine Scholl-Bürgi, MD, Jörn Oliver Sass, Dr. rer. nat., Wolfgang Sperl, MD, PhD, UlrichSchweigmann, MD, Jörg-Ingolf Stein, MD, and Daniela Karall, MD, Innsbruck and Salzburg, Austria, and Freiburg, Germany

50 Years Ago in The Journal of Pediatrics—Spontaneous Pneumothorax in the First Ten Days ofLife 197

James M. Greenberg, MD, Cincinnati, Ohio

GRAND ROUNDS

A Newborn Infant with Protracted Diarrhea and Metabolic Acidosis 198

Amir Bar, MD, Arieh Riskin, MD, Theodore Iancu, MD, Irena Manov, PhD, Ayala Arad, MD, and Ron Shaoul, MD, Haifa,Israel

CLINICAL AND LABORATORY OBSERVATIONS

Gaucher Disease: Progressive Mesenteric and Mediastinal Lymphadenopathy Despite EnzymeTherapy 202

T. Andrew Burrow, MD, Mitchell B. Cohen, MD, Ronald Bokulic, MD, Gail Deutsch, MD, Arabinda Choudhary, MD,Richard A. Falcone, Jr, MD, and Gregory A. Grabowski, MD, Cincinnati, Ohio

Recipient Twin Limb Ischemia with Postnatal Onset 207

Roland Spencer Broadbent, MB, ChB, Dunedin, New Zealand

Congenital Patent Ductus Venosus: An Association with the Hyper IgE Syndrome 210

Keren Sagiv-Friedgut, MD, Michaela Witzling, MD, Ilan Dalal, MD, Chana Vinkler, MD, Eli Someh, MD, and Arie Levine,MD, Holon and Tel Aviv, Israel

50 Years Ago in The Journal of Pediatrics—Classification and Etiological Factors in MentalRetardation 212

Pasquale J. Accardo, MD, Richmond, Virginia

LETTERS

The following section is available in the online version of The Journal.

Erythromycin for Treatment of Feeding Intolerance in Preterm Infants e30

Bai-Horng Su, MD, and Hung-Chih Lin, MD, Taichung, Taiwan

e30Sandra Mascarenhas, MRCPCH, and Mithilesh Lal, MRCPCH, Middlesborough, United Kingdom

Reply e31

Pracha Nuntnarumit, MD, Pakaphan Kiatchoosakun, MD, Wacharee Tantiprapa, MD, and Suppawat Boonkasidecha, MD,Bangkok, Khon Kaen, and Chiang Mai, Thailand

Cardiovascular Support in the Preterm: Treatments in Search of Indications e31

Istvan Seri, MD, PhD, Los Angeles, California

Reply e33

Keith J. Barrington, MD, Montreal, Quebec, Canada


READER SERVICES

Information for Readers 9A

Announcements 10A

Guide for Authors Available at www.jpeds.com


EDITORWilliam F. Balistreri

Cincinnati, Ohio

ASSOCIATE EDITORS

Stephen R. DanielsDenver, Colorado

Alan H. JobeCincinnati, Ohio

Sarah S. LongPhiladelphia, Pennsylvania

Thomas R. WelchSyracuse, New York

Robert W. WilmottSt. Louis, Missouri

MANAGING EDITOR

Alice M. LandwehrCincinnati, Ohio

EDITORIAL BOARD

Steven H. AbmanDenver, Colorado

Pasquale J. AccardoRichmond, Virginia

Hans C. AnderssonNew Orleans, Louisiana

James BrownEast Syracuse, New York

Ronald I. ClymanSan Francisco, California

Paul G. FisherPalo Alto, California

John G. FrohnaEditor, Current Best Evidence

Ann Arbor, Michigan

Thomas P. GreenChicago, Illinois

Terrie E. InderSt. Louis, Missouri

M. Susan JayMilwaukee, Wisconsin

Nancy F. KrebsDenver, Colorado

Mary B. LeonardPhiladelphia, Pennsylvania

Wayne J. MorganTucson, Arizona

James F. PadburyProvidence, Rhode Island

Janet H. SilversteinGainesville, Florida

Charles StanleyPhiladelphia, Pennsylvania

Bonita StantonEditor, AMSPDC Section

Detroit, Michigan

Marshall L. SummarNashville, Tennessee

Reginald L. WashingtonDenver, Colorado

Russell E. WareMemphis, Tennessee

INTERNATIONAL ADVISORY PANEL

Anita AperiaStockholm, Sweden

Michael J. LentzeBonn, Germany

Albert M. LiShatin, Hong Kong

Philip M. ShermanToronto, Canada




Information for ReadersCorrespondence. Communications regarding original articlesand editorial management should be addressed to William F.Balistreri, MD, The Journal of Pediatrics, Children’s HospitalMedical Center, 3333 Burnet Ave, MLC 3021, Cincinnati, OH45229-3039; telephone 513-636-7140.

Customer service. (orders, claims, online, change of address):Elsevier Periodicals Customer Service, 6277 Sea Harbor Drive,Orlando, FL 32887-4800. Tel: (800) 654-2452 (U.S. andCanada); (407) 345-4000 (outside U.S. and Canada). Fax: (800)225-6030 (U.S. and Canada); (407) 363-9661 (outside U.S. andCanada). E-mail: [email protected]

Address changes must be submitted four weeks in advance.

Yearly subscription rates. United States and possessions:Individual $189.00; Institution $520.00; Student/Resident: $95.00.All other countries (prices include airspeed delivery): Individual$235.00; Institution $566.00; Student/Resident $120.00. Single Issues$48.00. To receive student/resident rate, orders must be accompaniedby name of affiliated institution, date of term and the signature ofprogram/residency coordinator on institution letterhead. Orders willbe billed at the individual rate until proof of status is received.Current prices are in effect for back volumes and back issues.

Further information on this journal is available from the Publisheror from this journal’s website (http://www.elsevier.com/jpeds.Information on other Elsevier products is available through Elsevier’swebsite (http://www.elsevier.com).

Advertising information. Advertising orders and inquiries canbe sent to: USA, Canada, and South America, Pat Hampton, Elsevier,360 Park Avenue South, New York, NY, 10010; phone: 212-633-3181;fax: 212-633-3820. Classified advertising orders and inquiries can be sentto Tracy Peppers, Elsevier Inc., 360 Park Ave South, New York, NY10010-1710; phone: 212-633-3766; FAX 212-633-3820. Europe andthe rest of the world, Julie Toop; phone �44 (0) 1865 843016; fax �44(0) 1865 843976; e-mail [email protected].

Author inquiries. For inquiries relating to the submission ofarticles (including electronic submission where available), please visitElsevier’s Author Gateway at http://authors.elsevier.com. The AuthorGateway also provides the facility to track accepted articles and set upe-mail alerts to inform you of when an article’s status has changed, aswell as detailed artwork guidelines, copyright information, frequentlyasked questions, and more. Please see Information for Authors forindividual journal. Contact details for questions arising afteracceptance of an article, especially those relating to proofs, areprovided after registration of an article for publication.

�� The paper used in this publication meets the requirementsof ANSI/NISO Z39.48-1992 (Permanence of Paper).

Reprints. For queries about author offprints, [email protected]. To order 100 or more reprints foreducational, commercial, or promotional use, contact theCommercial Reprints Department, Elsevier Inc., 360 Park AvenueSouth, New York, NY 10010-1710. Fax: (212) 462-1935; [email protected]. Reprints of single articles available onlinemay be obtained by purchasing Pay-Per-View access for $30 per

article on the journal website, www.jpeds.com. Microform editionavailable from ProQuest Information and Learning, 300 N ZeebRd, Ann Arbor, MI 48106-1346.

© 2006 Mosby, Inc. All rights reserved.This journal and the individual contributions contained in it

are protected under copyright by Mosby, Inc., and the followingterms and conditions apply to their use:

Photocopying. Single photocopies of single articles may be madefor personal use as allowed by national copyright laws. Permission of thePublisher and payment of a fee is required for all other photocopying,including multiple or systematic copying, copying for advertising orpromotional purposes, resale, and all forms of document delivery. Specialrates are available for educational institutions that wish to makephotocopies for nonprofit educational classroom use.

Permissions may be sought directly from Elsevier’s RightsDepartment in Philadelphia, PA, USA: phone (�1) 215 239 3804,fax (�1) 215 239 3805, e-mail [email protected] may also be completed online via the Elsevier homepage(http://www.elsevier.com/locate/permissions).

In the USA, users may clear permissions and make paymentsthrough the Copyright Clearance Center, Inc., 222 RosewoodDrive, Danvers, MA 01923, USA; phone: (�1) (978) 7508400,fax: (�1) (978) 7504744, and in the UK through the CopyrightLicensing Agency Rapid Clearance Service (CLARCS), 90Tottenham Court Road, London W1P 0LP, UK; phone: (�44)20 7631 5555; fax: (�44) 20 7631 5500. Other countries mayhave a local reprographic rights agency for payments.

Derivative works. Subscribers may reproduce tables ofcontents or prepare lists of articles including abstracts for internalcirculation within their institutions. Permission of the Publisher isrequired for resale or distribution outside the institution.

Permission of the Publisher is required for all other derivativeworks, including compilations and translations.

Electronic storage or usage. Permission of the Publisher isrequired to store or use electronically any material contained inthis journal, including any article or part of an article.

Except as outlined above, no part of this publication may bereproduced, stored in a retrieval system or transmitted in any formor by any means, electronic, mechanical, photocopying, recordingor otherwise, without prior written permission of the Publisher.

Address permissions requests to: Elsevier Rights Department,at the fax and e-mail addresses noted above.

Notice. No responsibility is assumed by the Publisher for anyinjury and/or damage to persons or property as a matter ofproducts liability, negligence or otherwise, or from any use oroperation of any methods, products, instructions or ideascontained in the material herein. Because of rapid advances in themedical sciences, in particular, independent verification ofdiagnoses and drug dosages should be made.

Although all advertising material is expected to conform toethical (medical) standards, inclusion in this publication does notconstitute a guarantee or endorsement of the quality or value ofsuch product or of the claims made of it by its manufacturer.




ANNOUNCEMENTS

March 2007Miami Children’s Hospital 42nd Annual Pediatric

Postgraduate Course “Perspectives in Pediatrics” and Pe-diatric Board Review Dates. March 16-22, 2007, InterCon-tinental Hotel Downtown, Miami, FL. Adolescent Medicine,Allergy, Immunology, Cardiology, Dermatology, Gastroen-terology, Genetics, Hematology/Oncology, Infectious Dis-eases, Nephrology, Neurology, Neurosurgery, Nutrition,Opthalmology, Orthopedics, Pediatric Surgery, Plastic Sur-gery, Preventive Medicine/Community Pediatrics, Psychia-try, Radiology, and more. For more information, contactRiccardo Firmino, Conference Management Office, MiamiChildren’s Hospital; phone: 305-756-0791; E-mail:[email protected]; Website: www.ppgcpip.com.

The Spectrum of Developmental Disabilities XXIX.March 26-28, 2007, Johns Hopkins University School of Med-icine, Baltimore, MD. For more information, contact the Officeof Continuing Medical Education; phone: 410-955-2959; E-mail: [email protected]; Website: www.hopkinscme.net.

April 20072007 Regional Conference on Child Health Psychol-

ogy. April 26-28, 2006, Westin Hotel, Cincinnati, OH.Sponsored by the Division of Behavioral Medicine and Clin-ical Psychology at Cincinnati Children’s Hospital MedicalCenter in cooperation with the Society of Pediatric Psychol-ogy and the Ohio Chapter of the Society for Developmentaland Behavioral Pediatrics. For more information visit www.cincinnatichildrens.org/spp-conference.

May 2007The Programme for Global Paediatric Research Sym-

posium: “Global Childhood Diseases Which Can ImpairDevelopment” and Workshop: “Outcome Studies.” May 8-9,2007, Toronto. Sponsored by The Programme for Global Pae-diatric Research. PGPR’s fifth symposium will be held May 8,2007 in conjunction with the annual meeting of the PediatricAcademic Societies. To register for the PAS Meeting, at whichthis symposium will be held, please go to www.pas-meeting.org.To register for the follow-up workshop on May 9, 2007, pleasecontact Alvin Zipursky, Chair and Scientific Director; phone:416-813-8762; E-mail: [email protected]; Website:www.globalpaediatricresearch.org.

18th Annual Spring Conference on Pediatrics. May16-19, 2007, Marriott Frenchman’s Reef Beach Resort, St.Thomas, United States Virgin Islands. Sponsored by Sympo-sia Medicus. For more information, contact Symposia Medi-cus; phone: 925-969-1789, 800-327-3161; E-mail: [email protected]; Website: www.symposiamedicus.org.

July 200729th Annual Aspen Conference on Pediatric Gastro-

intestinal Disease: Pediatric Gastrointestinal Disease andIntestinal Transplantation. July 22-27, 2007, SnowmassConference Center, Snowmass (Aspen), CO. Sponsored byCincinnati Children’s Hospital Medical Center. The confer-ence will be devoted to the diagnosis and management ofgastrointestinal disease in children. Specific topics includeinflammatory bowel disease (diagnosis, complications, andmanagement), diarrhea (acute and chronic), celiac disease, h.pylori and peptic ulcer disease, gastroesophageal reflux, ab-dominal pain/irritable bowel syndrome, constipation, obesityand NAFLD/NASH, TPN-associated cholestasis, allergicgastroenteropathies, motility disorders, short gut syndrome/NEC, pancreatic disease/cystic fibrosis, and various aspects ofsmall bowel transplantation. For more information, contactLaura Werts, CME Office, Cincinnati Children’s HospitalMedical Center; phone 513-636-6732; E-mail: [email protected]; Website: www.cincinnatichildrens.org.

August 2007The 25th International Congress of Pediatrics. Au-

gust 25-30, 2007, Athens, Greece. For more information,visit www.icp2007.gr.

2007-2008 Certifying Examinations of theAmerican Board of Pediatrics

111 Silver Cedar Court, Chapel Hill, NC 27514-1513telephone: 919-929-0461 fax: 919-918-7114 or 919-929-9255 Website: www.abp.org

All applicants for certifying examinations must com-plete applications online during the registration periods. Thefinal month of each registration requires payment of a latefee. The requirements for online applications may be foundon the ABP Website (www.abp.org) or may be obtained bycontacting the ABP. Additional information including eligi-bility requirements and registration dates may be found on theABP Website.


Ibmaanchtm

npccsriaitmp

sMFpam

P2

FoM

RRS(

NOTES FROM THE ASSOCIATION OF MEDICALSCHOOL PEDIATRIC DEPARTMENT CHAIRS, INC.

Factors Influencing Community PediatricsTraining in Residency

CYNTHIA S. MINKOVITZ, MD, MPP, ANITA CHANDRA, DRPH, BARRY S. SOLOMON, MD, MPH,
LEE M. SANDERS, MD, MPH, HOLLY A. GRASON, MA, AND CAROL CARRACCIO, MD
d1wctmapcaUdik

rdohiitaoeo

aItiR

t�riwi

AM

SP

DC

ncreasingly, pediatricians are expected to work collabora-tively within their communities to address the complexenvironmental and social issues affecting children’s well-

eing.1 The American Academy of Pediatrics identifies com-unity pediatrics as “an integral part of the professional role

nd duty of the pediatrician.”2 Residency training is perceiveds an opportune time in which to equip pediatricians with theecessary skills to advocate effectively for children in theirommunities.3-6 Since 1997, pediatric residency programsave been required to provide structured educational oppor-unities, both didactic and experiential, in the areas of com-unity pediatrics and child advocacy.7

Most pediatric residency programs do provide commu-ity experiences for residents, with more than two thirds ofrograms in 2002 requiring involvement in four or moreommunity settings as well as teaching about issues such asultural competency, and the mental health and social serviceystems.8 Nonetheless, there is considerable variability in cur-icular offerings. Prior reports have suggested that residentnterest,9,4 faculty involvement,3,4 and the commitments ofcademic and community organizations3,4,10 are instrumentaln implementing community pediatrics curricula in residencyraining programs. Yet, there has been no systematic assess-ent of the factors that influence the ability of residency

rograms to implement community training.

METHODSTo address these issues, we conducted a Web-based

urvey of pediatric residency program directors betweenay and July 2005. The American Medical Association’s

ellowship and Residency Electronic Interactive Databaserovided contact information.11 The 11-item survey askedbout program identification, designation of having a pri-ary care track, and continuity clinic settings. Respon-

resented at the Pediatric Academic Societies’ Annual Meeting, Baltimore, MD, April 29,006.

rom the Johns Hopkins Bloomberg School of Public Health, and the Johns Hopkins Schoolf Medicine, Baltimore, Maryland; RAND, Arlington, Virginia; the University of Miami,iami, Florida; and the University of Maryland School of Medicine, Baltimore, Maryland.

eprint requests: Cynthia Minkovitz, MD, MPP, Department of Population, Family andeproductive Health, Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe

ctreet, E4636, Baltimore, MD 21205.E-mail: [email protected] Pediatr 2007;150:119-20)

ents indicated required or elective resident involvement in5 community settings (eg, Head Start program) andhether their programs provided education regarding 14

ommunity health topics (as didactic or practical instruc-ion). Respondents also reported the level of resident involve-ent (4-point Likert scale, 1 “not at all” to 4 “heavily”) in five

dditional activities: communicating with elected officials;roviding legislative testimony; participating in longitudinalommunity projects; conducting research in the community;nd addressing parents, teachers, or other community groups.sing a similar Likert scale, respondents reported on theegree to which the program emphasized resources and train-ng related to 11 content areas (eg, child advocacy) in mar-eting the program to prospective residents.

We developed a community orientation scale usingesponses to a question asking “In the last 3 years, to whategree have the following factors influenced your ability toffer residency training experiences in community childealth?” The scale included eight factors: community

nterest, faculty expertise and interest, departmental prior-ties, resident interest, institutional initiatives, implemen-ation of a competency-based curriculum, and resourcesnd money (Table I). For each program, the communityrientation score was calculated as the mean rating for allight factors, with higher scores indicating a more positiverientation to community pediatrics.

Data analysis was conducted using the Statistical Pack-ge for the Social Sciences version 11.5 (SPSS, Inc., Chicago,ll) using non-parametric tests to allow for non-normal dis-ribution of responses. Mean scores are reported for ease ofnterpretation. The Johns Hopkins Committee on Humanesearch approved the study.

RESULTSOf the 203 accredited programs, 161 program direc-

ors or their designees completed the survey (response rate79%). All but 12 respondents completed the question

egarding factors influencing community pediatrics train-ng with a resulting analytic sample of 149. The degree tohich selected factors positively influence a program’s abil-

ty to offer residency training experiences in community
hild health (“community orientation”) varied, with work
119

hm

mchrPCemwDo

g3fmwt(ooI

ppptcss

tembnerdtvci

grddcicsahrd

iruaeesicspaaisc

dDvp

0C

T(

CFFDRIIRW

Nro8c*p†

AM

SP

DC

1

ours restrictions having the lowest score (2.41) and com-unity interest having the highest (3.95) (Table I).

Using scaled scores, there were no differences in com-unity orientation by program size or presence of a primary

are track. However, those programs offering communityealth centers as one of the possible continuity clinic siteseported greater community orientation (mean 3.77 vs 3.45,

� .004). Among sites receiving support from the Dysonommunity Pediatrics Training Initiative ($500,000/year for

ach of 5 years to support training in community pediatrics),ean scores regarding the importance of resources and moneyere substantially higher than for their counterparts withoutyson support (3.83 vs 2.37, P � .001), as were overall scores

n community orientation (4.31 vs 3.55, P �.001).Greater community orientation was reported by pro-

rams with more required community settings, with scores of.38 for programs requiring fewer than four settings, 3.67 forour to six settings, and 3.80 for programs requiring seven orore settings (P � .006). Greater community orientation alsoas associated with more didactic training in 5 of the 14

opics and more practical experiences in 3 of the 14 topicsTable II; available at www.jpeds.com). Greater communityrientation was reported by programs with more involvementf residents in the five selected community activities (TableII; available at www.jpeds.com).

Community orientation also was associated with howrograms market themselves to prospective residents. Inarticular, higher community orientation scores were re-orted by programs that emphasized their resources andraining related to primary care, behavior and development,ommunity pediatrics, and advocacy but not to other areasuch as breadth of specialty services, clinical and lab re-

able I. Factors influencing community pediatricsn � 149)*

Factor Mean SD

ommunity interest 3.95 0.90aculty expertise 3.94 1.01aculty interest 3.93 1.08epartmental priorities 3.83 1.08esident interest 3.72 1.07

nstitutional initiatives 3.54 0.90mplementing a competency-based curriculum 3.47 0.96esources and money 2.49 1.19ork hours restrictions† 2.41 0.94

ote: A community-orientation score was created for each program using the sum ofesponses from the 8 factors/ total number of items (higher score � more positiverientation to community pediatrics). The Cronbach � reliability coefficient for the-item community-orientation scale (sum of items/8) was 0.83, indicating high internalonsistency. The scale range was 1.75 to 4.88 with a mean score of 3.61 (SD 0.69).Respondents identified the degree of influence (Likert scale, 1 “very negative” to 5 “veryositive”).Excluded in 8-item scale. All other factors were highly correlated (P � .05).

20 Minkovitz et al

DISCUSSIONThese findings suggest that multiple factors influence

he ability of pediatric residency programs to offer trainingxperiences related to community child health and thatany of these factors are mutable. Across all programs,

oth resources and money and work hours restrictionsegatively impacted training, whereas all other factorsxerted a positive influence. However, among programseceiving significant financial support for community pe-iatrics training, resources and money were viewed posi-ively; such resources can be used to support training pro-ided by community partners, faculty development, andoordinators responsible for scheduling resident activitiesn community settings.

Although work hours restrictions influenced pro-rams’ abilities to offer community health training, theseestrictions did not correlate with the other factors thatescribe community orientation. It is likely that respon-ents perceived the impact of work hours restrictions forommunity child health training to be comparable to itsmpact on other training opportunities, and not unique toommunity pediatrics. In a separate question asking re-pondents about the influence of work hours on theirbility to offer training experiences related to the care ofospitalized patients, scores for the impact of work hoursestrictions were comparable to those for community pe-iatrics.

This brief report offers a glimpse into the factors thatnfluence the ability of programs to offer community pediat-ics training experiences. Clearly, more work is needed tonderstand the role of other factors that may be importantnd to understand the full scope and quality of these experi-nces. However, this study does demonstrate a positive ori-ntation to community pediatrics training when there is in-titutional support as well as community, faculty, and residentnterest. Pediatrics departments that are interested in revisingurricular opportunities related to community pediatricshould assess the varied factors that likely influence theirrograms’ abilities to support these efforts. This informationlso may be important in informing the Residency Reviewnd Redesign Project of the American Board of Pediatrics ast looks to how the future of pediatric residency traininghould be structured to meet the evolving healthcare needs ofhildren.

We gratefully acknowledge participation of the programirectors and generous support from The Dyson Foundation. Theyson Initiative National Evaluation Advisory Committee re-

iewed survey content and provided critical input regardingreliminary findings.

References available at www.jpeds.com.

022-3476/$ - see front matteropyright © 2007 Mosby Inc. All rights reserved.

earch, or intensive care. 10.1016/j.jpeds.2006.10.047

The Journal of Pediatrics • February 2007

http://www.jpeds.com


http://www.www.jpeds.com

1b22p3Tf24tn5n26E

h1

7RAc

8p

9bp

1CP

1A

T

SWFPJM

SMHCLMIC

* nce tes

REFERENCES. Sanders LM, Robinson TN, Forster LQ, Plax K, Brosco JP, Brito A. Evidence-ased community pediatrics: building a bridge from bedside to neighborhood. Pediatrics005;115:1142-7.. American Academy of Pediatrics Committee on Community Health Services. Theediatrician’s role in community pediatrics. Pediatrics 2005;115;1092-4.. Shipley LJ, Stelzner SM, Zenni EA, Hargunani D, O’Keefe J, Miller C, et al.eaching community pediatrics to pediatric residents: strategic approaches and success-

ul models for education in community health and child advocacy. Pediatrics005;115;1150-7.. Palfrey JS, Hametz P, Grason H, McCaskill QE, Scott G, Chi GW. Educatinghe next generation of pediatricians in urban health care: the Anne E. Dyson Commu-ity Pediatrics Training Initiative. Acad Med 2004;79;1184-91.. The Future of Pediatric Education II: organizing pediatric education to meet theeeds of infants, children, adolescents, and young adults in the 21st century. Pediatrics000;105;163-212.. Ambulatory Pediatric Association Education Committee. Kittredge D, ed. APA

able II. Community orientation and health system

Community healthsystem topic

Didactic trainin

Yes No

ocial service system 3.66 (0.69) 3.42 (0.66)elfare system 3.76 (0.64) 3.44 (0.70)

oster care system 3.74 (0.61) 3.46 (0.74)ublic education system 3.70 (0.69) 3.52 (0.68)uvenile justice system 3.68 (0.68) 3.55 (0.69)ental health system forchildren/adolescents

3.65 (0.67) 3.29 (0.80)

ubstance abuse treatment 3.63 (0.66) 3.51 (0.82)anaged care 3.63 (0.69) 3.54 (0.68)ealthcare financing 3.66 (0.69) 3.38 (0.63)ultural competency 3.69 (0.65) 3.17 (0.74)egislative advocacy 3.75 (0.62) 3.33 (0.75)igrant healthcare 3.96 (0.66) 3.56 (0.68)

ndian Health Service 3.80 (0.63) 3.59 (0.69)hildren with specialhealthcare needs

3.63 (0.67) 3.32 (0.92)

Higher community-orientation scores reflect greater community orientation. Significa

ducational Guidelines for Pediatric Residency. 2004. Available online at: o

ttp://www.ambpeds.org/EGWEBNEW/index.cfm?CFID�203907&CFTOKEN�

9007223.

. Accreditation Council for Graduate Medical Education. Pediatrics Residencyeview Committee. Program Requirements for Residency Training in Pediatrics. 2006.vailable online at: http://www.acgme.org/acWebsite/RRC_320/320_prIndex.asp. Ac-

essed 7/17/06.

. Solomon BS, Minkovitz CS, Mettrick JE, Carraccio C. Training in communityediatrics: a national survey of program directors. Ambul Pediatr 2004;4:476-81.

. Kaczorowski J, Aligne CA, Halterman JS, Allan MJ, Aten MJ, Shipley LJ. Alock rotation in community health and child advocacy: improved competency ofediatric residency graduates. Ambul Pediatr 2004;4;283-8.

0. Sidelinger DE, Meyer D, Blaschke GS, Hametz P, Batista M, Salguero R, et al.ommunities as teachers: learning to deliver culturally effective care in pediatrics.ediatrics 2005;115;1160-4.

1. Fellowship and Residency Electronic Interactive Database (FREIDA). Chicago:merican Medical Association; 2004. Available at http://www.ama-assn.

cs, mean (SD) (n � 149)*

Practical experience

P value Yes No P value

.08 3.64 (0.69) 3.49 (0.69) .34

.003 3.64 (0.67) 3.56 (0.72) .58

.02 3.66 (0.64) 3.53 (0.76) .47

.13 3.67 (0.67) 3.44 (0.71) .06

.35 3.61 (0.62) 3.59 (0.76) .92

.14 3.63 (0.71) 3.56 (0.58) .44

.72 3.64 (0.66) 3.56 (0.74) .67

.56 3.65 (0.69) 3.51 (0.69) .24

.05 3.69 (0.72) 3.51 (0.65) .09

.007 3.66 (0.69) 3.32 (0.64) .03

.002 3.75 (0.59) 3.45 (0.77) .02

.022 3.94 (0.65) 3.56 (0.69) .03

.27 3.61 (0.82) 3.61 (0.69) .74

.41 3.62 (0.69) 3.46 (0.69) .46

ting reported using Mann-Whitney tests.

topi

g

rg/ama/pub/category/2997.html.

120.e1

http://www.ambpeds.org/EGWEBNEW/index.cfm?CFID=203907&CFTOKEN=19007223

http://www.ambpeds.org/EGWEBNEW/index.cfm?CFID=203907&CFTOKEN=19007223

http://www.acgme.org/acWebsite/RRC_320/320_prIndex.asp.%20Accessed%207/17/06



http://www.ama-assn.org/ama/pub/category/2997.html

http://www.ama-assn.org/ama/pub/category/2997.html

T

C

P

P

C

A

*

1

able III. Community orientation and resident activities (n � 149), mean (SD)*

Activity

Degree of involvement

P valueNot at all Somewhat Moderately Heavily

ommunicated withelected officials toadvocate on behalf ofchildren’s concerns

3.36 (0.77) 3.62 (0.62) 3.98 (0.55) 3.75 (0.83) .004

rovided legislativetestimony

3.49 (0.72) 3.92 (0.46) 3.84 (0.78) 4.0 (—) .01

articipated on alongitudinal projectproviding services in thecommunity

3.31 (0.78) 3.41 (0.64) 3.88 (0.56) 4.00 (0.55) �.001

onducted research inthe community

3.23 (0.78) 3.57 (0.61) 4.00 (0.55) 4.19 (0.55) �.001

ddress parents, teachers,or other communitygroups

3.30 (0.99) 3.40 (0.69) 3.69 (0.58) 4.14 (0.40) �.001

Higher community-orientation scores reflect greater community orientation. Significance testing reported using Kruskall-Wallis tests.

20.e2 Minkovitz et al The Journal of Pediatrics • February 2007

EDITORIALS

Is There a Role for �-Adrenergic Blockers in TreatingHypertension in Children?

Over the past decade, it has been recognized thathypertension exists in children and adolescents, thatpediatricians should screen for and evaluate it, and

that the ideal goal is to prevent the long-term cardiovascularconsequences of elevated blood pressure. Two parallel devel-opments have influenced the choice of antihypertensive agentfor children with persistent elevated blood pressure. The firstdevelopment is the concept that drug testing is needed toensure the efficacy and safety of any drug in children.1 Amechanism to accomplish this goal is patent extension toinduce a company marketing a drug under a patent to supportdrug trials that may lead to improved labeling.2 This exten-sion has been applied to antihypertensive agent studies.3 Thesecond development is the recommendations emanating froma report on the diagnosis, evaluation, and treatment of hy-pertension in children and adolescents recommending thatcertain classes of antihypertensive agents be used as therapy.4

In this issue of The Journal, Batisky et al5 evaluate a classof antihypertensive drugs recommended by that report: the car-dioselective �-blocker metoprolol.4 The importance of this studylies in the fact that it is well designed; previously, few well-conducted trials of �-blockers had been performed in children.5,6

The study evaluates metoprolol extended-release tablets, whichcan be given once daily. The findings are from a 4-week double-blind dose-range study and a 52-week open-label extensionstudy. Children age 6 to 16 years were evaluated.

The report provides evidence that extended-releasemetoprolol significantly reduced systolic blood pressure at 1.0and 2.0 mg/kg compared with placebo. Only the 2.0 mg/kgdose significantly reduced diastolic blood pressure. The drugwas well tolerated in the 52-week trial, and only 5% of studyparticipants had to drop out due to adverse events. Theresponse rate increased significantly, from 41% at the begin-ning of the study to 64% at week 52.

The principal value of this study lies in the assessmentof the safety and efficacy of a �-blocker in a short-term trial.This assessment is important because this class of agents hasnot been extensively evaluated.5,6 From these 2 studies, itappears that this agent is generally safe. However, a majorityof subjects in each trial did not achieve target blood pressurecontrol (55% and 59%, respectively) over the short term. Theopen-label study found that many more patients achieved thetarget blood pressure.5

A limitation of this study is that it fails to defineprecisely the circumstances in which a �-blocker would bechosen. All patients with hypertension secondary to coarcta-

tion of the aorta, pheochromocytoma, hyperthyroidism, andCushing syndrome were excluded, and most patients in bothtrials had essential hypertension. Patients with other forms ofhypertension were not generally evaluated.

The studies by Batisky5 and Sorof et al6 are placebo-controlled trials. Although the drug caused a greater drop inboth systolic and diastolic pressure than placebo, there was asubstantial placebo effect. This effect could be explained by 2factors: the Hawthorne effect and the oft-made observationthat hypertensive children may improve over time.7 TheHawthorne effect is a well-established phenomenon whereinthe very fact of entering a study has an effect on patientbehavior and may influence the symptoms or findings in apatient. Moreover, the data from these 2 trials and otherstudies indicate that children with mild primary hypertensionfollowed over time may experience normalization of bloodpressure without treatment. A clear implication is that suchpatients may need several months of observation before ini-tiation of pharmaceutical intervention.

The use of a placebo arm in childhood hypertensiontrials is controversial. Many hypertensive trials in bothchildren and adults do not have a placebo arm so as toavoid the well-established consequences of untreated hy-pertension.8 The International Conference on Harmonisa-tion (ICH) Harmonised Tripartite Guideline E10 con-cluded that “when a new treatment is tested for a conditionfor which no effective treatment is known, there is usuallyno ethical problem with a study comparing the new treat-ment to placebo.”9 The situation in antihypertensive trialsfalls outside this guideline, especially because so manyeffective agents are available.3,4,9 Flynn et al8 and Chesneyet al3 have made the argument that a placebo control can bejustified if the hyper-tension is mild, if thesubjects do not havehypertension-related tar-get organ damage, andif the study duration isshort.

In the study byBatisky et al,5 adverseeffects were uncommon,and even in the long-term open label com-ponent of the study,only 5 children dropped

See related article, p 134

Reprint requests: Russell W. Chesney, MD,Department of Pediatrics, University ofTennessee Health Science Center, 50 NDunlap, Room 304, Memphis, TN 38103.E-mail: [email protected].

J Pediatr 2007;150:121-2

0022-3476/$ - see front matter

Copyright © 2007 Mosby Inc. All rightsreserved.

10.1016/j.jpeds.2006.11.064

Editorials 121

out because of side effects. This represents just 5% of the 100patients in this phase of the trial. Reported side effects in-cluded fatigue, nightmares, anxiety, dizziness, and asthma. Tra-ditional contraindications to �-adrenergic inhibition, such asasthma and diabetes, also are relevant; however, a �1 receptorantagonist should largely affect the cardiogenic receptors. Thepresent study does not comment on the use of �-blockers incompetitive athletes or runners, in whom the need to raise theheart rate is pertinent.

A recent survey review of 438 North American pediatricnephrologists indicated that most used angiotension-convertingenzyme inhibitors (ACEIs), followed by calcium-channel block-ers (47% and 37%, respectively), in the first-line treatment ofprimary hypertension.10 �-blockers were used as a first-lineagent by only 7% of respondents, but in second-line therapy by17%. In children with hypertension associated with renal insuf-ficiency, ACEIs were used by 84% of respondents in first-linetherapy, and �-blockers were used only as second-line agents.The study of Batisky et al5 shows that the �-blocker metoprololcan be used safely in patients with primary hypertension, withthe target pressure reached in approximately 70% of subjects.This is valuable new information.

Russell W. Chesney, MDDeborah P. Jones, MD

Division of Pediatric Nephrology, Department of Pediatrics, and Center for PediatricPharmacokinetics and Therapeutics at the University of Tennessee Health Science Center

Children’s Foundation Research Center at Le Bonheur Children’s Medical CenterMemphis, TN

REFERENCES1. Christensen ML, Helms RA, Chesney RW. Is pediatric labeling really necessary?Pediatrics 1999;104:593-7.2. Chesney RW, Christensen ML. Changing requirements for evaluation of phar-macologic agents. Pediatrics 2004;113:1128-32.3. Chesney RW, Adamson P, Wells T, Wilson JT, Walson PD. The testing ofantihypertensive medications in children: report of the Antihypertensive Agent Guide-lines Subcommittee of the Pediatric Pharmacology Research Units. Pediatrics2001;107:558-61.4. National High Blood Pressure Education Program Working Group on HighBlood Pressure in Children and Adolescents. Fourth report on the diagnosis, evaluation,and treatment of high blood pressure in children and adolescents. Pediatrics2004;114:555-76.5. Batisky DL, Sorof JM, Sugg J, Llewellyn M, Klibaner M, Hainer JW, et al.Efficacy and safety of extended-release metoprolol succinate in hypertensive children 6to 16 years of age: a clinical trial experience. J Pediatr 2007;150:134-9.6. Sorof JM, Cargo P, Graepel J, Humphrey D, King E, Rolf C, et al. �-blocker/thiazide combination for treatment of hypertensive children: a randomized double-blind, placebo-controlled trial. Pediatr Nephrol 2002;17:345-50.7. Sorof JM, Urbina EM, Cunningham RJ, Hogg RJ, Moxey-Mims M, Eissa MA,et al. Screening for eligibility in the study of antihypertensive medication in children:experience from the Ziac Pediatric Hypertension Study. Am J Hypertens 2001;14:783-7.8. Flynn JT. Ethics of placebo use in pediatric clinical trials: the case of antihyper-tensive drug studies. Hypertension 2003;42:865-9.9. International Conference on Harmonisation of Technical Requirements for Reg-istration of Pharmaceuticals for Human Use. ICH Harmonised Tripartite GuidelineE10: Choice of control group and related issues in clinical trials. July 2000. Available athttp://www.ich.org/pdfICH/e10step4.pdf. Accessed September 2006.10. Woroniecki RP, Flynn JT. How are hypertensive children evaluated andmanaged? A survey of North American pediatric nephrologists. Pediatr Nephrol2005;20:791-7.

PET Scanning for Infants with HHI: A Small Step for Affected Infants,A Giant Leap for the Field

A ttempting to explain the pathophysiology of macro-somic infants with hypoglycemia has engaged phy-sician/scientists for some 50 years. First came the era

of careful observation and description suggesting a familial orgenetic basis and similarities in the symptoms to an overdoseof insulin.1,2 Next came the era of biochemistry, radioimmu-noassay, and glucose kinetics by means of stable isotopes thatdefined hyperinsulinemic hypoglycemia of infancy (HHI) asthe most common and potential neurologically devastatingform of neonatal hypoglycemia.3,4 The exponential leap for-ward with remarkable progress occurred in the past decade,since the elucidation of the molecular mechanisms responsiblefor insulin secretion and its modulation through the adeno-sine triphosphate–regulated potassium channel (KATP).5 This

channel converts the chemical energy of glucose or amino acidmetabolism to electrical signals that trigger voltage-gatedcalcium channels leading to insulin secretion. Inactivatingmutations in the actual channel pore (Kir 6.2), or its regula-tory subunit (SUR1),prevent channel opening,resulting in prolongedmembrane depolariza-tion, with ongoing andunregulated insulin se-cretion causing hypo-glycemia.5,6 Some ofthese mutations are fa-milial, mostly inheritedin an autosomal recessivemanner. In the generalpopulation, most casesof HHI are sporadic.

18F-L-DOPA 18F-fluoro-L-dopamineHHI Hyperinsulinemic hypoglycemia of infancyKATP Adenosine triphosphate–regulated potassium channel

See related article, p 140

Reprint requests: Mark A. Sperling, MD,Children’s Hospital of Pittsburgh, 3705Fifth Avenue, Pittsburgh, PA. E-mail:[email protected].

J Pediatr 2007;150:122-4



10.1016/j.jpeds.2006.11.058

122 Editorials The Journal of Pediatrics • February 2007

http://www.ich.org/pdfICH/e10step4.pdf

Dohipnoong

ecdahtmbauet

abssp(drtRd

dcbc

CN

MEDICALPROGRESS

Functional Echocardiography: An Emerging Clinical Tool for the Neonatologist

MARTIN KLUCKOW, FRACP, PHD, ISTVAN SERI, MD, PHD, AND NICK EVANS, DM, MRCPCH

espite significant advances in our ability to monitor complex and clinically relevant hemodynamic variables,1-3 in mostneonatal intensive care units (NICUs), cardiovascular function is assessed only by continuous heart rate, blood pressuremonitoring or poorly validated clinical signs such as capillary refill time.1 The use of indirect measures for assessment

f tissue perfusion is especially problematic in the very preterm neonate during the first postnatal days, when complexemodynamic changes occur during the transition to postnatal life.1 Although clearly these measurements give important

nformation, they provide only indirect and frequently limited insights into the complexities of cardiac function, changes ineripheral and pulmonary vascular resistance, intracardiac and extracardiac shunting, and the transitional circulation of theeonate. Indeed, the neonatologist is often faced with clinical dilemmas that are challenging to interpret and manage becausef a lack of basic hemodynamic physiologic information. Recent publications have highlighted the current uncertainty as to theptimal management of hypotension in the newborn infant4,5 and patency of the ductus arteriosus.6 The emerging field ofeonatal hemodynamics can offer a clearer understanding of the pathophysiology underlying these clinical presentations and helpuide treatment choices.2,3,7,8

In the neonate, echocardiography is commonly used to assess the structure and function of the heart. Obtaining anchocardiogram is usually dependent on the availability of pediatric cardiologists or echocardiographic technicians. The pediatricardiologist generally performs a single echocardiogram to rule out the presence of structural heart disease or global myocardialysfunction. However, this snapshot view is inadequate for the assessment of ongoing changes in the hemodynamic status suchs during the immediate postnatal transition or when cardiovascular compromise develops at a later stage during the neonate’sospital course. Longitudinal studies demonstrate a wide range of hemodynamic findings, both between different babies withhe same clinical scenario and in individual babies with time. These hemodynamic findings are often very different from whatight be assumed clinically, using conventional thinking. We propose the term “functional echocardiography” to describe the

edside use of echocardiography to longitudinally assess myocardial function, systemic and pulmonary blood flow, intracardiacnd extracardiac shunts, organ blood flow, and tissue perfusion. Functional echocardiography provides information about thenderlying hemodynamic function in real time and the changes in the cardiovascular status in response to treatment. Functionalchocardiography allows treatment to be targeted at actual rather than assumed functional status and for treatment effectivenesso be assessed over time.

Although the use of functional echocardiography has not yet been demonstrated toffect outcomes, its value for assessment of the rapidly changing hemodynamic status haseen embraced by many neonatologists.9,10 In Australia and New Zealand, a recent surveyhowed 40% of NICUs had at least one neonatologist with functional echocardiographickills,9 and several papers describe experience with neonatal functional echocardiogra-hy.7,9-12 The use of functional echocardiography is also widespread throughout Europepersonal communications) and a number of neonatal units in North America areeveloping the ability to provide point of care monitoring using functional echocardiog-aphy by appropriately trained neonatologists. More than 500 copies of a self-directedeaching program on echocardiography for the neonatologist using an interactive CD-OM13 are available throughout the world, again suggesting a widespread interest ineveloping these skills.

Until recently, functional echocardiography was primarily used as a research tool toefine the natural history of neonatal hemodynamics and relations to adverse out-omes.2,14-16 More recently, the hemodynamic effects of therapeutic interventions haveeen reported, but it remains to be shown whether the use of functional echocardiographyan modify neonatal outcomes.17,18

HD Congenital heart disease PDA Patent ductus arteriosus

From the Departments of Neonatal Medi-cine, Royal North Shore Hospital and RoyalPrince Alfred Mother’s and Babies, and Uni-versity of Sydney, Sydney, Australia; andUSC Division of Neonatal Medicine, De-partment of Pediatrics, Children’s HospitalLos Angeles and the Women’s and Chil-dren’s Hospital, LAC � USC Medical Cen-ter, Keck School of Medicine, University ofSouthern California, Los Angeles, California.

Submitted for publication Jun 21, 2006; lastrevision received Aug 19, 2006; acceptedOct 12, 2006.

Reprint requests: Dr Martin Kluckow, De-partment of Neonatal Medicine, RoyalNorth Shore Hospital, Pacific Highway, StLeonards, Sydney, NSW 2065 Australia.E-mail: [email protected].

J Pediatr 2007;150:125-30



ICU Neonatal Intensive Care Unit SBF Systemic blood flow
10.1016/j.jpeds.2006.10.056
125

tsiavm

TP

pAblpomstmLtefibrrSv

Ai

thcaeTncpibmfafirshe

tadstbLmadets

T(

bpHdsqobnlmvhlpcumaiaatmb

AW

stgpppfi

1

CLINICAL USES OF FUNCTIONALECHOCARDIOGRAPHY

The range of methodologies and clinical uses of func-ional echocardiography have been reviewed.2,15,16,19 Ithould be highlighted that all ultrasound measures have anntrinsic error ranging from about 10% for intraobserver vari-bility up to 15% to 20% for interobserver variability.16 Thisariance is similar to other noninvasive measures such aseasurement of cardiac output using thermodilution.

he Very Preterm Infant During the Transitionaleriod

The first 24 postnatal hours after birth of the veryreterm infant is a period of unique circulatory vulnerability.

significant number of babies develop not only hypotensionut also low systemic blood flow (SBF).16 During this period,ow SBF is often not recognized by measurement of bloodressure3 and has been associated with a range of adverseutcomes, both short and long term.20,21 The usual measure-ents of cardiac output used in older children and adults,

uch as the left ventricular output, are affected by the transi-ional circulation, so other measures of SBF such as measure-ent of superior vena cava flow have been developed.16,22

ow SBF also relates to larger ductal shunts, so assessment ofhe early constriction of the ductus arteriosus is important inarly echocardiographic assessments. After the transitionalrst 24 hours, where low SBF predominates, hypotensiveabies usually have normal or high SBF, indicating low pe-ipheral vascular resistance that is probably due to abnormalegulation of vascular tone.23 This change from low to higherBF has been implicated in the development of peri-intra-entricular hemorrhage in the very preterm neonate.20

ssessment and Monitoring of the Ductus Arteriosusn the Preterm Infant

In some preterm infants, the ductus arteriosus effec-ively constricts within a few hours of birth, although othersave a persisting large ductus arteriosus with no evidence ofonstriction. Many of the strategies for treating patent ductusrteriosus (PDA) are based on the incorrect assumption thatarly ductal shunting is of limited hemodynamic significance.he dominant direction of ductal shunting in the early post-atal period is left to right, and, in those ducts that fail toonstrict, large volumes of blood move from the systemic toulmonary circulation. The early left-to-right shunting resultsn consequences such as reduced systemic blood flow andlood pressure, increased ventilatory requirements, and pul-onary hemorrhagic edema.14,22,24 These hemodynamic ef-

ects may paradoxically be more important in the early hoursfter birth rather than later in the clinical course.25 Thesendings lend indirect support to the emerging suggestionsegarding early/prophylactic therapy of the PDA26 and sub-equent tolerance of the PDA in older infants who do notave cardiac failure.6 Functional echocardiography to assess
arly ductal constriction in infants during circulatory transi- c
26 Kluckow, Seri, and Evans

ion allows prediction of the likely closure of the ductusrteriosus and potential targeting of early treatment of theuctus arteriosus (or targeted prophylaxis) rather than non-pecific prophylaxis.2 Functional echocardiography provideshe clinician with further information to aid decision-makingeyond just the presence or absence of the ductus arteriosus.ongitudinal assessment of changes with time allows a judg-ent regarding the likely closure of the ductus arteriosus. In

ddition, measurement of hemodynamic effects such as theirection of diastolic flow in the descending aorta and anstimate of the pulmonary blood flow give an indication ofhe impact of the ductus arteriosus on the cardiovasculartatus.14,23

he Infant With Suspected Circulatory CompromiseOften Hypotensive)

It is a common assumption in neonatology that normallood pressure equates to normal SBF, and improving bloodressure means that blood flow must also have improved.owever, in very preterm infants during the first postnatal

ays, mean blood pressure does not correlate well with theimultaneously measured left ventricular output.3,27 Conse-uently, the ability to repeatedly assess SBF at the bedside isf considerable importance. Hypotension in the neonate cane due to several underlying scenarios with variable hemody-amic physiology.28,29 Very preterm infants may initially have

ow SBF and/or a large PDA. Term infants may have pooryocardial function as the result of asphyxia, pathological

asodilation in septic shock or asphyxia, or, less frequently,ypovolemia with cardiac underfilling caused by fluid or blood

oss. Each of these situations potentially results in low bloodressure. However, the appropriate management and logicalhoice of therapy in each varies.28,30 Echocardiography can besed to differentiate between these situations, combiningeasurement of cardiac output, assessment of cardiac filling,

nd myocardial function and even exclusion of life-threaten-ng pathology, such as a pericardial effusion tamponade fromn extravasation of a central line or from other causes. Finally,lthough detection of congenital heart disease (CHD) is nothe primary goal of functional echocardiography, its useeans that CHD can and will be diagnosed in some babies

efore it would have been suspected clinically.

ssessment and Response to Treatment of an Infantith High Oxygen Requirements

Babies with suspected persistent pulmonary hyperten-ion present the neonatal intensivist with a range of respira-ory and hemodynamic therapeutic challenges. As with theroup of infants with hypotension, functional echocardiogra-hy reveals a broad range of pathology underlying the clinicalresentation.31,32 These babies are often assumed to haveulmonary hypertension with right-to-left ductal shunting. Inact, such assumptions are often erroneous and can lead tonappropriate management. The ductus in such babies usually
onstricts early in the course and may close after the first 24

hpiwiaitdnThN

Ntnemirpmgiwabpoto2e

tmmrpdnetridhi“

itgatchsnstitfrp

scaoeesi

gmckabpise

Tne

F

ours. In addition, there is a complex interplay betweenulmonary and systemic blood flow, with both often measur-ng low.31 Because congenital heart disease can present in thisay, these infants must have this possibility excluded early on

n their course by a pediatric cardiologist. Another importantspect of the use of functional echocardiography in thesenfants is monitoring changes over time and responses toreatment such as vasopressor-inotropes, inotropes, and vaso-ilators. Variability in response to treatments such as inhaleditric oxide probably relates to this hemodynamic spectrum.33

his group of infants also underscores the importance ofaving functional echocardiography readily available in theICU.

POTENTIAL CONCERNS REGARDINGNEONATOLOGISTS DEVELOPING SKILLS IN

FUNCTIONAL ECHOCARDIOGRAPHYThe primary aim of functional echocardiography is

OT to diagnose and manage CHD but to provide frequent,imely, and relevant hemodynamic information in critically illeonates. The main concern regarding the use of functionalchocardiography by the neonatologist is the potential forisdiagnosis—either missing significant CHD or commenc-

ng inappropriate treatment in misdiagnosed CHD. These areeal risks. The development of training and accreditationrocesses for neonatal ultrasound should be aimed at mini-izing these risks. As a minimum, there should be agreed

uidelines for when consultation with a pediatric cardiologists mandatory. Beyond this, the appropriate safety mechanismsill vary according to the health care system and the resources

vailable34 (Table I). One example of a safety mechanism haseen developed by one of the authors at the Children’s Hos-ital Los Angeles and the Women’s and Children’s Hospitalf the LAC � USC Medical Center and the affiliated insti-utions. A review of the echocardiogram by a pediatric cardi-logist to rule out the presence of CHD is requested within4 hours of birth or the initiation of the use of functional

able I. Principles of safe practice foreonatologists using neonatal functionalchocardiography

● Neonatologists performing functional echocardiograms shouldreceive appropriate training and participate in an audit process.

● Functional echocardiography should be used as an adjunct to,not a replacement for, the appropriate clinical assessment ofthe infant.

● A pediatric cardiologist should be consulted if any evidence ofstructural heart disease is found on functionalechocardiography.

● If the primary question based on clinical signs or concerns is“has this baby got structural heart disease?,” this should beaddressed to a pediatric cardiologist.

● A pediatric cardiologist should be consulted before any specifictreatment or inter-hospital transfer for structural heart diseaseis instituted.

chocardiography by a neonatologist.

unctional Echocardiography: An Emerging Clinical Tool for the Neonat

Even in units experienced in the use of neonatal func-ional echocardiography, pediatric cardiologists will often addore detail to the initial screening echocardiographic findingsade by the neonatologist.9,35 Although it is important to

ecognize that CHD can be missed by neonatologists, theresence of a structural cardiac abnormality is almost alwaysetected, even if a full diagnosis is not made.7,10,11 Finally,eonatologists are not the only group of specialists usingchocardiography that can miss CHD in neonates—much ofhe published data on such echocardiographic misdiagnosiselates to adult cardiologists and radiologists.35,36 Anotherssue that has been raised is that the neonate will have beeneemed to have had a formal echocardiogram performed withigh-level exclusion of CHD.37 This can be avoided by stress-

ng to parents and other clinicians that the echocardiogram isfunctional,” or screening in nature.

TRAINING AND ACCREDITATIONAs clinicians realize the value of the hemodynamic

nformation that functional echocardiography can provide inhe neonate, the demand for access to functional echocardio-raphic data will increase. Already in Australia/New Zealandnd many European countries, clinicians who are not formallyrained as cardiologists or radiologists are using point-of-carelinical ultrasound to enhance their understanding of theemodynamic physiology underlying the clinical circum-tances they are managing.7,9 There is therefore a pressingeed to proactively plan for the dissemination of these ultra-ound skills by developing appropriate training and accredi-ation programs to limit the risks of untrained or unsupervisedndividuals using ultrasound in the NICU. In most countries,here are no recognized training and accreditation programsor neonatal ultrasound. This includes not only echocardiog-aphy but also cerebral ultrasound, which has been widelyerformed by neonatologists for many years.38

Neonatologists need to work in collaboration with thepecialties that currently provide these imaging services on aonsultative basis. For echocardiography, this is mainly pedi-tric cardiologists. It is our conviction that it is in the interestf pediatric cardiologists to encourage and support thismerging area to ensure appropriate use and limitations ofchocardiography by neonatologists. There are a number ofuccessful models of training and cooperation already operat-ng in Europe, Australia, and New Zealand.10

The components of a formal structured training pro-ram should include both theoretical teaching and develop-ent of practical skills alongside experienced neonatal echo-

ardiographers and/or interested pediatric cardiologists, anowledge of ultrasound equipment and modalities available,nd reinforcement of limitations with a close interactionetween the neonatologist and pediatric cardiologist. Audit oferformance and ongoing competency assessment are alsomportant by use of logbooks and review of scans on videoupervised by pediatric cardiologist(s) in collaboration withxperienced neonatal echocardiographers.

Issues regarding supervision and monitoring will vary

ologist 127

fdnddtpAshtocm

calStctawclatfuAmb

tsssetbdbfipttocit

icIattcsv

T

T

A

I

C

f

1

rom country to country and even between neonatal units,epending on the availability of skilled staff. There are aumber of ways to maintain quality control with the intro-uction of functional echocardiography. One model beingeveloped in Australia and New Zealand is through localraining and accreditation programs administered by an ex-erienced group of physicians. Other models used in Northmerica allow for the cardiologist to “sign off” on functional

tudies within a certain time period after the echocardiogramas been performed. This model allows careful supervision ofhe introduction of echocardiographic skills to a wider groupf individuals within the neonatal unit. In other units, echo-ardiographic technicians are utilized to provide timely he-odynamic monitoring.

Issues regarding appropriate training in a recognizedenter, certification of competence, and ongoing accreditationre all currently being addressed in Australia and New Zea-and through the main ultrasound body, the Australasianociety of Ultrasound in Medicine (ASUM). This organiza-ion facilitates the provision of focused clinician-centeredourses on point-of-care ultrasound, with ongoing accredita-ion and competence requirements. These courses are aimedt developing specific skills in particular clinician groups forhom ultrasound is an important skill but is not usually

entral to their training. Emergency physicians are able toearn skills of ultrasonic diagnosis of an abdominal aorticneurysm, obstetricians learn to assess women with first-rimester bleeding by ultrasound and perform “screening”etal ultrasound evaluations, and surgeons are taught to useltrasound for diagnosis of the acute abdomen or scrotum.39

ustralian neonatologists are currently investigating develop-ent of a similar model of training, which, if successful, could

able II. Benefits, barriers, and risks of functional ec

Benefi

raining and teaching ● Increased access to fECHO

udit ● Important in maintenance

ntroduction of fECHO to the NICU ● Facilitates documentationfunction in real time

linical use of fECHO in the NICU ● Assessment of need for tr● Assessment of response t● Longitudinal monitoring o● Detection of structural he

ECHO, Functional echocardiography.

e used as a model for other countries as well. p


BARRIERS TO PROVISION OF FUNCTIONALECHOCARDIOGRAPHY WITHIN AN NICU

There are many practical barriers to the development ofhese skills, not the least of which is the access to an ultra-ound machine, and creating opportunities for specialized andupervised training (Table II). These barriers are relativelytraightforward in comparison to the political issues thatmerge around access to ultrasound in the NICU by neona-ologists. It is an uncomfortable truth that most of thesearriers come from the consultative specialties that have tra-itionally provided these services. However, the extent of thearriers put up by the consultative specialties varies widelyrom country to country, and even between different centersn the same country. Each of the authors is indebted to localediatric cardiologists who have supported us and our func-ional echocardiography programs. Without such collabora-ion the introduction of fECHO programs cannot safelyccur. If the process of the dissemination of these skills is toontinue, then it is incumbent on specialist training programsn neonatology to incorporate appropriate accreditation struc-ures for ultrasound.

The best models of collaboration are in health care systemsn which the method of remuneration of the doctors encouragesooperation rather than competition in the provision of services.t is predictable and understandable that where departmentalnd/or personal income streams are threatened by a process,here will be barriers. Neonatologists involved in ultrasound needo be able to collaborate, and it must not be in their interest toompete for income. The authors of this review would notupport neonatologists developing ultrasound skills if the moti-ation were financial rather than clinical. In Australia, one pro-

rdiography

Barriers and risks

● Difficulties associated with recruitingexperienced pediatric cardiologistsinterested in teaching fECHO

● Lack of appropriately trained, experiencedneonatologists

● Lack of structured supervised trainingandards ● Lack of a structured training program in

most countriesmodynamic ● Access to expensive equipment

● Reluctance of some pediatric cardiologiststo participate

● Concerns about changes in income streamsentatmentodynamic changesisease

● Possibility of excessive treatment orintervention as a result of the test

● Availability of staff and machine at the timethey are needed

● Increased handling of the infant● Risk of misdiagnosis

hoca

ts

of st

of he

eatmo tref hemart d

osal being considered is that functional echocardiography


siii

ucdcattecmmEcdpstbt

vcsioifeanhbrtn

TvHoaard(o

1flp

2d13p4G526p27o8i9Z1t11g21NA1vN1v1o1m1cn1N2r2n2s12ta22n2v2LN2aJ2iN2p3m3n1

F

hould not be seen as a separately billed item but rather bencorporated into the current daily “provision of intensive care”tem as are other physiological monitoring techniques such asnvasive blood pressure.

Our experience has shown that clinical point-of-careltrasound complements rather than replaces the need foronsultative ultrasound. The reasons for the neonatologisteveloping echocardiographic skills are very different to theardiologist, and recognition of this allows development andppropriate use of functional echocardiography. The neona-ologist requires timely, longitudinal monitoring of changes inhe hemodynamic physiology of the sick neonate. This is anxtension of the clinical care of the infant rather than aonsultative diagnostic modality. It is our observation thatost pediatric cardiologists do not have the time or clinicalotivation to provide this longitudinal information 24/7.qually, the pediatric cardiologist provides an essential spe-

ialized consultative skill to the neonatologist, focused oniagnosis or exclusion of congenital heart disease and therovision of a confirmation of the hemodynamics, usually at aingle point in time. The significant distinction between thesewo different approaches means that the skills of each shoulde complementary rather than exclusive and can be usedogether to improve the clinical care of neonates.

CONCLUSIONThe area of functional echocardiography is being de-

eloped and driven by neonatologists as an extension of theirlinical skills and to enable them to assess the cardiovasculartatus of their patients. There is a wealth of hemodynamicnformation that can be derived from functional echocardi-graphy in the sick neonate. Often, this new hemodynamicnformation will provide clinical information that is differentrom the assumed underlying physiology.31 For functionalchocardiography to fulfil its clinical potential, it needs to bevailable at any time in the NICU. Because most NICUs doot have external diagnostic services to provide longitudinalemodynamic follow-up at the bedside, neonatologists shoulde able to develop echocardiographic skills in close collabo-ation with their cardiologist colleagues. Finally, the impact ofhe use of functional echocardiography on neonatal outcomeeeds to be assessed.

wo of the authors (NE and MK) are indebted to severalisionary pediatric cardiologists in the UK and at The Children’sospital, Westmead in Sydney, Australia who helped us develop

ur skills during our training and continue to provide support anddvice to our clinical services. The third author (IS) acknowledgesnd appreciates the enormous support he and his faculty haveeceived from the Division of Pediatric Cardiology at the Chil-ren’s Hospital Los Angeles, University of Southern CaliforniaUSC) in setting up the system of fECHO in the USC Divisionf Neonatology.

REFERENCES. Osborn DA, Evans N, Kluckow M. Clinical detection of low upper body blood
ow in very premature infants using blood pressure, capillary refill time, and central-eripheral temperature difference. Arch Dis Child Fetal Neonatal Ed 2004;89:F168-73. 3
unctional Echocardiography: An Emerging Clinical Tool for the Neonat

. Kluckow M, Evans N. Early echocardiographic prediction of symptomatic patentuctus arteriosus in preterm infants undergoing mechanical ventilation. J Pediatr995;127:774-9.. Kluckow M, Evans N. Relationship between blood pressure and cardiac output inreterm infants requiring mechanical ventilation. J Pediatr 1996;129:506-12.. Short BL, Van Meurs K, Evans JR. Summary proceedings from the Cardiologyroup on Cardiovascular Instability in Preterm Infants. Pediatrics 2006;117:34-9.

. Evans N. Which inotrope for which baby? Arch Dis Child Fetal Neonatal Ed006;91:F213-20.. Laughon MM, Simmons MA, Bose CL. Patency of the ductus arteriosus in theremature infant: is it pathologic? Should it be treated? Curr Opin Pediatr004;16:146-51.. Moss S, Kitchiner DJ, Yoxall CW, Subhedar NV. Evaluation of echocardiographyn the neonatal unit. Arch Dis Child Fetal Neonatal Ed 2003;88:F287-9.. Seri I. Hemodynamics during the first two postnatal days and neurodevelopmentn preterm neonates. J Pediatr 2004;145:573-5.. Evans N. Echocardiography on neonatal intensive care units in Australia and Newealand. J Paediatr Child Health 2000;36:169-71.

0. Skinner JR. Echocardiography on the neonatal unit: a job for the neonatologist orhe cardiologist? Arch Dis Child 1998;78:401-2.1. Whitehall J. Echocardiography by a neonatologist. Arch Dis Child 1999;80:580-1.2. Katumba-Lunyenya JL. Neonatal/infant echocardiography by the non-cardiolo-ist: a personal practice, past, present, and future. Arch Dis Child Fetal Neonatal Ed002;86:F55-7.3. Evans N, Malcolm G. Practical echocardiography for the neonatologist. Part 1.ormal 2D imaging and Doppler: An interactive multimedia CD-ROM. Sydney,ustralia: Royal Prince Alfred Hospital; 2006.

4. Evans N, Iyer P. Longitudinal changes in the diameter of the ductus arteriosus inentilated preterm infants: correlation with respiratory outcomes. Arch Dis Child Fetaleonatal Ed 1995;72:F156-61.

5. Evans N, Kluckow M. Early determinants of right and left ventricular output inentilated preterm infants. Arch Dis Child Fetal Neonatal Ed 1996;74:F88-94.6. Kluckow M, Evans N. Superior vena cava flow in preterm infants: a novel markerf systemic blood flow. Arch Dis Child Fetal Neonatal Ed 2000;82:182-7.7. Osborn D, Evans N, Kluckow M. Randomized trial of dobutamine versus dopa-ine in preterm infants with low systemic blood flow. J Pediatr 2002;140:183-91.

8. Noori S, Friedlich P, Wong P, Ebrahimi M, Siassi B, Seri I. Hemodynamichanges following low-dose hydrocortisone administration in vasopressor-treated neo-ates. Pediatrics 2006;118:1456-66.9. Groves AM, Kuschel CA, Skinner JR. The neonatologist as an echocardiographer.eoReviews 2006;7:e391-9.

0. Kluckow M, Evans N. Low superior vena cava flow and intraventricular haemor-hage in preterm infants. Arch Dis Child Fetal Neonatal Ed 2000;82:188-94.1. Hunt RW, Evans N, Rieger I, Kluckow M. Low superior vena cava flow andeurodevelopment at 3 years in very preterm infants. J Pediatr 2004;145:588-92.2. Evans N, Iyer P. Assessment of ductus arteriosus shunt in preterm infantsupported by mechanical ventilation: effects of interatrial shunting. J Pediatr994;125:778-85.3. Keller RL, Tacy TA, Fields S, Ofenstein JP, Aranda JV, Clyman RI. Combinedreatment with a nonselective nitric oxide synthase inhibitor (l-NMMA) and indometh-cin increases ductus constriction in extremely premature newborns. Pediatr Res005;58:1216-21.4. Kluckow M, Evans N. Ductal shunting, high pulmonary blood flow, and pulmo-ary hemorrhage. J Pediatr 2000;137:68-72.5. Evans N, Kluckow M. Early ductal shunting and intraventricular haemorrhage inentilated preterm infants. Arch Dis Child Fetal Neonatal Ed 1996;75:F183-6.6. Schmidt B, Davis P, Moddemann D, Ohlsson A, Roberts RS, Saigal S, et al.ong-term effects of indomethacin prophylaxis in extremely-low-birth-weight infants.Engl J Med 2001;344:1966-72.

7. Pladys P, Wodey E, Beuchee A, Branger B, Betremieux P. Left ventricle outputnd mean arterial blood pressure in preterm infants during the 1st day of life. EurPediatr 1999;158:817-24.8. Evans N, Seri I. Cardiovascular compromise in the newbornnfant. In: Taeusch HW, Ballard RA, Gleason CA, editors. Avery’s Diseases of the

ewborn. Philadelphia, Pa: WB Saunders & Co; 2004. p. 398-409.9. Noori S, Seri I. Pathophysiology of newborn hypotension outside the transitionaleriod. Early Hum Dev 2005;81:399-404.0. Osborn DA. Diagnosis and treatment of preterm transitional circulatory compro-ise. Early Hum Dev 2005;81:413-22.

1. Evans N, Kluckow M, Currie A. Range of echocardiographic findings in termeonates with high oxygen requirements. Arch Dis Child Fetal Neonatal Ed998;78:F105-11.
2. Skinner JR, Hunter S, Hey EN. Haemodynamic features at presentation in
ologist 129

pN3ch3f3i

3b3a3t3

1

ersistent pulmonary hypertension of the newborn and outcome. Arch Dis Child Fetaleonatal Ed 1996;74:F26-32.

3. Roze JC, Storme L, Zupan V, Morville P, Dinh-Xuan AT, Mercier JC. Echo-ardiographic investigation of inhaled nitric oxide in newborn babies with severeypoxaemia. Lancet 1994;344:303-5.4. Althouse LA, Stockman JA III. Pediatric workforce: a look at pediatric cardiology datarom the American Board of Pediatrics. J Pediatr 2006;148:384-5.
5. Ward CJ, Purdie J. Diagnostic accuracy of paediatric echocardiograms interpreted byndividuals other than paediatric cardiologists. J Paediatr Child Health 2001;37:331-6.
fc


6. Dorfman AL, Levine JC, Colan SD, Geva T. Accuracy of echocardiography in lowirth weight infants with congenital heart disease. Pediatrics 2005;115:102-7.7. Sholler G. Echocardiography in congenital heart disease: diagnosis, misdiagnosis,nd ownership. J Paediatr Child Health 2001;37:321-2.8. Levene MI, Fawer CL, Lamont RF. Risk factors in the development of intraven-ricular haemorrhage in the preterm neonate. Arch Dis Child 1982;57:410-7.9. Australasian Society for Ultrasound in Medicine: Certificate in Clinician Per-
ormed Ultrasound. Available from:http://www.asum.com.au/open/ccpu-1.htm. Ac-essed November 2006.
50 Years Ago in The Journal of PediatricsOBSERVATIONS ON THE CLINICAL USE OF v-CILLIN IN PEDIATRIC PRACTICE

Alvarez-Pagan M, Rees DC, Conway GF. J Pediatr 1957;50:39-43

In the 1940s, researchers had developed phenoxymethyl penicillin (penicillin V), an acid-stable form of penicillin thatcould resist inactivation by stomach acids. In view of the obvious advantages of oral administration of penicillin inchildren, the authors studied the use of penicillin V in a series of 84 patients between the ages of 5 months and 12 years.Subjects were included if they were diagnosed with an infection and were ill enough that they would usually be treatedwith a course of parenteral procaine penicillin G. None of the bacterial isolates were analyzed for antibiotic sensitivities.The patients were diagnosed with acute tonsillitis or pharyngitis (n � 53), otitis media (n � 20), or laryngotracheo-bronchitis or bronchopneumonia (n � 11). Depending on clinical features such as fever pattern, patient response wasclassified into 4 categories: excellent, good, satisfactory, and unsatisfactory. Sixty-five patients showed excellent or goodresponse, 12 satisfactory, and 7 unsatisfactory (therapeutic failure). The authors included descriptions of 2 case reportsthought to be typical of the study population: a 19-month-old boy with pneumonia and a 6-year-old boy with acutepharyngitis. Both appeared to respond to the course of penicillin V. The results were stated to be favorable whencompared with results expected from treatment with courses of twice-daily intramuscular penicillin G. The authorsconcluded that the oral suspension of penicillin V was proven to be effective in the treatment of respiratory tractinfections, as well as being well tolerated by children.

Much has changed in the past 50 years. Current studies with similar objectives would usually require more subjects(including suitable control subjects) and appropriate statistical analyses to convince pediatricians of the virtues of theantibiotic in question. More importantly, although penicillin resistance was reported shortly after it became widelyavailable for clinical use in the 1940s, it has become increasingly apparent only over the past 2 decades that antibioticresistance poses a major threat to our ability to combat bacterial infections. In contrast to what would currently beexpected, none of the subjects from whom Staphylococcus aureus was isolated showed unsatisfactory response. Theconvenience of treating most types of infections with the same antibiotic has long gone. Judicious use of antibiotics,sensitivity testing of suspected pathogens, rationalization of therapeutic regimes, and stringent infection control measuresare now important ways to help prevent infection of individuals, as well as reduce the spread of resistant organisms.

Hugh S. Lam, MRCPCHDepartment of Paediatrics

The Chinese University of Hong KongPrince of Wales Hospital, New Territories

Hong Kong10.1016/j.jpeds.2006.09.009


http://www.asum.com.au/open/ccpu-1.htm

TicToa

ti

2(

P

mp4s

yaom

P

ippb

ew

AA

SPECIALARTICLE

Pediatric Workforce: A Look at Pediatric Infectious Diseases Data fromthe American Board of Pediatrics

LINDA A. ALTHOUSE, PHD, AND JAMES A. STOCKMAN III, MD

his report, which is part of a series discussing workforce trends for general pediatrics and related subspecialty areas,highlights the American Board of Pediatrics’ (ABP) workforce data for pediatric infectious diseases. Readers areencouraged to read the initial report1 in the series, because it provides information about general pediatrics and summary

nformation about other ABP subspecialties. In 1994, pediatric infectious diseases became the 12th ABP subboard to offer aertification examination, with the first examination yielding 501 board-certified pediatric infectious diseases subspecialists.oday, approximately 1000 pediatricians have been certified by the ABP as pediatric infectious diseases physicians. The focusf this report is to provide a snapshot of the current ABP workforce data for this subspecialty. The full ABP workforce data arevailable on the ABP Web site at www.abp.org.

METHODSAs described in the initial report, the ABP uses 3 primary methods to collect and maintain data about workforce numbers:

racking of residents and fellows, examination application surveys, and continual maintenance of the ABP master database asndividuals become certified.

Tracking for first-year fellows began in 1995. By 1997-98, all subspecialty fellows in all training levels were tracked. In005, the ABP contacted all accredited pediatric infectious diseases training programs in the United States (n � 61) and Canadan � 8) to obtain tracking information. All programs contacted returned their tracking information, for a 100% response rate.

RESULTS

ediatric Infectious Diseases Fellow TrackingTable I provides the total number of fellows in training since the 1997-98 academic year, with a breakdown by sex and

edical school. The number of fellows enrolled in pediatric infectious diseases has been steadily increasing since 1997. Theercentage of women in pediatric infectious diseases training is currently 58.1%. This percentage has fluctuated from a low of8.1% in 1998 to a high of 61.0% in 2003. The number of American Medical School Graduates (AMG) fellows has increasedince 1997, from 51.6% to 63.7%, with a peak of 63.8% in 2002.

The Figure illustrates the number of fellows in training at each level. Since 1997-98, the average drop rate from trainingear 1 to training year 3 has been 13%. The decline may be attributed to many factors such as personal leave, visa restrictions,nd ABP-approved abbreviated training pathways. In addition, physicians who have completed fellowship training in Canadanly need 2 years of training to be certified by the Royal College of Physicians and Surgeons of Canada. These varying factorsake it difficult to ascertain whether the drop rate is a true reflection of those actually leaving the subspecialty.

ediatric Infectious Diseases Career DataThe ABP has 2 primary opportunities to gather information about career interest

n pediatric infectious diseases: a survey given to all first-time applicants for the generalediatrics certification examination and a survey given to all first-time applicants for theediatric infectious diseases certification examination. This section highlights results fromoth the 2005 general pediatrics and pediatric infectious diseases applications.

Of the 2994 first-time candidates applying for the general pediatrics certificationxamination in 2005, 866 (29%) indicated an interest in 1 of the subspecialty areas inhich the ABP awards or jointly awards certificates. Pediatric infectious diseases was

BP American Board of Pediatrics FOPE II Future of Pediatric Education II

From the American Board of Pediatrics,Chapel Hill, NC.

Correspondence: Linda A. Althouse, PhD,American Board of Pediatrics, 111 SilverCedar Ct, Chapel Hill, NC 27514. E-mail:[email protected].

J Pediatr 2007;150:131-3



MG American Medical School Graduates
10.1016/j.jpeds.2006.11.021
131

http://www.abp.org.

sm

ettfptda

C

stcw

p5(llyt

seytc

nspw1idCitp

fIfs

psbpdw

iidtwddao

T

111222222

Ft

1

elected by 6% of these 866 applicants, making it the eighthost selected pediatric subspecialty.

The infectious diseases certifying examination is givenvery 2 years. In 2005, there were 91 first-time applicants forhe pediatric infectious diseases certification examination. Ofhese applicants, 53.8% were women and 63.7% were AMGellows. Approximately 41.8% plan to practice exclusively inediatric infectious diseases in an academic setting. An addi-ional 2.2% plan to practice exclusively in pediatric infectiousiseases, but in a private practice or combined private practicend academic setting.

ertified DiplomatesAs a pediatric subspecialty, infectious diseases is the

eventh largest ABP discipline, with approximately 1000 cer-ified practitioners (as of 12/31/2005). The mean age ofertified pediatric infectious diseases physicians is 49.5 years,ith approximately 96% ranging from 31 to 65 years of age.

The ratio of current ABP-certified infectious diseaseshysicians-to-children younger than 18 years in each of the0 states and the District of Columbia is shown in Table IIavailable at www.jpeds.com). The population of childrenisted in Table II is based on the US Census Bureau Popu-ation Estimates and includes all children younger than 18ears.2 These numbers are based on a list of pediatric infec-

able I. Number of pediatric infectious diseases fello

Year Total Female

997–1998 125 51.2%998–1999 131 48.1%999–2000 138 50.7%000–2001 161 48.4%001–2002 158 55.7%002–2003 163 56.4%003–2004 164 61.0%004–2005 177 59.3%005–2006 179 58.1%

0

10

20

30

40

50

60

70

1997 1998 1999 2000 2001 2002 2003 2004 2005

Training Year

# o

f F

ello

ws

Training Level1

Training Level 2

Training Level 3

igure. Number of pediatric infectious diseases fellows in training at eachraining level.

ious diseases physicians with known addresses in 1 of the 50

32 Althouse and Stockman

tates or the District of Columbia. Pediatric infectious dis-ases physicians older than the average retirement age of 65ears were excluded. On the basis of these adjustments, theotal number of certified pediatric infectious diseases physi-ians categorized in Table II is 887.

Only 3 states (Alaska, Montana, and South Dakota) doot have a practicing ABP-certified infectious diseases phy-ician. Twenty-eight states have a pediatric infectious diseaseshysician-to-child ratio of at least 1 per 100,000 children,ith the District of Columbia having the largest ratio (6.4 per00,000), followed by Maryland (4.3 per 100,000). The 61nfectious diseases training programs in the United States areistributed across 27 of the 50 states and the District ofolumbia, as noted by the asterisk in Table II. The number

n parentheses denotes the number of training programs inhe state that were tracked during the 2005-06 trackingeriod.

DISCUSSIONAlthough many studies have projected physician work-

orce needs, it was not until the Future of Pediatric EducationI (FOPE II) task force report that a recent and detailed studyocused exclusively on pediatrics, both at the generalist andubspecialty level.3,4

In 2005, the number of pediatric infectious diseaseshysicians in training (training years 1-3) remained fairlytable from the previous year (an increase of only 2 fellows),ut it has more than doubled since 1997. The growing pro-ortion of women selecting pediatric infectious diseases as aiscipline supports the claim of increased involvement ofomen in pediatric subspecialties.5

Although the data in Table II provide the pediatricnfectious diseases physician-to-child ratio, the data do notndicate who is working full-time or part-time. General pe-iatrics research has shown an increasing trend toward part-ime work, particularly with the increase in the number ofomen entering pediatrics.5,6 However, there are no currentata to indicate that this is the case in pediatric infectiousiseases. Studies have reported that women in subspecialtiesre equally likely to work full time and treat an equal numberf patients as their male colleagues.6,7

n training since 1997

Male AMG IMG

48.8% 51.6% 48.4%51.9% 48.9% 51.1%49.3% 50.0% 50.0%51.6% 57.1% 42.9%44.3% 63.3% 36.7%43.6% 63.8% 36.2%39.0% 62.2% 37.8%40.7% 60.5% 39.5%41.9% 63.7% 36.3%

ws i

Although it is important to have an adequate number of



pddAFd

eorcpcn

tpoisn

wBiistc

1t2AA3n24v5a6i7s8

P

hysicians, where these physicians practice is just as critical inetermining whether appropriate care is available to all chil-ren. As aforementioned, currently 3 states do not have anBP-certified infectious diseases physician. In addition, theOPE II survey results indicate that only 4% of infectiousisease physicians practice in rural areas.3,4

Also contributing to a growing need for infectious dis-ases physicians, the FOPE II survey results indicate that 25%f infectious diseases physicians believe that the volume ofeferrals has increased and 39% also believe that the referralomplexity has increased. However, approximately 65% ofediatric infectious diseases physicians anticipate that theirommunities will not need additional subspecialists in theext 3 to 5 years.3,4

As Stoddard et al note, the FOPE II study provideshe supply-side perspective.4 The ABP data in this reportrovide the same perspective. These data are useful notnly to those studying workforce trends, but also to med-cal students and pediatric residents making career deci-ions. However, these data do not address or gauge theeed for medical services.

Although workforce studies are not new, attention to I

ediatric Workforce: A Look at Pediatric Infectious Diseases Data from

orkforce issues for pediatric subspecialties is relatively new.efore this study, the last large-scale workforce study specif-

cally for pediatric infectious diseases was in 1995.8 It ismportant that workforce research continues, from both theupply and demand perspective. Only then can we be sure thathe goal of providing all children with access to high-qualityare be met.

REFERENCES. Althouse LA, Stockman JA. Pediatric workforce: a look at general pediatrics data fromhe American Board of Pediatrics. J Pediatr 2006;148:166-9.. US Bureau of the Census. Population estimates by state. Revised July 1, 2004.vailable at http://www.census.gov/popest/states/asrh/SC-est2004-02.html. Accessedpril 18, 2005.

. The Future of Pediatric Education II: organizing pediatric education to meet theeeds of infants, children, adolescents, and young adults in the 21st century. Pediatrics000;105:163-212.. Stoddard JJ, Cull WL, Jewett EA, Brotherton SE, Mulvey HJ, Alden ER. Pro-iding pediatric subspecialty care: a workforce analysis. Pediatrics 2000;106:1325-33.. Freed GL, Nahra TA, Wheeler JR. Predicting the pediatric workforce: use of trendnalysis. J Pediatr 2003;143:570-5.. Brotheron SE, Mulvey HJ, O’Conner KG. Women in pediatric practice: trends andmplications. Pediatr Ann 1999;28:177-83.. Mayer ML, Preisser JS. The changing composition of the pediatric medical sub-pecialty workforce. Pediatrics 2005;116:833-40.. Feigin RD. The future of pediatric infectious diseases: manpower issues. Pediatr
nfect Dis J 1995;14:1023-5.
the American Board of Pediatrics 133

http://www.census.gov/popest/states/asrh/SC-est2004-02.html

T

AAAACCCDDFGHIIIIKKLMMMMMMMMNNNNNNNNOOOPRSSTTUVVWWWW

*a

1

able II. Number of American Board of Pediatrics-c

State

Number of ABP diplomin pediatrics infectiou

diseases

labama* (1) 13laska 0rizona 6rkansas* (1) 4alifornia* (8) 100olorado* (1) 16onnecticut* (1) 16elaware 4istrict of Columbia* (1) 7

lorida* (2) 41eorgia* (1) 30awaii 3

daho 1llinois* (2) 26ndiana 12owa 1ansas 2entucky* (1) 7ouisiana* (1) 15aine 3aryland* (3) 60assachusetts* (2) 41ichigan* (2) 20innesota* (2) 16ississippi 5issouri* (2) 18ontana 0ebraska* (1) 5evada 2ew Hampshire 2ew Jersey* (1) 36ew Mexico 4ew York* (9) 93orth Carolina* (1) 25orth Dakota 1hio* (4) 33klahoma 5regon 8

ennsylvania* (3) 36hode Island* (1) 4outh Carolina 7outh Dakota 0ennessee* (2) 27exas* (4) 76tah* (1) 1ermont 2irginia* (2) 23ashington* (1) 19est Virginia 5isconsin 5yoming 1

887

Note: States with an asterisk denote those with a pediatric infectious disease training procademic year.

ertified pediatric infectious diseases diplomates by state

atess Child

populationPhysician to child ratio(per 100,000 children)

1,094,533 1.2188,229 0

1,547,260 0.4676,550 0.6

9,596,463 11,178,889 1.4

838,788 1.9193,506 2.1109,547 6.4

4,003,290 12,332,567 1.3

298,693 1372,411 0.3

3,238,150 0.81,600,295 0.7

680,437 0.1683,491 0.3980,187 0.7

1,164,961 1.3282,129 1.1

1,394,808 4.31,464,189 2.82,533,439 0.81,240,280 1.3

749,569 0.71,384,542 1.3

208,093 0434,566 1.2603,596 0.3304,994 0.7

2,156,059 1.7492,287 0.8

4,572,363 22,118,492 1.2

138,955 0.72,779,212 1.2

859,870 0.6852,357 0.9

2,837,009 1.3243,813 1.6

1,024,700 0.7190,874 0

1,391,289 1.96,266,779 1.2

740,114 0.1134,894 1.5

1,804,900 1.31,486,020 1.3

384,641 1.31,307,986 0.4

116,932 0.9

73,277,998 1.2

gram. The number in parentheses indicates the number of programs tracked in the 2005–2006

33.e1 Althouse and Stockman The Journal of Pediatrics • February 2007

Od

SiC

R[as

Clc

CbfwmSh

tovccrts(c

p(t

A

1

Diagnosis and Localization of Focal Congenital Hyperinsulinism by18F-Fluorodopa PET Scan

OLGA T. HARDY, MD, PHD, MIGUEL HERNANDEZ-PAMPALONI, MD, PHD, JANET R. SAFFER, PHD, MARIKO SUCHI, MD, PHD,EDUARDO RUCHELLI, MD, HONGMING ZHUANG, MD, PHD, ARUPA GANGULY, PHD, RICHARD FREIFELDER, PHD, N. SCOTT ADZICK, MD,

ABASS ALAVI, MD, AND CHARLES A. STANLEY, MD

bjectives To assess the accuracy of 18F-fluoro-L-dihydroxyphenylalanine ([18F]-DOPA) PET scans to diagnose focal versusiffuse disease and to localize focal lesions in infants with congenital hyperinsulinism.

tudy design Twenty-four infants with hyperinsulinism unresponsive to medical therapy were studied. Patients werenjected intravenously with [18F]-DOPA, and PET scans were obtained for 1 hour. Images were coregistered with abdominalT scans.

esults The diagnosis of focal or diffuse hyperinsulinism was correct in 23 of the 24 cases (96%) and equivocal in 1 case.18F]-DOPA PET identified focal areas of high uptake of radiopharmaceutical in 11 patients. Pathology results confirmed thatll 11 had focal adenomatosis, and the locations of these lesions matched the areas of increased [18F]-DOPA uptake on the PETcans in all of the cases.

onclusions [18F]-DOPA PET scans were 96% accurate in diagnosing focal or diffuse disease and 100% accurate inocalizing the focal lesion. These results suggest that [18F]-DOPA PET imaging should be considered in all infants withongenital hyperinsulinism who need to have pancreatectomy. (J Pediatr 2007;150:140-5)

ongenital hyperinsulinism, the most common cause of persistent hypoglycemia in infants and children, is most oftenassociated with recessive mutations of the �-cell ATP-sensitive potassium (KATP) channel. The channel is encoded bytwo adjacent genes on chromosome 11p15.1, SUR1 and Kir6.2. In cases of diffuse disease, patients have mutations of

oth alleles encoding the KATP channel resulting in dysregulation of insulin secretionrom all �-cells.1,2 Recessive KATP mutations may also cause focal hyperinsulinism inhich there is an area of �-cell adenomatosis caused by loss of heterozygosity for theaternal 11p region and expression of a paternally derived KATP channel mutation.3,4

urgical intervention is often necessary to control hypoglycemia in both forms of KATPyperinsulinism but is only curative in the cases of focal disease.

Surgical treatment of infants with congenital hyperinsulinism depends on being ableo distinguish between focal and diffuse disease and to locate focal lesions. Functional testsf insulin responses to secretagogues before surgery are unable to reliably distinguish focalersus diffuse disease, in part because some disease-causing mutations of the KATPhannel retain partial function.5-7 Focal lesions are rarely identifiable at surgery andannot be detected by using conventional imaging techniques such as computed tomog-aphy, magnetic resonance imaging, transabdominal or intraoperative ultrasound, or byhe use of radiolabeled octreotide scans.8 Interventional radiologic techniques such aselective pancreatic arterial calcium stimulation with hepatic vein insulin samplingASVS) and transhepatic portal venous insulin sampling (THPVS) are invasive, techni-ally difficult, and are also not reliable in either diagnosing or localizing focal lesions.3,7

Neuroendocrine cells have an affinity for taking up and decarboxylating amino acidrecursors.9,10 Consequently, amino acid precursors such as L-dihydroxyphenylalanine

L-DOPA) may be taken up by these cells and become decarboxylated to dopaminehrough the action of aromatic amino acid decarboxylase (AADC).11-13 Neuroendocrine

SVS Selective pancreatic arterial calciumstimulation with hepatic vein insulin

[18F]-DOPA 18F-fluoro-L-dihydroxyphenylalanineTHPVS Transhepatic portal venous insulin sampling

See editorial, p 122

From the Division of Endocrinology, De-partment of Radiology, Department of Pa-thology, and Department of Surgery, TheChildren’s Hospital of Philadelphia, Philadel-phia, Pennsylvania; and the Department ofRadiology and Department of Genetics,University of Pennsylvania School of Medi-cine, Philadelphia, Pennsylvania.

Supported in part by grants from the Na-tional Institutes of Health: RO1-DK-56268(C.A.S.) and MO1-RR-00240. O.T.H. wassupported by National Institutes of Healthtraining grant T32-DK63688 (C.A.S.).

Submitted for publication Feb 8, 2006; lastrevision received Jul 5, 2006; accepted Aug12, 2006.

Reprint requests: Charles A. Stanley, MD,Division of Endocrinology, The Children’sHospital of Philadelphia, 34th Street andCivic Center Boulevard, Philadelphia, PA19104. E-mail: [email protected].



sampling

40

10.1016/j.jpeds.2006.08.028

tsp

tcpcadcsfcsl

S

tPflphfoaptaepMawri

SS

wlisDplnFao

wt

I

vD(stidmPpacsctzdphkmilcPe6atfditPfiwattb

ACPt

I

b

D

umors such as carcinoids and pheochromocytomas have beenuccessfully localized with the use of 18F-fluoro-L-dihydroxy-henylalanine ([18F]-DOPA) PET.14,15

A preliminary report from Otonkoski et al16 suggestedhat [18F]-DOPA PET scans could image focal lesions inongenital hyperinsulinism. Two small series of cases sup-orting this suggestion have been recently reported.17-19 Theurrent study was undertaken to test prospectively both theccuracy of [18F]-DOPA PET scans in diagnosing focalisease and its ability to localize focal lesions in children withongenital hyperinsulinism. The study was designed to have aample size of 50 cases, of which half were expected to beocal. This preliminary report provides analysis of the first 24ases and suggests that the accuracy of [18F]-DOPA PETcanning is better than anticipated for both diagnosis andocalization of focal hyperinsulinism.

METHODS

ubjectsAll of the patients included in this study were referred

o the Hyperinsulinism Center at the Children’s Hospital ofhiladelphia between December 2004 and November 2005

or surgical treatment of medically uncontrollable hyperinsu-inism. Patients not requiring surgery and patients who hadrevious pancreatectomies were excluded. The diagnosis ofyperinsulinism was based on previously described criteria:asting hypoglycemia accompanied by inadequate suppressionf plasma insulin, inappropriately low plasma free fatty acidnd plasma �-hydroxybutyrate concentrations, and an inap-ropriate glycemic response to glucagon injection.20,21 Two ofhe infants underwent ASVS testing before surgery. Mutationnalysis of KATP genes was performed by sequencing codingxons and flanking intronic regions of genomic DNA fromeripheral blood leukocytes (Athena Diagnostics, Worcester,A). In all cases, results of mutation analyses were not

vailable before surgery. Loss of heterozygosity in focal lesionsas identified by haplotype analysis through the use of mic-

osatellite markers and by absence of p57KIP2 immunostainingn paraffin-embedded surgical specimens.3,4,22,23

urgical Technique and Intraoperative Frozenection Evaluation

All of the infants underwent pancreatectomy between 2eeks and 18 months of age. PET scans were performed at

east 12 hours before surgery (6 half-lives). The results of thenterpretation of the PET scan were made available to theurgeon to help in the identification of potential focal lesions.uring surgery, biopsy specimens from three areas of the

ancreas were obtained and examined for �-cells with en-arged nuclei suggestive of diffuse disease. The absence ofuclear enlargement indicated the presence of a focal lesion.indings from the PET scan were compared with the findingst pathology, based on permanent sections, as described by
thers.24-25 The determination of focal versus diffuse disease n
iagnosis and Localization of Focal Congenital Hyperinsulinism by 18F-Fl

as made on the permanent histologic sections by two pa-hologists who were masked to the results of the PET scan.

maging Technique[18F]-DOPA PET scans were performed at the Uni-

ersity of Pennsylvania PET Imaging Facility. The [18F]-OPA was administered under an Investigational New Drug

#48923) reviewed by the Radiation Safety Committee, In-titutional Committees, and the Food and Drug Administra-ion. The Investigational New Drug was modified to allownclusion of children, and the Food and Drug Administrationosimetry studies are part of that application. The isotope wasanufactured by the Cyclotron Facility at the University ofennsylvania on the day of the test, using the procedurereviously published.26 Typical yields were 10 to 15 mCi for500 micro-amp-minute irradiation. Specific activity of the

ompound is approximately 500 to 1000 Ci/mol at end ofynthesis. Patient injection occurred about 45 minutes afterompletion of the synthesis. Medications that could poten-ially interfere with pancreatic �-cell function, such as dia-oxide, octreotide, and glucagon, were discontinued 5 days, 2ays, and 12 hours, respectively, before the procedure. Allatients received intravenous dextrose infusion for control ofypoglycemia and to enhance elimination of isotope from theidneys. Plasma glucose was monitored before and every 60inutes during the procedure. Rates of intravenous glucose

nfusion were adjusted as needed to maintain plasma glucoseevels greater than 70 mg/dL. In one case, the patient wasatheterized to facilitate bladder emptying. For obtainingET images, patients were intubated and sedated with gen-ral anesthesia. Patients were injected intravenously with 3 toMBq/kg (0.08 to 0.16 mCi/kg) of [18F]-DOPA, which is

pproximately one-tenth of the total adult dose. After injec-ion, five or six consecutive 10-minute-long scans were per-ormed. The images were acquired through the use of aedicated brain PET camera based on Anger-logic gadolin-um oxyorthosilicate (GSO) detectors, designed and built byhe Physics and Instrumentation Group at the University ofennsylvania.27 This instrument has both axial and transverseelds of view of 25 cm and is therefore suited for imaging thehole body in infants. An abdominal CT was obtained sep-

rately to define the anatomy of the pancreas and adjacentissue before surgery. The CT image was coregistered withhe PET scan to assist in defining the location of focal lesionsy using standard software that is available for this purpose.

The study protocol was approved by the Food and Drugdministration and the institutional review boards of thehildren’s Hospital of Philadelphia and the University ofennsylvania. Written informed consent was obtained from

he parents of the patients.

mage InterpretationThe image set for each patient was visually interpreted

y one of the investigators at the completion of the exami-
ation. Images were reviewed in all three planes as well as
uorodopa PET Scan 141

mpaetqrgva

S

PwwStoi9

fstpp

brstip

ormbec

rd

wmtptrwcD

wpptstp

otipcmt

t1sdidlliota

Kdwedoscsa

cp

Thb

NSLAAA

1

aximum intensity projection views, and the presence and theattern of uptake in the pancreas were taken into consider-tion in generating final reports for these scans. The imagexamination was considered positive for focal disease whenhe uptake of the radiotracer in part of the pancreas wasualitatively higher when compared with the uptake in theemaining pancreatic tissue and other surrounding back-round organs. In contrast, when the entire pancreas wasisualized with nearly uniform [18F]-DOPA uptake, the ex-mination was considered to represent diffuse disease.

tatistical AnalysisBased on previous reports on ASVS and THPVS from

aris and Philadelphia, we assumed that 50% of surgical casesould have focal disease and that at least 70% of focal lesionsould be correctly localized by [18F]-DOPA PET test.16,19

ample size estimates based on these assumptions indicatedhat 52 total cases (26 focal cases) would provide an estimatef test accuracy between 52% and 88% (95% confidencenterval, CI). The efficient-score method was used to calculate5% CIs for proportions.28

RESULTSTable I summarizes the clinical features of the 24 in-

ants with medically unresponsive hyperinsulinism who weretudied with [18F]-DOPA PET scans before surgery. Most ofhe infants in both groups were large for gestational age,resented with hypoglycemia at birth, and were referred forancreatectomy within a few months after birth.

ASVS testing was done in patient 13 and patient 15efore surgery. In both, ASVS showed a step-up in insulinelease after calcium infusion in the gastroduodenal artery,uggesting a focal lesion in the head of the pancreas. Al-hough the ASVS test was correct in diagnosing focal diseasen both cases, it was correct in localizing the lesion only inatient 13 (Table II).

After intravenous injection of [18F]-DOPA, the uptakef isotope into pancreas and other tissues, such as liver, wasapid and then remained essentially constant from 10 to 60inutes. There was also uptake early in the kidneys and

ladder that decreased with time as radiopharmaceutical wasxcreted in the urine. Drainage of urine through a Foley

able I. Clinical features of infants withyperinsulinism studied with [18F]-DOPA PET scanefore surgery [median (range)]

Characteristic Diffuse disease Focal disease

o. 12 12ex, M/F 6/6 8/4arge for gestational age 83% 83%ge at first symptoms (d) 1 (1-4) 1 (1-90)ge at PET (wk) 5 (2-72) 7 (2-56)SVS performed 0 2

atheter helped decrease the intensity of uptake in the bladder c

42 Hardy et al

egion in one patient. No patients became hypoglycemicuring the PET scan.

Figures 1 and 2 show illustrative PET scans of casesith diffuse and focal hyperinsulinism. Figure 1A shows theaximum intensity projection image from a diffuse case (pa-

ient 4) with uniform [18F]-DOPA uptake throughout theancreas. The intensity of uptake in the pancreas was greaterhan in the liver and surrounding tissue but less than in theenal calyces. Figure 1 (B through D) shows the same caseith the abdominal CT scan and PET scan separate and

oregistered, which confirms that the diffuse uptake of [18F]-OPA was in the pancreas.

Figure 2 shows a focal case (patient 23) in which thereas a discrete area of [18F]-DOPA uptake in the head of theancreas. There was also uptake in the neighboring normalancreatic tissue, but it was less intense than in the head ofhe pancreas (Figure 2A). Coregistration of the PET and CTcans from patient 23 (Figure 2, B through D) demonstrateshat the location of the focal uptake was in the head of theancreas.

Diffuse uptake of [18F]-DOPA was observed through-ut the pancreas of patient 17, with areas of intense uptake inhe head, body, and tail of the pancreas. This scan wasnterpreted as being either consistent with diffuse disease or,ossibly, an extensive focal lesion. The latter possibility wasonfirmed at surgery, where an extensive area of islet adeno-atosis occupied 80% to 90% of the pancreas and extended

hroughout the body, tail, and most of the head of the gland.Table II summarizes the results of the PET scans and

he histologic diagnoses in the 24 cases studied. In all of the2 diffuse cases, the preoperative interpretation of the PETcan was the same as the histologic diagnosis of diffuseisease. In 11 of the 12 cases of focal disease, focal uptake ofsotope on PET correctly diagnosed the presence of focalisease. As noted above, the one case with an extensive focal

esion was suggested but not definitively diagnosed as a focalesion by PET. However, in all of the 12 cases of focal disease,ncluding the case with focal adenomatosis over 80% to 90%f the gland, the PET scan correctly indicated the location ofhe lesions. There were no side effects observed that could bettributed to the radiopharmaceutical.

As shown in Table II, mutation analysis of the twoATP genes identified two mutations consistent with diffuseisease in 7 of the 11 diffuse cases in whom mutation analysisas completed. In two other diffuse cases, only one of the

xpected two mutations could be identified. In two of theiffuse cases, no mutation could be found in either maternalr paternal allele. Among the 12 focal cases, mutation analysishowed a paternal-only mutation in 11 cases. Immunohisto-hemical staining for p57KIP2 and haplotype analysis demon-trated loss of heterozygosity for the maternal 11p region inll of the 12 focal cases.

In addition to the 24 cases who fit the study inclusionriteria, [18F]-DOPA PET scans were done in 3 additionalatients. In one of these patients who did not fit inclusion
riteria, medical treatment with octreotide was sufficiently

eeoflwaftsufffjtbditt

P

T

D123456789111F111111122222

N

Fudgpppcoregistration. H indicates head of pancreas; T, tail of pancreas.

D

ffective so that surgery was not considered mandatory. How-ver, a focal lesion was suspected on the basis of the presencef a single paternal KATP mutation. The PET scan showedocal uptake in the head of the pancreas suggesting a focalesion; surgery was elected that revealed a focal lesion, whichas excised. In a second patient, two prior pancreatectomies

t another institution had removed 98% of the pancreas butailed to improve hypoglycemia. A focal lesion was detected inhe head of the pancreas but not completely excised in theecond surgery. [18F]-DOPA PET scan showed areas of focalptake not only in the remaining head of the gland but also inour ectopic extrapancreatic areas. At surgery, the residualocal lesion in the head of the pancreas was resected, as wereour ectopic focal adenomatosis lesions in the wall of theejunum. (Details of this case will be reported separately.) Inhe third case, the parents declined surgical treatment on theasis of the interpretation of the [18F]-DOPA PET scan asiffuse disease. Although these three patients could not be

ncluded in our study of the [18F]-DOPA PET scan accuracy,he test did correctly localize the focal lesions in two of thehree patients.

DISCUSSIONThese preliminary results suggest that [18F]-DOPA

able II. Results of preoperative [18F]-DOPA PET sc

Patient No. [18F]-DOPA PET diagnosis Histologic dia

iffuse diseaseDiffuse DiffuseDiffuse DiffuseDiffuse DiffuseDiffuse DiffuseDiffuse DiffuseDiffuse DiffuseDiffuse DiffuseDiffuse DiffuseDiffuse Diffuse

0 Diffuse Diffuse1 Diffuse Diffuse2 Diffuse Diffuseocal disease3 Head Head and ne4 Head Head5 Body Body6 Head Head7 Extensive lesion Extensive les8 Tail Tail9 Tail Tail0 Head Head1 Tail Tail2 Head Neck3 Head Head4 Head Head

D, not done; LOH, loss of heterozygosity.

ans in 24 infants with congenital hyperinsulinism

gnosis

Kir6.2 or SUR1 mutations LOH by haplotyping andimmunostainingPaternal Maternal

none noneg3992-9a G173Rnone none31aa insertion g1630�1tdel20aa noneR837X 15aa insertionQ219X noneR620C E501LQ444H Q444HND NDg3992-9a 1388delFR999X R999X

ck E282K none LOHD310V none LOHg1630�1t none LOHg2041�21a none LOH

ion R1461C none LOHD1472N none LOHE490X none LOHg3992�9a none LOHG954X none LOHND ND LOHdel6aa none LOHt1176�2c none LOH

ET is accurate in both the diagnosis of focal or diffuse


igure 1. [18F]-DOPA PET of patient 4 with diffuse disease. A, Diffuseptake of [18F]-DOPA is visualized throughout the pancreas on thisepth-weighted maximum intensity projection sagittal image. Intensity isreater than that observed in the liver and surrounding tissue. Also notehysiologic distribution in the kidneys. Transverse views show B, normalancreatic tissue on abdominal CT; C, diffuse uptake of [18F]-DOPA inancreas; and D, confirmation of pancreatic uptake of [18F]-DOPA with


ha(P9

iscuthd[t[atfiDtc

cOwnfiwdt

wbca

tdl1cfPrTl

lwmcEiwbpeeDmid

[ilts8Dit

WdoDtHOMt

1l2

FapTnppp

1

yperinsulinism and the localization of the focal lesion. Di-gnosis was correct in 23 of 24 cases and equivocal in 1 case96%; CI, 77% to 98%). Of the 12 focal lesions, [18F]-DOPAET localized the focal lesion in all of the cases, providing a5% CI between 70% and 100%.

In assessing the accuracy of this new technique, it ismportant to differentiate diagnosis of the form of hyperin-ulinism from localization of the focal lesion. There may belinical situations in which surgery would not be considerednless the patient has focal disease. In such cases, it is crucialo know the accuracy of the PET scan in diagnosing focalyperinsulinism so that patients with focal disease can un-ergo curative surgery. In the current series, the sensitivity of18F]-DOPA PET in diagnosing focal disease was 92% andhe specificity was 100%. The positive predictive value of18F]-DOPA in diagnosing focal disease was 100% (11/11)nd the negative predictive value was 92% (12/13). Localiza-ion is a separate issue because successful surgery depends onnding and removing the focal lesion. For example, if [18F]-OPA PET definitively identifies the lesion in the head of

he pancreas, the surgeon can perform a proximal pancreate-tomy with Roux-en-Y and preserve the tail of the pancreas.

The accuracy of [18F]-DOPA PET in our series isonsistent with data reported in two smaller series of cases.tonkoski and colleagues17 reported that [18F]-DOPA PETas able to correctly diagnose focal versus diffuse disease inine patients and accurately localized the lesion in all of theirve focal cases. Ribeiro et al19 showed that in nine patientsho had surgical treatment, [18F]-DOPA PET accuratelyiagnosed focal versus diffuse disease and correctly localized

igure 2. [18F]-DOPA PET of patient 23 with focal disease. A, Discreterea of increased [18F]-DOPA uptake is visualized in the head of theancreas on this depth-weighted maximum intensity projection image.he intensity of this area is greater than that observed in the liver andeighboring normal pancreatic tissue. Transverse views show B, normalancreatic tissue on abdominal CT; C, focal uptake of [18F]-DOPA inancreatic head; and D, confirmation of [18F]-DOPA uptake in theancreatic head with coregistration.

he lesion in all five focal cases. Both series included cases in13

44 Hardy et al

hich the [18F]-DOPA PET scan suggested diffuse disease,ut surgery to confirm the diagnosis was not done. This is inontrast to our study, in which histopathologic diagnosis wasvailable for all of the 24 cases.

The current data suggest that [18F]-DOPA PET hashe potential to be more accurate than ASVS or THPVS iniagnosing focal versus diffuse disease and localizing the focal

esion. ASVS correctly diagnosed diffuse disease in only 4 of3 cases and localized the focal lesion in only 73% of 33ases.7 Of 45 cases of focal disease, THPVS localized theocal lesion correctly in only 75%.3 In addition, even thoughET scans require general anesthesia and a minimal dose of

adioactivity, the method is much less invasive than ASVS orHPVS and does not require exposure to hypoglycemia or

arge volumes of blood.The results of mutation analysis and testing of focal

esions for loss of heterozygosity in Table II are consistentith the concept that diffuse disease is caused by recessiveutations of the KATP channel and that focal disease is

aused by expression of a paternally derived channel mutation.vidence for loss of heterozygosity for 11p was demonstrated

n all of the focal cases. Mutations in Kir6.2 or SUR1 genesere found in both focal and diffuse cases. However, as haseen noted in previous series, sequencing of coding regions oferipheral blood genomic DNA failed to uncover all of thexpected mutations.4,7 Mutations could have been present butscaped detection if, for example, they were in regions ofNA not sequenced. Because of the possibility of undetectedutations, DNA sequencing can be informative for diagnos-

ng diffuse disease (mutant maternal allele) but not focalisease.

On the basis of the preliminary results of our study, the18F]-DOPA PET scan was 96% accurate (CI, 70% to 100%)n diagnosis of focal hyperinsulinism and 100% accurate inocalizing the focal lesion. When combined with data fromhe two other recent reports, the overall accuracy of PETcans has been 100% in localizing focal lesions (21/21; CI,1% to 100%). These encouraging results suggest that [18F]-OPA PET imaging should be strongly considered in all

nfants with congenital hyperinsulinism who fail medicalherapy and need to have pancreatectomy.

e are grateful for the help of the Children’s Hospital of Phila-elphia Department of Anesthesiology, the staff at the Universityf Pennsylvania PET Imaging, and Cyclotron Facilities and theepartment of Radiology Nuclear Medicine Section (Nancy Win-

ering). We are also indebted to the expert nursing team of theyperinsulinism Center (Laura Wanner, Lori Halaby, Susan’Rourke, Marie Smolenski) and to Nkecha Hughes, Andreaatter, and Courtney MacMullen for their expert assistance with

hese studies.

REFERENCES. Glaser B, Thornton P, Otonkoski T, Junien C. Genetics of neonatal hyperinsu-inism. Arch Dis Child Fetal Neonatal Ed 2000;82:F79-86.. Stanley CA. Hyperinsulinism in infants and children. Pediatr Clin North Am
997;44:363-74.. de Lonlay-Debeney P, Poggi-Travert F, Fournet JC, Sempoux C, Vici CD,

B14e1p5Ds6GrJ7ebs8ap9W1i1pn1y11c11Wo1PR

1P1EH1ep1a(d1aL2h2a2sl2p22iN2Hl2r62ad

D

runelle F, et al. Clinical features of 52 neonates with hyperinsulinism. N Engl J Med999;340:1169-75.. Verkarre V, Fournet JC, de Lonlay P, Gross-Morand MS, Devillers M, Rahier J,t al. Paternal mutation of the sulfonylurea receptor (SUR1) gene and maternal loss of1p15 imprinted genes lead to persistent hyperinsulinism in focal adenomatous hyper-lasia. J Clin Invest 1998;102:1286-91.. Grimberg A, Ferry RJ Jr, Kelly A, Koo-McCoy S, Polonsky K, Glaser B, et al.ysregulation of insulin secretion in children with congenital hyperinsulinism due to

ulfonylurea receptor mutations. Diabetes 2001;50:322-8.. Henwood MJ, Kelly A, Macmullen C, Bhatia P, Ganguly A, Thornton PS, et al.enotype-phenotype correlations in children with congenital hyperinsulinism due to

ecessive mutations of the adenosine triphosphate-sensitive potassium channel genes.Clin Endocrinol Metab 2005;90:789-94.. Stanley CA, Thornton PS, Ganguly A, MacMullen C, Underwood P, Bhatia P,t al. Preoperative evaluation of infants with focal or diffuse congenital hyperinsulinismy intravenous acute insulin response tests and selective pancreatic arterial calciumtimulation. J Clin Endocrinol Metab 2004;89:288-96.. Adzick NS, Thornton PS, Stanley CA, Kaye RD, Ruchelli E. A multidisciplinarypproach to the focal form of congenital hyperinsulinism leads to successful treatment byartial pancreatectomy. J Pediatr Surg 2004;39:270-5.. Sandler MP. Diagnostic Nuclear Medicine. 4th edition. Philadelphia: Lippincott

illiams & Wilkins; 2003.0. Sundin A, Eriksson B, Bergstrom M, Langstrom B, Oberg K, Orlefors H. PETn the diagnosis of neuroendocrine tumors. Ann N Y Acad Sci 2004;1014:246-57.1. Aguanno A, Afar R, Albert VR. Tissue-specific expression of the nonneuronalromoter of the aromatic L-amino acid decarboxylase gene is regulated by hepatocyteuclear factor 1. J Biol Chem 1996;271:4528-38.2. Borelli MI, Villar MJ, Orezzoli A, Gagliardino JJ. Presence of DOPA decarbox-lase and its localisation in adult rat pancreatic islet cells. Diabetes Metab997;23:161-3.3. Lindstrom P, Sehlin J. Aromatic amino acids and pancreatic islet function: aomparison of L-tryptophan and L-5-hydroxytryptophan. Mol Cell Endocrinol986;48:121-6.4. Hoegerle S, Altehoefer C, Ghanem N, Koehler G, Waller CF, Scheruebl H, et al.

hole-body 18F dopa PET for detection of gastrointestinal carcinoid tumors. Radiol-gy 2001;220:373-80.5. Hoegerle S, Nitzsche E, Altehoefer C, Ghanem N, Manz T, Brink I, et al.
heochromocytomas: detection with 18F DOPA whole body PET: initial results.adiology 2002;222:507-12.
2p


6. Otonkoski T, Veijola R, Huopio H, Nanto-Salonen K, Tapanainen P, deLonlay, Fekete C, et al. Diagnosis of focal persistent hyperinsulinism of infancy with8F-fluoro-L-DOPA PET. In: Krzisnik C, editor. 42nd Annual Meeting of theuropean Society for Paediatric Endocrinology (ESPE); 2003; Ljubljana (Slovenia):ormone Research; 2003. p. 2.

7. Otonkoski T, Nanto-Salonen K, Seppanen M, Veijola R, Huopio H, Hussain K,t al. Noninvasive diagnosis of focal hyperinsulinism of infancy with [18F]-DOPAositron emission tomography. Diabetes 2006;55:13-8.8. de Lonlay P, Simon-Carre A, Ribeiro MJ, Boddaert N, Giurgea I, Laborde K, etl. Congenital hyperinsulinism: pancreatic [18F]fluoro-L-dihydroxyphenylalanineDOPA) positron emission tomography and immunohistochemistry study of DOPAecarboxylase and insulin secretion. J Clin Endocrinol Metab 2006;91:933-40.9. Ribeiro MJ, De Lonlay P, Delzescaux T, Boddaert N, Jaubert F, Bourgeois S, etl. Characterization of hyperinsulinism in infancy assessed with PET and 18F-fluoro--DOPA. J Nucl Med 2005;46:560-6.0. Finegold DN, Stanley CA, Baker L. Glycemic response to glucagon during fastingypoglycemia: an aid in the diagnosis of hyperinsulinism. J Pediatr 1980;96:257-9.1. Stanley CA, Baker L. Hyperinsulinism in infancy: diagnosis by demonstration ofbnormal response to fasting hypoglycemia. Pediatrics 1976;57:702-11.2. Glaser B, Ryan F, Donath M, Landau H, Stanley CA, Baker L, et al. Hyperin-ulinism caused by paternal-specific inheritance of a recessive mutation in the sulfony-urea-receptor gene. Diabetes 1999;48:1652-7.3. Kassem SA, Ariel I, Thornton PS, Hussain K, Smith V, Lindley KJ, et al.57(KIP2) expression in normal islet cells and in hyperinsulinism of infancy. Diabetes001;50:2763-9.4. Rahier J, Guiot Y, Sempoux C. Persistent hyperinsulinaemic hypoglycaemia ofnfancy: a heterogeneous syndrome unrelated to nesidioblastosis. Arch Dis Child Fetal

eonatal Ed 2000;82:F108-12.5. Suchi M, MacMullen C, Thornton PS, Ganguly A, Stanley CA, Ruchelli ED.istopathology of congenital hyperinsulinism: retrospective study with genotype corre-

ations. Pediatr Dev Pathol 2003;6:322-33.6. Namavari M, Bishop A, Satyamurthy N, Bida G, Barrio JR. Regioselectiveadiofluorodestannylation with [18F]F2 and [18F]CH3COOF: a high yield synthesis of-[18F]Fluoro-L-dopa. Int J Rad Appl Instrum [A] 1992;43:989-96.7. Karp JS, Surti S, Daube-Witherspoon ME, Freifelder R, Cardi CA, Adam LE, etl. Performance of a brain PET camera based on anger-logic gadolinium oxyorthosilicateetectors. J Nucl Med 2003;44:1340-9.
8. Newcombe RG. Two-sided confidence intervals for the single proportion: com-arison of seven methods. Stat Med 1998;17:857-72.

Od

SacH

R(mr

Csw

DSYie

cspswludH(wlst

ABC

1

Longitudinal Lipid Screening and Use of Lipid-Lowering Medications inPediatric Type 1 Diabetes

DAVID M. MAAHS, MD, MA, R. PAUL WADWA, MD, KIM MCFANN, PHD, KRISTEN NADEAU, MD, MATTHEW R. WILLIAMS, BS,ROBERT H. ECKEL, MD, AND GEORGEANNA J. KLINGENSMITH, MD

bjective Because cardiovascular disease (CVD) is the leading cause of death in patients with type 1 diabetes (T1D) andyslipidemia is an important CVD risk factor, we investigated dyslipidemia and its treatment in children with T1D.

tudy design Subjects had T1D (n � 360), repeated lipid measurements (n � 1095; mean, 3.04 � 0.94; range, 2 to 11),nd were seen between 1994 and 2004. Total cholesterol (TC), high-density lipoprotein cholesterol (HDL), and non-HDLholesterol (non-HDL) were categorized on the basis of published guidelines. Age, diabetes duration, sex, body mass index,bA1c, and lipid-lowering medication use were recorded. Predictors of TC, HDL, and non-HDL were determined.

esults Sustained abnormalities existed for TC >200 mg/dL (16.9%); HDL <35 mg/dL (3.3%); and non-HDL >130 mg/dL27.8%), >160 mg/dL (10.6%), and >190 mg/dL (3.3%). Lipid-lowering medications were started on 23 patients. In mixedodel longitudinal data analyses, HbA1c was significantly related to TC and non-HDL. Body mass index z-score was inversely

elated to HDL.

onclusions In this retrospective, longitudinal study of pediatric patients with T1D with repeated lipid measurements,ustained abnormal levels for TC, HDL, and non-HDL were present. Prospective longitudinal data for dyslipidemia in youthith T1D are needed. (J Pediatr 2007;150:146-50)

espite guidelines for the management of dyslipidemia in children1-3 and longitudinal studies of serum lipids in thegeneral pediatric population,4-6 there are fewer data on lipids in pediatric subjects with type 1 diabetes (T1D).7-9 Theantecedents of adult cardiovascular disease (CVD), the primary cause of death in T1D,10 are present in children.4,11,12

tudies including the Bogalusa Heart Study,13-15 the Pathological Determinants of Atherosclerosis in Youth study,11,12 and theoung Finns Study16,17 demonstrate tracking of CVD risk factors into adulthood. Furthermore, the relation of CVD risk factors

n childhood relate to abnormalities in surrogate markers of atherosclerosis16,17 and/or atherosclerotic lesions in pathologyvaluations.11,15 Dyslipidemia is a significant CVD risk factor,18 and CVD is the leading cause of death in patients with T1D.10

We have previously reported that 18.6% of children with T1D had abnormal totalholesterol (TC) or high-density lipoprotein (HDL) levels in a retrospective cross-ectional analysis,19 but published longitudinal data on lipids in the T1D pediatricopulation are limited.1,20,21 In this study, we used electronic medical records to retro-pectively examine TC, HDL, and non-HDL levels in 360 pediatric patients with T1Dith repeated lipid measurements and determine the use and effectiveness of lipid-

owering medications during a 10-year period (1994 to 2004). We hypothesized that (1)ntreated sustained abnormal TC, HDL, and non-HDL levels would not normalizeespite standard diet and lifestyle counseling and attempts to control HbA1c; (2) TC,DL, and non-HDL would be significantly related to HbA1c and body mass index

BMI) over time; (3) in the time period studied (1994 to 2004), lipid-lowering medicationould be rarely used in children with T1D with dyslipidemia; and (4) children taking

ipid-lowering medication would have no significant adverse events and would haveignificant improvements in their repeat TC, HDL, and non-HDL after pharmacologicreatment.

DA American Diabetes AssociationMI Body mass index

HDL High-density lipoproteinTC Total cholesterol

From the Barbara Davis Center for Child-hood Diabetes and the Department ofMedicine, University of Colorado HealthSciences Center, Aurora, Colorado.

Support for this study was provided by NationalInstitutes of Health grants T32 DK063687-03(Dr Maahs), K23RR020038-02 (Dr Nadeau),and K12 DK063722-03 (Dr Wadwa), andDiabetes Education Research Center GrantP30 DK57516.

Submitted for publication Mar 9, 2006; lastrevision received Sep 16, 2006; acceptedOct 16, 2006.

Reprint requests: Dr David M. Maahs, Bar-bara Davis Center for Childhood Diabetes,University of Colorado Health SciencesCenter, 1775 Ursula St, PO Box 6511, MailStop A140, Aurora, CO 80045. E-mail:[email protected].



VD Cardiovascular disease T1D Type 1 diabetes

46

10.1016/j.jpeds.2006.10.054

AaDhown

ofccsTpiCdFmtmaatAf

�1Adma

T�iTt

mitagobvtc

o0smtMwna(abswlacd

T

AMDHBTHNMTHNNNL

N

L

METHODSStudy subjects were diagnosed with T1D (as defined by

merican Diabetes Association [ADA] criteria)22 �21 yearst their initial visit at the Barbara Davis Center for Childhoodiabetes between January 1, 1994, and June 30, 2004, and

ad screening lipids measured. Screening lipid levels werebtained as part of routine diabetes care. Since fasting statusas not routinely recorded, only TC, HDL, and calculatedon-HDL levels (TC minus HDL) were analyzed.

To ensure that only patients with T1D were studied,nly subjects either with autoantibody determinations positiveor diabetes-associated autoimmunity or specifically physi-ian-diagnosed as T1D were included. Furthermore, to ex-lude the lipid abnormalities of untreated diabetes, thoseubjects with serum lipids measured within the first month of1D diagnosis were excluded. During the study period, 2173atients 2 to 21 years of age with a specific diagnosis of T1Dn our database were seen at the Barbara Davis Center forhildhood Diabetes (mean age, 10.3 years; mean diabetesuration, 2.2 years; mean hemoglobin A1c [HbA1c], 8.5%).or study inclusion, subjects were required to have had lipidseasured at least twice during the study period (n � 1095

otal lipid measurements; mean, 3.04 � 0.94; range, 2 to 11easures/patient), leaving 360 subjects for analysis (170 male

nd 190 female). The cohort included patients with T1D,ges 2 to 21 years, followed for a mean of 2.9 years (range, 0.1o 9.6 years). All subjects eligible for inclusion are reported.ge, BMI, diabetes duration, HbA1c, and sex were obtained

or each clinical visit at which lipids were determined.HDL levels were categorized as either �35 mg/dL or

35 mg/dL, and TC levels were categorized as �170 mg/dL,70 to 200 mg/dL, or �200 mg/dL, based on currentDA1,3 and American Heart Association guidelines2 for chil-ren. Non-HDL levels were categorized as either �130g/dL or �130 mg/dL.23 A sustained abnormality is defined

able I. Baseline characteristics

Variable n � 360 subject

ge, years 13.6 � 4.1ale/female, % male 190/170iabetes duration, years 4.5 � 0.3bA1c, % 8.8 � 1.6MI z-score 0.62 � 1.00C, mg/dL 176.8 � 39.5DL cholesterol, mg/dL 54.8 � 15.9on-HDL cholesterol, mg/dL 122.0 � 38.1ean follow-up time, years 2.9 � 2.1C �200 mg/dL, % 18.9%DL cholesterol �35 mg/dL, % 4.2%on-HDL �130 mg/dL, % 34.2%on-HDL �160 mg/dL, % 11.4%on-HDL �190 mg/dL, % 5.8%

ipid-lowering medications use, % 6.4%

o significant differences between sexes.

s being abnormal for a category more than once over time. s

ongitudinal Lipid Screening and Use of Lipid-Lowering Medications in P

he proportion of subjects with non-HDL �160 mg/dL and190 mg/dL is also reported. TC and HDL were measured

n commercial laboratories as part of standard clinical practice.he Colorado Multiple Institutional Review Board approved

his study.The electronic database was queried for lipid-lowering

edication usage. In addition, each patient’s electronic med-cal record was reviewed to determine lipid-lowering medica-ion usage, dosage, start and stop dates, and whether anydverse effects were documented. Records were reviewed be-inning one visit before the first lipid screen through the endf the study period for each patient. Finally, individual dia-etologists were queried, and stored paper charts were re-iewed for selective confirmation of lipid-lowering medica-ion status. No discrepancies between electronic and paperharts were found.

Data were obtained on 360 patients (Table I), for a totalf 1095 observations (mean observations per patient, 3.04 �.94; range, 2 to 11). TC, HDL, and non-HDL were con-idered to be changed if they increased or decreased by one orore category from baseline at the patient’s last visit. Longi-

udinal data analysis was performed with the use of SAS Procixed with SP (POW) structure for unequally spaced data,

ith diabetes duration, HbA1c (log transformed due to non-ormal distribution), and BMI z-score as independent vari-bles and TC, HDL, and non-HDL as dependent variablesTable II). This longitudinal analysis method fully utilizes allvailable data points and accounts for different time intervalsetween measurements. Since age and diabetes duration areo closely correlated in children with T1D, separate modelsere constructed, using age and diabetes duration. The use of

ipid-lowering medications (yes/no) was included as a covari-te. Therefore, the direction and magnitude of the �-coeffi-ient for the lipid-lowering medication variable signifies theifference in lipid level for medication status. The Wilcoxon

Males (n � 190) Female (n � 170)

13.5 � 4.0 13.6 � 4.152.8% 47.2%

4.4 � 3.4 4.7 � 3.58.8 � 1.5 8.8 � 1.6

0.61 � 1.11 0.62 � 0.90174.3 � 38.5 179.2 � 40.453.8 � 17.4 55.7 � 14.4

120.5 � 35.7 123.4 � 40.22.7 � 2.0 3.1 � 2.119.4% 18.4%4.1% 4.2%

34.1% 34.2%11.2% 11.6%6.5% 5.3%5.3% 7.7%

s

igned-rank or Kruskal-Wallis test was used to evaluate dif-

ediatric Type 1 Diabetes 147

fdifaettsc

Tb6hm(�2.12b

(m

(n(shST(Hmflon�m

tsHH(Tds(tate

3(nwOmt�kwwswtmmwaweh.PP

Tcd

T

H

N

NT

1

erences in continuous variables, and the �2 test of indepen-ence was used to evaluate differences in discrete character-

stics. Results were considered significant at � � 0.05. Finally,or patients on lipid-lowering medications (n � 23, Table III;vailable at www.jpeds.com), the following are reported forach patient: lipid measurements immediately before initia-ion of medication, the most-improved lipid measurements,he time interval between premedication and final lipid mea-urements, the final lipid measurements, and the percenthange in lipid.

RESULTSBaseline characteristics of study patients are shown in

able I. There were no statistically significant differencesetween male and female subjects at baseline. At the first visit,8 of 360 (18.9%) patients had TC �200 mg/dL, 15 (4.2%)ad HDL �35 mg/dL, 123 (34.2%) had non-HDL �130g/dL, 41 (11.4%) had non-HDL �160 mg/dL, and 21

5.8%) had non-HDL �190 mg/dL. Those patients with TC200 mg/dL had worse glycemic control (HbA1c � 9.7% �

.1%, P � .0001) and were older (age: 14.5 � 3.9 years, P �0013) than those with normal cholesterol (HbA1c � 8.5% �.2%; age: 13.8 � 4.2 years) or cholesterol between 170 and00 mg/dL (HbA1c � 8.8% � 1.4%; age: 12.6 � 3.9 years)ut were similar in BMI z-score and duration of diabetes.

On at least one visit during the study period, 12334.2%) had TC �200 mg/dL, 34 (9.4%) had HDL �35

able II. Longitudinal predictors of totalholesterol, high-density lipoprotein, and non–high-ensity lipoprotein

Variable �-coefficient (SE) P value

C (n � 358)Intercept 95.61 (17.39) �.0001Diabetes duration 0.0454 (0.3239) .89Lipid medication: No �39.53 (4.35) �.0001Lipid medication: Yes 0Log10A1c 123.94 (17.18) �.0001BMI z-score 0.25 (1.08) .82DL (n � 358)Intercept 43.53 (7.03) �.0001Diabetes duration 0.0255 (0.1291) .84Lipid medication: No 3.57 (1.73) .04Lipid medication: Yes 0Log10A1c 9.35 (6.97) .18BMI z-score �0.99 (0.43) .02on-HDL (n � 358)Intercept 49.27 (16.64) .0033Diabetes duration �0.0265 (0.3024) .93Lipid medication: No �43.89 (4.02) �.0001Lipid medication: Yes 0Log10A1c 118.27 (16.53) �.0001BMI z-score 1.19 (1.01) .24

ote: Parameter estimates obtained using SAS Proc Mixed.wo patients lacked complete data for analysis.

g/dL, 172 (47.8%) had non-HDL �130 mg/dL, 68 s

48 Maahs et al

18.9%) had non-HDL �160 mg/dL, and 35 (9.7%) hadon-HDL �190 mg/dL. An abnormality of any of theseTC, HDL, or non-HDL) on at least one visit during thetudy period was found in 191 (53.1%), whereas 148 (41.1%)ad either abnormal TC �200 mg/dL or HDL �35 mg/dL.ustained abnormalities existed for 61 (16.9%) patients withC �200 mg/dL, 12 (3.3%) with HDL �35 mg/dL, 100

27.8%) with non-HDL �130 mg/dL, 38 (10.6%) with non-DL �160 mg/dL, and 12 (3.3%) with non-HDL �190g/dL. Lipid measurements were not related to time of

ollow-up (data not shown). Among patients with baselineipid abnormalities, 30 of 68 (44%) with TC �200 mg/dL, 6f 15 (40%) with HDL �35 mg/dL, 77 of 123 (63%) withon-HDL �130 mg/dL, 17 of 41 (41%) with non-HDL160 mg/dL, and 6 of 21 (29%) with non-HDL �190g/dL were also abnormal at their final visit.

After adjustment for the use of lipid-lowering medica-ions, mixed-model longitudinal data analyses demonstratedignificant relations between HbA1c and TC (P � .0001) andbA1c and non-HDL (P � .0001) but not HbA1c andDL (P � .18). BMI z-score had a negative relation to HDL

P � .02), but there was no relation between BMI z-score andC (P � .82) or non-HDL (P � .24) (Table II). Diabetesuration was not related to TC, HDL, or non-HDL. In aeparate model, age was not related to HDL (P � .77), TCP � .63), or non-HDL (P � .96). As expected, patients notaking lipid-lowering medications had significantly lower TCnd non-HDL and higher HDL compared with patientsaking lipid-lowering medications, as indicated by the �-co-fficients in Table II.

Lipid-lowering medications were started in 23 of the60 (6.4%) patients, including 19 exclusively taking statinsTable III). Of the 23, 12 had non-HDL �190 mg/dL, 7 hadon-HDL between 160 and 189 mg/dL (including one whoas �12 years of age), and 4 had non-HDL �160 mg/dL.f the 7 patients with non-HDL between 160 and 189g/dL, all had LDL �130 mg/dL and/or hypertension. Of

he 4 patients who had non-HDL �160 mg/dL, 2 had LDL130 mg/dL, 1 had hypertension, and 1 had a BMI of 36

g/m2. One patient was prescribed ezetimibe (No. 21), oneas started on a statin and then continued in combinationith ezetimibe (No. 14), another was started on a fibrate but

witched to a statin and then ezetimibe (No. 4), and anotheras prescribed a statin and later a fibrate (No. 17). Of the 23

aking medications, 10 had TC �200 mg/dL on the finaleasurement, whereas 13 did not. There were no recordedajor adverse effects from lipid-lowering medications. Thereere no significant differences between those who improved

nd those who did not improve on lipid-lowering medicationsith respect to age, duration of diabetes, BMI, or HbA1c. As

xpected, those placed on lipid-lowering medications hadigher TC (248 � 64 mg/dL vs 172 � 32 mg/dL; P �

0001), non-HDL (195 � 67 mg/dL vs 117 � 30 mg/dL;� .0001), and HbA1c (10.1% � 2.1% vs 8.7% � 1.5%,� .0005) than the rest of the study population. Table III

hows lipid levels at the visit before beginning lipid-lowering



mltmm2lmi

wAtHPmublhivHOopwdfannmlmg

ecatrispcwvlmcr

eTspsctibe1icg�ootsbap

irmddut

WtC

1C2Hb3CA4h5lE6pf17a8c29

L

edications. Although more than half of the patients takingipid-lowering medication did not have TC �200 mg/dL athe final measurement as desired for pharmacologic treat-ent, their most-improved lipid measurement showed aean decrease in TC (–25% � 17%) and non-HDL (–30% �

0%) and an increase in HDL (�8% � 17%) while takingipid-lowering medication (Table III). This suggests that

ore rigorous therapy as well as patient education on themportance of continued therapy may be needed.

DISCUSSIONSince the use of lipid-lowering medication in children

ith T1D is a relatively new recommendation from theDA,1,3 lipid-lowering medications were started in few pa-

ients; however, these patients had a higher mean TC, non-DL, and HbA1c than the rest of the study population.atients started on lipid-lowering medications had improve-ent in lipid levels, although these improvements were not

niformly sustained. Compliance with medications could note assessed with this study. Similarly, although diet andifestyle counseling as well as attention to glycemic controlave long been part of our center’s multidisciplinary approach,

ndividual patient response to this counseling cannot be pro-ided. A limitation of this report is that patients’ TC andDL were not performed in a standard research laboratory.nly a subset of patients had their lipids measured more than

nce introducing a selection bias. There are also no data onubertal stage. Although we report no serious adverse eventsith lipid-lowering medications similar to studies in adults,24

ata on the safety of these agents in T1D youth requireurther study, including the long-term use and potential ter-togenic issues for female adolescents who may become preg-ant.25,26 Monitoring of liver function tests (on which we doot report data) and musculoskeletal symptoms are recom-ended.1 Our data also suggest that life-long attention to

ifestyle modification in addition to pharmacologic treatmentay be required to meet and sustain recommended lipid

oals.Although numerous longitudinal reports in childhood

xist, data on lipids in children with T1D are limited andonsist of cross-sectional7-9,19,27 and short-duration (2 weeksnd 6 months) longitudinal reports that support the impor-ance of glycemic control20,21 for lipid health. The studyeported by Schwab et al9 is noteworthy for reporting dyslip-demia (TC, LDL, or HDL) in 28.6% in a large, cross-ectional study of young (�26 years) German and Austrianatients. In this report, only 0.4% reported receiving pharma-ologic treatment. We report that 40% to 63% of childrenith abnormal lipids remain abnormal, but others follow aariable course from their first to last visit. Although repeatipid determination before pharmacologic treatment is recom-

ended,3 the intrasubject variability of lipids over time inhildren with T1D cannot be fully addressed with these
etrospective data.
ec

ongitudinal Lipid Screening and Use of Lipid-Lowering Medications in P

Current clinical guidelines in pediatrics are based onxpert opinion and extrapolation from the adult literature.1-3

he ADA recommendations for children with T1D are tocreen for dyslipidemia in children �2 years of age in theresence of a positive or unknown family history; otherwise,creening should occur at �12 years of age (once glycemicontrol has been obtained in the newly diagnosed patient) andhen repeated every 5 years if normal. Treatment with med-cation is recommended for LDL �160 mg/dL. In addition,lood glucose control and dietary and exercise counseling aremphasized. In children with LDL levels between 130 and59 mg/dL, glucose control and dietary and exercise counsel-ng are recommended for 6 months, with medication to beonsidered if the LDL remains �130 mg/dL. The treatmentoals are LDL �100 mg/dL, HDL �35 mg/dL, and TG150 mg/dL.1,3 Ideally, screening blood samples should be

btained in the fasting state. However, given the difficulties ofbtaining fasting samples (which include safety issues in pa-ients with T1D), the Adult Treatment Panel III (ATP-III)uggests screening with nonfasting TC and HDL, followedy a complete fasting lipoprotein panel if screening results arebnormal.23 Non-HDL is used as a secondary target in adults,articularly among patients with elevated triglycerides.23

Our findings support the importance of regular screen-ng for dyslipidemia in children with T1D. The variabilityeported indicates the importance of repeated lipid measure-ents. Future research needs include prospective longitudinal

ata on the natural history of dyslipidemia, safety and efficacyata from clinical trials of lipid-lowering medications, andltimately the long-term relation of dyslipidemia and itsreatment to future health outcomes in youth with T1D.

e thank Alex Brown and the Information Technology staff athe Barbara Davis Center for Childhood Diabetes in Denver,olorado.

REFERENCES. Management of dyslipidemia in children and adolescents with diabetes. Diabetesare 2003;26:2194-7.

. Kavey RE, Daniels SR, Lauer RM, Atkins DL, Hayman LL, Taubert K. Americaneart Association guidelines for primary prevention of atherosclerotic cardiovascular disease

eginning in childhood. J Pediatr 2003;142:368-72.. Silverstein J, Klingensmith G, Copeland K, Plotnick L, Kaufman F, Laffel L, et al.are of children and adolescents with type 1 diabetes: a statement of the American Diabetesssociation. Diabetes Care 2005;28:186-212.

. Lauer RM, Lee J, Clarke WR. Factors affecting the relationship between child-ood and adult cholesterol levels: the Muscatine Study. Pediatrics 1988;82:309-18.. Webber LS, Srinivasan SR, Wattigney WA, Berenson GS. Tracking of serumipids and lipoproteins from childhood to adulthood: the Bogalusa Heart Study. Am Jpidemiol 1991;133:884-99.. Porkka KV, Viikari JS, Taimela S, Dahl M, Akerblom HK. Tracking andredictiveness of serum lipid and lipoprotein measurements in childhood: a 12-yearollow-up: the Cardiovascular Risk in Young Finns study. Am J Epidemiol994;140:1096-110.. Cruickshanks KJ, Orchard TJ, Becker DJ. The cardiovascular risk profile ofdolescents with insulin-dependent diabetes mellitus. Diabetes Care 1985;8:118-24.. Polak M, Souchon P-F, Benal K, Tubiana-Rufi N, Czernichow P. Type 1 diabetichildren have abnormal lipid profiles during pubertal years. Pediatr Diabetes001;1:74-81.. Schwab KO, Doerfer J, Hecker W, Grulich-Henn J, Wiemann D, Kordonouri O,
t al. Spectrum and prevalence of atherogenic risk factors in 27,358 children, adoles-ents, and young adults with type 1 diabetes: cross-sectional data from the German
ediatric Type 1 Diabetes 149

d21Rao1sPA1omV1ch1eoc1Wi11ei1Rt

1eE1Td2Ei2sw2t2PB2IJ2ar2st2i

1

iabetes documentation and quality management system (DPV). Diabetes Care006;29:218-25.0. Libby P, Nathan DM, Abraham K, Brunzell JD, Fradkin JE, Haffner SM, et al.eport of the National Heart, Lung, and Blood Institute-National Institute of Diabetes

nd Digestive and Kidney Diseases Working Group on Cardiovascular Complicationsf Type 1 Diabetes Mellitus. Circulation 2005;111:3489-93.1. McGill HC Jr, McMahan CA, Malcom GT, Oalmann MC, Strong JP. Effects oferum lipoproteins and smoking on atherosclerosis in young men and women: theDAY Research Group: Pathobiological Determinants of Atherosclerosis in Youth.rterioscler Thromb Vasc Biol 1997;17:95-106.

2. McGill HC Jr, McMahan CA, Malcom GT, Oalmann MC, Strong JP. Relationf glycohemoglobin and adiposity to atherosclerosis in youth: Pathobiological Deter-inants of Atherosclerosis in Youth (PDAY) Research Group. Arterioscler Thrombasc Biol 1995;15:431-40.

3. Bao W, Srinivasan SR, Wattigney WA, Berenson GS. Persistence of multipleardiovascular risk clustering related to Syndrome X from childhood to young adult-ood. Arch Intern Med 1994;154:1842-7.4. Bao W, Srinivasan SR, Valdez R, Greenlund KJ, Wattigney WA, Berenson GS,t al. Longitudinal changes in cardiovascular risk from childhood to young adulthood inffspring of parents with coronary artery disease: the Bogalusa Heart Study [seeomments]. JAMA 1997;278:1749-54.5. Berenson GS, Srinivasan SR, Bao W, Newman WP III, Tracy RE, Wattigney

A, et al. Association between multiple cardiovascular risk factors and atherosclerosisn children and young adults: the Bogalusa Heart Study [see comments]. N Engl J Med998;338:1650-6.6. Raitakari OT, Juonala M, Kahonen M, Taittonen L, Laitinen T, Maki-Torkko N,t al. Cardiovascular risk factors in childhood and carotid artery intima-media thicknessn adulthood: the Cardiovascular Risk in Young Finns Study. JAMA 2003;290:2277-83.7. Juonala M, Jarvisalo MJ, Maki-Torkko N, Kahonen M, Viikari JS, Raitakari OT.
isk factors identified in childhood and decreased carotid artery elasticity in adulthood:
he Cardiovascular Risk in Young Finns Study. Circulation 2005;112:1486-93. y

50 Maahs et al

8. Soedamah-Muthu SS, Chaturvedi N, Toeller M, Ferriss B, Reboldi P, Michel G,t al. Risk factors for coronary heart disease in type 1 diabetic patients in Europe: theURODIAB Prospective Complications Study. Diabetes Care 2004;27:530-7.9. Maahs DM, Maniatis AK, Nadeau K, Wadwa RP, McFann K, Klingensmith GJ.otal cholesterol and high-density lipoprotein levels in pediatric subjects with type 1iabetes mellitus. J Pediatr 2005;147:544-6.0. Lopes-Virella MF, Wohltmann HJ, Mayfield RK, Loadholt CB, Colwell JA.ffect of metabolic control on lipid, lipoprotein, and apolipoprotein levels in 55

nsulin-dependent diabetic patients: a longitudinal study. Diabetes 1983;32:20-5.1. Attia N, Touzani A, Lahrichi M, Balafrej A, Kabbaj O, Girard-Globa A. Re-ponse of apolipoprotein AIV and lipoproteins to glycaemic control in young peopleith insulin-dependent diabetes mellitus. Diabet Med 1997;14:242-7.2. American Diabetes Association. Clinical practice recommendations 2002. Diabe-es Care 2002;25(Suppl 1):S1-147.3. Executive Summary of the Third Report of the National Cholesterol Educationrogram (NCEP) Expert Panel on Detection, Evaluation, and Treatment of Highlood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285:2486-97.4. Graham DJ, Staffa JA, Shatin D, Andrade SE, Schech SD, La Grenade L, et al.ncidence of hospitalized rhabdomyolysis in patients treated with lipid-lowering drugs.AMA 2004;292:2585-90.5. Wiegman A, Hutten BA, de Groot E, Rodenburg J, Bakker HD, Buller HR, etl. Efficacy and safety of statin therapy in children with familial hypercholesterolemia: aandomized controlled trial. JAMA 2004;292:331-7.6. Clauss SB, Holmes KW, Hopkins P, Stein E, Cho M, Tate A, et al. Efficacy andafety of lovastatin therapy in adolescent girls with heterozygous familial hypercholes-erolemia. Pediatrics 2005;116:682-8.7. Kershnar A, Daniels S, Dolan L, Hamman R, Palla S, Pettitt D, et al. Lipid levelsn youth with diabetes mellitus: preliminary results from the SEARCH for diabetes in
outh study. Diabetes 2005;54:A569.

Table III. Change in total cholesterol, high-density lipoprotein, and non–high-density lipoprotein in patients (n � 23) taking lipid-loweringmedications

Premedication* Most improved‡ Final

Patientno.

Age, ySex (M/F) TC HDL

Non-HDL TC %� HDL %�

Non-HDL%�

Timeinterval, y TC %� HDL %�

Non-HDL%�

1 13.7 313 45 268 147 56 94 2.3 147 53 94F �53% �24% �65% �53% �18% �65%

2 21.5 298 55 243 232 67 172 6.2 291 59 232M �22% �22% �29% �2% �7% �5%

3 24.5 251 46 205 144 47 98 1.2 204 47 157M �43% �2% �52% �19% �2% �23%

4 18.0 441 45 396 306 62 260 1.9 310 45 265F �31% �38% �34% �30% 0% �30%

5 23.5 259 50 209 183 57 132 2.4 246 57 189F �29% �14% �37% �5% �14% �10%

6 14.8 242 41 201 159 53 117 2.7 244 53 191F �34% �29% �42% �1% �29% �5%

7 19.0 188 66 122 169 53 116 0.4 169 53 116F �10% �20% �5% �10% �20% �5%

8 14.7 211 38 173 106 45 76 2.6 210 45 165M �50% �18% �56% �0.4% �18% �5%

9 11.3 235 82 153 160 58 102 3.6 160 58 102M �32% �29% �33% �32% �29% �33%

10 14.9 370 59 311 193 64 145 3.4 193 48 145M �48% �8% �53% �48% �19% �53%

11 17.3 240 64 176 221 70 157 1.4 221 62 159F �8% �9% �11% �8% �3% �10%

12 10.7 226 42 184 201 41 160 0.6 201 41 160F �11% �2% �13% �11% �2% �13%

13 18.6 242 66 176 172 67 105 0.3 172 67 105F �29% �2% �40% �29% �2% �40%

14 12.1 216 60 156 170 65 105 1.3 170 65 105F �21% �8% �33% �21% �8% �33%

15 12.8 271 46 225 323 59 274 3.4 333 59 274F �19% �28% �22% �23% �28% �22%

16 17.4 243 52 191 195 58 145 0.8 195 50 145M �20% �12% �24% �20% �4% �24%

17 12.8 282 59 223 203 68 153 0.8 369 68 301F �28% �16% �31% �31% �16% �35%

18 16.5 208 45 163 194 43 160 2.0 194 34 160M �7% �4% �2% �7% �24% �2%

19 16.7 232 55 177 143 43 100 1.3 143 43 100F �38% �22% �44% �38% �22% �44%

20 13.4 267 62 205 200 63 137 0.3 200 63 137

LongitudinalLipidScreening

andU

seof

Lipid-Lowering

Medications

inPediatric

Type

1D

iabetes150.e1

Table III. Continued

Premedication* Final

Patientno.

Age, ySex (M/F) TC HDL

Non-HDL y TC %� HDL %�

Non-HDL%�

M �25% �2% �33%21 14.8 189 44 145 117 41 76

F �38% �7% �48%22 20.0 227 53 174 205 54 151

M �10% �2% �13%23 15.6 246 43 203 226 46 180

M �8% �7% �11%Mean � SD 16.3 � 3.7 256 � 57 52 � 11 203 � 5 �16% � 20% �1% � 16% �19% � 23%

*Before medication start, patient may have had additional previous lipid measurem†Medications used are statins (HMG CoA reductase inhibitors), except for patien (statin then fibrate), and No. 21 (ezetimibe).‡“Most improved” lipid levels are for each individual measurement (TC, HDL, an ost improved” columns.

150.e2M

aahset

alT

heJournalof

Pediatrics•February

2007

Most improved‡

TC %� HDL %�Non-HDL

%�Time

interval,

�25% �2% �33%117 41 76 0.8

�38% �7% �48%205 54 151 0.3

�10% �2% �13%226 46 180 0.5�8% �7% �11%

9 �25% � 17% �8% � 17% �30% � 20% 1.8 � 1.4

ents. TC, HDL, and non-HDL in mg/dL.t No. 4 (fibrate, then statin, then ezetimibe), No. 14 (statin, then ezetimibe), No. 17d non-HDL); therefore TC minus HDL does not always equal non-HDL in the “M

Oo

Sm

RaPnsi

Cnns

Cialisbc

Ca

CsC

u

BCCCD

Cranial Ultrasound Scanning and Prediction of Outcome in Newborns withCongenital Cytomegalovirus Infection

GINA ANCORA, MD, MARCELLO LANARI, MD, PHD,* TIZIANA LAZZAROTTO, MD, VALENTINA VENTURI, MD,ELISABETTA TRIDAPALLI, MD, FABRIZIO SANDRI, MD, MADDALENA MENARINI, MD, EMANUELA FERRETTI, MD, AND

GIACOMO FALDELLA, MD

bjective To report the accuracy of ultrasound scanning (US) in predicting neurodevelopmental and sensorineuralutcome in patients with congenital cytomegalovirus (CMV) infection.

tudy design Fifty-seven neonates with congenital CMV infection underwent brain US and were observed prospectively forotor skills, developmental quotient, and hearing function.

esults Abnormal results on US were found in 12 of 57 neonates. US lesions were more frequent in newborns with clinicalnd laboratory signs of congenital CMV infection at birth (10/18) than in newborns who had no symptoms at birth (2/39;< .001). At least 1 sequela developed in all neonates with symptoms who had abnormal US results, whereas none of the

eonates with symptoms who had normal US results had long-term sequelae (P < .001). In the population without symptoms,ensorineural hearing loss developed in 3 of 37 (8.1%) neonates with normal US results, whereas severe sequelae developedn 1 of 2 neonates with abnormal US results.

onclusions A good correlation was found between cerebral US abnormalities and the prediction of outcome inewborns who were congenitally infected with CMV and had symptoms at birth. US could be performed as the firsteuroimaging study in these newborns. Data are insufficient to permit any suggestions for the population withoutymptoms. (J Pediatr 2007;150:157-61)

ongenital cytomegalovirus (CMV) infection can cause a wide spectrum of brain damage related to inflammatory andteratogenetic effects, including meningoencephalitis, calcifications, microcephaly, disturbance of neuronal migration,germinal matrix cysts, ventriculomegaly, and cerebellar hypoplasia.1 Computed tomography (CT), magnetic resonance

maging (MRI), and cerebral ultrasound scanning (US) are well-documented means for detecting brain lesions and othernomalies of the central nervous system (CNS).2-5 CT has been used to detect CNSesions, predicting neurodevelopmental outcome,3,6 and to identify infants at risk ofmpaired hearing and those who could benefit from gancilovir treatment.7,8 US is theafest means to image the neonatal brain and, unlike CT, is readily available at theedside. The use of US to predict the neurodevelopmental and sensorineural outcome ofongenital CMV infection has not been reported.

We undertook a systematic prospective US study of all neonates in whom congenitalMV infection was diagnosed, correlating cerebral US features with neurodevelopmental

nd sensorineural sequelae.

METHODSBetween January 1997 and September 2003, 57 newborns in whom congenital

MV infection was diagnosed were referred to our tertiary care hospital, where we haveet up a multidisciplinary team specializing in the treatment of patients with congenitalMV infection.

Congenital CMV infection was diagnosed on the basis of isolation of the virus fromrine within the first 2 weeks of life in neonates born from mothers with a suspected or

AER Brainstem auditory evoked responsesMV CytomegalovirusNS Central nervous systemT Computed tomography

LSV Lenticulostriate vasculopathyMRI Magnetic resonance imagingOR Odds ratioSNHL Sensorineural hearing loss

From the Institute of Neonatology, De-partment of Gynecological, Obstetrical, andPediatric Sciences, Sant’Orsola Hospital,University of Bologna, Bologna, Italy; De-partment of Clinical and Experimental Medi-cine, Microbiology Division, Sant’Orsola Hos-pital, University of Bologna, Bologna, Italy;Department of Rehabilitative Medicine,Sant’Orsola Hospital, University of Bologna,Bologna, Italy; and Neonatal Intensive CareUnit, University of Sherbrooke, Quebec,Canada.

*Gina Ancora, MD and Marcello Lanari,MD, PHD contributed equally to this article.

Submitted for publication Feb 1, 2006; lastrevision received Sep 16, 2006; acceptedNov 3, 2006.

Reprint requests: Gina Ancora, Via Massa-renti 11, 40138 Bologna, Italy. E-mail:[email protected].



Q Developmental quotient US Ultrasound scanning
10.1016/j.jpeds.2006.11.032
157

atcdtrsniClnetifitUscevmltvpnwiwpfsp

na6sicBefthh

mccaDwe

drtbsdmriq

i2af

l

wsfi

tctsfcwfioaapNafibctamid

m1d

oe

1

scertained CMV infection during pregnancy.9 Virus isola-ion was performed in cell cultures with the “shell vial” pro-edure.10 Newborns congenitally infected with CMV wereefined as symptomatic when they had clinical signs, labora-ory signs, or both at birth, including intrauterine growthetardation, microcephaly, seizures, chorioretinitis, hepato-plenomegaly, petechiae, elevated serum transaminase levels,eutropenia, and thrombocytopenia. US was obtained by us-

ng an Esaote AU5 with a 7.5-MHz sector probe transducer.ranial US was performed in all cases within the first week of

ife and repeated at 1 and 3 months of age when results wereegative. Neonates with abnormal US findings were re-xamined weekly with US during the first month of life andhen monthly until they were 6 months old. Two independentnvestigators (G.A. and F.S.), who were blinded to clinicalndings, reviewed all scans and obtained the same results forhe presence of lesions related to CMV congenital infection.S results were classified as normal or abnormal. Pathologic

cans were defined as having periventricular/parenchymal cal-ifications, ventriculomegaly, cysts, cerebellar lesions (hyper-chogenicity, hypoplasia), or “candlestick” lenticulostriateasculopathy (LSV).1,4,5 Ventricular size was measured at theidbody of the lateral ventricles on a sagittal view. Ventricu-

omegaly was classified as mild, moderate, or severe, accordingo Allan.11 Scans showing only mild ventriculomegaly (ie,entricular size between 3 and 5 mm) were not consideredathologic. In addition, isolated LSV (ie, increased echoge-icity at the level of vessels in the basal ganglia and thalamus)as not considered pathologic. LSV is not a specific marker of

ntrauterine infection and has also been reported in neonatesith chromosomal trisomies, alcohol and drug exposure, as-hyxia, heart disease, or respiratory distress, and in disease-ree newborns.12,13 Moreover, longitudinal studies havehown that this US finding is not in itself associated with aoor prognosis.14,15

Ophthalmologic examination was performed in theeonatal period. Hearing function was assessed by brainstemuditory evoked responses (BAER) in the neonatal period, atand 12 months of age, and yearly audiometric tests until

chool age. The BAER threshold was defined as the lowestntensity level at which wave V could be detected and repli-ated. Sensorineural hearing loss (SNHL) was defined as aAER threshold �25 dB detected in at least 2 differentvaluations or by an air conduction threshold at 1 or morerequencies �20 dB. Tympanometry was routinely performedo exclude middle ear disorders, and children with conductiveearing loss in the absence of SNHL were not considered toave hearing loss.

Psychomotor development was assessed at 6, 12, and 24onths of age by using the Brunet-Lezine test rating posture,

oordination, speech and socialization.16 The global score wasalculated as a ratio between observed developmental age andnagraphic age (developmental quotient [DQ]). Scores ofQ �0.85 were considered to be abnormal in accordanceith Grossman.17 At 3 months of age and at every follow-up

xamination, an expert physiatrist assessed the psychomotor i

58 Ancora et al

evelopment of the infants with a 3-axis grid including neu-ovegetative, motor, and relational/behavioral items combininghe Milani-Comparetti neuroevolutive assessment,18 Brazeltonehavioral assessment,19 and Prechtl general movement as-essment.20 This comprehensive assessment method allowedetection of suspicion for abnormal psychomotor develop-ent, to permit early referral (before 4-6 months of age) for

ehabilitation. Motor delay was defined as functional deficitsn motor skills and unachieved developmental milestones re-uiring rehabilitation.

Cranial CT, MRI, or both was performed in all surviv-ng infants with abnormal US results. MRI was performed in

cases during the neonatal period and in 6 cases between 4nd 12 months of age. CT was done in the first trimester ofetal life.

Informed consent was obtained from the parents oregal guardians of the infants studied.

Data were recorded in an Excel database and analyzedith the SPSS 5.0 software (SPSS, Chicago, IL). The chi-

quare statistic was used to test the relationship between USndings and outcome.

RESULTSTwelve of 57 (21.0%) neonates referred to our institu-

ion with a diagnosis of congenital CMV infection showederebral US abnormalities typical of congenital CMV infec-ion. Eighteen of 57 (31.6%) showed clinical or laboratoryigns of congenital infection at birth. US lesions were morerequent in newborns with clinical and laboratory signs ofongenital CMV infection at birth (10/18) than in newbornsho had no symptoms at birth (2/39; P � .000). Clinical data

or patients with normal and abnormal US results are reportedn Table I. In most cases, US showed a varying combinationf calcifications, ventriculomegaly, cysts, cerebellar anomalies,nd LSV. Lesions were detected at the first US examinationnd remained stable in all cases, with the exception of aeriventricular cyst that enlarged during the first week of life.o infant with normal US results at birth developed lesion(s)

t subsequent evaluations. Further investigations were per-ormed in the 11 surviving infants with pathologic US find-ngs: 3 underwent CT, 4 underwent MRI, and 4 underwentoth CT and MRI. CT results confirmed US findings in allases, and no additional abnormalities were detected. Al-hough the MRI missed calcifications in 2 cases, it discloseddditional findings in 6 cases compared with US, includingigrational disorders, leukodystrophy, and delayed myelin-

zation. In 1 case, US missed a temporal horn cyst, 5 mm iniameter, that was evident on MRI.

Follow-up data until patients were a minimum of 12onths old (mean age � SD at last follow-up visit, 42.3 �

1.3 months) were available for 56 of 57 patients; 1 infantied in the neonatal period of aortic thrombosis.21

Of the 11 surviving neonates with abnormal US results,nly 1, who had moderate ventriculomegaly and a hyper-choic lesion at the level of the germinal matrix that evolved
nto a cyst, had a normal outcome at follow-up. Strabismus

dsbmidlmat

mce

tAtnrtnd(dbpnw.I

tic

racfiha

htwqtdwinnnoornp

iailinsbbwu

Ts

MMMNC

N*

C

eveloped in 1 infant whose US and CT scans showed aingle calcification at the caudate level, and SNHL and stra-ismus developed in another infant with US and CT signs ofultiple calcifications and ventriculomegaly. The other 8

nfants, who had a varying combination of ventriculomegaly,iffuse calcifications, cerebellar hypoplasia, and periventricu-

ar cysts at birth, had poor psychomotor development (DQedian, 76.8; range, 67-79) associated with SNHL in 5 cases

nd motor delay in 6 cases (the latter were those with addi-ional MRI findings compared with US; Table II).

Forty-two of 45 patients without cerebral US abnor-alities had normal psychomotor and sensorineural out-

omes; hypoacusia developed in 3 children, mild and unilat-ral in 1 case and bilateral in the other 2 (Table II).

Statistical analysis of data was performed separately forhe 2 groups of patients with or without symptoms at birth.t least 1 sequela developed in all 10 newborns with symp-

oms who had abnormal US results or they died, whereasone of the 8 newborns with symptoms who had normal USesults showed long term sequelae (P � .001). In the asymp-omatic population, SNHL developed in 3 of 37 (8.1%)ewborns with normal US results, whereas severe sequelaeeveloped in 1 of the 2 newborns with abnormal US resultsodds ratio [OR], 11.3; 95% CI, 1.0-141.6; P � .2). Whenata for infected newborns, with or without symptoms atirth, were pooled, at least 1 sequela developed in 91.7% ofatients with abnormal results on US performed in the neo-atal period or they died, compared with 6.7% of the groupith normal US results (OR, 154; 95% CI, 17.3-1219.6; P �

001). The individual risk for each sequela is reported in TableII.

DISCUSSIONCMV infection is the most common intrauterine infec-

ion in developed countries, with a prevalence of 0.2% to 2.2%n live neonates. Cerebral lesions, including meningoen-

able I. Clinical data of neonates with congenital cytcanning findings

ChildrenUS resu

ale/Femaleean (SD) gestational age in weeks 38ean (SD) birth weight in grams 30umber of neonates with symptoms at birth (%)linical and laboratory signsIntrauterine growth retardation, n (%)Petechiae, n (%)Hepatosplenomegaly, n (%)Microcephaly, n (%)ChorioretinitisOther, n (%)*Death, n (%)

S, Not statistically significant.Elevated serum transaminase level, neutropenia, thrombocytopenia.

ephalitis, calcifications, microcephaly, disturbance of neu- s

ranial Ultrasound Scanning and Prediction of Outcome in Newborns w

onal migration, germinal matrix cysts, ventriculomegalynd cerebellar hypoplasia, may develop in patients who areongenitally infected. The severity of the neuropathologicndings at birth correlates with poor outcome, includingearing loss, mental retardation, cerebral palsy, seizures,nd chorioretinitis.1

Early identification of neonates at risk is important toelp healthcare providers counsel parents and give adequatereatment and follow-up care. Neonatal clinical signs correlateith neurologic outcome in congenital CMV infection: se-uelae can develop in as many as 80% of infants with symp-oms, whereas available data suggest that either audiologic orevelopmental problems develop in 5% to 15% of newbornsithout symptoms in the first year of life.22-26 These findings

ndicate that the presence of symptoms at birth does notecessarily differentiate children in whom sequelae will or willot develop. The addition of CT scans improves the prog-ostic accuracy in patients with clinical signs: as many as 90%f infected infants with symptoms who have cerebral lesionsn CT examination have neurodevelopmental or sensorineu-al sequelae.3 CT may also reveal rarer anomalies in infectedewborns without symptoms who have poor developmentalerformance.24

Although CT is the most widely used and acceptedmaging technique for investigating the brain in neonates whore congenitally CMV infected, we aimed to find the leastnvasive prognostic indicator in this population.27 Most of theesions related to fetal infection are readily disclosed by US,ncluding calcifications, ventriculomegaly, cysts in the germi-al matrix, and cerebellar atrophy.1,4,5 In addition, US is theafest imaging technique for the evaluation of the newbornrain. It is also useful as a screening tool, available at theedside, and feasible even in the most critical situations,hereas CT is not usually available in neonatal intensive carenits, and newborns must be transported. Further, US is the

galovirus infection according to neural ultrasound

h normaln � 45)

Children with pathologicalUS results (n � 12) P value

5/7 NS.5) 38.0 (2.2) NS76) 2875 (767) NS8) 10 (83) �.001

) 2 (17) NS) 8 (67) �.001

4 (33) .0016 (50) �.001

) 0 NS1) 8 (67) �.001

1 (8) NS

ome

witlts (

18/27.3 (2

54 (68 (1

1 (22 (400

1 (25 (10

ame imaging tool used in the perinatal period, making US-

ith Congenital Cytomegalovirus Infection 159

bc

wbwhwnTpniaCUb

sfiwltfTwo

cctatSci

istilm

danntihwac

Ts

DDDDSISN

Tc

DMSD

P* p data

1

ased prognostic data useful to physicians involved in prenatalare.

Postnatal cerebral US was performed in both patientsith symptoms and patients without symptoms in our cohortecause cerebral lesions have been detected even in patientsithout symptoms.24,25 All newborns with symptoms whoad abnormal US results showed at least 1 sequela or died,hereas none of the newborns with symptoms who hadormal US results had any long-term sequelae (P � .001).hese findings are in agreement with those reported by Bop-ana et al, who used cerebral CT scanning to establish prog-osis in a population of neonates with symptoms.3 Our data

ndicate a good correlation between cerebral US abnormalitiesnd the prediction of outcome in newborns with congenitalMV infection who have symptoms at birth, suggesting thatS may be performed as the first diagnostic study to detectrain lesions in these patients.

Only 2 of 39 neonates who were infected but had noymptoms had cerebral US abnormalities in our study. Therst was a term newborn with ventriculomegaly associatedith a globular hyperechogenicity at the germinal matrix

evel. At the time of diagnosis, we were not able to differen-iate between a germinal matrix hemorrhage and a microin-arction of the germinal matrix suggestive of CMV infection.he finding was classified as abnormal; the patient’s outcomeas normal. Severe psychomotor and auditory deficits devel-

able II. Outcome in neonates with congenital cytomcanning findings and presence or absence of sympto

Outcomes (n)

Normal US re

Symptoms(n � 8)

eath —Q �85 and motor delay —Q �85 � SNHL —Q �85 � SNHL � motor delay —

NHL � strabismus —solated SNHL —trabismus —ormal 8

able III. Value of cranial ultrasound scanning in preytomegalovirus infection

Number (%) of newborns with poo

Normal US results Pathological

Q �85 0/45 8/11 (7otor delay 0/45 6/11 (5

NHL 3/45 (6.7%) 6/11 (5eath or any sequela 3/45 (6.7%) 11/12 (9

PV, Positive predictive value; NPV, negative predictive value; NE, could not be estimOne newborn with pathological US results died during the neonatal period. Follow-u

ped in the second newborn, who had ventriculomegaly, p

60 Ancora et al

alcifications, and cysts. Adverse psychomotor outcome asso-iated with SNHL occurred in 1 of 39 infants without symp-oms (2.6%). US permitted identification of cerebral lesions,ssociated with poor outcome, early in this patient’s life. Inhe asymptomatic population, US failed to predict isolatedNHL in 3 patients. Our data are too limited to reach anyonclusions on the usefulness of cerebral US in congenitallynfected newborns without symptoms at birth.

Although we did not find changes in the course of timen calcifications, ventriculomegaly, cerebellar hypoplasia, andubependymal cysts, follow-up US can serve as a means torack poroencephalic cysts or hemorrhages. None of the casesn which the first US results were negative showed furtheresions at subsequent examinations, indicating that repeat US

ay not be warranted in these children.MRI may be a particularly useful imaging modality for

etecting white matter lesions or gyral abnormalities, whichre relatively common in congenital CMV infection after theeonatal period.2,28,29 MRI also is considered to be the besteuroimaging technique for studying abnormalities of migra-ion.30 US has a low sensitivity in detecting gyral abnormal-ties or myelinization deficits because the conic shape of USinders exploration of the cortical surface and the grey andhite matter are difficult to distinguish.31 MRI detected gyral

nd white matter disorders not disclosed by US in 6 of ourases. As a complementary examination, MRI may be best

lovirus infection according to neural ultrasoundat birth

(n � 45) Pathologic US results (n � 12)

o symptoms(n � 37)

Symptoms(n � 10)

No symptoms(n � 2)

— 1 —— 3 —— 2 —— 2 1— 1 —3 — —

— 1 —34 — 1

ing outcome in 57 patients with congenital

come Odds ratio(95% CI) P value PPV NPVesults*

) NE �.001 72.7% 100%) NE �.001 54.5% 100%) 16.8 (3.2–89.0) �.001 54.5% 93.3%) 154 (17.3–1219.6) �.001 91.7% 93.3%

were available for 11 of the 12 patients who lived.

egams

sults

N

dict

r out

US r

2.7%4.5%4.5%1.7%

ated.

erformed after the first 6 months of life to study the mye-


lMi

1NS2i13Nw4cA5Ga6ec7GpA8ad29cc1i1ni1d1oi

1i1i1P1A1a1M2s2a22Pc2c2An12i92ac2i2wv22PtR3A3

C

inization pattern. It has yet to be defined whether and whenRI should be performed in infants with congential CMV

nfection who have no symptoms.

REFERENCES. Volpe JJ. Viral, protozoan, and related intracranial infections. In: Volpe, JJ, ed.eurology of the newborn. 3rd ed. Philadelphia, London, Toronto, Montreal,ydney, Tokyo: Saunders Company; 1995. p. 675-729.. Barkovich AJ, Lindan CE. Congenital cytomegalovirus infection of the brain:maging analysis and embryologic considerations. AJNR Am J Neuroradiol994;15:703-15.. Boppana SB, Fowler KB, Vaid Y, Hedlund G, Stagno S, Britt WJ, et al.euroradiographic findings in the newborn period and long-term outcome in childrenith symptomatic congenital cytomegalovirus infection. Pediatrics 1997;99:409-14.. Malinger G, Lev D, Zahalka N, Ben Aroia Z, Watemberg N, Kidron D, et al. Fetalytomegalovirus infection of the brain: the spectrum of sonographic findings. AJNRm J Neuroradiol 2003;24:28-32.

. de Vries LS, Gunardi H, Barth PG, Bok LA, Verboon-Maciolek MA,roenendaal F. The spectrum of cranial ultrasound and magnetic resonance imaging

bnormalities in congenital cytomegalovirus infection. Neuropediatrics 2004;35:113-9.. Noyola DE, Demmler GJ, Nelson CT, Griesser C, Williamson WD, Atkins JT,t al. Early predictors of neurodevelopmental outcome in symptomatic congenitalytomegalovirus infection. J Pediatr 2001;138:325-31.. Whitley RJ, Cloud G, Gruber W, Storch GA, Demmler GJ, Jacobs RF, et al.anciclovir treatment of symptomatic congenital cytomegalovirus infection: results of a

hase II study. National Institute of Allergy and Infectious Diseases Collaborativentiviral Study Group. J Infect Dis 1997;175:1080-6.

. Kimberlin DW, Lin CY, Sanchez PJ, Demmler GJ, Dankner W, Shelton M, etl. Effect of ganciclovir therapy on hearing in symptomatic congenital cytomegalovirusisease involving the central nervous system: a randomized, controlled trial. J Pediatr003;143:16-25.. Lanari M, Lazzarotto T, Venturi V, Papa I, Gabrielli L, Guerra B, et al. Neonatalytomegalovirus blood load and risk of sequelae in symptomatic and asymptomaticongenitally infected newborns. Pediatrics 2006;117:e76-83.0. Lazzarotto T, Varani S, Guerra B, Nicolosi A, Lanari M, Landini MP. Prenatalndicators of congenital cytomegalovirus infection. J Pediatr 2000;137:90-5.1. Allan WC, Hold PJ, Sawyer LR, Tito AM, Meade SK. Ventricular dilation aftereonatal periventricular-intraventricular hemorrhage: natural history and therapeutic

mplications. Am J Dis Child 1982;136:589-93.2. Coley BD, Rusin JA, Boue DR. Importance of hypoxic/ischemic conditions in theevelopment of cerebral lenticulostriate vasculopathy. Pediatr Radiol 2000;30:846-55.3. Cabanas F, Pellicer A, Morales C, Garcia-Alix A, Stiris TA, Quero J. New pattern
f hyperechogenicity in thalamus and basal ganglia studied by color Doppler flowmaging. Pediatr Neurol 1994;10:109-16.
gD

ranial Ultrasound Scanning and Prediction of Outcome in Newborns w

4. Wang HS, Kuo MF, Chang TC. Sonographic lenticulostriate vasculopathy innfants: some associations and a hypothesis. AJNR Am J Neuroradiol 1995;16:97-102.5. Weber K, Riebel T, Nasir R. Hyperechoic lesions in the basal ganglia: anncidental sonographic finding in neonates and infants. Pediatr Radiol 1992;22:182-6.6. Brunet O, Lézine I. Le développement psychologique de la petite enfance. Paris:UF; 1965.7. Grossman HJ. Manual on terminology and classification. Washington: Americanssociation on Mental Deficiency; 1973.

8. Milani-Comparetti A, Gidoni FA. Routine developmental examination in normalnd retarded children. Dev Med Child Neurol 1967;9:631-8.9. Brazelton TB, Nugent JK. Neonatal behavioral assessment scale. 3rd ed. London:ac Keith Press; 1995.

0. Prechtl HFR. State of the art of a new functional assessment of the young nervousystem. An early prediction of cerebral palsy. Early Hum Dev 1997;50:1-11.1. Lanari M, Lazzarotto T, Papa I, Venturi V, Bronzetti G, Guerra B, et al. Neonatalortic arch thrombosis as a result of congenital cytomegalovirus infection. Pediatrics001;108:E114.2. Stagno S, Pass RF, Cloud G, Britt WJ, Henderson RE, Walton PD, et al.rimary cytomegalovirus infection in pregnancy: incidence, transmission to fetus, andlinical outcome. JAMA 1986;256:1904-8.3. Saigal S, Lunyk O, Larke B, Chernesky MA. The outcome in children withongenital cytomegalovirus infection. Am J Dis Child 1982;136:896-901.4. Williamson W, Percy AK, Yow MD, Gerson P, Catlin FI, Koppelman ML, et al.symptomatic congenital cytomegalovirus infection: audiologic, neuroradiologic, andeurodevelopmental abnormalities during the first year. Am J Dis Child990;144:1365-8.5. Williamson W, Demmler GJ, Percy AK, Catlin FI. Progressive hearing loss innfants with asymptomatic congenital cytomegalovirus infection. Pediatrics 1992;0:862-6.6. Fowler KB, McCollister FP, Dahle AJ, Boppana S, Britt WJ, Pass RF. Progressivend fluctuating sensorineural hearing loss in children with asymptomatic congenitalytomegalovirus infection. J Pediatr 1997;130:624-30.7. Huda W, Chamberlain CC, Rosenbaum AE, Garrisi W. Radiation doses tonfants and adults undergoing head CT examinations. Med Phys 2001;28:393-9.8. Haginoya K, Ohura T, Kon K, Yagi T, Sawaishi Y, Ishii KK, et al. Abnormalhite matter lesions with sensorineural hearing loss caused by congenital cytomegalo-irus infection: retrospective diagnosis by PCR using Guthrie cards. Brain Dev002;24:710-4.9. van der Knaap MS, Vermeulen G, Barkhof F, Hart AA, Loeber JG, Weel JF.attern of white matter abnormalities at MR imaging: use of polymerase chain reaction

esting of Guthrie cards to link pattern with congenital cytomegalovirus infection.adiology 2004;230:529-36.0. Barkovich AJ, Chuang SH, Norman D. MR of neuronal migration anomalies.m J Radiol 1988;150:179-87.

1. Pellicer A, Cabanas F, Perez-Higueras A, Garcia-Alix A, Quero J. Neural mi-
ration disorders studied by cerebral ultrasound and color Doppler flow imaging. Archis Child Fetal Neonatal Ed 1995;73:F55-61.
ith Congenital Cytomegalovirus Infection 161

Oc

StmMt

Rcta

Cw

Toabata

cbwFi(afAbbm

CEFFGIL

1

Comparative Efficacy and Cost of Asthma Care in Children with AsthmaTreated with Fluticasone Propionate and Montelukast

DAVID A. STEMPEL, MD, DENISE T. KRUZIKAS, PHD, AND RANJANI MANJUNATH, MSPH

bjective To assess the comparative efficacy of fluticasone propionate (FP) and montelukast (MON), using administrativelaims for pediatric asthma in a clinical setting.

tudy design This retrospective observational study used the PharMetrics Integrated-Outcomes Database. Children age 4o 17 years with an ICD-9-CM 493.xx for asthma, therapy with an inhaled corticosteroid in the 12 months before the indexedications, and an index claim for FP or MON between January 2001 and December 2003 were studied. FP- andON-treated children were propensity-matched based on health care utilization. Asthma-related parameters studied included

reatment failure, hospitalizations, and total cost of care.

esults The children treated with MON were more likely to experience treatment failure (odds ratio [OR] � 2.55; 95%onfidence interval [CI] � 2.19 to 2.96) and to be admitted to the hospital for asthma-related care (OR � 1.99; 95% CI � 1.15o 3.44) compared with those treated with FP. Furthermore, the children treated with MON incurred significantly highersthma-related treatment costs compared with those treated with FP (parameter estimate � 0.418; P < .0001).

onclusions In children with asthma, treatment with FP is associated with better outcomes and lower cost than treatmentith MON. (J Pediatr 2007;150:162-7)

he National Asthma Education and Prevention Program (NAEPP)1 recommends inhaled corticosteroids (ICSs) as thepreferred treatment for children with mild persistent asthma. Specifically, the NAEPP reports that for the long-termmanagement of asthma in children, “studies comparing inhaled corticosteroids to cromolyn, nedocromil, theophylline,

r leukotriene receptor antagonists are limited, but available evidence shows that none of these long-term control medicationsppear to be as effective as inhaled corticosteroids in improving asthma outcomes.”1 This recommendation is further supportedy the 2005 Global Initiative for Asthma (GINA).2 In addition, observational studies in children and adults have demonstrateddecreased risk of major exacerbations associated with the use of ICS.3-5 The Cochrane Review on initial pediatric controller

herapy concluded that ICS therapy is the treatment of choice for mild persistent asthma compared with leukotriene receptorntagonists (LTRAs).6 Since the publication of that report, 3 additional clinical trials have compared ICS and LTRA.7-10

The Childhood Asthma Research and Education Network7,8 conducted a studyomparing fluticasone propionate (FP) and montelukast (MON) in children with meanaseline forced expiratory volume in 1 second (FEV1) of 94% to 96% predicted. MONas associated with more treatment failures requiring oral corticosteroids (P � .019), andP produced a significantly greater improvement in multiple measures of lung function,

ncluding FEV1 (6.8% and 1.9%; P � .001), and a greater reduction in exhaled nitric oxideeNO) (P � .0028).7,8 A second direct comparison of FP and MON in children withsthma and baseline FEV1 of 75% to 76% predicted demonstrated the superiority of FPor lung function, decreased rescue albuterol use, and reduced nighttime symptom scores.9

third study comparing FP and MON in children with mild persistent asthma and aaseline FEV1 of 87% to 88% affirmed the primary endpoint of noninferiority for MONut demonstrated an increased relative risk of 1.56 (95% CI � 1.17 to 2.06) for rescueedication use (excluding short-acting �-agonist use), greater use of �-receptor agonists

I Confidence intervalD Emergency departmentEV1 Forced expiratory volume in 1 secondP Fluticasone proprionateINA Global Initiative for Asthma

CS Inhaled corticosteroid

LTRA Leukotriene receptor antagonistMON MontelukastNAEPP National Asthma Education and Prevention

ProgramOCS Oral corticosteroidOR Odds ratio

From Infomed Northwest, Bellevue, WA;Department of Pediatrics, University ofWashington, Seattle, WA; NDCHealth,Phoenix, AZ; and Global Health Out-comes, GlaxoSmithKline, Research TrianglePark, NC.

Supported by GlaxoSmithKline. Dr Stem-pel is a consultant to GlaxoSmithKline,Dr Kruzikas is an employee of NDCHe-alth, and Ms Manjunath is an employee ofGlaxoSmithKline. All 3 authors were in-volved in study design, analysis, writing, andthe decision to submit this manuscript.

Submitted for publication May 31, 2006;last revision received Aug 1, 2006; ac-cepted Oct 26, 2006.

Reprint requests: David A. Stempel, MD.E-mail: [email protected].



ABA Long-acting beta agonist SABA Short-acting beta agonist

62

10.1016/j.jpeds.2006.10.069

(i1dbi

ilLdsicm

D

cc7cdrNqidhfse

aapsc

S

lMapFoEmcb

gc

etpwcacp(mc9(nt

S

mwmoalpae(tswpc

M

cgCabrm

mv(rttm1

C

P � .003), 68% greater use of oral corticosteroids, andncreased relative risk of an asthma attack (1.26 [95% CI �.04 to 1.52]) for MON compared with FP.10 These studiesemonstrate the consistent superiority of FP over MON for aroad range of both objective and subjective clinical endpointsn children with a range of baseline lung function.

Along with clinical studies, several observational studiesn a general population of asthmatics have demonstratedower cost and better outcomes11-13 with ICSs compared withTRAs. Only 1 study was not able to demonstrate theseifferences between FP and MON.14 We designed a databasetudy comparing health outcomes and costs to gain insightnto the treatment differences between FP and MON inhildren age 4 to 17 years with a diagnosis of asthma receivingaintenance treatment with controller medications.

METHODS

atabaseThe study was based on medical and pharmaceutical

laims in the PharMetrics Patient-Centric Database, whichovers roughly 36 million unique patients from approximately5 health plans across the United States. The database in-ludes both inpatient and outpatient diagnoses and proce-ures, as well as both standard and mail order prescriptionecords; available data on prescription records include theational Drug Code (NDC), as well as days supplied and

uantity dispensed. All medical and pharmaceutical claimsnclude dates of service. Additional data elements includeemographic variables (eg, age, sex, geographic region),ealth plan type (eg, health maintenance organization, pre-erred provider organization), payer type (eg, commercial,elf-pay), provider specialty, and start and stop dates for plannrollment.

The database is in compliance with the Health Insur-nce Portability and Accountability Act (HIPAA) of 1996nd contains encrypted unique patient identifiers to integrateharmacy and medical claims. All analyses were based onecondary databases in which patient information was en-rypted and patient privacy protected.

ample SelectionPatients were selected for inclusion based on the fol-

owing criteria: (1) an index prescription claim for FP orON between January 1, 2001 and December 31, 2003; (2)

ge 4 to 17 years at the time of the index claim; (3) at least 1rescription claim for ICS within 12 months before the indexP or MON claim; (4) a documented asthma diagnosis basedn codes of the International Classification of Diseases, Ninthdition, Clinical Modification (ICD-9-CM; code 493.xx) 6onths before and/or 12 months after the index FP or MON

laim; and (5) continuous plan enrollment for 12 monthsefore and 12 months after the index prescription claim.

Exclusion criteria were applied to ensure a more homo-eneous population and allow the evaluation of patient out-
omes based on asthma treatment regimen. Patients were a
omparative Efficacy and Cost of Asthma Care in Children with Asthma

xcluded if they had a switch or augmentation of FP or MONherapy within 60 days of the index claim or if they hadrescription claims for a non-ICS asthma-controller medicationithin 12 months before the index FP or MON claim. Medi-

ations considered non-ICS controllers included LTRAs, long-cting beta agonists (LABAs), methylxanthines, andromones. Patients were excluded from analyses if they hadrescription claims for omalizumab, high-dose FP/salmeterol500 �g/50 �g), or high-dose FP (220 �g) during the 12onth pre- or post-period, or diagnoses for the following

onditions during the 2-year study: cystic fibrosis (ICD--CM codes 277.xx), chronic obstructive pulmonary diseaseICD-9-CM codes 491.xx, 494.xx, 496.xx), bronchopulmo-ary dysplasia (ICD-9-CM codes 770.7x), or respiratory dis-ress syndrome (ICD-9-CM codes 769.xx).

tudy GroupsTwo treatment cohorts of interest were studied: FP

onotherapy and MON monotherapy. Propensity matchingas used to match the FP and MON cohorts. Propensityatching is a methodology used to match study cohorts based

n specific characteristics before conducting multivariatenalyses, to reduce the potential for selection bias and patient-evel differences between comparison groups. FP and MONatients were propensity-matched based on pre-index patientge, previous health care utilization (hospitalizations andmergency department [ED] visits), use of oral corticosteroidsOCSs) and short-acting beta agonists (SABAs), season ofhe index prescription date, US Census region, physicianpecialty, and diagnosis of allergic rhinitis. Physician specialtyas defined as either primary care (ie, pediatrician, generalractitioner, family practitioner or internal medicine,) or spe-ialist (ie, allergist, pulmonologist, ear/nose/throat).

etricsPatient cohorts were compared with respect to patient

haracteristics to confirm the comparability of treatmentroups. The patient characteristics included age, sex, USensus region, and payer. Comparability was further evalu-

ted by analyzing health care utilization during the 12 monthsefore the index prescription claim, focusing on asthma-elated hospitalizations, ED visits, noncontroller asthmaedication use, and asthma-related total treatment costs.

Three outcomes were defined for evaluation: (1) treat-ent failure (defined as asthma-related hospitalizations, ED

isits, OCS use, or switch/augmentation of index therapy);2) asthma-related hospitalization alone; and (3) asthma-elated costs. Patients were followed for up to 12 months afterhe index prescription date or until switch/augmentation ofherapy, whichever occurred first. Costs were calculated as perember per month and annualized to provide estimates of

2-month costs if patients were discontinued due to switch/
ugmentation.
Treated with Fluticasone Propionate and Montelukast 163

S

crtc

freamtcawmmppiu

i(cawpbdMppMP

shvrrIcitoiM

DT

m

pmcrphhFwwcmp.

spClg[

Tp

nAFP

C

P

R

P

SM

*†

1

tatistical AnalysisBivariate analyses compared FP and MON treatment

ohorts with respect to patient demographics (age, sex, oregion), plan type, and previous health care utilization. Con-inuous variables were analyzed using Student t-tests, andategorical variables were evaluated using �2 analyses.

Multivariate regression techniques controlled for con-ounding factors while evaluating the impact of treatmentegimens on patient outcomes. Logistic models were used toxamine separately the likelihood of treatment failure as wells asthma-related hospitalizations, and generalized linearodel regression analyses were used to model asthma-related

otal treatment costs. Nonlinear transformations of treatmentosts were done to account for heteroscedasticity in the vari-nces. Because direct interpretation of log-transformed resultsas not possible, smearing methods were used to estimate theagnitude of the effects of independent variables. For allultivariate analyses, treatment cohort was the main inde-

endent variable of interest. All models further controlled foratient age, sex, US Census region, payer type, season ofndex prescription date, comorbidities, previous health caretilization, and previous asthma-related total treatment costs.

RESULTSA total of 4974 children were available for analyses,

ncluding 2961 (60%) patients treated with FP and 201340%) patients treated with MON. From this group, 3647hildren were propensity-matched with 50% in each of the FPnd MON treatment cohorts. The mean age of the sampleas approximately 10 years, and approximately 40% of theopulation was female. Table I shows the similarities inaseline demographics between the 2 cohorts. Significantifferences by payer are seen. Compared with the FP cohort,ON patients were more likely to be covered by commercial

lans (68% vs 58%; P � .0001) or to pay for expenses out ofocket (8% to 5%; P � .0002), and were less likely to haveedicaid or other/unknown payer coverage (13% vs 21%;� .0001 and 12% vs 16%; P � .0003, respectively).

Analyses of previous health care utilization confirmedimilarities between treatment cohorts with regard to previousospitalizations and ED visits, physician specialty, and pre-ious treatment with OCSs, SABAs, and claims for allergichinitis and antihistamine products. During the baseline pe-iod, the patients indexed to FP had significantly higher use ofCSs during the 12-month pre-period (3.36 claims vs 1.92laims; P � .0001), whereas the patients indexed to MONncurred significantly higher pre-period asthma-related totalreatment costs ($833 vs $671; P � .0005). The mean numberf dispensings of the index medication (standard deviation [SD])n the post-index year was significantly higher (P � .0001) for

ON, 4.03 (3.58), compared with 2.46 (2.77) for FP.

ifferences in Asthma-Related Outcomes Based onreatment Regimen

Significant differences in outcomes between the 2 treat-
ent groups were found before adjusting for any potential a
64 Stempel, Kruzikas, and Manjunath

atient-level differences, as shown in Table II. There wereore treatment failures in the MON cohort than in the FP

ohort (49% vs 26%; P � .0001). In the pre-period (whileeceiving ICS therapy), the 2 indexed treatment groups ex-erienced a similar number of hospitalizations for asthma;owever, in the post-period, significantly more children wereospitalized in the MON group (2.4%) compared with theP group (1.5%; P � .0445). The percentages of childrenith ED visits and claims for OCS in the post-index periodere 7% and 12%, respectively, for both the FP and MON

ohorts (P � not significant). Furthermore, annualized asth-a-related mean costs were significantly higher for the MON

atients ($1316) compared with the FP patients ($861; P �0001).

These observed differences between the cohorts per-isted after adjusting for confounding variables, such as age,ayer, previous utilization, and previous asthma-related costs.hildren in the MON group were more than 2.5 times more

ikely to experience treatment failure than those in the FProup (odds ratio [OR] � 2.55; 95% confidence intervalCI] � 2.19 to 2.96). In addition, the risk of a switch in or an

able I. Patient baseline characteristics andre-period health care utilization

FP MON P

1,837 1,837ge, mean (SD)* 9.6 (4.0) 9.5 (4.3) .5811emale, # (%) 663 (36) 687 (37) .4115revious health care

utilization, # (%)Hospitalizations* 91 (5) 99 (5) .5015ED visits* 209 (11) 212 (12) .8765

laim for allergic rhinitis*# (%)

505 (28) 516 (28) .1988

revious medicationclaims, mean (SD)

ICS 3.4 (2.4) 1.9 (1.5) .0001OCS* 0.3 (0.7) 0.3 (0.7) .6384SABA* 2.9 (2.9) 2.9 (3.0) .6562Antihistamines 1.7 (2.7) 1.6 (2.6) .6942

egion, # (%)*Northeast 387 (21) 416 (23) .2470Midwest 1,118 (61) 1,091 (59) .3630South 164 (9) 162 (9) .9076West 168 (9) 168 (9) 1.000

ayer type, # (%)Commercial 1,065 (58) 1,245 (68) .0001Medicaid 391 (21) 230 (13) .0001Self-pay 94 (5) 150 (8) .0002Other/unknown 287 (16) 212 (12) .0003

pecialty care*, † (%) 655 (36) 689 (38) .2442ean asthma-related

total treatment costs(SD)

$671 ($1000) $833 ($1216) .0005

Criteria used for propensity matching.Defined as patients. treated by allergist, pulmonologist, and ear/nose/throat specialists.

ugmentation of therapy was 3 times more likely with MON


ciMt(

mMmeof1a8c

aaa0at$coF

aocmotTdomci

t

agyefitMciF

eupIapmcur

lncwfraaa

stps

Tp

NT

HA

T

MPMSPPP

*Ucoa

C

ompared with FP (OR � 3.20; 95% CI � 2.75 to 3.76). Anncreased risk of treatment failure was also associated with

edicaid coverage, self-pay, previous ED care, previous OCSherapy, and higher pre-period asthma treatment costsTable III).

When hospitalizations were examined as a separateeasure, adjusted multivariate analyses confirmed that theON-treated children were nearly twice as likely to be ad-itted to the hospital for asthma-related care after the index

vent (OR � 1.99; 95% CI � 1.15 to 3.44) An increased riskf hospitalization was also associated with other independentactors, including Medicaid coverage (OR � 2.38; 95% CI �.27 to 4.46), self-pay (OR � 2.66; 95% CI � 1.06 to 6.68),nd previous hospitalizations (OR � 4.32; 95% CI � 2.18 to.58). The likelihood of hospitalization decreased with in-reasing age (OR � 0.92; 95% CI � 0.86 to 0.99).

In addition to the increased risk for treatment failurend hospitalization seen in the MON cohort, these childrenlso incurred significantly higher asthma-related costs afterdjusting for independent factors (parameter estimate �.418; P � .0001) (Table IV). Smearing techniques revealedn adjusted mean cost of $910 (SD � $364) for the FP-reated children compared with an adjusted mean cost of1382 (SD � $552) for the MON-treated children. Afterontrolling for confounding factors, asthma-related costs weren average 52% higher in MON-treated patients than inP-treated patients.

DISCUSSIONThis observational study in children with a diagnosis of

sthma who were treated with FP or MON has assessed bothutcome measures and costs and demonstrates that FP whenompared to MON is associated with a lower risk of treat-ent failure (ED/hospitalization and switch in/augmentation

f therapy), a decreased risk of asthma-related hospitaliza-ions, and significantly lower cost of asthma care in children.hese findings are consistent with those of several previousouble-blind controlled clinical trials demonstrating betterutcomes associated with FP compared with MON,7-10 2eta-analyses demonstrating better outcomes with ICSs

ompared with LTRAs,6,15 and recommendations of the ev-dence-based NAEPP and GINA guidelines.1,2

The strength of our study is that patients in each

able II. Asthma-related outcomes in the 12-monthost-period

FP MON P

umber of children 1,837 1,837reatment failure, mean(%)

485 (26) 904 (49) .0001

ospitalization, # (%) 28 (1.5) 45 (2.4) .0445nnualized meanasthma-related totaltreatment costs (SD)

$861 ($1804) $1316 ($3851) .0001

reatment cohort were well matched on various demographic s


nd utilization-related factors. This is particularly importantiven the complexities of care in pediatric asthma, becauseounger age of the child,16 previous hospitalization and EDvents, and previous use of OCSs and SABAs are all identi-ed as important variables that may increase the risk ofreatment failure or subsequent health care utilization.17

edicaid and self-pay are also important variables that in-rease the risk of treatment failure compared with commercialnsurance. The higher rate of Medicaid insurance in theP-treated children may have biased the results against FP.

Asthma control may be highly variable,18 and there isvidence indicating that appropriate therapy with consistentse of ICS-containing controller drugs is important in im-roving outcomes.19-21 Our study suggests that continuingCS treatment in the year after the claim was associated withdecreased risk of treatment failure and hospitalization com-ared with changing from ICS to MON. The summeronths are commonly reported as a period of improved

ontrol with lower rates of hospitalization.22 Children whosed ICSs during the summer months may be at decreasedisk for asthma exacerbations in the fall.23

An additional strength of our study was the use of aarge commercial claims database that provides an opportu-ity to analyze cost data and determine the costs of asthmaare. The improved outcomes associated with FP in our studyere also associated with lower costs, making FP the pre-

erred medication in this comparison. Some concern has beenaised about using pharmacy costs, due to varying contractualrrangements.24 However, the data contained in this analysisre from more than 75 plans across the country and report thectual costs paid by the plans inclusive of co-payments.

We also reported that adherence with both controllers wasuboptimal, although it was significantly greater with MONhan with FP. The greater adherence to MON therapy did notrotect against the greater risk for hospitalization or the need towitch or augment controller therapy for asthma. The finding of

able III. Likelihood of treatment failure*

Referentgroup

Oddsratio

95% confidencelimits

Lower Upper

ON FP 2.55 2.19 2.96atient age — 1.03 1.01 1.05edicaid Commercial 1.42 1.16 1.76

elf-Pay Commercial 1.58 1.17 2.14revious ED visits — 1.20 1.03 1.40revious OCS use — 1.55 1.29 1.86revious asthma-related totalcosts

— 1.20 1.10 1.30

Statistically significant (P � .05) results are reported. The model also controlled for sex,S Census region, other payer (other/unknown), season of index prescription claim,

omorbidities (acute respiratory infection, otitis media, pneumonia, allergic rhinitis,ther respiratory disease), previous hospitalizations, and previous ICS, SABA, andntihistamine use.

uperior outcome associated with FP compared with MON is


sps

esdnbptwTabcaeotdMc

tctatLntabia

t

spgsrbptetmtmetnIo

1wrmpvc

1GJ2vs13td4

T

MPFSMSOSPPPP

*ou

1

upported by additional pediatric reports and recommendationsublished after the guidelines7-10,25 and previous observationaltudies in the general population.12,13,26,27

Our results differ from those reported by Allen-Rameyt al.14 Although both that study and our study used propen-ity matching, the 2 studies had several differences in studyesign. The Allen-Ramey et al study did not require a diag-osis of asthma, but allowed patients to be included on theasis of 2 or more prescriptions of either MON or FP; 19% ofatients were included based on pharmacy claims only, andhis may have included some patients in the MON group whoere being treated for other diseases, such as allergic rhinitis.here is some uncertainty as to whether the results can be

ttributed to asthma or to asthma therapy. In a second studyy these same authors, they again did not require a medicallaim for asthma.28 Both studies do not report the cost ofsthma care. In addition, patients in both of the Allen-Rameyt al studies were excluded from the analysis if they switchedr augmented therapy. However, changes in medication pat-erns can be termed a treatment failure. In our study, it wasemonstrated that nearly twice as many children treated withON required an additional or alternative controller medi-

ation compared with those treated with FP.Furthermore, our study required active ICS therapy in

he year before the index medication, along with a diagnosticode for asthma, which strongly increased the likelihood thathe therapies were being compared for the treatment ofsthma. Alternative controllers were not selected, becausehere is minimal use of cromones and theophylline, andTRAs may be prescribed for both asthma and allergic rhi-itis. The modest use of SABAs in both cohorts suggests thathis population of children had a milder form of persistentsthma. The higher use of ICSs by the FP cohort in theaseline year may suggest that this cohort had greater diseasentensity, but yet the results demonstrate less treatment failurend better outcomes at lower cost with FP.

One of the limitations of observational claims data is

able IV. Asthma-related total treatment costs*

Referent group

ON FPatient age —emale Maleouth Northeastedicaid Commercial

elf-pay Commercialther/unknown payer Commercial

pring Winterrevious hospitalizations —revious ICS use —revious antihistamine use —revious asthma-related total costs —

Statistically significant results (P � .05) are reported. The model also controlled for patief index prescription date, comorbidities (acute respiratory infection, otitis media, pneumse.

he lack of information on disease severity. However, propen-I1

66 Stempel, Kruzikas, and Manjunath

ity matching based on resource utilization ensures that pre-eriod ED, inpatient, and OCS use were similar betweenroups and serves as a surrogate for establishing comparableeverity or asthma control in the study population. The 5%ate of hospitalization and 11% rate of ED visits in theaseline period are representative of the general pediatricopulation.16 Another limitation associated with observa-ional claims data is the inability to demonstrate cause andffect of an intervention. However, resource utilization andhe pattern of medication dispensing in pediatric practice areore generalizable than those observed in controlled clinical

rials, in which the study subjects are often highly selected anday be atypical of practice settings. For example, the differ-

nces in acute exacerbations may be minimized in a clinicalrial design that allows for patients to use additional or alter-ative controllers, but are captured in an observational design.n general, the results from controlled clinical trials andbservational studies are comparable.29,30

In conculsion, this study found that in children age 4 to7 years with a diagnosis of asthma and previous treatmentith ICS, treatment with FP resulted in significantly lower

isk of asthma-related hospitalization, decreased risk of treat-ent failure, and a lower cost of asthma-related care com-

ared with treatment with MON. The results of this obser-ational study are consistent with the findings of randomizedlinical trials.

REFERENCES. National Asthma Education and Prevention Program, Expert Panel Report.uidelines for the diagnosis and management of asthma update on selected topics, 2002.Allergy Clin Immunol 2002;110:s141-219.. Global Initiative for Asthma. Global strategy for asthma management and pre-ention. NIH Publication 02-3659, issued January 1995, updated 2002; managementegment (chap 7) updated 2005. Available at www.ginasthma.org. Accessed December, 2005.. Wennergren G, Kristjánsson S, Stannegård IL. Decrease in hospitalizations forreatment of childhood asthma with increased use of anti-inflammatory treatment,espite an increase in the prevalence of asthma. J Allergy Clin Immunol 1996;97:742-8.. Donahue JG, Weiss ST, Livingston JM, Goetsch MA, Greineder DK, Platt R.

arameter estimate Standard error P

0.418 0.039 �.00010.014 0.005 .0036

�0.079 0.036 .0287�0.195 0.089 .0275

0.225 0.053 �.00010.401 0.076 �.00010.140 0.054 .0100

�0.102 0.042 .0153�0.280 0.074 .0002

0.089 0.040 .02430.127 0.037 .00060.346 0.021 �.0001

, sex, US Census region, payer (commercial, Medicaid, self-pay, other/unknown), seasonprevious hospitalizations and ED visits, and previous OCS, SABA, and antihistamine

P

nt ageonia),

nhaled corticosteroids and the risk of hospitalization for asthma. JAMA997;227:887-91.


http://www.ginasthma.org

5l6cC7ec8Ra9Ra1cm1om1aA1aA1sA1m1S1o

1c1l2s22a2v22aJ2ma2fiI2cJ2om2cc2c3

C

. Suissa S, Ernst P, Kezouh A. Regular use of inhaled corticosteroids and theong-term prevention of hospitalisation for asthma. Thorax 2002;57:880-4.. Ducharme FM, Di Salvio F. Anti-leukotriene agents compared to inhaled corti-osteroids in the management of recurrent and/or chronic asthma in adults and children.ochrane Database Syst Rev 2, 2005. Accessed January 31, 2006 through DVID.

. Szefler SJ, Phillips BR, Martinez FD, Chinchilli VM, Lemanske RF, Strunk RC,t al. Characterization of within-subject responses to fluticasone and montelukast inhildhood asthma. J Allergy Clin Immunol 2005;115:233-42.. Zeiger RS, Szefler SJ, Phillips BR, Schatz M, Martinez FD, Chinchilli VM, et al.esponse profiles to fluticasone and montelukast in mild-to-moderate persistent childhood

sthma. J Allergy Clin Immunol 2006;117:45-52.. Ostrum NK, Decotis BA, Lincourt WR, Edwards LD, Hanson KM, Carranzaosenzweig JR, et al. Comparative safety and efficacy of low-dose fluticasone propionate

nd montelukast in children with persistent asthma. J Pediatr 2005;147:213-20.0. Garcia Garcia ML, Wahl U, Gilles L, Swern A, Tozzi CA, Polos P. Montelukast,ompared with fluticasone, for control of asthma among 6- to 14- year-old patients withild asthma: the MOSAIC study. Pediatrics 2005;116:360-69.

1. Stempel DA, Mauskopf, J, McLaughlin T, Yazdani C, Stanford RH. Comparisonf asthma costs in patients starting fluticasone propionate compared to patients startingontelukast. Resp Med 2001;95:227-234.

2. Stempel DA, Pinto L, Stanford RH. The risk of hospitalization in patients withsthma switched from an inhaled corticosteroid to a leukotriene receptor antagonist. Jllergy Clin Immunol 2002;110:39-41.

3. Orsini L, Limpa-Amara S, Crown WH, Kamal K. Asthma hospitalization risknd costs for patients treated with fluticasone propionate vs montelukast. Ann Allergysthma Immunol 2004;92:523-9.

4. Allen-Ramey FC, Duong PT, Riedel AA, Markson LE, Weiss KB. Observationaltudy of the effects of using montelukast vs fluticasone in patients matched at baseline. Annllergy Asthma Immunol 2004;93:373-80.

5. Sin DD, Man J, Sharpe H, Gan WQ, Paul Man SF. Pharmacological manage-ent to reduce exacerbations in adults with asthma. JAMA 2004;292:367-76.

6. Mannino DM, Homa DM, Akinbami LJ, Moorman JE, Gwynn C, Redd SC.urveillance for asthma: United States, 1980-1999. MMWR 2002;51:1-14.
7. Blais L, Ernst P, Biovin J-F, Suissa S. Inhaled corticosteroids and the preventionf readmission to hospital for asthma. Am J Respir Crit Care Med 1998;158:126-32. s

8. Stempel DA, McLaughlin TP, Stanford RH, Fuhlbrigge AL. Patterns of asthmaontrol: a 3-year analysis of patient claims. J Allergy Clin Immunol 2005;115:935-9.9. Suissa S, Ernst P, Kezouh A. Regular use of inhaled corticosteroids and theong-term prevention of hospitalization for asthma. Thorax 2002;57:880-4.0. Schatz M, Cook EF, Nakahiro R, Petitti D. Inhaled corticosteroids and allergypecialty care reduce emergency hospital use for asthma. J Allergy Clin Immun003;111:503-8.1. Lanes SF, Garcia Rodriquez LA, Huerta C. Respiratory medications and risk ofsthma death. Thorax 2002;57:683-7.2. Pendergraft TB, Stanford RH, Beasley R, Stempel DS, McLaughlin T. Seasonalariation in asthma-related hospital and intensive care unit admissions. J Asthma005;42:265-71.3. Johnston NW, Johnston SL, Duncan JM, Greene JM, Kebadze T, Keith PK, etl. The September epidemic of asthma exacerbations in children: a search for etiology.Allergy Clin Immunol 2005;115:132-8.4. Luskin A, Bukstein D, Sazonov Kocevar V, Yin DD. Asthma rescue and allergyedication use among asthmatic children with prior allergy prescriptions who initiated

sthma controller therapy. Ann Allergy Asthma Immunol 2005;95:129-36.5. Kelly HW. Inhaled corticosteroids versus leukotriene receptor antagonists asrst-line therapy for mild persistent asthma in children? Pediatr Asthma Allergymmunol 2005;18:259-63.6. Stempel DA, Meyer JW, Stanford RH, Yancey SW. One year claims analysisomparing inhaled fluticasone propionate with zafirlukast for the treatment of asthma.Allergy Clin Immunol 2001;107:94-8.7. Stempel DA, Mauskopf J, McLaughlin T, Yazdani C, Stanford RH. Comparisonf asthma costs in patients starting fluticasone propionate compared to patients startingontelukast. Resp Med 2001;95:227-34.

8. Allen-Ramey FC, Anstatt DT, Sajjan SG, Markson LE. Asthma-related healthare resource use among patients starting fluticasone or montelukast therapy. Pharma-otherapy 2005;25:1752-60.9. Benson K, Hartz AJ. A comparison of observational studies and randomizedlinical trials. New Engl J Med 2000;342:1887-92.0. Concato J, Shah N, Horowitz RI. Randomized, controlled trials, observational
tudies, and the hierarchy of research designs. New Engl J Med 2000;342:1878-86.

Oc

Spc

Rtaamwba

Chp(

Hochatddif

iaw

AHEG

1

Clinical Spectrum and Histopathologic Features of Chronic Hepatitis CInfection in Children

PARVATHI MOHAN, MD, CAMILLA COLVIN, MPH, BS, CHEVELLE GLYMPH, MPH, ROMA R. CHANDRA, MD, DAVID E. KLEINER, MD,KANTILAL M. PATEL, PHD,† NAOMI L.C. LUBAN, MD, AND HARVEY J. ALTER, MD

bjective To define the natural history and outcomes of children infected with hepatitis C virus (HCV) at birth or in earlyhildhood.

tudy design This retrospective, prospective study identified 60 HCV-infected children through a transfusion look-backrogram (group 1) and by referrals (group 2). Perinatal/transfusion history, clinical course, and laboratory studies wereorrelated with findings from 42 liver biopsy specimens.

esults Mean age at infection was 7.1 months, and duration of infection 13.4 years. The sources of infection were bloodransfusion (68%), perinatal transmission (13%), and both (7%). Most patients were asymptomatic; three referral patients haddvanced liver disease at presentation. Mean alanine aminotransferase level was normal in 25%, 1 to 3 times normal in 62%,nd greater than 3 times normal in 13%. Liver biopsy specimens showed minimal to mild inflammation in 71%, absent orinimal fibrosis in 88%, and bridging fibrosis in 12%. Age at infection and serum gamma-glutamyltranspeptidase correlatedith fibrosis; serum alanine aminotransferase correlated with inflammation unless complicated by comorbidity. Repeatiopsies within 1 to 4 years in four patients showed no significant progression in three and cirrhosis in one. Two patients diedfter liver transplantation.

onclusions Children with chronic HCV infection are generally asymptomatic. By 13 years after infection, 12% of patientsad significant fibrosis. Patients enrolled by referral had more severe liver disease than those identified through the look-backrogram, demonstrating the importance of selection bias in assessing the long-term outcome of HCV infection.J Pediatr 2007;150:168-74)

epatitis C virus (HCV) progresses insidiously and incrementally and results inpotentially serious complications such as cirrhosis and hepatocellular carcinomain approximately 20% and 4% of adult patients, respectively.1-4 However, data

n the natural history and histopathology of HCV-related liver disease in children areonflicting.5-13 Studies from Japan and Europe point to relatively benign clinical andistopathologic liver disease,14-20 whereas studies from the United States suggest a moreggressive course, with development of early fibrosis.21-23 Geographic variation in geno-ypes and diversity of the infected population studied may account for some of theseifferences, in addition to as-yet-unknown viral/host factors. The lack of uniformity in theescriptions of the natural history, clinical presentation, and histologic features of HCV

nfection in children prompted us to evaluate a cohort of 60 HCV-infected childrenollowed in our institute over a 5-year period.

The broad objectives of this study were to identify children with HCV who werenfected perinatally or during early childhood, follow their natural history throughdolescence, correlate clinical and laboratory data with liver histology, identify childrenho would benefit from medical therapy, and follow treatment or natural outcomes.

LT Alanine aminotransferaseCV PCR Hepatitis C polymerase chain reactionIA Enzyme immune assay

HAI Histologic Activity IndexLKM Liver kidney microsomal antibodiesRIBA Recombinant immunoblot assay

From the Departments of Gastroenterol-ogy and Nutrition and Laboratory Medicineand Pathology, Center for Health Servicesand Community Research, Children’s Re-search Institute, Children’s National Medi-cal Center, the George Washington Uni-versity School of Medicine, Washington,DC; and the Department of TransfusionMedicine, the National Institutes of Health,and the Laboratory of Pathology, NationalCancer Institute, Bethesda, Maryland.

The study was funded in part by a grant(RO1 HL 56060) from the National Insti-tutes of Health.

Submitted for publication Dec 29, 2005;last revision received Oct 5, 2006; acceptedNov 9, 2006.

Reprint requests: Parvathi Mohan, MD, De-partment of Gastroenterology and Nutri-tion, Children’s National Medical Center,111 Michigan Avenue NW, Washington,DC 20010. E-mail: [email protected].†Deceased.



GTP Gamma-glutamyltranspeptidase ULN Upper limit of normal

68

10.1016/j.jpeds.2006.11.037

S

tttjcHedepob(vjicvasfwidmdbtr

L

eAsCAtmMVIgpbftitADm

ARtHc

L

pvpwrMuMtsaciilb

S

sostwaletmgHtt1c

D

fsTiw

C

METHODS

tudy SubjectsChildren (n � 2100) who received blood/blood product

ransfusions between 1982 and 1992 were tested for HCVhrough a look-back program established at Children’s Na-ional Medical Center, Washington, DC. The primary ob-ective was to study the natural history of HCV infection inhildren who had no other risk factors for liver disease thanCV acquired through transfusions in the newborn period or

arly in life. Patients with sickle cell disease, hemophilia, renalisease, and malignancy within 5 years of treatment werexcluded from the study. Study subjects included 30 HCV-ositive children who were transfused before the availabilityf anti-HCV testing, identified, and recalled by the look-ack program (group 1). In addition, we studied 30 childrengroup 2), identified through referrals for evaluation of ele-ated liver enzymes and/or a known history of hepatitis C oraundice. Included in the referral group were 6 adopteesdentified as HCV-positive during evaluation by their primaryare physician. Potential sources of HCV infection were in-estigated through a questionnaire administered to parentsnd to children above the age of 14 years. Parental bloodamples were obtained for anti-HCV testing for all patientsrom group 1 and those patients from group 2, when thereere no obvious risk factors for HCV infection. The onset of

nfection was considered to be the date of a transfusion or theate of any surgery before 1992 or the date of birth when theother was anti-HCV positive; in the case of adopted chil-

ren, the age at infection was presumed to be the date ofirth. Patients were recruited and followed from 1996hrough 2001; the study was approved by the institutionaleview board of our institute.

aboratory DataHCV status was evaluated through the use of an HCV

nzyme immune assay (EIA, second and third generation,bbott Laboratories, Chicago, IL) and confirmed with a

econd-generation recombinant immunoblot assay (RIBA 2,hiron Strip Immunoblot assay, Chiron, Emmeryville, CA).nti-HCV positive or indeterminate patients were further

ested for qualitative and quantitative HCV RNA by poly-erase chain reaction (RT-PCR and Cobas Amplicor HCVonitor v2.0 assay, Roche Diagnostics, Branchburg, NJ).

iral genotype was identified by line-probe assay (InnoLiPa,nnogenetics, Belgium). Alanine aminotransferase (ALT),amma-glutamyltranspeptidase (GGTP), creatinine phos-hokinase, albumin, and serum iron were measured in serumy Ektachem 500 (Ortho Diagnostics, Raritan, NJ). Alpha-etoprotein was measured by EIA (Immuno I, Bayer, Tarrey-own, NY). At the initial visit, other causes of liver diseasesncluding alpha-1-antitrypsin deficiency, Wilson disease, au-oimmune hepatitis, and coinfections with HIV and hepatitis

and B were investigated. An abdominal sonogram withoppler study of portal venous flow was performed to identify
ass lesions in the liver, splenomegaly, and portal hypertension. r
linical Spectrum and Histopathologic Features of Chronic Hepatitis C I

Patients had follow-up visits at 4- to 6-month intervals.nti-HCV EIA/RIBA, HCV PCR, and quantitative HCV-NA were done once or twice a year unless signs and symp-

oms prompted more frequent analyses. A negative test forCV RNA on at least two samples tested within 1 year was

onsidered as evidence for viral clearance.

iver BiopsyAfter informed consent was obtained, liver biopsy was

erformed if the ALT was 2 times the age-related normalalue on two or more visits or if autoimmune markers wereresent. If the ALT was less than 11⁄2 times normal, a biopsyas deferred during the initial 6 to 12 months but was

ecommended subsequently, within the 5-year study period.ost of the liver biopsy specimens were obtained by a closed,

ltrasound-guided procedure, using 16- to 18-gauge Bardonopty needles. Transjugular biopsies were performed if

he patients had thrombocytopenia. All specimens weretained by hematoxylin and eosin, Masson trichrome, periodiccid-Schiff, Perl stain for iron, and Rhodamine stain foropper. Two independent pathologists, blinded to the clinicalnformation, evaluated the biopsy specimens, using the scor-ng systems of Batts and Ludwig24 and the Knodell Histo-ogic Activity Index (HAI).25 The ALT levels at the time ofiopsy were recorded.

tatistical AnalysisThe continuous data were summarized as mean �

tandard deviation (SD), range, and minimum and maximumbservations. The nominal, categorical, or ordinal data wereummarized by count and/or proportion (%). With the use ofhe Wilcoxon and Kruskall-Wallis statistics, the two groupsere compared with respect to the averages of variables such

s age an infection, age at biopsy, duration of infection,aboratory values, and biopsy grade and stage.26 To assess theffects of these covariates on the dependent variables, namelyhe histologic grade and stage, cumulative logistic regressionodels with proportional odds ratios were fitted to each

roup and to the two groups of patients pooled together. TheAI grade was compressed into three ordinal values to reduce

he number of proportional odds ratio assumptions needed inhe model, that is, HAI grade 0 for the original grades 0 to 4,

for 5 to 8, and 2 for 9 and above. A value of P � .05 wasonsidered significant.

RESULTS

emography and Clinical FeaturesDemographics, source of infection, and follow-up data

or the 60 patients are given in Table I. The majority ofubjects were infected in infancy at a mean age of 7.1 months.he mean age at diagnosis was 11.9 years, and duration of

nfection to the end of the study was 13.4 years. Transfusionas the primary source of infection, occurring in 68%. The

easons for HCV testing included a history of transfusion in

nfection in Children 169

4psatf(mpdb

B

sSwltTw((ppitpa(nw

m1ov

f5bsStttVggvbm

R

vsig((s

T

SAADDES

C

PAADD

1

8%, elevated aminotransferase levels in 28%, HCV-positivearent in 8%, screening of adoptees in 8%, and blood donorcreening and acute jaundice in 2% each. Patients were mostlysymptomatic. Symptoms, when reported, were mild andransient and were similar in both groups. They includedatigue in 17 patients (28%), diffuse abdominal pain in 1017%), nosebleeds in 7 (12%), poor weight gain in 7 (12%),uscle aches in 5 (8%), and headaches in 2 (3%). Two

atients from group 2 presented with acute, self-limited jaun-ice and two others with edema of the legs, shortness ofreath, and laboratory evidence of chronic liver failure.

iochemical and Virologic IndexesThe ALT varied widely among the patients and in the

ame patient at different times during the study period (mean �D � 98.0 � 115.7 U/L). The mean values from each patientere expressed as normal, 1 to 11⁄2 times age-adjusted upper

imit of normal (UNL), 11⁄2 to 3 times ULN, and above 3imes the ULN (Table II; available at www.jpeds.com).wenty-eight of 60 (47%) patients had mean ALT levelsithin the normal range or below 1.5 times the ULN, 24

40 %) had mean ALT levels 11⁄2 to 3 times the ULN, and 813%) had mean values above 3 times the ULN. A fewatients skewed the data with inordinately high ALT atresentation, although ALT levels generally normalized dur-ng the follow-up period. They included an infant with ver-ically acquired HCV (1786 U/L), a patient with myopathy27

resumed to be secondary to HCV infection (1005 U/L), andtransfused infant coinfected with HIV and cytomegalovirus

801 U/L). Serum bilirubin, albumin, and iron levels wereormal in all patients except in two, who presented initially

able I. Population characteristics of 60 HCV-infecte

Demographics (mean � SD, *P < .05)

ex (male/female)ge at infection (mo)ge at diagnosis (y)uration of follow-up (y) (n � 60)uration of infectionthnicity (white/black/Hispanic/Russian/others/mixed)ource of infectionTransfusion alonePerinatalTransfusions � perinatalUnknown

oexisting conditionsSequelae of prematurityCardiac/congenital anomaliesHematologic/past malignancyViral hepatitis/HIV

atients lost to follow-upge at infection (mean in mo)ge at diagnosis (mean in y)uration of infection (mean in y)uration of follow-up (mean in y)

ith acute obstructive jaundice and early liver failure. The v

70 Mohan et al

ean ALT during the follow-up period (group 1 � 86.3 �13.6 U/L vs group 2 � 109.6 � 95.4 U/L) and at the timef liver biopsy (Table IV) was significantly higher in group 2ersus group 1 patients.

The majority of patients (n � 50, 83%) tested positiveor antibodies to HCV and HCV RNA throughout the-year study period. One patient was HCV RNA–positiveut EIA-negative on two consecutive tests 6 months apart butubsequently converted to seropositivity after 24 months.28

even anti-HCV and HCV RNA–positive patients were losto follow-up after the initial testing. One patient each fromhe two groups was HCV RNA–positive at enrollment buthereafter remained negative for the duration of the study.iral genotype was performed on 45 patients (group 1 � 26,roup 2 � 19); 80% of group 1 and 78% of group 2 were ofenotype 1 and the rest were of genotype 2 and 3. The meaniral load at the time of liver biopsy did not differ significantlyetween groups 1 and 2, and only 15% had viral loads �2illion copies/mL.

adiologic Findings

Abdominal sonogram with Doppler study of the portalenous system was performed in 43 of 60 (72%) patients. Thetudy was normal in 27 children (63%). Abnormal findingsncluded mild splenomegaly in 5 patients (12%), gallstones/all bladder sludge in 3 (7%), and abnormal echogenicity in 25%). Splenomegaly was found only in the referral groupgroup 2). Abdominal CT scans in 2 of 5 patients withplenomegaly showed nodularity of the liver and esophageal

ildren

otal Group 1 (30) Group 2 (30)

0/30 16/14 16/14� 16.9 6.92 � 17.8� 7.3 � 16.1� 14.1 11.53 � 1.4� 8.74 � 5.0� 1.7 2.99 � 1.5 3.45 � 2.1� 4.12 14.9 � 2.9 11.8 � 4.6

3/1/1/0 10/8/1/6/3/2

(68%) 28 13(13%) — 8(7%) 2 2(12%) — 7

9 23 69 7 24 0 43 0 3(17%) 7 (12%) 3 (5%)1.62 0.82 3.022.0 11.25 13.55.7 14.71 17.502.0 1.83 2.25

d ch

T

37.1111.93.3

13.3515/1

41847

2

10

11

arices.


H

HbsCecPgtfl9wasli1sisfty

sopmoao

CL

iprtga.hfsgsh

C

iulpdlnefrn

tHotitateotof

THn

P

L

P

T

S

*

C

istologyForty-two initial 1iver biopsy specimens were scored by

AI25; the results are summarized in Table III. These 42iopsy specimens and 2 additional biopsy specimens were alsocored by using the classification of Batts and Ludwig.24

orrelation between the two systems of classification wasxcellent, both for the grade and stage of histology (Spearmanorrelation for grade: r � 0.75, P � .0001; for stage: r � 0.62,� .0001; Pearson correlation linear regression analysis for

rade: r � 0.78; and stage r � 0.79). Overall, the inflamma-ory changes (the sum of portal, periportal, and lobular in-ammation) were mild (grade 0 to 8) in 71%, moderate (gradeto 12) in 24%, and severe (grade 13 to 18) in 5%. Fibrosisas absent or mild, with only periportal expansion in 88%,

nd bridging fibrosis was seen in 12%. All of the patients withevere inflammation (5%) and bridging fibrosis (12%) be-onged to the referral group. None of the initial biopsy spec-mens showed cirrhosis. Lymphoid aggregates were present in8 of 41 (44%). Mild steatosis was identified in four biopsypecimens (10%) and bile duct involvement (Poulson lesions)n three (7%). Stainable iron was detected in the biopsypecimens from two patients who had received multiple trans-usions for acute leukemia and thalassemia major. Four pa-ients had repeat liver biopsy procedures over a span of 1 to 4

able III. Histopathologic findings on liver biopsy byAI scoring system (group 1: n � 19; group 2:� 23)

Score Group 1 Group 2 Total (%)

eriportal inflammationNone 0 1 2 (7)Mild 1 12 7 (45)Moderate 3 6 10 (39)Marked 4-6 0 4 (10)

obular inflammationNone 0 0 0 (0)Mild 1 7 3 (24)Moderate 3 5 6 (24)Marked 4 7 14 (50)

ortal inflammationNone 0 5 3 (19)Mild 1 13 12 (60)Moderate 3 1 8 (21)Marked 4 0 0 (0)

otal grade (0-18)*Minimal 0-4 6 4 (24)Mild 5-8 10 10 (47)Moderate 9-12 3 7 (24)Marked 13-18 0 2 (5)

tage fibrosis (0-4)None 0 11 6 (40)Periportal/expansion 1 8 12 (48)Bridging fibrosis 3 0 5 (12)Cirrhosis 4 0 0 (0)

Sum of portal, periportal, and lobular inflammation.

ears. Three biopsy specimens showed no histologic progres- o


ion; one showed progression to early cirrhosis over the coursef 4 years. Bridging fibrosis was seen in five of group 2atients. Combining the initial and follow-up biopsy speci-ens, bridging fibrosis or early cirrhosis was found in 6 (14%)

f 42 patients undergoing percutaneous biopsy procedure. Inddition, two patients were found to have cirrhosis at the timef liver transplantation.

orrelation Between Histology, Demography, andaboratory Parameters

When the two groups were pooled together, age atnfection and serum GGTP levels were significantly andositively associated with HAI stage (P � .002 and P � .032,espectively). When the groups were examined separately,here were no statistically significant predictors of stage inroup 1, but age at infection was significantly and positivelyssociated with the stage of liver disease in group 2 (P �041). ALT at the time of biopsy correlated positively withistologic grade in group 1 (P � 0.021). No covariate wasound to affect the grade significantly in group 2. In summary,erum ALT was positively associated with histologic grade inroup 1, and age at infection was positively associated withtage in group 2. Table IV compares the laboratory andistologic findings between the two groups of patients.

linical OutcomeThe clinical outcome of the 60 patients is summarized

n the Figure. Of the 50 patients who had continuous follow-p, 47 (94 %) had clinical and laboratory evidence of mildiver disease. Two patients underwent liver transplantation forortal hypertension and chronic liver failure; both of themied of complications from the transplantation. The native

iver of one of them showed evidence of hepatocellular carci-oma. A third patient whose repeat biopsy specimen showedvidence of cirrhosis continued to exhibit deterioration in liverunction. The only patients with cirrhosis were from theeferral group and had acquired HCV infection through peri-atal transmission.

DISCUSSIONWe describe the demographic, virologic, and his-

opathologic data on 60 children and adolescents with chronicCV infection, evaluated during a 5-year period. The onset

f infection was reasonably established in 88%, based onransfusion history and/or perinatal transmission from HCV-nfected mothers. The mean duration of infection, based onhe history of exposure, was 13 years. Histologic data werevailable for 60% of the children, and only 13% had treatmenthat would alter the natural history of the infection. Wevaluated patients in two clinical settings. The primary focusf our study was HCV-infected children (group 1) who wereransfused at our institution in the years before the availabilityf testing for HCV. Group 2 included patients referred to usor evaluation of established liver disease. Although inclusion
f these patients introduced a referral bias, a prospective study

oocoopasfitaOillc2i

iagdTpdUplhoh

hfri

T

AADAGAVHH

*

F

1

f sufficient length to analyze the full spectrum of HCVutcomes in children is not feasible and hence we chose toombine a retrospective, prospective cohort analysis with thatf an equal number of referral cases in whom more severeutcomes were expected and indeed found. This dual ap-roach aimed to achieve a balanced approximation of the fullnd varied spectrum of HCV outcomes, as demonstrated intudies of HCV infection in adults.1-4 We were surprised tond that only 1.5% of the 2100 transfusion recipients inves-igated in the look-back study were HCV-infected despite thebsence of donor screening at the time of their transfusions.ther look-back studies in children29-31 have reported HCV

nfection in 0.3% to 6% of blood transfusion recipients. Theower-than-expected prevalence of HCV positivity in ourook-back study may have been due in part to our entryriteria that excluded children with concurrent diseases. Onlyof the 30 patients in group 1 appeared to have cleared HCV

nfection spontaneously.

able IV. Comparison of cohort and referral groups

VariableGroup 1 (cohort)

n � 19

ge at infection (mo) 4.6 � 11.8ge at biopsy (y) 15 � 1.8*uration of infection (y) 14.1 � 2.8*LT at biopsy 61 � 59.5GTP 34.1 � 21.4FP 3.7 � 3.7iral load (million) 1.7 � 2.8AI grade 5.1 � 2.1AI stage 0.42 � 0.5

P � .05.

igure. Clinical outcomes of 60 patients over the 5-year study period.

Clinical and laboratory evaluation of the identified cases c

72 Mohan et al

n our study showed an excellent correlation between ALTnd the histologic grade on liver biopsy in group 1. Patients inroup 2 had concurrent pathology such as hepatitis B andecompensated cirrhosis that may have affected the ALT.wenty-six of 28 children (93%) with normal ALTs or levelsersistently less than 11⁄2 times the ULN had only mild liverisease by biopsy. Patients with ALT greater than 3 times theLN generally had more severe pathology; however, twoatients with persistently normal ALTs also had advanced

iver disease. Although ALT was an excellent correlate of liveristology, discrepancy between ALT and histology in 7% ofur patients attests to the continued value of liver biopsy, asas been supported by other pediatric and adult studies.14,16,32-34

The characteristic features of chronic HCV-associatedepatitis found in adults, namely, lymphoid aggregates andollicles, macrovesicular steatosis, and bile duct damage, wereelatively rare in our series.35 Early cirrhosis was documentedn only one patient on a repeat biopsy procedure, and frank

he time of liver biopsy

Group 2 (referral)n � 23

Kruskal-Wallistest

7.3 � 16.5 0.1578.8 � 5.0 0.00018.3 � 4.5 0.0002

154.7 � 255.5* 0.001114.1 � 197.7 0.4926.2 � 78.2 0.172.0 � 3.5 0.097.4 � 3.4* 0.0121.2 � 1.1* 0.012

at t

irrhosis was found in two other patients at the time of liver


tcaaWithe(bwgcffitrvewacp7p

Ofa61rtatbibefR4thrwthsairict

metla7cystorfadlcc

tycdHitfit(

efitcprctfuoogaqpdcphtmdpa

C

ransplantation. Among four patients with repeat biopsy pro-edure, only one showed fibrosis progression; this patientlready had bridging fibrosis in the initial biopsy specimennd then advanced to early cirrhosis over the course of 4 years.

hen the biopsy specimens from the two groups were exam-ned separately, we found that the ALT level at the time ofhe liver biopsy procedure, and the grade and the stage ofistology were significantly higher in group 2 than in group 1,ven though the duration of infection was longer in the latterTable IV). Sixty percent of those with no (stage 0) fibrosiselonged to group 1, and all of the patients with stage 3-4ere from group 2. The increased severity of liver disease inroup 2 reflects, at least in part, a referral bias wherein theseases represented patients with established liver disease re-erred for further evaluation. In these patients, the degree ofbrosis correlated with the age at which they were infected:he older the age at infection, the worse the fibrosis. This ageelation was not seen in group 1 patients, possibly because theast majority was infected before age 2. In contrast, Guidot al36 studied the progression of fibrosis in untreated patientsho had acquired HCV infection in infancy and found that

ge at biopsy procedure and duration of infection positivelyorrelated with stage of fibrosis. In their study, 13 of 112atients had a repeat biopsy procedure at a mean interval of.9 years; 54% showed a 1-2 stage increase in fibrosis. Therogression was slower than reported in adult patients.37

Long-term cohort studies in children have been limited.ne of the earliest longitudinal studies involved 77 children

rom Europe who were HCV-positive after transfusion-relatednd community-acquired infection.13 When followed over a-year period, 27% had active but mild histologic liver disease,0% achieved biochemical remission, and only 3% had cir-hosis.13 The same group recently reported that 7% to 10% ofhese children had subsequent development of anti-LKMntibodies that were associated with a more severe outcome ofheir liver disease.38 Vogt et al18 emphasized the relativelyenign course of liver disease in persons who acquire HCVnfection early in life. They evaluated 458 children who hadeen transfused for cardiac surgery approximately 20 yearsarlier at a mean age of 2.8 years.18 Fifteen percent wereound to be anti-HCV–positive, but only 55% were HCVNA–positive, suggesting a spontaneous viral clearance in5%. Liver biopsy specimens in 17 chronically infected pa-ients showed mild histologic disease in (82%); only 1 patientad cirrhosis attributable to HCV infection.18 Casiraghi et al5

eported mild liver disease with slow progression in personsho acquired HCV infection through mini-transfusions in

he newborn period; 35 years after exposure, 9 of 11 (82%)ad minimal or no inflammatory activity or fibrosis and 2 hadtage 3-4 fibrosis. Repeat biopsy procedures in five patientsfter 5 years showed progression in only one patient.5 Studiesn young adults also demonstrate a high spontaneous recoveryate and predominantly mild histologic changes. Two studiesnvolving young women inadvertently exposed to HCV-ontaminated Rh immune globulin revealed cirrhosis in less
han 4% and bridging fibrosis in only 10% to 15% approxi- H

ately 25 years after the onset of HCV infection.4,39 Minolat al6 studied 392 patients with post-transfusion HCV infec-ion and calculated that the median time to develop end-stageiver disease was 33 years for those infected between ages 21nd 30 years and 16 years for those infected after age 40. In7 patients infected below the age of 20 years, only 4% hadirrhosis.6 These and other studies demonstrate that theounger the individual at the time of infection, the less theeverity of HCV-related liver disease.4,5,6,18,37 The reasons forhis observation are not known but may relate to more vig-rous immune responses to an infection acquired early in life,educed fibrogenic mechanisms in children and/or the con-ounding effects of alcohol, and obesity and coinfections indult patients.6,40 Although children tend to have more in-olent HCV infection than adults, the development of severe

iver disease can be accelerated in the presence of comorbidonditions such as thalassemia,9 iron overload,9,21 childhoodancer7,8,22,23 and HIV coinfection.40

Retrospective studies involving patients referred to ter-iary care centers for identified liver diseases demonstrateet another population of patients with more severe out-omes.21,22 Illustrative of these retrospective studies, Badiza-egan et al21 examined 50 liver biopsy specimens from 40CV-infected children ages 2 to 18 years and found signif-

cant fibrosis in 78%, including bridging fibrosis and archi-ectural distortion in 58% and cirrhosis in 8%. The degree ofbrosis correlated positively with age and duration of infec-ion. These study subjects are comparable to our group 2referral) patients.

We conclude that children who acquire HCV infectionarly in life generally manifest only mild liver disease over therst two decades of their infection. However, it is apparenthat severe liver disease may develop in select groups ofhildren infected with HCV. Despite the inherent bias ofatient selection, referral studies prove that cirrhosis andarely hepatocellular carcinoma are potential sequelae ofhildhood HCV infection and emphasize the need for iden-ification and careful follow-up of those with potential riskactors. However, only a very long-term, prospective study onnselected cohort groups can provide an accurate accountingf the relative proportion of children who manifest severeutcomes. It is uncertain at present whether the slow pro-ression of hepatitis C in immuno-competent children willccelerate as they grow older or will remain stable. Thisuestion has considerable implications for therapy. A slowlyrogressive course of HCV infection in children would favoreferred treatment, given that the multiple side effects of theurrently available therapy observed in adults might be com-ounded by growth retardation in children. On the otherand, if pediatric trials show that children tolerate antiviralherapy well with high sustained response rates, early treat-ent would be preferred over the potential for serious liver

isease decades later. Our natural history study and otherediatric studies10-20 suggest that treatment can be deferred todolescence or later as long as patients are followed closely.
owever, it is important that this premise be tested by

lpr

Wwp

1i2Lh234a35Lt6iB7h8Pl9oha1HJ1h21h1a11Sd1P1h11Cc1

ci1fc2hc2o12e92eo22A2Fa2I2hd2p2t3OI37323lC3Pc33h23Dh3H(3hs4

1

onger-term natural history studies and by carefully conductedediatric drug treatment trials to assess toxicity and sustainedesponse rates.

e thank Dr Albert Hoang and Dr Robert McCarter for their helpith the statistical analyses of our data during the revision of this

aper, following the demise of Dr Kantilal Patel.

REFERENCES. Alter HJ, Seeff LB. Recovery, persistence and sequelae in hepatitis C virusnfection: a perspective on long-term outcome. Semin Liv Dis 2000;20:17-35.. Seeff LB, Hollinger FB, Alter HJ, Wright EC, Cain CMB, Buskell ZL, et al.ong-term morbidity and mortality of transfusion-associated non-A, non-B, and type Cepatitis: a National Heart, Lung, and Blood Institute collaborative study. Hepatology001;33:455-63.. Seeff LB. Natural history of chronic hepatitis C. Hepatology 2002;36:S35-46.. Kinney-Walsh E. Clinical outcomes after hepatitis C infection from contaminatednti-D immune globulin: Irish Hepatology Research Group. N Engl J Med 1999;40:1228-33.. Casiraghi MA, Paschale MD, Romano L, Biffi R, Assi A, Binelli G, et al.ong-term outcome (35 years) of hepatitis C after acquisition of infection through mini

ransfusions of blood given at birth. Hepatology 2004;39:90-6.. Minola E, Prati D, Suter F, Maggiolo F, Caprioli F, Zogni A, et al. Age atnfection affects the long term outcome of transfusion-associated chronic hepatitis C.lood 2002;99:4588-91.. Cesaro S, Petris MG, Rossetti F, Cusinato R, Pipan C, Guido M, et al. Chronicepatitis C virus infection after treatment for pediatric malignancy. Blood 1997;90:1315-20.. Locasciulli A, Testa M, Pontisso P, Benvegnu L, Fraschini D, Corbetta A, et al.revalence and natural history of hepatitis C infection in patients cured of childhood

eukemia. Blood 1997;90:4628-33.. Lai ME, DeVirgilis S, Argiolu F, Farci P, Mazzoleni AP, Lisci V, et al. Evaluationf antibodies to hepatitis C virus in a long-term prospective study of post transfusionepatitis among thalassemic children: comparison between first and second generationssay. J Pediatr Gastroenterol Nutr 1993;16:458-64.0. Mohan P, Glymph C, Baxter C, Chandra RR, Kleiner DE, Luban NLC, et al.istopathology and epidemiology of hepatitis C in children: a single institution study.Pediatr Gastroenterol Nutr 2000;31:S115.1. Schwimmer JB, Balistreri WF. Transmission, natural history and treatment ofepatitis C virus infection in the pediatric population. Semin Liv Dis000;20:37-46.2. Aach RD, Yomtovian RA, Hack M. Neonatal and pediatric post transfusionepatitis C: a look back and a look forward. Pediatrics 2000;105:836-42.3. Bortolloti F, Jara P, Diaz C, Vajro P, Hierro L, Giacchino R, et al. Posttransfusionnd community-acquired hepatitis C in childhood. J Pediatr Gastroenterol Nutr994;18:279-83.4. Matsuoka S, Tatara K, Hayabuchi Y, Taguchi Y, Mori K, Honda H, et al.erologic, virologic, and histologic characteristics of chronic phase hepatitis C virusisease in children infected by transfusion. Pediatrics 1994;94:919-22.5. Kage M, Fujisawa T, Shiraki K, Tanaka T, Fugisawa T, Kimura A, et al.athology of chronic hepatitis C in children. Hepatology 1997;26:771-75.6. Guido M, Rugge M, Jara P, Hierro L, Giacchino R, Larrauri J, et al. Chronicepatitis C in children: the pathological and clinical spectrum. Gastroenterology998;115:1525-9.7. Garcia-Monzone C, Jara P, Fernandez-Bermejo M, Hierro L, Frauca E,amarena C, et al. Chronic hepatitis C in children: a clinical and immunohisto-

hemical comparative study with adult patients. Hepatology 1998;28:1696-701.8. Vogt M, Lang T, Frosner G, Klingler C, Sendl AF, Zeller A, et al. Prevalence and

Lc

74 Mohan et al

linical outcome of HCV infection in children who underwent cardiac surgery before themplementation of blood-donor screening. N Engl J Med 1999;341:866-70.9. Hoshiyama A, Kimura A, Fujisawa T, Kage M, Kato H. Clinical and histologiceatures of chronic hepatitis C virus infection after blood transfusion in Japanesehildren. Pediatrics 2000;105:62-5.0. Jara P, Resti M, Hierro L, Giacchino R, Barbera C, Zancan L, et al. Chronicepatitis C virus infection in childhood: clinical patterns and evolution in 224 whitehildren. Clin Infect Dis 2003;36:275-80.1. Badizadegan K, Jonas MM, Ott MJ, Nelson SP, Perez-Atayde AR. Histopathol-gy of the liver in children with chronic hepatitis C viral infection. Hepatology998;28:1416-23.2. Strickland DK, Reily CA, Patrick CC, Jones-Wallace D, Boyett JM, Waters B,t al. Hepatitis C infection among survivors of childhood cancer. Blood 2000;5:3065-70.3. Castellino S, Lensing S, Reily C, Rai SN, Davila R, Hayden RT, et al. Thepidemiology of chronic hepatitis C infection in survivors of childhood cancer: an updatef the St Jude Children’s Research Hospital hepatitis C seropositive cohort. Blood004;103:2460-6.4. Batts KP, Ludwig J. Chronic hepatitis: an update on terminology and reporting.m J Surg Pathol 1995;19:1409-17.

5. Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N, et al.ormulation and application of a numerical scoring system for assessing histologicalctivity in asymptomatic chronic active hepatitis C. Hepatology 1981;1:431-5.6. SAS Institute Inc. SAS Procedures Guide, Release 6.03 Ed, Cary, NC: SASnstitute Inc, 1988: The CORR Procedure, 125-150.7. Mohan P, Chandra RS, Escolar DM, Luban NLC. Inflammatory myopathy andepatitis C in a pediatric patient: role of liver biopsy in evaluating the severity of liverisease. Hepatology 2001;34:851-2.8. Mohan P, Chandra RS, Kleiner DE, Luban NLC. An unusual presentationerinatally transmitted hepatitis C. Arch Dis Child 2003;88:160-1.9. Roberts EA, King SM, Fearon M, McGee N. Hepatitis C in children afterransfusion: assessment by look-back studies. Acta Gastroenterol Belg 1998;61:195-7.0. Davoren A, Dillon AD, Power JP, Donnellan J, Quinn JM, Willis JW, et al.utcome of an optional HCV screening program for blood transfusion recipients in

reland. Transfusion 2002;42:1501-6.1. Goldman M, Long A. Hepatitis C lookback in Canada. Vox Sang 2000;8(suppl 2):249-52.2. Hoofnagle J. Hepatitis C: the clinical spectrum of disease. Hepatology 1997;6(suppl 1):15S-20S.3. McCormick SE, Goodman ZD, Maydonovitch CL, Sjogren MH. Evaluation ofiver histology, ALT elevation, and HCV RNA titer in patients with chronic hepatitis. Am J Gastroenterol 1996;91:1516-22.

4. Pradat P, Alberti A, Poynard T, Esteban JI, Weiland O, Marcellin P, et al.redictive value of ALT levels for histologic findings in chronic hepatitis C: a Europeanollaborative study. Hepatology 2002;36:973-7.5. Ishak KG. Pathologic features of chronic hepatitis. Am J Clin Pathol 2000;113:405.6. Guido M, Bortolotti F, Leandro G, Jara P, Hierro L, et al. Fibrosis in chronicepatitis C acquired in infancy: is it only a matter of time? Am J Gastroenterol003;98:660-3.7. Poynard T, Badossa P, Opolon P (OBSVIRC, METAVIR, CLINIVIR &OSVIRC groups). Natural history of liver fibrosis progression in patients with chronic

epatitis C. Lancet 1997;349:825-32.8. Bortolliti F, Muratori I, Jara P, Hierro L, Verucchi G, Giacchino R, et al.epatitis C virus infection associated with liver-kidney microsomal antibody type 1

LKM 1) autoantibodies in children. J Pediatr 2003;142:185-90.9. Weise M, Grñgreiff K, Gt̃hoff W, Lafrenz M, Oesen U, Porst H. Outcome in aepatitis C (genotype 1b) single source outbreak in Germany: a 25-year multicentertudy. J Hepatol 2005;43:590-8.0. Benhamou Y, Bochet M, DiMartino V, Charlotte F, Azria F, Coutellier A, et al.
iver fibrosis progression in human immunodeficiency virus and hepatitis C virusoinfected patients. Hepatology 1999;30:1054-8.

T

AAPAAA

*

C

able II. Comparison of ALT values among study co

ALT (U/L) (mean � SD) Total

LT during study period (n � 60) 98.0 � 115.LT at biopsy (n � 42) 131.5 � 22atients with normal ALT (%) 15 (25%)LT �1 to 11⁄2 times normal 13 (22%)LT �11⁄2 to 3 times normal 24 (40%)LT �3 times normal 8 (13%)

P � � .05.

horts

Group 1 Group 2

7 86.3 � 133.6 109.6 � 95.4*61.0 � 59.5 154.7 � 255.5*

11 45 8

12 124 4

linical Spectrum and Histopathologic Features of Chronic Hepatitis C Infection in Children 174.e1

Ow

Ssh

Rdad

Cc

Jbp

(sptdt

Fddsbefio

cWam

BF

Fetal Alcohol Spectrum Disorders in Young Adulthood

HANS-LUDWIG SPOHR, MD, JUDITH WILLMS, MD, AND HANS-CHRISTOPH STEINHAUSEN, MD, PHD

bjective To test the hypothesis that fetal alcohol syndrome (FAS), with the full phenotype, and fetal alcohol effect (FAE),ith some but not all of the features, can be combined under the umbrella term fetal alcohol spectrum disorders (FASD).

tudy design We investigated the long-term sequelae of intrauterine alcohol exposure using physical examination, psycho-ocial interviews, and a behavioral checklist in a 20-year follow-up study of 37 patients with FASD originally diagnosed asaving FAS or FAE in infancy and childhood.

esults Although the characteristic craniofacial malformations of FAS/FAE diminish over time, microcephaly, a poorlyeveloped philtrum and a thin upper lip, and, to a lesser degree, short stature and underweight (in boys) persist. In females,dult body weight increases. Persistent mental handicaps, including intellectual disability, limited occupational options, andependent living, are the major sequelae, and the scores for various behavioral problems are significantly increased.

onclusions The devastating effects of intrauterine exposure to alcohol persist into early adulthood and severely limitareers and independent living. (J Pediatr 2007;150:175-9)

ones and Smith were the first to describe a distinct pattern of craniofacial abnormalities and central nervous systemdysfunction in 11 children whose mothers were chronic alcoholics.1 These authors coined the term fetal alcohol syndrome(FAS) to describe an entity characterized by craniofacial abnormalities, growth retardation, delayed psychomotor maturation,and impaired intellectual development. This syndrome also was independently identified by French authors,2 and FAS

ecame recognized as a major cause of intellectual disability and was studied intensively with regard to such issues asathogenesis, development, and intervention. The findings have been documented in numerous publications.3-6

The teratogenic effects of alcohol on the developing fetus represent a continuum known as fetal alcohol spectrum disordersFASD). Updated methodologies for diagnosing FASD have been published recently,7 but traditionally, various diagnosticchemes have been used to categorize these adverse outcomes. Originally, FAS identified those children who displayed the fullhenotype, with fetal alcohol effects (FAE) applied to individuals demonstrating some, but not all of the features of FAS.8 Theerms FAS and FAE do not exactly correlate with some more recently updated diagnoses for the FASD. In this report, theiagnostic categories FAS and FAE are used, because these represent the diagnostic classifications available 20 years ago, whenhis study cohort was developed.

The estimated incidence of FAS is at approximately 1 per 1000 live births; that ofAE, in 3 to 5 per 1000 live births.9,10 FASD is a leading cause of marked developmentalelay, including intellectual disability and other psychopathologies, such as attentioneficit hyperactivity disorder (ADHD).3-6 Few studies have addressed the long-termequelae of FAS/FAE.6,12-15 More extensive studies include the secondary disability studyy Streissguth et al16,17 from the United States, the 21-year longitudinal analysis of theffects of prenatal alcohol exposure on young adult drinking by Baer et al,18 and our ownndings from more than 10 years of follow-up.19 Here we report on more than 20 yearsf follow-up of our original cohort.

METHODSIn our previous report on the 10-year follow-up, we presented findings on 60

hildren and adolescents living in Berlin and various other locales throughout Germany.19

e were planning for a long-term follow-up at approximately 20 years so as to studydaptation in young adulthood in patients who had been in care of the first author overany years. Of these 60 subjects, 52 qualified for a 20-year follow-up between the years

MI Body mass index FAS Fetal alcohol syndrome

From the Department of Pediatrics, Hospi-tal of the German Red Cross, Berlin, Ger-many and Department of Child and Ado-lescent Psychiatry, University of Zurich,Zurich, Switzerland.

None of the authors has any conflict ofinterest associated with this study.

Submitted for publication May 12, 2006;last revision received Oct 4, 2006; acceptedNov 17, 2006.

Reprint requests: Hans-Ludwig Spohr, MD,Department of Pediatrics, Hospital of theGerman Red Cross, DRL-Kliniken West-end, Spandauer Damm 130, D-14050 Ber-lin, Germany. E-mail: [email protected].



AE Fetal alcohol effect FASD Fetal alcohol spectrum disorders
10.1016/j.jpeds.2006.11.044
175

2cCTtwd[

fsbri

cnfwPm(d5ns

cRgcddcicvmh

waamr

f

1

2

3

o3scw

wia

loedf

T

PMDHMSTMLM

N

1

000 and 2003. Of these 52 subjects, 3 had died, 7 refused toooperate, and 5 had unknown addresses or changed names.onsequently, we were able to evaluate 37 former patients.here were no significant differences in the sample charac-

eristics between the 2 clinical samples (Table I; available atww.jpeds.com). All subjects were Caucasian, and the meanuration of follow-up was 20.5 years (standard deviationSD] � 2.8 years).

Intelligence was tested at initial examination after re-erral or during the early school-age years. There were 12ubjects with normal intelligence (IQ �85), another 10 withorderline intelligence (IQ 70 to 85), and 15 with mentaletardation (IQ �70) based on standardized intelligence test-ng using the Wechsler Intelligence Test.

To test any potential bias resulting from attrition, weompared the 12 subjects who refused to cooperate or wereot available at follow-up with the 37 participants in thisollow-up study. The 12 dropouts did not differ significantlyith regard to age (mean age, 26.3 years; SD, 2.1; t � 1.91;� not significant [NS]), sex distribution (7 males, 5 fe-ales; �2 � .07; P � NS), FAS versus FAE distribution

7 FAS vs 5 FAE; �2 � .01; P � NS), and intelligenceistribution (3 normal intelligence, 4 borderline intelligence,mentally retarded; �2 � .29; P � NS). Therefore, there waso evidence of any major bias resulting from attrition to theample.

The initial FAS diagnosis was made in accordance withriteria in use at the time of examination, as defined by theesearch Society on Alcoholism4: (a) prenatal or postnatalrowth retardation (height and weight below the 10th per-entile for age or gestational age; (b) central nervous systemysfunction (ie, any neurologic abnormality, developmentalelay, or intellectual impairment); and (c) characteristicraniofacial abnormalities, including at least 2 of the follow-ng: microcephaly (ie, head circumference below the 3rd per-entile), microphthalmia, short palpebral fissures, poorly de-eloped philtrum, thin upper lip, and flattening of theaxillary area. According to the Research Society on Alco-

able II. Signs and symptoms at initial examination

Initialexamination

n %

ostnatal growth deficiency 33 89.2icrocephaly 36 97.3evelopmental delay 33 89.2yperactivity 27 73.0uscular hypotonia 24 64.9

hort upturned nose 23 62.2hinned upper lip 31 83.8axilla hypoplasia/flat midface 23 62.2

imited joint movement 3 8.1inor external genital anomalies 14 37.8

S, not significant.

olism diagnostic criteria, the designation FAE is applied a

76 Spohr, Willms, and Steinhausen

hen a child shows 2 of criteria (a), (b), and (c). Maternallcohol abuse during pregnancy was ascertained by obtainingcareful history from the biological mother and/or familyembers, including the analysis of all available medical

ecords.At follow-up, the assessment procedure included the

ollowing:

. Physical examination (performed by H.L.S.), with specialemphasis on the characteristic craniofacial malformationsand measurement of growth variables according to Praderet al,20 leading to the calculation of the body mass index(BMI), including transformation into percentiles based onGerman norms according to Kromeyer-Hauschild et al.21

. A coded item list based on an interview (performed byJ.W.) dealing with academic and occupational career, do-mestic arrangements, and independent living.

. The Young Adult Behavior Checklist (YABCL),22 a ques-tionnaire for assessing emotional and behavioral problemsin young adults.

The interview was conducted with the subject’s caregiverr closest relative: a parent (in 8% of cases), a foster parent (in5% of cases), or a caring person in the institution where theubject lived at the time of the examination (41%). In 16% ofases, the subject’s partner was interviewed. The same informantho gave the interview filled out the YABCL.

Informed consent was obtained from all patients whoere competent, or from caregivers otherwise. Statistical tests

ncluded t tests, �2 tests, McNemar tests, and multivariatenalysis.

RESULTSSigns and symptoms at initial examination and fol-

ow-up are listed in Table II. Despite a substantial reductionf symptoms across time, a sizeable number of subjects stillxhibited growth retardation, microcephaly, developmentalelay, and hyperactivity. In contrast, the very marked cranio-acial features present early in life had mostly disappeared,

at follow-up

Follow-up

n % �2 P

14 37.8 41.6 �.00118 48.6 35.1 �.00115 40.5 36.4 �.00114 37.8 19.5 �.0014 10.8 112.3 �.0012 5.4 497.9 �.001

34 91.9 NS17 45.9 NS2 5.4 NS- 5.4 76.2 �.001

and

lthough an elongated philtrum and a thin upper lip were still



pscrjh

owwwsbrtt(.Bw.ne

.sh.(

ats

lfjsd[

vohofi

F

F

rominent. Various other physical and organ defects had beenurgically corrected before follow-up; operations includedleft palate repair (in 2 subjects), ventricular septal defectepair (in 3 subjects), complex heart defect repair (in 1 sub-ect), and minor operations for phimosis, cryptorchidism, andernias.

Many subjects, even some formerly severely affectednes, exhibited some degree of catch-up growth in height andeight and normalization of BMI (Table III; available atww.jpeds.com). As expected, much less catch-up growthas recorded for head circumference (Fig 1). Of the 31

ubjects (83.8%) with microcephaly (ie, head circumferenceelow the 3rd percentile) on first assessment, 17 (45.9%)emained microcephalic at follow-up. Comparing the propor-ions of subjects scoring below and above the 3rd percentile inhe total sample revealed significant changes in heightMcNemar exact P � .01), weight (McNemar exact P � �001), head circumference (McNemar exact P � .001), andMI (McNemar exact P � � .001). In males, these changesere significant only by trend for height (McNemar exact P �

07), significant for weight (McNemar exact P � .004), sig-ificant only by trend for head circumference (McNemar

igure 1. Follow-up of morphometric variables.

xact P � .06), and significant for BMI (McNemar exact P � r

etal Alcohol Spectrum Disorders in Young Adulthood

001). In females, the changes in the proportion of subjectscoring below the 3rd percentile were not significant foreight, but were significant for weight (McNemar exact P �

03), head circumference (McNemar exact P � .02), and BMIMcNemar exact P � 0).

Additional comparisons of the continuous BMI at firstssessment and follow-up are given in Table IV. A strongrend is seen for normalization in the total sample and the 2ex groups; however, this trend is less strong in males.

Furthermore, there is a clear correlation between intel-ectual disability at first assessment and reduced head circum-erence at follow-up, as shown in Table V (available at www.peds.com). The prevalence of microcephaly at follow-up wasignificantly greater in those subjects who were intellectuallyisabled at first assessment (�2 � 10.3; degrees of freedomdf] � 2; P � .006).

The psychosocial and career development interview re-ealed that 18 subjects (49%) had received special educationnly, 14 (38%) had passed primary school, and only 5 (13%)ad a secondary school education. By occupational status,nly 5 subjects (13%) had ever held an “ordinary” job. Thesendings were disappointing, because 25 subjects (69%) had
eceived at least some preparatory job training and 21 (58%)
177




hida

hdw8p

Ftgfa(0s1

traoep

ocAoip1oitwb

ln

nctatg

btitbSfllaw

pbcpAsa

F-

T

TMF

N

1

ad either started or progressed to formal occupational train-ng. However, 7 (19%) terminated job training prematurelyue to a mismatch between mental and personal capacitiesnd vocational skill requirements.

Assessment of domestic arrangements in young adult-ood showed that 27% lived in institutions, 35% were in aependent-living situation with assistance from others, 14%ere able to live independently alone, 8% lived with a partner,% had their own family, and 8% lived with his or her fatherlus a mother surrogate.

The YABCL profiles of the 8 subscales are shown inig 2. Because of the lack of local norms, the raw scores were

ransformed into z scores (mean � 0; SD � 1). For the totalroup, significant deviations from the mean are seen on theollowing subscales: thought disorder (t � 2.79; P � .01),ttention problems (t � 5.57; P � .001), intrusive behaviort � 3.52; P � .01), and aggressive behavior (t � 3.48; P �.1). The profiles for the 2 sex subgroups did not differignificantly on multivariate analysis (Wilks’ � � .63; F �.68; df � 8/23; P � NS).

DISCUSSIONIn contrast to previous cross-sectional studies,6,12,13,16,17

his prospective longitudinal and long-term outcome studyeports the consequences of prenatal alcohol exposure in ange-homogenous sample of young adults diagnosed with FASr FAE in childhood. The 20-year follow-up allows a definitestimation of remaining handicaps, because all subjects in theresent study are past childhood and adolescence.

Despite the physical changes, including disappearancef some of the features of FAS/FAE, many young adults stillan be identified based on their residual signs and symptoms.lthough some catch-up growth occurred, a large proportionf the subjects had growth deficiency. The apparent normal-zation of BMI is due predominantly to the fact that a largeroportion of the sample is getting fatter. As in our previous0-year follow-up,13 and in accordance with the follow-upbservations of Streissguth et al,16 males seemed to be morempaired than females in all aspects of growth deficiency. Athe same time, the development of growth variables in malesas less consistent, whereas females showed a strong postpu-ertal increase in body weight.

There was a significant association between early intel-ectual disability and persistent microcephaly, although a mi-

able IV. Comparison of BMI at first assessment and

Firstassessment Follo

Mean SD Mean

otal sample (n � 37) 10.9 4.1 21ales (n � 20) 10.2 3.1 19.9

emales (n � 17) 11.7 5.0 22.4

S, not significant.

ority of subjects with microcephaly had borderline or even p

78 Spohr, Willms, and Steinhausen

ormal intelligence. The association between intelligence inhildhood and head circumference at follow-up also matcheshe generally relatively high stability of intelligence over timend the clinically and developmentally useful classifications ofhe 3 major groups of normal intelligence, borderline intelli-ence, and mental retardation.

The vocational development of our cohort was markedy a large group of subjects who had received special educa-ion at school and later experienced unemployment or workedn sheltered workshops only. Despite the preparatory jobraining and occupational programs in which the subjects hadeen enrolled, a high rate of vocational disability was noted.imilarly, the handicapping effects of FAS/FAE were re-ected in the high proportion of young adults in assisted-

iving conditions and the small number of subjects who wereble to live independently. In contrast to Streissguth et al,16

e did not find major crime and incarceration in our sample.Finally, an assessment of emotional and behavioral

roblems showed that FAS/FAE subjects were characterizedy features predominately of an externalizing or disruptiveharacter and independent of sex effects. Most notably, theredominance of attention problems matches the findings ofDHD symptoms in FAS/FAE children found in our earlier

tudies23,24 and also reported by others.25,26 These problemsnd higher scores for aggressive and delinquent problems

igure 2. Profile of emotional and behavioral problems. – � –, males;-’- -, females; — � —, total.

low-up

SD t df P r P

3.5 �14.5 36 .000 .38 .023.4 �9.6 19 .000 .01 NS3.3 �11.0 16 .000 .61 .009

fol

w-up

oint to a high persistence from childhood to young adult-


hhiysiibtp

cpsiafsqhaffhtfwlfen

dnsimot

tsp

1L2a

3a415C6a7G8a9ev1i11Pc1aG1A1M1aa1oeS1RJ1lA1ds2b2ms2B12nP2i12pC2d

F

ood, as shown in the study by Streissguth et al,17 whereas theigher scores for thought problems and intrusive behavior

ndicate some newly emerging additional mental problems inoung adulthood. Additional analyses not presented here havehown that the specific features of the emotional and behav-oral problem profiles in these subjects were independent ofntellectual impairment and whether or not the subjects hadeen diagnosed with FAS or FAE. Thus, the subjects withhe minor variant FAE experienced similar handicaps androblems as those with FAS.

These findings are similar to those identified in theross-sectional study of Streissguth et al,13 who identified 8rotective factors that contributed to a less impaired life in aubsample of their subjects. Our longitudinal sample exhib-ted 7 of these 8 protective factors: (1) being diagnosed beforege 6 years (in 32 of 37 subjects); (2) living in a stable homeor over 70% of one’s lifetime; (3) staying in each livingituation for more than 2.8 years; (4) experiencing a good-uality home environment between age 8 and 12 years; (5)aving applied for therapeutic help and assistance; (6) havingdiagnosis of FAS rather than FAE, making early diagnosis

easible (in 22 of 37 subjects); and (7) having basic needs metor at least 13% of one’s lifetime. There was only a slightlyigher rate of subjects who had experienced violence againsthemselves in early childhood (3 of 37 subjects) (the eighthactor), whereas in the study by Streissguth et al,16 no subjectho experienced violence was included in the subsample with

ess impaired life. Although almost all of the 8 protectiveactors were also present in our study, the outcome in thentire sample was significantly impaired. Thus, our data doot corroborate the protective function of these factors.16

These differences may be due in part to transnationalifferences in patient selection, health care systems for diag-osing and treating FASD, and social systems, includingpecial education and rehabilitation. A limitation of our studys the relatively small sample of referred patients. Further-

ore, we know very little about any potential postnatal effectf the social environment on ultimate functioning other thanhe limited effects of rehabilitation.

This long-term follow-up study documents the devas-ating effects of intrauterine exposure to alcohol. Despiteubstantial efforts at rehabilitation, only a very small group ofatients was able to live a normal adult life. 11

REFERENCES. Jones KL, Smith DW. Recognition of the fetal alcohol syndrome in early infancy.ancet 1973;11:999-1001.. Lemoine P, Harousseau H, Boteyru JP, Menuet JC. Les enfant des parents
lcooliques: anomalies observées à propos de 127 cas. Quest Médicale 1968;25:476-82. A
etal Alcohol Spectrum Disorders in Young Adulthood

. Clarren SK, Smith DW. The fetal alcohol syndrome: experience with 65 patientsnd a review of the world literature. N Engl J Med 1978;298:1063-7.. Majewski F. Alcohol embryopathy: experience in 200 patients. Dev Brain Dysfunc993(6):248-65.. Spohr H-L, Steinhausen H-C. Alcohol, Pregnancy, and the Developing Child.ambridge, UK: Cambridge University Press; 1996.

. Streissguth AP, Aase JM, Clarren SK, Randels SP, La Due RA, Smith DF. Fetallcohol syndrome in adolescents and adults. JAMA 1991;265:1961-7.. Department of Health and Human Sciences (2004), Fetal Alcohol Syndrome:uidelines for Referral and Diagnosis. Washington, DC.

. Steinhausen HC, Winkler Metzke C, Spohr HL. Behavioural phenotype in fetallcohol syndrome and fetal alcohol effects. Dev Med Child Neurol 2003;45:179-82.. Sampson PD, Streissguth AP, Bookstein FL, Little RE, Clarren SK, Dehaene P,t al. Incidence of fetal alcohol syndrome and prevalence of alcohol-related neurode-elopmental disorders. Teratology 1997;56:317-26.0. Schoneck U, Spohr HL, Willms J, Steinhausen HC. Alcohol drinking andntrauterine dystrophia: effects and significance in infancy. Monatsschr Kinderheilkd992;140:34-41.1. Hoyme HE, May PA, Kalberg WO, Kodituwakkn P, Gossage JP, TrujilloM, et al. A practical clinical approach to diagnosis of fetal alcohol spectrum disorders:larification of the 1996 institute of medicine criteria. Pediatrics 2005;115:39-47.2. Aronson M, Hagberg B. Neuropsychological disorders in children exposed tolcohol during pregnancy: a follow-up study of 24 children to alcoholic mothers inöteburg, Sweden. Alcohol Clin Exp Res 1998;22:321-4.

3. Spohr HL, Willms J, Steinhausen HC. The fetal alcohol syndrome in adolescence.cta Paediatr Suppl 1994;404:19-26.

4. Autti-Ramö I. Twelve-year follow-up of children exposed to alcohol in utero. Deved Child Neurol 2000;42:406-11.

5. Autti-Ramö I, Fagerlung A, Ervalahti N, Loimu L, Korkman M, Hoyme E. Fetallcohol spectrum disease in Finland: clinical delineation of 77 older children anddolescents. Am J Med Genet 2006;140A:137-43.6. Streissguth AP, Barr HM, Kogan J, Bookstein FL. Understanding the occurrencef secondary disabilities in clients with fetal alcohol syndrome (FAS) and fetal alcoholffects (FAE). Final Report to the Centers for Disease Control and Prevention (CDC).eattle: University of Washington, Fetal Alcohol and Drug Unit; 1996.7. Streissguth AP, Bookstein FL, Barr HM, Sampson PD, O’Malley K, Kogan J.isk factors for adverse life outcomes in fetal alcohol syndrome and fetal alcohol effects.Dev Behav Pediatr 2004;25:228-38.8. Baer JS, Sampson PD, Barr HM, Connor PD, Streissguth AP. A 21-yearongitudinal analysis of the effects of prenatal alcohol exposure on young adult drinking.rch Gen Psychiatry 2003;60:377-85.

9. Spohr H-L, Willms J, Steinhausen H-C. Prenatal alcohol exposure and long-termevelopmental consequences: a 10-year follow-up of 60 children with fetal alcoholyndrome (FAS). Lancet 1993;341:907-10.0. Prader A, Largo RH, Molinari L, Issler C. Physical growth of Swiss children fromirth to 20 years of age. Helv Paediatr Acta 1989;52:1-125.1. Kromeyer-Hauschild K, Wabitsch M, Kunze D, et al. Perzentilen für den bodyass index für das kindes und jugendalter unter heranziehung verschiedener deutscher

tichproben. Monatsschr Kinderheilkd 2001;149:807-18.2. Achenbach TM. Manual for the Young Adult Self-Report and Young Adultehavior Checklist. Burlington, VT: University of Vermont, Department of Psychiatry;997.3. Steinhausen HC, Willms J, Spohr HL. Long-term psychopathological and cog-itive outcome of children with fetal alcohol syndrome. J Am Acad Child Adolescsychiatry 1993;32:990-4.4. Steinhausen H-C, Willms J, Spohr H-L. Correlates of psychopathology andntelligence in children with fetal alcohol syndrome. J Child Psychol Psychiatry994;35:323-31.5. Aronson M, Hagberg B, Gillberg C. Attention deficits and autistic spectrumroblems in children exposed to alcohol during gestation: a follow-up study. Dev Medhild Neurol 1997;39:583-7.

6. Hill S, Lowers L, Locke-Wellman J, Shen SA. Maternal smoking and drinkinguring pregnancy and the risk for child and adolescent psychiatric disorders. J Stud
lcohol 2000;61:661-8.
179

T

A

DS

*F � 7.39; df � 1; P � 0.1.

T

H

W

H

B

�25th percentile (n � 4) — — — 3

Ta

N

B

M

179.e1 Spohr, Willms, and Steinhausen

2) FAE (n � 15) Total (n � 37)

2.3 � 1.4 2.8 � 2.321.6 � 3.3 23.4 � 3.719.3 � 3.3 20.6 � 2.8

8 207 17

able V. Association of intelligence at firstssessment with head circumference at follow-up

Head circumference

<3rdpercentile

>3rdpercentile

n % n %

ormal intelligence, IQ� 85 (n � 12)

4 33 8 67

orderline intelligence,IQ 71 to 85 (n � 10)

2 20 8 80

ental retardation,IQ � 70 (n � 15)

12 80 3 20

able I. Sample characteristics

FAS (n � 2

ge, years (mean � SD)At first assessment 3.2 � 2.8At follow-up 24.7 � 3.5uration of follow-up, years (mean � SD)* 21.5 � 2.0

ex, nMales 12Females 10

able III. Follow-up of morphometric variables

Follow-up

First assessment <3rd <10th <25th >25th

eight�3rd percentile (n � 23) 13 5 2 3�10th percentile (n � 6) — 2 2 2�25th percentile (n � 2) — — — 2�25th percentile (n � 6) — — 2 4eight�3rd percentile (n � 4) 9 4 3 8�10th percentile (n � 8) — 1 1 6�25th percentile (n � 4) — — — 4�25th percentile (n � 1) — — — 1ead circumference�3rd percentile (n � 31) 17 12 2 —�10th percentile (n � 6) 1 2 1 2�25th percentile — — — —�25th percentile — — — —

MI�3rd percentile (n � 31) 5 2 8 17�10th percentile (n � 1) — — — 1�25th percentile (n � 1) — — — 1


Osw

Sf

RiisrC

Cti

Rnrs

ttitwacs((w

1

AC

1

Clinical and Laboratory Features, Hospital Course, and Outcome of RockyMountain Spotted Fever in Children

STEVEN C. BUCKINGHAM, MD, GARY S. MARSHALL, MD, GORDON E. SCHUTZE, MD, CHARLES R. WOODS, MD, MS,MARY ANNE JACKSON, MD, LORI E. R. PATTERSON, MD, AND RICHARD F. JACOBS, MD, AS THE TICK-BORNE INFECTIONS IN CHILDREN

STUDY GROUP

bjectives To describe the clinical characteristics and course of children with laboratory-diagnosed Rocky Mountainpotted fever (RMSF) and to identify clinical findings independently associated with adverse outcomes of death or dischargeith neurologic deficits.

tudy design Retrospective chart review of 92 patients at six institutions in the southeastern and southcentral United Statesrom 1990 to 2002. Statistical analyses used descriptive statistics and multiple logistic regression.

esults Children with RMSF presented to study institutions after a median of 6 days of symptoms, which most commonlyncluded fever (98%), rash (97%), nausea and/or vomiting (73%), and headache (61%); no other symptom or sign was presentn >50% of children. Only 49% reported antecedent tick bites. Platelet counts were <150,000/mm3 in 59% of children, anderum sodium concentrations were <135 mEq/dL in 52%. Although 86% sought medical care before admission, only 4 patientseceived anti-rickettsial therapy during this time. Three patients died, and 13 survivors had neurologic deficits at discharge.oma and need for inotropic support and intravenous fluid boluses were independently associated with adverse outcomes.

onclusions Children with RMSF generally present with fever and rash. Delays in diagnosis and initiation of appropriateherapy are unacceptably common. Prognosis is guarded in those with hemodynamic instability or neurologic compromise atnitiation of therapy. (J Pediatr 2007;150:180-4)

ocky Mountain spotted fever (RMSF) is a potentially life-threatening tickborne infection caused by the intracellulargram-negative coccobacillus Rickettsia rickettsii. The geographic connotation of its name notwithstanding, this illness isendemic to much of the western hemisphere—including nearly the entire conti-

ental United States.1-3 A recent seroprevalence survey suggests that infection with Rickettsii may occur much more commonly among children in the southeastern andouthcentral United States than was previously believed.4

Although the peak incidence of RMSF occurs among children 5 to 9 years of age,2,3

he clinical and laboratory manifestations of this disease have been described morehoroughly for adults than for children. The only large descriptive series of cases of RMSFn children were both published in the 1970s.5,6 Contemporary data are needed regardinghe epidemiology, clinical and laboratory manifestations, and hospital course of childrenith RMSF. Such data may be valuable to clinicians in deciding, for example, when to

dminister empiric anti-rickettsial therapy to children. This retrospective study wasonducted to describe the characteristics of children with laboratory-diagnosed RMSF atix medical centers in the southcentral United States. Specific objectives were threefold:1) to describe the demographic, clinical, and laboratory findings at the time of admission;2) to describe the hospital course, complications, and outcomes; and, (3) to determinehich patient characteristics are independently associated with adverse outcomes.

METHODSStudy patients were children (�18 years of age) treated for RMSF between January

, 1990 and December 31, 2002 at any of six institutions in the southeastern and

OR Adjusted odds ratio RMSF Rocky Mountain spotted fever

From the Departments of Pediatrics, Uni-versity of Tennessee Health Science Centerand Children’s Foundation Research Cen-ter at Le Bonheur Children’s Medical Cen-ter, Memphis, Tennessee; University ofLouisville and Kosair Children’s Hospital,Louisville, Kentucky; University of Arkansasfor Medical Sciences, College of Medicineand Arkansas Children’s Hospital, LittleRock, Arkansas; Wake Forest UniversitySchool of Medicine, Winston-Salem, NorthCarolina; Children’s Mercy Hospital, KansasCity, Missouri; and East Tennessee Chil-dren’s Hospital, Knoxville, Tennessee.

Dr Jacobs receives support from theHorace C. Cabe Foundation.

No reprints available from authors.Presented in part at the Pediatric AcademicSocieties’ meeting, Washington, DC, May14-17, 2005.

Submitted for publication Feb 27, 2006; lastrevision received Aug 19, 2006; acceptedNov 1, 2006.



I Confidence interval

80

10.1016/j.jpeds.2006.11.023

sbirIvdTmR

sfp“spawtuaasfiacai

ptfM

awfawiaftpm

C

poawvvaa

tnp

Tc

D

MAR

R

RDH

*†Rp‡

Tas

FA

NHMADCALPHPSMSPC

*

C

outhcentral United States. Each site’s institutional reviewoard approved the study design and waived requirements fornformed consent. Possible RMSF cases were identified byeviewing hospital databases for patients with dischargeCD-9 codes for rickettsioses (i.e., 082.0-083.9) and by re-iewing records of each site’s laboratory, pediatric infectiousisease consultation service, and infection control service.hese medical records were reviewed, and only patients whoet previously established criteria for confirmed or probableMSF3,7 were included in the final study population.

Demographic, clinical, and laboratory data were ab-tracted from medical records using a standard case reportorm. Clinical features were considered present only if theirresence was documented in the medical record. The termadmission” refers to a patient’s first presentation to one of thetudy institutions, regardless of whether the patient was hos-italized at that time. Altered mental status was defined asny disturbance in the subject’s level of consciousness. Comaas considered present if a patient was described as “coma-

ose” or “unresponsive,” and if this state could not be attrib-ted to pharmacologic interventions. Seizures were defined byny notation of a “seizure,” “convulsion,” or “seizure-likectivity.” An intravenous fluid bolus was defined as an infu-ion of �10 mL/kg of colloid or crystalloid. The results of therst complete blood count with differential obtained afterdmission were recorded, as were the results of the first serumhemistry and coagulation tests obtained within 72 hours ofdmission. Some clinical and laboratory data from 15 patientsncluded in this report have been presented elsewhere.8

Case report forms were forwarded to one site (Mem-his), where data were analyzed using Statview (SAS Insti-ute, Cary, NC). Bivariate statistical comparisons were per-ormed using Fisher’s exact or Wilcoxon’s rank sum tests.

able I. Demographic and historical features of 92hildren with Rocky Mountain spotted fever

Characteristic Value*

isease classificationConfirmed† 34 (37)Probable‡ 58 (63)ale sex 43 (47)ge, years 5.8 (3.7-9.1)aceAfrican-American 7 (8)Caucasian 85 (92)

eported tick bite 45 (49)Days from bite to symptom onset 4 (2-9)

eported exposure to wooded area 31 (34)ays of symptoms at admission 6 (4-8)ealthcare visit before admission 80 (86)No. of visits before admission 1 (1-2)

Median (interquartile range) or no. of patients (%).Confirmed by fourfold rise in R rickettsii antibody titer (29 patients), demonstration of

rickettsii antigen by immunostaining of tissue specimens (5 patients), or both (1atient).Single reactive R rickettsii antibody titer and clinically compatible symptom complex.

ultiple logistic regression was used to identify clinical vari- f

linical and Laboratory Features, Hospital Course, and Outcome of Roc

bles independently associated with adverse outcome, whichas defined as death or discharge from the hospital with a

unctionally significant neurologic deficit. Independent vari-bles present in at least 10% of the study population and forhich bivariate P values were �.10 were included in the

nitial multiple logistic regression model. Independent vari-bles were then removed via backward elimination; only thoseor which multivariate P was �.05 were retained. After aentative final model was created, previously eliminated inde-endent variables were individually forced back into theodel and retained if P was then �.05.

RESULTS

linical and Laboratory CharacteristicsDemographic and historical characteristics of the 92

atients with RMSF are listed in Table I. Ninety percentf cases were diagnosed between April 1 and August 31,nd 61% in May, June, and July (Figure 1; available atww.jpeds.com). Most patients were taken to healthcare pro-iders before admission; one child had five separate healthcareisits. Nonetheless, only 4 patients were prescribed annti-rickettsial antibiotic (in all cases, doxycycline) beforedmission.

Data regarding symptoms and physical findings at theime of admission are listed in Table II. Traditionally recog-ized symptom constellations had poor sensitivity: 58% ofatients had fever, rash, and headache noted,2,3 and 45% had

able II. Symptoms and signs before or duringdmission in 92 children with Rocky Mountainpotted fever

Characteristic Number of Patients (%)

ever 89 (98)ny rash 89 (97)Maculopapular only 32 (35)Any petechial component 57 (62)Involving palms and soles 60 (65)ausea or vomiting 67 (73)eadache* 52 (61)yalgia* 38 (45)bdominal pain* 31 (36)iarrhea 32 (34)onjunctival injection 30 (33)ltered mental status 30 (33)ymphadenopathy 27 (29)eripheral edema 23 (25)epatomegaly 21 (23)hotophobia 17 (18)eizure 16 (17)eningismus 15 (16)

plenomegaly 15 (16)eriorbital edema 11 (12)oma 9 (10)

Excludes children �2 years of age; denominator is 85 for these characteristics.

ever, rash, and a history of tick attachment noted.5,9

ky Mountain Spotted Fever in Children 181


Tlcaop

C

mhuptcrrpcctc

rvt

d(Pcn

aupwoprbrpavbt

figtoctphom

at

P

n

Tw

B

HP

S

BCAABAPPC

*†‡

Ffad

1

The results of initial laboratory studies are shown inable III. Most patients had thrombocytopenia; 6 had plate-

et counts of �30,000/mm3. No patient had documentederebrospinal fluid hypoglycorrhachia. Fifty-one patients hadchest radiograph obtained within 48 hours of admission; 17f these (33%) were interpreted as showing an infiltrate orneumonia.

ourse and OutcomeAs illustrated in Figure 2, children were first brought to

edical care providers after a median of 2 days of symptoms;owever, they did not receive effective anti-rickettsial therapyntil after a median of 7 days of symptoms. Eighty-nineatients were hospitalized, and 3 were managed as outpa-ients. Thirty-three patients (36%) spent time in an intensiveare unit, 15 (16%) were mechanically ventilated, 38 (41%)eceived at least one intravenous fluid bolus, and 16 (17%)eceived inotropic medications. After admission, all but 3atients received anti-rickettsial therapy, consisting of: doxy-ycline only (66 patients), chloramphenicol only (10), tetra-ycline only (7), doxycycline plus chloramphenicol (5), andetracycline plus chloramphenicol (1). Since 1994, only doxy-ycline has been used for anti-rickettsial therapy.

The time from symptom onset to initiation of anti-ickettsial therapy ranged from 0 to 15 days. Children whoisited a healthcare provider within the first 2 days of symp-

able III. Initial laboratory findings* in 92 childrenith Rocky Mountain spotted fever

Characteristic Value†

lood leukocytes/�L 9500 (5500-14,700)No. with �15,000/mm3 22/92 (24)No. with �4000/mm3 8/92 (9)emoglobin, g/dL 11.2 (10.4-12.6)latelets/mm3 128,000 (72,000-228,000)No. with �150,000/mm3 54/92 (59)No. with �100,000/mm3 38/92 (41)

odium, mEq/dL 133 (129-137)No. with �135 mEq/dL 48/91 (52)

lood urea nitrogen, mg/dL 10 (7-15)reatinine, mg/dL 0.6 (0.4-0.7)lanine transaminase, U/L 55 (36-78)spartate transaminase, U/L 83 (44-125)ilirubin, mg/dL 0.6 (0.4-1.0)lbumin, mg/dL 3.0 (2.4-3.4)rothrombin time, sec 12.7 (11.9-14.1)artial thromboplastin time, sec 30.9 (28-37.9)erebrospinal fluid‡Leukocytes/mm3 25 (3-38)Percent neutrophils 26 (16-56)Glucose, mg/dL 61 (56-70)Protein, mg/dL 68 (35-126)

Obtained within 72 hours of admission.Median (interquartile range) or no. of subjects (%).Cerebrospinal fluid was examined in 38 patients.

oms started anti-rickettsial therapy significantly later than i

82 Buckingham et al

id those who first sought care later in their course of diseasemedian time from visit to starting therapy, 4.5 days vs 1 days;

� .007). No other clinical variable was significantly asso-iated with time to initiation of anti-rickettsial therapy (dataot shown).

Three children died: a 4-year-old girl, a 6-year-old boy,nd a 15-year-old boy. All were Caucasian. One had annderlying seizure disorder and was admitted with a rapidlyrogressive illness on the day of symptom onset. The othersere previously healthy and were admitted after 3 and 6 daysf symptoms, respectively. These last two visited healthcareroviders 2 and 3 days before admission, respectively, andeceived therapy for presumed streptococcal pharyngitis (withenzathine penicillin and cephalexin, respectively); neithereceived prescriptions for anti-rickettsial antibiotics. All threeatients had fever, respiratory failure, renal insufficiency, andltered mental status on admission and required mechanicalentilation, inotropic cardiac support, and transfusions oflood products. The cause of death was multiorgan failure inwo patients, and diffuse cerebral edema in the third patient.

Thirteen of 89 surviving patients (15%) were dischargedrom the hospital with documented neurologic deficits, whichncluded: speech and/or swallowing dysfunction (6 patients),lobal encephalopathy (4 patients), ataxia or other gait dis-urbances (4 patients), and cortical blindness (1 patient). Twof these children were discharged to inpatient rehabilitationenters, and 4 others continued with physical or occupationalherapy after discharge. Follow-up data were available for 5atients who were discharged with neurologic deficits. Onead weakness that had resolved by 10 days after discharge; thether 4 had persistent abnormalities documented from 2onths to 4 years after discharge.

Two patients had digital necrosis. One suffered auto-mputation of one finger and mummification of three others;he other had dry necrosis of tips of multiple toes.

redictors of Adverse OutcomesAll 16 patients who died or were discharged with focal

eurologic deficits had altered mental status and required care

igure 2. Timeline of events in children with Rocky Mountain spottedever. Arrows indicate median values, vertical edges of boxes indicate 25th

nd 75th percentiles, and parentheses include the number of patients withata available for given events.

n an intensive care unit; thus, these variables had undefined


oraaoiC1ai

ibvstaoa2a

o

CP

w4wm.(msad3w

lRvtncthbpp

hv

o1�srbpnmwaptha

wteussloSsfpeoatbRc1breppttfto

tsfcvupR

C

dds ratios and could not be analyzed using multiple logisticegression. Several clinical variables were significantly associ-ted with adverse outcome in bivariate analysis (Table IV;vailable at www.jpeds.com); however, in multivariate analysisnly coma (adjusted odds ratio [AOR], 51.3; 95% confidencenterval [CI], 1.5-1770), inotropic support (AOR, 50.1; 95%I, 6.0-419), and receipt of an intravenous fluid bolus (AOR,3.0; 95% CI, 1.1-157) were independently associated withdverse outcome. The presence of at least two of these find-ngs was 81% sensitive and 97% specific for adverse outcome.

The median number of days from symptom onset tonitiation of anti-rickettsial therapy did not differ significantlyetween patients with and without adverse outcomes (6 dayss 7 days, P � .19). Furthermore, the proportion of childrenuffering adverse outcomes did not differ significantly be-ween those who received anti-rickettsial therapy before orfter day 5 of illness in bivariate (4 of 24 vs 11 of 55, P � .99)r multivariate analyses. The 3 patients who died receivednti-rickettsial therapy after 0, 5, and 6 days of symptoms; thewho suffered digital necrosis received anti-rickettsial ther-

py after 7 and 9 days of symptoms.Four patients received sulfonamide-containing antibi-

tics after admission. None suffered an adverse outcome.

omparisons of Patients with Confirmed versusrobable Disease

Compared with patients with probable disease, thoseith confirmed disease were older (median age, 6.5 years vs.8 years, P � .02); more frequently reported exposure to aooded area (17 of 34 vs 14 of 58, P � .02); had loweredian platelet counts (87,000/�L vs 158,000/�L, P �

001); and had lower median serum concentrations of sodium132 mEq/dL vs 135 mEq/dL, P � .02) and albumin (2.5g/dL vs 3.1 mg/dL, P � .006). These groups did not differ

ignificantly in any of the other clinical and laboratory char-cteristics listed in Tables I, II, and III; in the timing of theisease course (Figure 2); or in adverse outcome rates. Two ofpatients who died had confirmed disease; the third died

ithout convalescent sera collected.

DISCUSSIONThe results of this study reaffirm some established be-

iefs regarding the epidemiology and clinical manifestations ofMSF in children and call others into question. As in pre-

ious studies, more than 90% of patients were admitted be-ween April and September—although cases were also diag-osed in colder months.2,3 Although RMSF has beenonsidered to affect young males predominantly,2,3,5 morehan half of the patients in this study were female. Fewer thanalf of the present study’s patients reported an antecedent tickite, a rate even lower than the 56% to 66% reported inrevious studies.2,3,9,10 Similarly, exposure to a wooded arearoved to be a poorly sensitive predictor of RMSF.

As in previous studies, nearly all patients in this studyad a rash,5,6 which frequently became petechial9,10 and in-
olved the palms and/or soles.2,9 However, the onset of rash a

ccurred after a shorter interval in the present study (median,day of illness) than was noted by Helmick et al for patients14 years of age (3 days) or children (2 days).9 In the present

tudy, every child had at least one of the findings of fever,ash, or headache; however, as noted previously,2,3 the com-ination of all three was present in less than two-thirds ofatients. Although children in the present study experiencedumerous other signs and symptoms, none was present inore than half of patients. Furthermore, as reported else-here,5,6,9,10 laboratory findings generally were nonspecific

nd difficult to differentiate from those that would be ex-ected, for example, in viral syndromes. These data suggesthat no constellation of clinical and laboratory abnormalitiesas adequate sensitivity for their absence to exclude the di-gnosis of RMSF in a child.

This study has several limitations. First, only data thatere recorded in medical records was analyzed. Because his-

orical and physical examination findings were only consid-red present if they were documented, the results mightnderestimate the true rates of some clinical findings. Aecond possible limitation reflects the imperfect specificity oferologic testing for RMSF. The authors’ previous seropreva-ence study found R rickettsii antibody titers of �1:64 in 12%f children in the southeastern and southcentral Unitedtates.4 Thus, it is possible that some patients in the presenttudy had elevated RMSF antibody titers but were acutely illrom a different cause and were mistakenly classified as havingrobable RMSF. Patients who met criteria for probable dis-ase, however, had clinical findings, hospital courses, andutcomes similar to those of patients with confirmed disease,s was previously reported by Dalton et al.2 Thus, it is unlikelyhat the present study’s results have been substantially skewedy inappropriately including patients with diseases other thanMSF. On the contrary, by using more stringent enrollment

riteria than were used in the pediatric studies from the970s,5,6 many patients who truly did have RMSF must haveeen excluded. Another limitation, then, is that this study’sesults can only be considered representative of patients inndemic areas who are ill enough to have serologic testserformed, and they cannot necessarily be generalized to allatients with RMSF. Finally, owing to serologic cross-reac-ivity among spotted fever group rickettsiae, it is conceivablehat some patients in this series were infected with spottedever rickettsiae other than R rickettsii4; of note, this limita-ion must also apply to all previous serologically based studiesf RMSF.

The diagnosis of RMSF and initiation of appropriateherapy were delayed substantially in most patients in thistudy. Delays occurred not because patients with RMSFailed to seek medical attention, but because their treatinglinicians failed to consider RMSF. Such diagnostic misad-entures are understandable, given the nonspecificity of thesual presenting clinical and laboratory features of RMSF. Inrevious studies, factors associated with delayed diagnosis ofMSF included: absence of history of tick attachment2,9;

bsence of rash11 or delay in appearance of rash12; absence of

ky Mountain Spotted Fever in Children 183


hiacfafdi

orbToaRwsifptact

pdpicacsciiwott

tRt

pCecT

TTr

1d2Ns13R24ea25D6c7Ao8H9l1c1o1F1Ps11J1s1Lo21t

1

eadache2; illness outside peak months of tick activity11;llness with complaints other than fever, rash, or headache9;nd presentation to a healthcare provider early in the diseaseourse.11 In the present study, however, only the last of theseactors was significantly associated with delayed initiation ofnti-rickettsial therapy. These results suggest that cliniciansail to recognize and prescribe appropriate therapy for chil-ren with RMSF in general, and not just in those with

ncomplete constellation of symptoms.Numerous clinical factors were associated with adverse

utcomes in bivariate analyses; however, multiple logisticegression revealed only coma, inotropic support, and fluidoluses as independently associated with adverse outcomes.hese results, and the findings that all patients with adverseutcomes required care in an intensive care unit and hadltered mental status, suggest that neurologic outcomes inMSF are closely tied to patients’ disease severity, particularlyith regard to level of consciousness and to hemodynamic

tability. Similarly, Conlon et al13 reported that neurologicnvolvement was independently associated with mortalityrom RMSF; in that study, elevated serum creatinine atresentation was the only other factor independently predic-ive of mortality. Because only 4 patients received sulfon-mides (and none was exposed before admission), this studyould not evaluate the reputed association between exposureo this class of antibiotics and adverse outcome in RMSF.14,15

Experts recommend that patients with suspected RMSFromptly receive anti-rickettsial therapy, preferably withoxycycline,2,7,16-18 because delays in the initiation of appro-riate therapy have been associated with increased mortalityn adults.2,9,11,12 Although this study did not demonstrate alear association between timing of anti-rickettsial therapynd outcome, such therapy should still be prescribed forhildren once the diagnosis of a tick-borne rickettsial illness istrongly considered.17,18 It would be inappropriate to con-lude, based on this study’s results, that timing of therapy isrrelevant because only 25% of patients received therapy dur-ng the first 4 days of illness—a time frame clearly associatedith decreased mortality in previous studies.2,9,11,12 More-ver, 2 of 3 children who died sought medical care early inheir illness courses but were not prescribed anti-rickettsialherapy at these visits.

Clinicians frequently are presented with opportunitieso start therapy early in the disease course in children withMSF. This diagnosis should be considered in children in

he first few days of their illness who have any single com-ep

84 Buckingham et al

atible finding, especially during the spring and summer.hildren generally improve markedly after starting anti-rick-

ttsial therapy—as demonstrated by, on average, deferves-ence within 2 days and hospital discharge within 5 days.hese outcomes certainly are worth achieving.

he authors wish to acknowledge the assistance of Nancy C.ucker, RN, and Amy Hayden, MD, in reviewing medical

ecords for this study.

REFERENCES. Raoult D, Roux V. Rickettsioses as paradigms of new or emerging infectiousiseases. Clin Microbiol Rev 1997;10:694-719.. Dalton MJ, Clarke MJ, Holman RC, Krebs JW, Fishbein DB, Olson JG, et al.ational surveillance for Rocky Mountain spotted fever, 1981-1992: epidemiologic

ummary and evaluation of risk factors for fatal outcome. Am J Trop Med Hyg995;52:405-13.. Treadwell TA, Holman RC, Clarke MJ, Krebs JW, Paddock CD, Childs JE.ocky Mountain spotted fever in the United States, 1993-1996. Am J Trop Med Hyg000;63:21-6.. Marshall GS, Stout GG, Jacobs RF, Schutze GE, Paxton H, Buckingham SC,t al. Antibodies reactive to Rickettsia rickettsii among children living in the southeastnd south central regions of the United States. Arch Pediatr Adolesc Med003;157:443-8.. Bradford WD, Hawkins HK. Rocky mountain spotted fever in childhood. Am Jis Child 1977;131:1228-32.

. Haynes RE, Sanders DY, Cramblett HG. Rocky Mountain spotted fever inhildren. J Pediatr 1970;76:685-93.. Holman RC, Paddock CD, Curns AT, Krebs JW, McQuiston JH, Childs JE.nalysis of risk factors for fatal Rocky Mountain spotted fever: evidence for superiorityf tetracyclines for therapy. J Infect Dis 2001;184:1437-44.. Hayden AM, Marshall GS. Rocky Mountain spotted fever at Kosair Children’sospital, 1990-2002. J Ky Med Assoc 2004;102:209-14.

. Helmick CG, Bernard KW, D’Angelo LJ. Rocky Mountain spotted fever: clinical,aboratory, and epidemiological features of 262 cases. J Infect Dis 1984;150:480-8.0. Kirk JL, Fine DP, Sexton DJ, Muchmore HG. Rocky Mountain spotted fever: alinical review based on 48 confirmed cases, 1943-1986. Medicine 1990;69:35-45.1. Kirkland KB, Wilkinson WE, Sexton DJ. Therapeutic delay and mortality in casesf Rocky Mountain spotted fever. Clin Infect Dis 1995;20:1118-21.2. Hattwick MAW, Retailliau H, O’Brien RJ, Slutzker M, Fontaine RE, Hanson B.atal Rocky Mountain spotted fever. JAMA 1978;240:1499-1503.3. Conlon PJ, Procop GW, Fowler V, Eloubeidi MA, Smith SR, Sexton DJ.redictors of prognosis and risk of acute renal failure in patients with Rocky Mountainpotted fever. Am J Med 1996;101:621-6.4. Harrell GT. Rocky Mountain spotted fever. Medicine 1949;28:333-70.5. Steigman AJ. Rocky Mountain spotted fever and the avoidance of sulfonamides.Pediatr 1977;91:163-4. Letter.6. Fishbein DB, Frontini MG, Giles R, Vernon LL. Fatal cases of Rocky Mountainpotted fever in the United States, 1981-1988. Ann N Y Acad Sci 1990;590:246-7.7. American Academy of Pediatrics. Rocky Mountain spotted fever. In: PickeringK, Baker CJ, Long SS, McMillan JA, eds. 2006 Red Book: Report of the Committeen Infectious Diseases. 27th ed. Elk Grove Village, IL: American Academy of Pediatrics;006:570-2.8. Centers for Disease Control and Prevention. Diagnosis and management ofick-borne rickettsial diseases in the United States—Rocky Mountain spotted fever,
hrlichioses, and anaplasmosis: a practical guide for physicians and other health care andublic health professionals. MMWR 2006;55(No. RR-4):1-28.

F(

Tat

AIICMISBABBCMTPPHPAA

C

igure 1. Monthly distribution of confirmed (solid bars) and probablehatched bars) cases of Rocky Mountain spotted fever in children.

*


able IV. Clinical variables significantly associated withdverse outcome in children with Rocky Mountain spot-ed fever: results of bivariate analyses

Characteristic

CrudeOddsRatio

95%Confidence

Interval

ltered mental status Undefined –ntensive care unit admission Undefined –notropic support 105 19.1-581oma 75.0 8.3-679echanical ventilation 39.1 9.1-168

ntravenous fluid bolus* 34.6 4.3-277eizure 31.2 7.8-126lood product transfusion 31.2 7.8-126lbumin �3.0 mg/dL 17.6 2.1-147ilirubin �1.5 mg/dL 14.0 2.8-69.1lood urea nitrogen �15 mg/dL 8.5 2.5-28.5reatinine �1.1 mg/dL 8.5 1.7-43.2eningismus 6.6 1.9-22.6riad of fever, rash, and headache 6.6 1.4-31.2artial thromboplastin time �45 sec 6.4 1.2-33.7latelet count �150,000/�L 6.3 1.3-29.6eadache 6.3 1.3-29.6etechial rash 5.4 1.1-25.3lanine aminotransferase �55 U/L 4.9 1.2-20.0spartate aminotransferase �100 U/L 3.5 1.1-11.7

Any infusion of �10 mL/kg or �1L of colloid or crystalloid.

ky Mountain Spotted Fever in Children 184.e1

OSw(u

Re2[

Cin

ClnhgPP

oecbsoaffidafmpmpl

9B

Parental Depression Following the Early Diagnosis of Cystic Fibrosis: AMatched, Prospective Study

CLAIRE GLASSCOE, PHD, GILLIAN A. LANCASTER, PHD, ROSALIND L. SMYTH, MD, AND JONATHAN HILL, FRCPSYCH

bjective To assess risks for parental depression following the diagnosis of cystic fibrosis (CF) in a child.

tudy design Matched cohort study in NW England; 45 parental couples with a child diagnosed with CF were comparedith 45 control couples matching for age, sex, and position in the family of the index child. The Beck Depression Inventory

BDI-II) with a clinical cut-off >13 for dysphoria (mild depression) was the main outcome. A stratified analysis was conductedsing the Mantel-Haenszel risk-ratio estimator (RRMH) with eight strata for each of the matching variable combinations.

esults Heterogeneity was found within the dataset. Parents with a child with CF <9 months of age at baseline had anlevated prospective risk of depression (mothers RRMH [95% confidence interval(CI)] � 2�6[1�05,6�42], fathers RRMH [95%CI] �

�26 [0�97,5�28]). The absence of a group effect for depression at follow-up after adjusting for the matching (mothers RRMH

95%CI] � 1�1 [0�59,2�05], fathers RRMH [95%CI] � 1�42 [0�66,3�08]) masked this heterogeneity.

onclusion This hypothesis-generating finding suggests that parents may be more vulnerable to depression when their childs diagnosed with a life-shortening condition during the first few months of life. Mood in parents of infants diagnosed earlyeeds to be monitored longitudinally and preventative strategies devised. (J Pediatr 2007;150:185-91)

ystic fibrosis (CF) affects 1 in 2500 live births in those of European descent.1 Early diagnosis is important for survival,2,3

and national neonatal screening programs for CF have been established in many parts of the Western world. Previousstudies have indicated that the majority of parents adapt well to having a child with cystic fibrosis,4-6 although elevated

evels of depression and reduced partner satisfaction have been reported.7-10 Cross-sectional studies of parental depression mayot, however, adequately reflect the multiple ways that the illness impacts on the family. Parents, the affected child, and siblingsave to come to terms with a diagnosis that implies a premature death, to respond over many years to a physical condition thatenerally fluctuates markedly, and to cope day to day with the practical demands of administering treatment in the home.arental mental health therefore is likely to fluctuate over time, and it may be influenced by multiple interacting risk factors.rospective, longitudinal studies with representative samples, defined in terms of time since diagnosis, are needed.

Studies of depression in the general population find that subclinical levels increase risk for further dysfunction and episodesf depression predict further episodes in an escalating cycle of vulnerabilities.11 Maternal depression is known to influencemerging parent-child attachment status.12 It not only influences maternal reports ofhildhood behavioral difficulties13 but is also associated with cognitive,14 emotional,15 andehavioral difficulties in developing children.12,16 Studies of fathers’ role or the relation-hip between fathers’ mood and child development are scarce. Although an early diagnosisf CF avoids delay the start of treatment, which may reduce anxiety in parents,17 potentialdverse effects of the news of a life-shortening condition in a newborn infant on theormation of the parent-child relationship are now cause for concern.18,19 Preliminaryndings from a longitudinal study suggest elevated risk for depression associated with aiagnosis given in infancy.20 This study aimed to identify the persistent risk to mothers’nd fathers’ mood associated with a clinical diagnosis for CF that might have implicationsor an early diagnosis through screening. The parameters and demands of doing thisethodologically were explored with a design that had two principal features; it was

rospective and matched with possible sources of heterogeneity anticipated from theatching variables. The ultimate aim was to identify a population-based approach for the

revention of mood disorder in parents dealing with the new challenge of a chronicife-shortening condition in a child diagnosed in infancy.

5% CI 95% confidence interval CF Cystic fibrosis

From the University of Liverpool, ChildMental Health Unit, Royal Liverpool Chil-dren’s Hospital; the University of Liverpool,Centre for Medical Statistics and HealthEvaluation; and the University of Liverpool,Institute of Child Health, Royal LiverpoolChildren’s Hospital, UK.

Dr Glasscoe was funded by a NHS NorthWest Region PhD Training Fellowship.

Submitted for publication Nov 4, 2005; lastrevision received Aug 26, 2006; acceptedNov 1, 2006.

Reprint requests: Dr Claire Glasscoe, RoyalLiverpool Children’s Hospital – Alder Hey,Child Mental Health Unit, 1st Floor, Mul-berry House, Eaton Road, Liverpool L122AP, UK. E-mail: [email protected].



DI-II Beck Depression Inventory RRMH Mantel Haenszel risk-ratio
10.1016/j.jpeds.2006.11.022
185

D

twTmwasBufac

R

cUEaonaf(cpcaLHwgaopiao(ci

ohwiccppm

Wndwan

M

smatdabacitr(sc

O

etmytgi1cKc8cdqhi

D

stsmtv

1

METHODS

esignThe aim of the design was to ensure as far as possible

wo groups for comparison that differed only with respect tohether a couple had had a child recently diagnosed with CF.he groups were matched for the index child’s age (�3onths), sex, and position in the family (first/not first) andere compared on a further six parameters including parental

ge, length of couple relationship, family composition, familyize, income, and family stress21 over the preceding 12 months.aseline interviews in the CF group of parents were sched-led to take place within 3 months of the diagnosis andollow-up interviews 9 months later; thus the study periodimed to capture the first year following the diagnosis. Theontrol group participants were seen twice, 9 months apart.

ecruitmentThis study was a consecutive sample of incident CF

ases from two main centers in the North West of England,K. Ethical approval was obtained from Liverpool Children’sthics Committee, Salford and Trafford Ethics Committee,

nd Wirral Ethics Committee, and full written consent wasbtained from each participant. All parents of children diag-osed with CF between September 1998 and January 2002ttending the centers and their peripheral clinics were selectedor this study. Families in which two primary caretakersbiological, step, foster, or adoptive) were living with the firsthild to receive a diagnosis of CF were approached for theirarticipation by the consultant pediatrician responsible for thehild’s care or their specialist CF nurse. Participating mothersnd fathers in the CF group were recruited from Royaliverpool (n � 16) and Booth Hall (n � 9) Children’sospitals and 11 peripheral clinics (n � 20). Control familiesith no diagnosis of CF in a child were recruited from threeeneral practice (GP) clinics: one large clinic covering tworeas in the city of Liverpool (n � 30) and two smaller clinicsn the Wirral peninsular (n � 7 and n � 8). These generalractices served a broad socioeconomic range with Jarmanndexes (Jarman score: a zero value represents the nationalverage deprivation score with a national range of �50 to 75)f between 20�39 and �13�42 (Liverpool) and 4�66 and zeroWirral). A systematic matching procedure was used to selectontrol families. Each selected family received a letter ofnvitation from their child’s family doctor.

Couples were excluded from both groups where (a) onlyne parent was living with the index child; (b) either parentad a psychotic illness where the interview and questionnairesould have unsettled them; (c) either parent had a physical

llness that rendered him or her incapable of providing childare; (d) there had been a recent child death; (e) there wereomplex social problems with statutory services involved andarticipation in the study would disrupt child protectionlans; (f) they had moved and were untraceable by reasonable
eans; or, (g) English was not the first or second language. m
86 Glasscoe et al

here two children were being assessed for CF simulta-eously in the same family then the first child to receive aiagnosis was regarded as the index child. Where the childas diagnosed in utero then the date of the birth was regarded

s the date of diagnosis because that point marked the begin-ing of treatment.

atching VariablesThe matching variables reflected three anticipated

ources of heterogeneity: (1) age of the child at baseline (�3onths) because the demands of a newborn baby may be

ssociated with postnatal depression in mothers22; (2) sex ofhe child because male and female children are known to haveifferent physical vulnerabilities and prognostic outcome23;nd (3) position in the family of the child (first/not first)ecause birth order and the number of siblings in the familyre known to be associated with different psychosocial out-omes.24 A query was entered into the GP databases for eachdentified child with CF, to produce a list of matches based onhe child’s age (�3 months) and sex. This list was thenefined to include children matched for position in the familyfirst/not first). One child was randomly selected from tenuitable matches until each identified child had one matchedontrol for comparison.

utcome MeasureThe primary outcome was depression reported by par-

nts with the second edition of the Beck Depression Inven-ory (BDI-II).25 This is a 21-item self-report instrument foreasuring severity of depression in adults and adolescents 13

ears of age and older. The BDI-II is a widely used measurehat has been extensively validated.25 It is reported to haveood internal consistency25 and discriminant validity.26 Clin-cal cutoffs for the BDI-II of 0 to 12 for nondepressed, 13 to9 for dysphoria, and 20 to 63 for dysphoric or depressed werealculated by Dozois et al26 using the procedure described byendall et al.27 These empirically derived cutoffs had overall

orrespondence rate of 91% with the original (sensitivity �1%; specificity � 92% .and � coefficient of 0�7).26 A lowlinical cutoff of 13 was selected to indicate dysphoria (mildepression), and the term dysphoria is preferred as this was auestionnaire-based assessment. Versions of this inventoryave been utilized as a screening measure for depressive illness

n the general population.28,29

ata AnalysesEmphasis was placed on estimation of parameter effect

izes and 95% confidence intervals (95% CI).30 To comparehe risk-ratio of depression in the two matched groups atratified analysis using the Mantel Haenszel risk-ratio esti-ator (RRMH)31 was used and is presented with 95% CI. For

he purpose of these analyses, the continuous age matchingariable was transformed into a dichotomous variable (�9
onths/�10 months). This cutoff was chosen because it

rb

svbfsctdn

T

ftpuFiotmm

fr

C

pmmwmodlmwcfpt

P

mbfcpTi1

dcwpr2

E

tirtcata(spCra

FCpsp

P

epresented the median age of children in the CF group ataseline.

The method for dealing with a stratified dataset de-cribed by Greenland31 (pp. 283-5) was followed. This in-olved examination of the dataset for stratum-specific effectsefore pooling across strata. Analyses were conducted usingormulae 15.20 and 15.22, p. 271.31 Where the data wereparse and zero dysphoric participants in one or other groupreated one empty cell, then 0�5 was added to all four cells inhe 2 � 2 table for that stratum.32 Where there were noysphoric participants in both groups, the stratum RRMH wasot computed.

RESULTS

he SampleThe Figure shows the recruitment rate was (50/56) 89%

or the CF group and (58/81) 72% for the control group. Ofhe four families excluded in the CF group there was one lonearent, in two families child protection investigations werenderway, and in one other the parents did not speak English.orty-five matched couple pairs were identified and entered

nto the analyses and of these 39 pairs of fathers and 41 pairsf mothers completed the study at both time points. Sixty-wo percent of couples in the CF group were seen within 3onths, 82% within 4 months, and all were seen within 9

igure. Recruitment flow chart illustrating the number of couples in theF group (left) and control group (right) who were eligible to takeart in the study, the number included, the number consented, and theubsequent number of mothers and fathers in each group thatroduced data at baseline and follow-up 9 months later.

onths of the diagnosis. A sensitivity analysis was conducted t

arental Depression Following the Early Diagnosis of Cystic Fibrosis: A

or the main analysis, with and without the eight late recruitsevealing no change in the main trends reported.

haracteristics at BaselineThe sample, described in Table I, comprised 45 matched

airs of parents with the identified child ranging in age from 3onths to 11 years, 4 months. The median age at diagnosis inonths (interquartile range) was 6 (2�5,31�5). Matching variablesere examined within the two groups to assess the success of theatching process. Children in the CF group had a median age

f 1.5 months younger than children in the control group (me-ian difference [95%CI] �1�5 [�2, �1]). The medians for the

ower age stratum were 5 months for the CF group and 6onths for the control group yielding a median group differenceith 95%CI of �1[�3,0]. The groups did not differ signifi-

antly with respect to the other two matching variables, or as aunction of six potential confounders summarized in Table I asarental age, length of the couple relationship, family composi-ion, size, income, and stress.21

revalence of DysphoriaThe prevalence of dysphoria (Table II) reported by

others in the whole group was similar in both groups ataseline and follow-up. Thirty-three percent of CF groupathers reported dysphoria at baseline compared with 18% ofontrol fathers, and 18% of CF group fathers reported dys-horia at follow-up compared with 10% of control fathers.he paired percentage difference [95%CI] for group compar-

sons of fathers showed no clear differences (baseline5�5%[�4%,34%] and follow-up 7�7%[�8%,24%]).

The proportion of couples where both partners wereysphoric in the CF couples was 22% compared with 11% ofontrol couples at baseline, and 15% of CF couples comparedith 5% of control couples at follow-up. The percentageaired group differences [95%CI] were in the predicted di-ection, although confidence intervals were wide (baseline0% [�2%,39%], follow-up 13% [�4%,34%]).

ffect Measure Modification and Risk for DepressionA stratified analysis assumes homogeneity of effect and

hus the stratum-specific effects were examined before pool-ng as detailed in the Greenland protocol.31 This procedureevealed a heterogeneous pattern of elevated risk estimates inhe strata for those with children �9 months of age inontrast to those �9 months of age at baseline (not shown)nd at follow-up (Table III). Despite matching for age, fa-hers in the CF group were more likely to experience dysphoriat baseline with a child �9 months than in the control groupRRMH [95%CI] � 3�51 [1�17,10�52]) and the effect for mothershowed a similar trend (RRMH [95%CI] � 2�1 [0�81,5�44]). Thisattern persisted at follow-up (Table III), when mothers in theF group relative to controls were more likely to report dyspho-

ia with a child �9 months (RRMH [95%CI] � 2�6 [1�05,6�42]nd fathers with a child �9 months showed an almost iden-
ical trend in the same direction (RRMH [95%CI] � 2�26
Matched, Prospective Study 187

[air

n[

T

MA

S

P

PA

L

F

F

F

F

*†‡§

T

BMFBFMFB

D

1

0�97,5�28]). These estimates varied very little in a sensitivitynalysis excluding the eight late recruits, and they suggest anmportant finding of heterogeneity for age that needs to be

able I. Sample characteristics at baseline

CF group

atching variablesge of childNumber of months (median, interquartile range) 10 (5�9 months (n, %) 20 (4�10 months (n, %) 25 (5

ex of childMale (n, %) 21 (4Female (n, %) 24 (5

osition in the familyFirst (n, %) 27 (6Not first (n, %) 18 (4

otential confounding variablesge of parentsMother (mean, SD) 31.13 (5Father (mean, SD) 33.64 (6

ength of couple relationship�5 years (n, %) 13 (26-10 years (n, %) 14 (3�11 years (n, %) 18 (4

amily compositionIntact nuclear (n, %) 37 (8Reconstituted (n, %) 8 (1

amily sizeNumber of children (median, interquartile range) 2 (1

amily income�10,000 (n, %) 4 (810,000-24,999 (n, %) 23 (525,000-35,999 (n, %) 12 (2�35,000 (n, %) 6 (1

amily Stress*Mother (median, interquartile range) 12 (7Father (median, interquartile range) 10 (4

Parenting Stress Index (Abidin, 1995) - subscale family stress with possible range 0-7Paired mean difference.Paired median difference.Paired percentage difference.

able II. Prevalence of dysphoria (mild depression) i

Dysphoria (BDI-II) Clinical cutoff

CF Group

n Count %

aseline (�13)others 45 20 44

athers 45 15 33oth partners in a couple 45 10 32ollow-up (�13)others 41 15 36

athers 39 7 17oth partners in a couple 39 6 15

ysphoria (BDI-II) - Beck Depression Inventory - Beck, 1996.

eported rather than treated as bias to be eliminated.31 1

88 Glasscoe et al

By contrast to the stratum-specific analysis there waso prospective pooled main effect for mothers (RRMH

95%CI] � 1�1 [0�59,2�05]) or fathers (RRMH [95%CI] �

� 45) Control group (n � 45) Difference [95%CI]

4) 12 (7, 35.5) �1.5 [�2, �1]‡) 16 (35.6%) 9% [�0.4%, 18%]§) 29 (64.4%)

) 21 (46.7%) —) 24 (53.3%)

27 (60%) —18 (40%)

31.2 (6.17) �0.07 [�2.3, 2.2]†33.87 (5.39) �0.22 [�2.17, 1.73]†

) 13 (28.8%) —) 9 (20%) 11% [�8%, 29%]§

23 (51.1%) �11% [�29%, 8%]§

) 38 (84.4%) 2% [�12%, 16%]§) 7 (15.6%)

2 (1, 2) —

5 (11.1%) —17 (37.8%) 9% [�11%, 28%]§

) 14 (31.1%) �4% [�23%, 15%]§) 9 (20%) �4% [�19%, 10%]§

7.5) 10 (5, 16) 2 [�1, 5]‡10 (5, 15) 1 [�1.5, 4]‡

e CF and control groups at baseline and follow-up

Control Group CF/Control Paired Difference

n Count % Percentage Difference [95% CI]

45 16 35.6 8.9% [�11%, 28%]45 8 17.8 15.5% [�4%, 34%]45 5 11.1 20% [�2%, 39%]

41 13 31.7 4.9% [�18%, 27%]39 4 10.3 7.7% [�8%, 24%]39 2 5.1 13% [�4%, 34%]

(n

.5, 34.4%5.6%

6.7%3.3%

0%)0%)

.45)

.08)

8.8%1.1%0%)

2.2%7.8%

, 2)

.9%)1%)6.7%3.3%

.5, 1, 6)

9.

n th

.4

.3

.2

.6

.9

.4

�42 [0�66,3�08]). Thus, no pooled main effect was established


fa

ondaamdfiC

odWam

mofhficnw

mbastuttsrb

Tas

M

4

48F

4

48

RL*†

P

or dysphoria in the CF compared with the control group aswhole.

DISCUSSIONThe overall finding was that there was no elevated rate

f depression in the CF group of parents. However, exami-ation of effects within predetermined strata for age revealedifferences in those parents with a child of �9 months of aget baseline in contrast to parents with a child �10 months ofge. This hypothesis-generating finding suggests that parentsay be more vulnerable to depression when their child is

iagnosed with a life-shortening condition during the firstew months of life. Preliminary results from a study of mother-nfant relationships in the context of newborn screening forF support this contention.20

This study was designed to recruit a consecutive samplef incident CF case patients and their families within aefined time since diagnosis and to follow them prospectively.

e aimed to assess families within 3 months of diagnosis,nd we achieved this in 62%, with 82% assessed within 4

able III. Prospective effect of CF/control group medjusting for matching variables and sparse data. Ovtrata at follow-up are displayed

Pooled Effects BDITime B

Number of participantswith dysphoria

N

CF Group Control Group CF

othersfemale/not 1 st/�9 mo 4 1female/1 st/�9 mo 2 1male/not 1 st/�9 mo 2.5 0.5male/1 st/�9 mo 3 1lower age stratafemale/not 1 st/�10 mo 1 2female/1 st/�10 mo 1 5male/not 1 st/�10 mo 1 2male/1 st/�10 mo 1 1upper age strataaggregated strataathersfemale/not 1 st/�9 mo 1.5 0.5female/1 st/�9 mo 1.5 0.5male/not 1 st/�9 mo 1.5 0.5male/1 st/�9 mo 1.5 0.5lower age stratafemale/not 1 st/�10 mo 0 0female/1 st/�10 mo 2 3male/not 1 st/�10 mo 1 1male/1 st/�10 mo 0 0upper age strataaggregated strata

RMH - Mantel Haenszel risk-ratio; Rothman and Greenland, 1998.L & UL - lower and upper limits for 95% confidence intervals.

RRMH - pooled Mantel Haenszel risk-ratio.Dysphoria - Beck Depression Inventory cut-off �13; Beck, 1996.

onths. In practice, some of these children had complicated i

arental Depression Following the Early Diagnosis of Cystic Fibrosis: A

edical procedures that involved travel to specialist units inther parts of the country, which effectively precluded theamilies’ involvement in this study until the child’s conditionad stabilized. Although 3 months was set for the time of therst interview, there was no reason to think that 3 months waslearly preferable to 4 months. A sensitivity analysis showedo difference in the effects observed when the late recruitsere excluded from the analyses.

The design enabled effect-measure modification by theatching variables to be examined. This has not been done

efore, and only summary effects for the whole group arevailable elsewhere.7-9 The stratum-specific findings pre-ented are not based on a priori considerations, but neither arehey post hoc. Stratified analyses carry an assumption ofniform effects. To consider this assumption the first step iso examine stratum specific effects—if heterogeneity is found,hen it should be reported as a finding and considerationhould be given to whether the strata should be pooled oregarded separately. Small group sizes meant the data coulde quite sparse in the eight strata created by the three match-

rship on dysphoria† in mothers and fathers afterestimates and subgroups for upper and lower age

er participants inach stratum Stratum-specific

Effects RRMH

GroupEffectsRRMH* LL ULp Control Group

4 4.003 1.202 3.335 2.50

2.60 1.05 6.424 0.50

12 0.246 0.605 1.25

0.46 0.17 1.291.1 0.59 2.05

4 3.003 1.802 2.004 2.40

2.26 0.97 5.283

12 0.806 1.205

0.90 0.24 3.411.42 0.66 3.08

mbeerall

umbe

Grou

4536

41054

4535

31054

ng variables. The advantage of using the Mantel Haenszel

Matched, Prospective Study 189

soegp

rcucwnrpAcettppfta

efspapbinhw

ildiacclaCcwmrCs

TgW

Galc

1as2fi3J4J5aM6a7i18fi9is1t1s71do131sP1p1c1Gc1m11g2c1212f12fi2a2A2D2

1

tatistic here is that it remains robust in sparse data. Theverall effect for the pooled strata was not significant; how-ver, we have shown the importance of assessing the hetero-eneity in the strata because this may be masked throughooling.

Matching on age was important, not only because theisks may be different in infancy, but also because it ensuredomparability over a large age distribution up to 11 years. Wesed a caliper-matching strategy for age based on �3 monthriterion within a 6-month age span. Where the CF infantas �3 months of age at diagnosis, the available age span wasecessarily reduced. This and the length of time taken toecruit a control child once he or she had been identified as aotential match resulted in a marginally older control group.lthough the differences between the groups were small, we

annot rule out the possibility that this age difference influ-nced the effect in the parents of the younger children. Weook care to establish as far as possible comparability acrosshe groups by selecting controls with a systematic matchingrocedure from three general practice lists. Moreover, com-arison across six key parameters showed no significant dif-erence between the groups. It remains possible, however, thathe groups did differ on other parameters not measured suchs a history of depression in the parents.

A low clinical cutoff for mild depression allowed forxamination of the spread of effect and risk for further dys-unction. The inclusion of fathers was a strong point of thistudy and has rarely been done. We did not specify biologicalarents, and step, foster, and adoptive parents were includeds were those parents who separated before follow-up. Inractice, only one family was excluded in the CF groupecause of being a single parent. If lone parents had beenncluded in the sample at baseline, then we would haveeeded to control for one versus two parents as lone parent-ood represents very different risks for depression. Thisould have placed a heavier burden on the analyses.

The results from this study indicate a need for furthernvestigation to understand the mechanisms at play and theong-term implications for adaptation in the families of chil-ren with cystic fibrosis. It is well established that depression

n mothers is disadvantageous to the parent-child bonding12

nd subsequent development in children.12,14-16 This adverseombination is even more important to identify when thehild is sick and the illness is progressive, as there are potentialong-term implications for the psychological health of parentsnd their children. With the advent of newborn screening forF in place in several parts of the Western world, and due to

ommence shortly in the UK, the vast majority of childrenith CF will be diagnosed in infancy and this may mean thatore parents are vulnerable to depression. Any risks to pa-

ental mental health associated with newborn screening forF need to be documented longitudinally and preventive

trategies need to be developed to manage them.

he authors are grateful to all the families in this study whoenerously gave their time to participate in a collective endeavour.
e would like to thank staff within the hospital CF centers and
rR

90 Glasscoe et al

P clinics located in the North West of England for their supportnd help in recruiting the families that took part. We would alsoike to thank our three anonymous reviewers for their helpfulomments on the draft manuscript.

REFERENCES. Bobadilla JL, Macek Jr M, Fine JP, Farrell PM. Cystic fibrosis: A worldwidenalysis of CFTR mutations - correlation with incidence data and application tocreening. Hum Mutat 2002;19:575-606.. Murray J, Cuckle H, Taylor G, Littlewood J, Hewison J. Screening for cysticbrosis. Health Technol Assess 1999;3:1-97.. Southern KW. Newborn screening for cystic fibrosis: the practical implications.R Soc Med 2004;97:1-3.. McCollum AT, Gibson LE. Family adaptation to the child with cystic fibrosis.Pediatr 1970;77:571-8.. Cowen L, Corey M, Keenan N, Simmons R, Arndt E, Levison H. Familydaptation and psychosocial adjustment to cystic fibrosis in the preschool child. Soc Sci

ed 1985;20:553-60.. Sabbeth BF, Leventhal JM. Marital adjustment to chronic childhood illness:critique of the literature. Pediatrics 1984;73:762-8.

. Goldberg S, Morris P, Simmons R, Fowler RS, Levison H. Chronic illness innfancy and parenting stress: a comparison of three groups of parents. J Pediatr Psychol990;15:347-58.. Quittner AL, DiGirolamo AM, Michel M, Eigen H. Parental response to cysticbrosis: a contextual analysis of the diagnosis phase. J Pediatr Psychol 1992;17:683-704.. Quittner AL, Espelage DL, Opipari LC, Carter B, Eid N, Eigen H. Role strainn couples with and without a child with a chronic illness: associations with maritalatisfaction, intimacy, and daily mood. Health Psychol 1998;17:112-24.0. Thompson RJ, Gustafson KE. Adaptation to Chronic Childhood Illness. Washing-on, DC: American Psychological Association; 1996.1. Lewinsohn PM, Klein DN, Durbin EC, Seeley JR, Rohde P. Family study ofubthreshold depressive symptoms: risk factor for MDD? J Affect Disord 2003;7:149-57.2. Murray L, Fiori-Cowley A, Hooper R, Cooper PJ. The impact of postnatalepression and associated adversity on early mother-infant interactions and later infantutcome. Child Dev 1996;67:2512-26.3. Puckering C. Annotation: maternal depression. J Child Psychol Psychiatry 1989;0:807-17.4. Hay DF, Pawlby S, Sharp D, Asten P, Mills A, Kumar R. Intellectual problemshown by 11-year-old children whose mothers had postnatal depression. J Child Psycholsychiatry 2001;42:871-89.5. Hay DF, Pawlby S. Prosocial development in relation to children’s and mothers’sychological problems. Child Dev 2003;74:1314-27.6. Hay DF, Pawlby S, Angold A, Harold GT, Sharp D. Pathways to violence inhildren of mothers who were depressed postpartum. Dev Psychol 2003;39:1083-94.7. Mérelle ME, Huisman J, Alderden-van der Vecht A, Taat F, Bezemer D,riffioen RW, et al. Early versus late diagnosis: psychological impact on parents of

hildren with cystic fibrosis. Pediatrics 2003;111:346-50.8. Simmons RJ, Goldberg S, Washington J, Fischer-Fay A, Maclusky I. Infant-other attachment and nutrition in children with cystic fibrosis. J Dev Behav Pediatr

995;16:183-6.9. Farrell MH, Farrell PM. Newborn screening for cystic fibrosis: ensuring moreood than harm. J Pediatr 2003;143:707-12.0. Tluczek A, Clark R, Koscik RL, Farrell PM. Mother-infant relationships in theontext of neonatal CF diagnosis: preliminary findings. Pediatr Pulmonol 2005;(suppl 28):79-80.1. Abidin RR. Parenting Stress Index. Psychological Assessment Resources, Inc.;995.2. Murray L, Stanley C, Hooper R, King F, Fiori-Cowley A. The role of infantactors in postnatal depression and mother-infant interactions. Dev Med Child Neurol996;38:109-19.3. Rosenfeld M, Davis R, FitzSimmons S, Pepe M, Ramsey B. Gender gap in cysticbrosis mortality. Am J Epidemiol 1997;145:794-803.4. Quittner AL, Opipari LC. Differential treatment of siblings: interview and diarynalyses comparing two family contexts. Child Dev 1994;65:800-14.5. Beck AT, Steer RA, Brown GK. Beck Depression Inventory (BDI-II) - Manual. Sanntonio: The Psychological Corporation, Harcourt Brace & Company; 1996.

6. Dozois DJA, Dobson KS, Ahnberg JL. A psychometric evaluation of the Beckepression Inventory - II. Psychol Assess 1998;10:83-9.

7. Kendall PC, Hollon SD, Beck AT, Hammen CL, Ingram RE. Issues and
ecommendations regarding use of the Beck Depression Inventory. Cognitive Therapyes 1987;11:289-99.

2Ia2Cp

3C31

L

8. Steer RA, Cavalieri TA, Leonard DM, Beck AT. Use of the Beck Depressionnventory for primary care to screen for major depression disorders. Gen Hosp Psychi-try 1999;21:106-11.9. Lasa L, Ayuso-Mateos JL, Vázquez-Barquero JL, Díez-Manrique FJ, Dowrick
F. The use of the Beck Depression Inventory to screen for depression in the generalopulation: a preliminary analysis. JAffect Disord 2000;57:261-5.
3f

i

0. Altman DG, Machin D, Bryant TN, Gardner MJ. Statistics with Confidence:onfidence Intervals and Statistical Guidelines. Bristol: BMJ Books; 2000.1. Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia: Lippincott-Raven;998.
2. Greenland S, Robins JM. Estimation of a common effect parameter from sparseollow-up data. Biometrics 1985;41:55-68.
50 Years Ago in The Journal of PediatricsINTRATHECAL HYDROCORTISONE IN THE TREATMENT OF TUBERCOLOUS MENINGITIS

Choremis C, Papadatos C, Gargoulas A, and Drosos C. J Pediatr 1957;50:138-44

Tuberculosis remains a global health hazard, with an annual incidence of 8.8 million cases worldwide and 1.8 milliondeaths.

Meningitis is a particularly severe form of tuberculosis, with death and severe neurological deficit being reported ina significant proportion of patients despite treatment. Rupture of a subependymal or parameningeal tubercle into thesubarachnoid space initiates a cascade of inflammatory response that leads to complications such as intracranial vasculitisand communicating hydrocephalus. Adjunctive therapies to improve the clinical outcome of tuberculous meningitis havebeen studied for decades. The anti-inflammatory effect of corticosteroids remains the most biologically plausible way toabort the pathologic sequelae. Choremis et al reported in this journal 50 years ago their experience with the use ofintrathecal corticosteroid in treating 82 children with tuberculous meningitis. They compared three different treatmentregimens. In the first group of 24 patients, “standard” therapy of intramuscular and intrathecal streptomycin and isoniazidwas given. The second group of 29 patients was given, in addition to the standard therapy, cortisone intramuscularly. Inthe third group of 29 patients, intrathecal streptomycin was replaced by intramuscular and/or intrathecal cortisone, whichwas given in addition to intrathecal and intramuscular isoniazid and intramuscular streptomycin. The three groups,though not formally controlled, were comparable to each other. Using the patient’s general condition, resolution ofelectroencephalographic abnormalities, and changes in sugar and protein levels in the cerebral spinal fluid as outcomemeasures, the authors concluded that patients who received the third treatment regimen had the best clinical improve-ment, even though no statistical analysis was presented. The authors also pointed out that in two cases allocated to thethird treatment regimen, early blockage in the cerebral spinal fluid circulation was relieved by the cisternal infusion ofhydrocortisone.

The treatment for tuberculous meningitis has changed since the study was published. The use of anti-tuberculousdrugs in the intrathecal route is rarely indicated nowadays, and adjunctive dexamethasone given systemically rather thanintrathecally has been reported as effective. A recently published prospective, randomized, placebo-controlled trialinvolving 545 patients over 14 years of age with tuberculous meningitis in two hospitals in Vietnam showed thatadjunctive treatment with dexamethasone significantly reduced mortality rates.1 The burden of tuberculosis remains aglobal challenge, and research work similar to that by Choremis et al and Thwaites et al are essential in guiding us inthe right direction.

Albert M. Li, MBDepartment of PaediatricsPrince of Wales Hospital

The Chinese University of Hong KongShatin, Hong Kong

10.1016/j.jpeds.2006.09.007

REFERENCES1. Thwaites GE, Nguyen DB, Nguyen HD, et al. Dexamethasone for the treatment of tuberculous meningitis in adolescents and adults. N Engl J Med 2004;351:1741-51.

191

O(

S(efv

Rre

Cptg

Pchsps

ppQcH

gv

2Ci

AEFP

1

Prolonged QTc Intervals and Decreased Left Ventricular Contractility inPatients with Propionic Acidemia

DANIELA BAUMGARTNER, MD, SABINE SCHOLL-BÜRGI, MD, JÖRN OLIVER SASS, DR. RER. NAT., WOLFGANG SPERL, MD, PHD,ULRICH SCHWEIGMANN, MD, JÖRG-INGOLF STEIN, MD, AND DANIELA KARALL, MD

bjective To investigate electrophysiological and functional signs of myocardial damage in patients with propionic acidemiaPA), an inborn error of metabolism caused by deficiency of propionyl CoA carboxylase (PCC).

tudy design In an observational longitudinal study 10 patients with PA (6 boys and 4 girls) ranging between 2.5 and 20.2median 9.0) years of age at last follow-up were investigated over a period of up to 20 (mean 7.4) years using 12-leadlectrocardiograms (ECGs), 24-hour continuous ECG recordings, bicycle exercise testings, and echocardiography with specialocus on repolarization abnormalities such as corrected QT interval (QTc) prolongation, ventricular dysrhythmias, and leftentricular systolic function.

esults QTc interval was prolonged (>440 ms) in 70% of patients beyond infanthood. Continuous ECG recordings revealedhythm disturbances in 20% of patients. M-mode echocardiographic left ventricular function was reduced (fractional short-ning [FS] <30%) in 40%. One patient showed signs of dilated cardiomyopathy.

onclusions The majority of patients with PA (even in clinically stable situations) have disturbances in cardiac electro-hysiology that can contribute to cardiac complications. Possible mechanisms include effects of toxic metabolites or depriva-ion of essential substrates. To avoid life-threatening complications, we recommend regular cardiological evaluations in thisroup of patients. (J Pediatr 2007;150:192-7)

ropionic acidemia (PA) is a recessive disorder caused by a deficiency of propionyl CoA carboxylase (PCC), an enzymeinvolved in the catabolism of valine, isoleucine, methionine, threonine, cholesterol and odd-carbon numbered fatty acids,thymine, and uracil.1 Mutations of both the PCCA gene mapped to chromosome 13q32 and the PCCB gene mapped to

hromosome 3q13.3-q22 have been described.2,3 Accumulating propionyl-CoA or its metabolites may result in hypoglycemia,yperammonemia, and hyperglycinemia.4-6 Different long-term complications occur and involve mainly the central nervousystem, feeding difficulties, and metabolic crises in catabolic situations.7,8 Because of the great genetic heterogeneity2 the clinicalicture varies between severe early-onset forms manifesting during the first days of life, and mild late-onset variations, whichhow first signs during adulthood.

Cardiomyopathy and sudden cardiac death have been described as complications of several metabolic disorders.9,10 Inatients with PA, these life-threatening complications are frequent.11-15 Deficiencies in carnitine or selenium and acidosis areossible causes. Furthermore, electrophysiological changes such as prolongation of theT interval, which has been observed in patients with PA,11,16 can occur in patients with

ardiomyopathy and can predispose them to life-threatening ventricular arrhythmias.11,12

owever, electrophysiological changes have not been reported in this patient group.The aim of this study was to characterize electrocardiographic and echocardio-

raphic changes in patients with PA before to life-threatening arrhythmia or loss ofentricular function.

METHODSTen patients with PA followed by our institutions between July 2000 and December

005 (Department of Pediatrics, Innsbruck Medical University, Austria, 9 patients;hildren=s Hospital, Private Medical University, Salzburg, Austria, 1 patient) were

ncluded in this study. Age ranged between 1 day and 12 months at diagnosis, and

SD Atrial septal defectCG ElectrocardiogramS Fractional shortening

PCC Propionyl CoA carboxylaseQTc Corrected QT intervalVEB Ventricular ectopic beats

From the Clinical Departments of PediatricCardiology and General Pediatrics, Inns-bruck Medical University, Innsbruck, Aus-tria; the Laboratory of Metabolism, Univer-sity Children’s Hospital Freiburg, Freiburg,Germany; and the Children’s Hospital, Pri-vate Medical University, Salzburg, Austria.

Submitted for publication Apr 5, 2006; lastrevision received Sep 19, 2006; acceptedNov 17, 2006.

Reprint requests: Dr Daniela Karall (for-merly Skladal), Innsbruck Medical Univer-sity, Department of Pediatrics, Anichstrasse35, A-6020 Innsbruck, Austria. E-mail:[email protected].



A Propionic acidemia

92

10.1016/j.jpeds.2006.11.043

blTycifiry(ps3epsanamatodsppa7pd

cibeaI

trimtEa

oMmd(ptfl

itdaatrvarmsidewri

dpMwpebnr

g(iwoEuatpeerl

c(WmaEewtdp

P

etween 2.5 and 20.2 (median 9.0) years at the latest fol-ow-up investigation (Table I; available at www.jpeds.com).ime of follow-up ranged between 1.0 and 20.0 (mean 7.4)

ears. Diagnosis was established on the basis of urinary ex-retion of the characteristic organic acids in all patients, andt was confirmed by reduced enzymatic activity of PCC inbroblasts (Dr Regula Baumgartner, Department of Pediat-ics, University of Basel, Switzerland) and/or mutational anal-sis of the PCCA and PCCB genes in all but one patientTable I).2,3,17 Three different homozygous mutations wereresent. Patients 2 and 3, 4 and 5, 6 and 7, and 8 and 9 areiblings, whereas patients 8 and 9 are cousins of patients 2 and. The mutation of patients 1 through 3 and 8 and 9 (refer-nce 3) and of patients 7 and 8 (reference 2) have beenublished elsewhere.2,3 In all cases therapy consisted of re-tricted natural protein intake (1.1 to 1.5 g/kg daily), specialmino acid supplementation (free of valine, leucine, methio-ine, and threonine, to achieve adequate protein intake forge), L-carnitine (50-100 mg/kg body weight/day), and vita-in supplementation. All patients grew well (between 10th

nd 97th weight percentile). Seizures in patients 4 and 5 werereated with oral valproic acid. Patients 1 and 2 had a historyf syncope in their teenage years, which was probably vaso-epressor or neurally mediated. Patient 4 had a surgical clo-ure of a secundum atrial septal defect at 8 years of age. Threeatients died, after data acquisition for this study was com-leted: patient 2 at 16 years of age because of a bicycleccident (an underlying syncope cannot be excluded), patientat 6 years of age because of pneumococcal meningitis, and

atient 10 at 3 years of age after a prolonged metabolicecompensation.

According to our study protocol, we performed seriallinical investigations during routine follow-up examinationsn yearly intervals and in some patients during hospitalizationecause of metabolic decompensation. All patients and par-nts gave written informed consent. The study protocol waspproved by the institutional committee on human research innnsbruck.

We evaluated all patients according to our study pro-ocol as described below by standard 12-lead electrocardiog-aphy (ECG) and echocardiography on several occasions dur-ng yearly routine follow-up. Twenty-four-hour Holter

onitoring was performed in all patients at least once; inhose with arrhythmias up to seven recordings were done.xercise testing was performed in five patients, who were

ppropriate for bicycle ergometry (�6 years of age).Resting 12-lead ECGs were recorded at a paper speed

f 50 mm per second with a Corina recorder (General Electricedical Systems, Information Technologies, Freiburg, Ger-any) in supine position. All prospective recordings were

igitally stored. One investigator (D.B.) analyzed 2 to 15mean 6.8) ECGs per patient, in total, 64 ECGs, 44 of themrospectively. Twenty ECGs, which were registrated beforehe study started in July 2000, were analyzed retrospectivelyrom paper prints. RR and QT intervals were measured in
ead II from five nonconsecutive beats, and the corrected QT d
rolonged QTc Intervals and Decreased Left Ventricular Contractility in

nterval (QTc) was calculated by dividing the QT interval byhe square root of the RR interval (Bazett=s formula).18 Ad-itionally, we performed morphologic analysis of ST intervals,nd measurement of QT-interval dispersion, which is defineds the difference between the maximum and minimum QTime occuring in any of the 12 ECG leads, so far it can beeliably measured.19 For each patient the median QTc inter-al of all his or her standard ECG recordings (median QTc)s well as the maximum QTc interval of all standard ECGecordings (maximum QTc) and the QT dispersion deter-ined from the first ECG recording during the prospective

tudy period are shown in Table II. Interobserver reproduc-bility of QT and QTc measurements, calculated as the stan-ard deviation of the difference between measurements andxpressed as the percentage of the mean of the measurements,as determined after re-evaluation of randomly selected ECG

ecordings by a second investigator (U.S.) blinded to thenitial results.

Twenty-four-hour continuous ECG recordings wereone in all patients at least once (maximum seven recordingser patient) using a Lifecard CF Holter recorder system (Delar Reynolds Medical, Hertford, UK). Follow-up recordings

ere performed in yearly intervals if rhythm disturbances wereresent. After an initial automatic processing of the data, anxperienced observer reviewed the classification of abnormaleats (total number of ventricular ectopic beats [VEB], and totalumber of VEB occuring in couplets). Besides analysis ofhythm disorders, QTc at maximum heart rate was determined.

In patients �6 years of age, we performed upright,raded exercise testing on an ERG-900 bicycle ergometerErgoline, Bitz, Germany). The test protocol scheduled annitial 2-minute period with approximately 0.5 W/kg bodyeight (15, 20 or 25 W, respectively), followed by incrementsf 0.5 W/kg body weight every 2 minutes, until exhaustion.ach patient was verbally encouraged to continue exercisingntil his or her maximum voluntary exercise capacity wasttained. After exercise, patients were immediately placed inhe supine position to monitor ECG, heart rate, and bloodressure for 10 minutes. QTc was determined before, duringxercise at the end of each workload level, and after exercisevery 2 minutes for at least 10 minutes until heart rateeturned to the resting value. The maximum QTc was se-ected from these measurements.

All patients underwent a complete transthoracic echo-ardiographic examination using a System Five echo machineGeneral Electric Vingmed Ultrasound, Horten, Norway).

e performed two-dimensional guided M-mode measure-ents of the left ventricle to measure cardiac performance

ccording to recommendations of the American Society ofchocardiography20; therefore we determined left ventricular

nd-diastolic, end-systolic, luminal, and wall diameters, ande calculated fractional shortening (FS). Anatomy and func-

ion of the mitral and aortic valve were assessed by two-imensional echocardiography, color flow mapping, andulsed-wave and continuous-wave Doppler recordings. Car-
iomyopathy was classified as dilated on the basis of mor-
Patients with Propionic Acidemia 193


pbiS

p1pt37iir7bw((

sfQ(bl

eII

.E

lect

roca

rdio

grap

hic

and

echo

card

iogr

aphi

cch

arac

teri

stic

s

tien

tm

ber

Sta

ndar

dE

CG

Max

(med

ian)

QT

c(m

s)

Sta

ndar

dE

CG

repo

lari

zati

on

abno

rmal

itie

s

Sta

ndar

dE

CG

QT

disp

ersi

on

(ms)

Exe

rcis

ete

stin

gM

axQ

Tc

(ms)

Ho

lter

EC

GM

axQ

Tc

(ms)

Ho

lter

EC

Gar

rhyt

hmia

s/sy

nco

pes

Ech

oca

rdio

grap

hyL

Vfr

acti

ona

lsh

ort

enin

g(%

)M

in-M

ax(m

edia

n)E

cho

card

iogr

aphi

cab

norm

alit

ies

147

0(4

20)

pret

erm

Tin

v30

440

470

sync

opes

30-3

5(3

2)2

460

(445

)pr

eter

mT

inv

4050

045

5sy

ncop

es26

-31

(28.

5)3

500

(470

)pr

eter

mT

inv

4050

545

024

-35

(29)

MI,

dila

ted

LV4

460

(445

)pr

eter

mT

inv

5050

051

2co

uple

t,si

ngle

VEB

,si

nus

brad

ycar

dia,

sinu

sar

rest

(2.3

s)

28-4

0(3

2)M

I,A

SDcl

osur

e,m

itral

prol

aps

549

0(4

60)

pret

erm

Tin

v40

550

590

VEB

,sin

usbr

adyc

ardi

a30

-38

(33)

646

0(4

50)

4044

033

-38

(34)

744

0(4

20)

4044

030

(30)

843

0(4

30)

4043

034

-36

(35)

948

0(4

40)

pret

erm

Tin

v40

454

31-3

4(3

2.5)

1045

0(4

00)

flat

T40

450

18-2

7(2

0)di

late

dca

rdio

myo

path

y

,atr

ials

epta

ldef

ect;

LV

,lef

tve

ntri

cula

r;M

ax,m

axim

um;M

I,m

itral

insu

ffici

ency

;Min

,min

imum

;pre

term

Tne

g,pr

eter

min

alT

inve

rsio

n;V

EB

,ven

tric

ular

ecto

pic

beat

s.

F(pqTQs

194

hology and function of the ventricles, when dilatation ofoth ventricles and poor pump function were present.21 Allmages were digitally stored as raw data with the EchoPacystem, version 6.4.1 (General Electric Vingmed).

RESULTSMaximum QTc was abnormal (�440 ms) in 70% of

atients beyond infanthood (patients 1-6, 10; Table II, Figure, and Figure 2; available at www.jpeds.com), and clearlyrolonged (�460 ms) in 60% of patients (patients 1-6) inheir standard 12-lead ECGs. In 2 of the 6 patients (patients

and 5), prolonged QTc at the initial examination at 3 andyears of age remained increased during follow-up, whereas

n the other four patients an initially normal QTc (�440 ms)ncreased to prolonged values at 3, 13, 14, and 16 years of age,espectively (Figure 3). The four younger patients (patients-10) showed QTc values �460 ms at all investigationseyond infanthood. At the last follow-up investigation, QTcas normal (410-440 ms) in 40% of patients, mildly increased

440-460 ms) in another 40%, and significantly prolonged�460 ms) in 20%.

Interobserver reproducibility, which was calculated ineven patients, was 1.14% for QT measurements, and 1.11%or QTc determinations. No correlation between prolongedTc values and routinely investigated biochemical variables

electrolytes including calcium and magnesium, blood gas

igure 1. Box plot of maximum QTc (QTcmax) and median QTcQTcmedian) of all 64 standard 12-lead electrocardiograms of the 10atients with PA. Median (black line inside the box), first and thirduartile (margins of the box), minimum and maximum QTc are shown.he dotted line represents the upper limit of normal QTc. Note that theTcmax of most patients is above normal, and that QTcmedian of all

equential ECG recordings is normal in approximately half of the patients.

analysis, ammonia concentrations, plasma aminoacids, andTa

Pa

nu ASD

Baumgartner et al The Journal of Pediatrics • February 2007


an5

ht(cvwtwcomQigAaipp

o(Q

EBSsrnshswh

cs

wsa

totit3dataIcnmoi

hEtnv

itQsff1bapbdttds

dlirhiab

F(il

P

cylcarnitine profile) was found. QT dispersion was withinormal ranges in all patients. Note the largest dispersion of0 ms in one of the older patients (patient 4, Table II).

To unmask prolongation of repolarization at highereart rates, QTc was determined during and after exerciseesting in the five older patients. In 4 of these 5 patientspatients 2-5) QTc exceeded resting QTc and values werelearly pathological (�500 ms). We observed maximum QTcalues during the first or second step of graded exercising at aorkload of about 20 W to 50 W in all four patients, before

he heart rate increased so much, that the T wave mergedith the following P wave and shorter QTc values were

alculated. During recovery, the highest QTc values werebtained at about 6 to 8 minutes after exercise, when theerging of T and P waves disappeared. Individual maximumTc values during exercise and during recovery were similar

n all five patients. In the five younger patients, bicycle er-ometry could not be performed because of technical reasons.lternatively, QTc was determined manually in all patients—

lso the young ones—from 24-hour continuous ECG record-ngs during episodes of maximum heart rate as a result ofhysical activity. However, QTc values of the five youngatients did not exceed resting QTc measurements.

Preterminal T wave inversion in leads V4 to V6 as signf abnormal repolarization was present in 60% of patientspatients 1-5, 9; Figure 2), and was associated with maximumTc �460 ms in standard ECG in all cases.

Rhythm disturbances were documented on continuousCG recordings in 20% of patients (siblings 4 and 5; Table II).oth siblings were treated with valproic acid because of seizures.ingle VEB, couplets, and episodes of sinus bradycardia andinus arrest (maximum length of 2.3 s) in those two did notequire treatment. Patients 1 and 2 reported syncope, which wasot shown on ECG. Notably, all patients with a history ofyncope and all patients with documented rhythm disturbancesad a maximum QTc �460 ms and preterminal T wave inver-ion on standard ECG. In the two patients (patients 7 and 10),ho died in the hospital, no arrhythmias were noted during theirospitalization.

Two-dimensional echocardiography showed normalardiac structure in all but one patient. Patient 4 had an atrial

igure 3. QTc values from standard electrocardiograms of patients 4left) and 2 (right) were plotted against age. The trend lines show thencrease of QTc length with age. The arrows indicate the upper normalimit of QTc.

eptal defect (ASD) at 3 months of age, which was associated m


ith mitral valve prolapse and mild mitral regurgitation. Afterurgical ASD closure at 8 years of age, mitral valve prolapsessociated with mild regurgitation remained.

M-mode echocardiographic determination of left ven-ricular systolic function revealed a reduced FS �30% at anyccasion beyond the neonatal period in 40% of patients (pa-ients 2-4, 10; Table II). This reduction of FS was moderaten patients 2 to 4 (minimum 26%, 24%, and 28%, respec-ively), but more severe in patient 10 (minimum 18%). Patienthad borderline dilatation of the left ventricular enddiastolic

iameter with reduced FS values on 4 of 6 occasions, associ-ted with mild mitral regurgitation. Patient 10 showed biven-ricular dilatation and diminished FS values at all investigations,nd therefore was diagnosed as having dilated cardiomyopathy.n the girl with ASD (patient 4), FS normalized after its surgicallosure at 8 years of age. In patient 2, an initially reduced FSormalized at 14 years of age without any specific cardiac treat-ent. Left ventricular systolic function was normal (FS � 30%)

n all occasions in 6 of 10 patients, and, at the last follow-upnvestigation, showed normal values in 8 of 10 patients.

All four patients with reduced left ventricular contractilityad prolonged maximum QTc values �450 ms on standardCG and Holter recordings. On the other end of the spectrum,

wo young patients (patients 7 and 8) were characterized byormal resting and exercise QTc values (440 ms) and normal leftentricular systolic function.

DISCUSSIONIn 1993, Massoud et al11 reported cardiac complications

n a group of six patients with propionic acidemia. With 50%he related mortality was alarming. In one patient a prolongedT interval was noted. Several case reports have followed

ince then.12-16 An 8-year-old girl, after a 2-week history ofeeling “poorly” became acutely ill and succumbed to heartailure and ventricular fibrillation in 12 hours.12 Another 4/2–year-old girl became comatose during a simple infectionecause of a metabolic decompensation and died 2 days afterdmission as a result of arrhythmia.14 A 23-year-old womanresented with a 1-week history of shortness of breath causedy an adult-onset cardiomyopathy.13 Three further cases ofilated cardiomyopathy were reported, one of the case pa-ients underwent liver transplantation, which also improvedhe cardiac complication.15 A 7-year-old girl fulfilled theiagnostic criteria of both PA and asymptomatic long-QTyndrome.16

We followed a cohort of 10 pediatric patients with threeifferent genotypes of PA longitudinally. The electrophysio-

ogical investigations confirmed the finding of prolonged QTcnterval (�440 ms) in 70% of them, which is an independentisk factor for sudden cardiac death.22,23 Throughout child-ood, QTc prolongation seems to fluctuate, but it tends to

ncrease (Figure 3), and affected all patients �4 years of age,lthough in 3 of the 4 younger patients QTc intervals rangedelow the upper normal limit.

In view of the two reported cases with fatal arrhyth-
ia,12,14 the overall incidence of ventricular dysrhythmia in

osHtsrsr

pVwgfp

moctatQdcdmwifmpn

mmiSarpcpoicmei

glKnyg

opfwte

cl2t

Wp

1ma22Ca3Wa4b15BD6nB7Nc8a9s12111sM1H1P21rM1p1l111m2t

1

ur young patient group was low (20%) and no patienthowed sustained ventricular tachycardia during the 24-hourolter ECG. Sinus arrest and sinus bradycardia occurred in

wo siblings (patients 4 and 5) treated with valproic acid foreizures. The relationship between sinus arrest, which has notequired treatment, and valproic acid remains unclear. In theame two patients, we observed isolated VEB, which wereegarded as normal, and a couplet of VEB was detected once.

The most frequent ECG abnormality apart from QTcrolongation was preterminal T wave inversion in leads V4 to6, a sign of repolarization abnormality, which was associatedith QTc prolongation and older patient age. Because ofenetic heterogeneity and phenotypic variability in PA, therequency of ECG alterations may be different in other studyopulations.

Our observations are not obviously connected to theetabolic state of the patients, as most of the investigations in

ur cohort were performed in stable out patient under controlonditions. We observed no obvious correlation between rou-inely investigated biochemical variables (blood gas analysis,mmonia concentrations, plasma amino acids, and acylcarni-ine profile) and the electrophysiological results. However,Tc prolongation in patients with PA may be related in some

irect way to the underlying metabolic defect. Some authorslaim that deficiency of carnitine may induce electromyocar-ial changes.24 It seems possible that this is an underlyingechanism, although all of our patients were supplementedith L-carnitine and regular determinations of free carnitine

n blood were normal (data not shown). Notably, total andree carnitine concentrations were reported to be low in heartuscle of a patient with PA and cardiac hypertrophy at

ostmortem despite carnitine supplementation and repeatedlyormal plasma carnitine levels.12

Three further mechanisms seem possible. First, thereight be a direct toxic effect from one or a combination ofetabolites present in the metabolism of these patients, caus-

ng prolonged repolarization of the myocardium over time.imilar mechanisms are described for several drugs.25 Second,n intracardiac depletion of an essential—that is, an anaple-otic—substrate in the intermediary metabolism, with trap-ing of another, might cause energy deficiency in the myo-ardium and lead to pathology. A third possibility for therolongation of QTc intervals could be the inhibition of thexidative phosphorylation in mitochondria by propionyl CoAn similar to the way methylmalonic acid affects respiratoryhain complex II and causes nephropathy.26 The myocardiumight be sensitive for energy deficiency. So far, these hypoth-

ses cannot be demonstrated. Investigations of tissue requir-ng cardiac biopsy would be necessary.

The possibility that some patients with PA have aenetic abnormality, that is, a variant of the congenitalong-QT syndrome, as proposed by Schwartz et al27 andakavand et al,16 seems unlikely, as the QT prolongation isot permanently present in the patients with PA, and theounger patients seem less affected than the older ones, sug-
esting that this is an ongoing progressive process. On the
s2

96 Baumgartner et al

ther hand, it is possible, that the presence of ion-channelolymorphisms causing minor alterations in ion-channelunction, like the D85N polymorphism in the KCNE1 gene,hich has been associated with drug-induced QTc prolonga-

ion, plays a modifying role.22 Further specific studies tolucidate the mechanism involved would be of interest.

We recommend regular electrocardiographic and echo-ardiographic investigations in all these patients including ateast yearly ECGs with determination of QTc and regular4-hour Holter monitoring to detect life-threatening, butreatable ventricular arrhythmias.

e wish to thank all colleagues involved in the medical care of ouratients.

REFERENCES. Fenton WA, Gravel WA, Rosenblatt DS. Disorders of propionate and methyl-alonate metabolism. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The Metabolic

nd Molecular Bases of Inherited Diseases. 8th ed. New York: McGraw-Hill,001:2165-93.. Perez B, Desviat LR, Rodriguez-Pombo P, Clavero S, Navarrete R, Perez-Cerda, et al. Propionic acidemia: identification of twenty-four novel mutations in Europe

nd North America. Mol Genet Metab 2003;78:59-67.. Muro S, Rodriguez-Pombo P, Perez B, Perez-Cerda C, Desviat LR, Sperl

, et al. Identification of novel mutations in the PCCB gene in European propioniccidemia patients. Human Mutat 1999;14:89-90.. Lehnert W, Sperl W, Suormala T, Baumgartner R. Propionic acidaemia: clinical,iochemical and therapeutic aspects. Experience in 30 patients. Eur J Pediatr994;153:68-80.. Sweetman L, Williams JC. Branched chain organic acidurias. In: Scriver CR,eaudet AL, Sly WS, Valle D, eds. The Metabolic and Molecular Basis of Iinheritedisease. 8th ed. New York: McGraw-Hill; 2001:2125-63.

. Ogier de Baulny H, Saudubray JM. Branched-chain organic acidurias. In: Fer-andes J, Saudubray JM, van den Berghe G, eds. Inborn Metabolic Diseases. 3rd ed.erlin, Heidelberg, New York: Springer; 2000:196-212.. North KN, Korson MS, Gopal YR, Rohr FJ, Brazelton TB, Waisbren SE, et al.eonatal-onset propionic acidemia: neurologic and developmental profiles, and impli-

ations for management. J Pediatr 1995;126:916-22.. Sass JO, Hofmann M, Skladal D, Mayatepek E, Schwahn B, Sperl W. Propioniccidemia revisited: a workshop report. Clin Pediatr 2004;43:837-43.. Kohlschütter A, Hausdorf G. Primary (genetic) cardiomyopathies in infancy. Aurvey of possible disorders and guidelines for diagnosis. Eur J Pediatr 1986;145:454-59.0. Guertl B, Noehammer C, Hoefler G. Metabolic cardiomyopathies. Int J Exp Path000;81:349-72.1. Massoud AF, Leonard JV. Cardiomyopathy in propionic acidaemia. Eur J Pediatr993;152:441-5.2. Mardach R, Verity MA, Cederbaum SD. Clinical, pathological, and biochemicaltudies in a patient with propionic acidemia and fatal cardiomyopathy. Mol Genet

etab 2005;85:286-90.3. Bhan AK, Brody C. Propionic acidemia: a rare cause of cardiomyopathy. Congesteart Fail 2001;7:218-9.

4. Lücke T, Perez-Cerda C, Baumgartner M, Fowler B, Sander S, Sasse M, et al.ropionic academia: unusual course with late onset and fatal outcome. Metabolism004;53:809-10.5. Ou P, Touati G, Fraisse A, Sidi D, Kachaner J, Saudubray JM, et al. Une causeare de cardiomyopathie de l’enfant: l’acidemie propionique. A propos de 3 cas. Arch

al Cœur 2001;94:531-3.6. Kakavand B, Schroeder VA, Di Sessa TG. Coincidence of long QT syndrome andropionic acidemia. Pediatr Cardiol 2006;27:160-1.7. Sperl W, Murr C, Skladal D, Sass JO, Suormala R, Wendel U. Odd-numberedong-chain fatty acids in propionic acidaemia. Eur J Pediatr 2000;159:54-8.8. Bazett HC. An analysis of the time-relations of electrocardiograms. Heart920;7:353-70.9. Day CP, McComb JM, Campbell RWF. QT dispersion: an indication of arrhyth-ia risk in patients with long QT intervals. Br Heart J 1990;63:342-4.

0. Sahn DJ, DeMaria A, Kissolo J, Weyman A. Recommendations regarding quan-itation in M-mode echocardiography: results of a survey of echocardiographic mea-
urements. Circulation 1978;58:1072-83.1. Goodwin JF. The frontiers of cardiomyopathy. Br Heart J 1982;48:1-18.

2d2Ja2Rp

2d2mm2

P

2. Moss AJ. QTc prolongation and sudden cardiac death. The association is in theetail. J Am Coll Cardiol 2006;47:368-9.3. Straus SMJM, Kors JA, De Bruin ML, van der Hooft CS, Hofman A, Heeringa, et al. Prolonged QTc interval and risk of sudden cardiac death in a population of olderdults. J Am Coll Cardiol 2006;47:362-7.4. Rijlaarsdam RS, van Spronsen FJ, Bink-Boelkens MTE, Reijngoud DJ, Wanders
JA, Niezen-Koning KE, et al. Ventricular fibrillation without overt cardiomyopathy as firstresentation of organic cation transporter 2-deficiency in adolescence. Pace 2004;27:675-6.
P1


5. Mitcheson JS, Chen J, Lin M, Culberson C, Sanguinetti MC. A structural basis forrug-induced long QT syndrome. Proc Natl Acad Sci USA 2000;97:12329-33.6. Kölker S, Schwab M, Hörster F, Sauer S, Hinz A, Wolf NI, et al. Methyl-alonic acid, a biochemical hallmark of methylmalonic acidurias but no inhibitor ofitochondrial respiratory chain. J Biolog Chem 2003;278:47388-93.

7. Schwartz PJ, Stramba-Badiale M, Segantini A, Austoni P, Bosi G, Giorgetti R, et al.
rolongation of the QT interval and the sudden infant death syndrome. N Engl J Med998;338:1709-14.
50 Years Ago in The Journal of PediatricsSPONTANEOUS PNEUMOTHORAX IN THE FIRST TEN DAYS OF LIFE

Howie VM, Weed AS. J Pediatr 1957;50:6-15

By the 1950s, the transition of early newborn care from home to hospital in the United States was well established.Hospital care afforded ready access to newer technologies such as radiographic imaging, which in turn was a means ofrevealing pathologic findings to clinicians in a new context. Extra-pleural air accumulation, once considered an autopsydiagnosis, was now being identified in viable neonates, raising new questions about treatment and prognosis.

Howie and Weed’s retrospective case series provided clinicians with an initial perspective of pneumothorax in the earlyneonatal period. The authors subdivided their patients into 4 categories—mantle, alternating, bilateral or tensionpneumothorax—on the basis of radiographic and clinical criteria. They noted a wide range of reported incidence, likelyrelated to selection bias and lack of standardized techniques for newborn infant chest radiography. Howie and Weed’sseries focused on symptomatic term and late pre-term infants and preceded the era of positive pressure assisted ventilationfor the management of newborn respiratory distress. However, their case series demonstrated the importance ofidentifying and treating tension pneumothorax through “prompt aspiration and continuous drainage by underwater trap.”They also reported on several infants successfully treated by using direct needle aspiration of the extra-pleural aircollection. Although they also cite oxygen therapy as a component of management, they did not specifically refer to thenotion of nitrogen washout.

Howie and Weed did note an association between pneumothorax and other pulmonary pathologies, includingatelectasis and interstitial emphysema. Later case series reflected this as well, proposing needle aspiration, chest tubeplacement, nitrogen washout, or careful observation, depending on the severity and the presence of tension phenomena.1

By the 1970s, the epidemiology of extra-pleural air leak shifted from term to pre-term infants, as treatments for hyalinemembrane disease and mechanical ventilation techniques became more widespread.2

In the 21st century, we still understand pneumothorax in the early neonatal period as a protean phenomenon that mayrequire anything from careful observation to chest tube placement. Because of the lack of prospective trials to date, Howieand Weed’s approach remains a relevant component of our management.

James M. Greenberg, MDDivision of Neonatology

Cincinnati Children’s Hospital Medical CenterCincinnati, OH

10.1016/j.jpeds.2006.07.039

REFERENCES1. Chernick V, Avery ME. Spontaneous alveolar rupture at birth. Pediatrics 1963;32:816-24.2. Yu YVH, Liew SW, Robertson NRC. Pneumothorax in the newborn: changing pattern. Arch Dis Child 1975;50:449-53.


Fl

T

�

f

1

igure 2. The standard ECG of patient 1 at 20 years of age shows borderline prolongation of QTc interval (447 ms) and preterminal T wave inversion ineads II, aVF, and V4 to V6.

able I. Demographic characteristics and disease course

Patientnumber

Age atlast

follow-up(years)

Sex(f/m) Mutation Gene

PCCactivity in

fibroblasts�

Age atonset of

symptomsAge at

diagnosisWeight

percentile Clinical course

1 20.2 f V205D pccB 0.9% 1 month 2 months 25th Syncopes2� 15.0 f V205D pccB 1.1% 11 months 12 months 90th Syncopes, deceased

at age 16 years3� 13.9 m V205D pccB 0.8% Asymptomatic 3 months 90th Uneventful4† 19.2 f M348K pccA 1.2% 1 month 4 months 10th Seizures5† 15.6 m M348K pccA 1.3% 1 week 14 days 10th Seizures6‡ 4.1 f IVS11-2A¡G pccB 0.4% 3 days 5 days 97th Recurrent infections7‡ 2.6 m IVS11-2A¡G pccB 2.1% Asymptomatic 1 day 25th Recurrent infections,

deceased at age 6 years8§ 3.1 m V205D pccB 0.8% 5 days 5 days 10th Uneventful9§ 2.6 m V205D pccB n.d. Asymptomatic 2 days 25th Severe metabolic

decompensation (age 9months)

10 2.5 m n.d. n.d. n.d. 20 days 30 days 10th Dilated cardiomyopathy,deceased at age 3 years

,†,‡, and § indicate siblings of families 1, 2, 3, and 4. Patients 8 and 9 are cousins of patients 2 and 3., female; m, male; n.d., not determined; PCC, propionyl CoA carboxylase.

97.e1 Baumgartner et al The Journal of Pediatrics • February 2007

Aemdf

bsdat

isr

M

aamc(

nimlgnaemt

MPP

1

GRANDROUNDS

A Newborn Infant with Protracted Diarrhea and Metabolic Acidosis

AMIR BAR, MD, ARIEH RISKIN, MD, THEODORE IANCU, MD, IRENA MANOV, PHD, AYALA ARAD, MD, AND RON SHAOUL, MD

term baby girl was born via cesarean section because of a nonprogressive labor, with Apgar score of 10 at 1 and 5minutes. Her birth weight was 3310 g. Meconium-stained amniotic fluid was noted, but the baby was vigorous and hadnormal physical examination at birth. She is the first-born child to parents who are first-degree cousins of Arab

thnicity. The mother is known to have factor V-Leiden deficiency, and she had a history of four unexplained early-stageiscarriages. She was treated with low-molecular-weight heparin during pregnancy. There was no report on polyhydramnion

uring the pregnancy. The father had a history of cow’s milk allergy. There are no known metabolic or genetic disorders in theiramilies.

INITIAL PRESENTATIONThe baby was doing well for the first 3 days. She received cow’s milk–based formula and gradually switched to exclusive

reast feeding on the third day. On her fourth day of life, she presented with multiple yellowish-watery stools. She had clinicaligns of mild dehydration without fever or vomiting, and her vital signs were normal. The abdomen appeared neither tender noristended, and the rest of her physical examination was normal. The laboratory findings revealed hypernatremia of 151 mEq/L,nd severe metabolic acidosis: blood pH, 7.11; HCO3

�, 8.2 mEq/L; and base excess of �21.5. There was no leukocytosis, and

he serum level of C-reactive protein was normal.

HOSPITAL COURSEWorking Diagnosis: Infectious Diarrhea and Dehydration-Induced Metabolic Acidosis?The infant was treated with intravenous fluids as well as antibiotics after obtaining blood and stool cultures, assuming an

nfectious etiology. After 24 hours of rehydration therapy, the diarrhea, dehydration, and hypernatremia resolved; however, theevere metabolic acidosis persisted: blood pH, 7.27; HCO3

�, 9.5 mEq/L, and base excess, �17.5. This discrepancy between theapid response to rehydration and the intractable metabolic acidosis raised the possibility of alternative diagnoses.

etabolic AcidosisThe differential diagnosis of metabolic acidosis is wide, and the first diagnostic step should be the calculation of serum

nion gap (sAG; calculated as: [Na� � (HCO3� � Cl�)]). Increased sAG indicates overload of an acid with an additional

nion, suggesting clinical conditions such as lactic acidosis, ketoacidosis, and amino or organic acidemia. A normal sAGetabolic acidosis implies a decreased H� excretion or excessive HCO3

� losses, as a result of either renal or gastrointestinalauses. In children and adults the upper limit of normal sAG is 12 mEq/L, whereas a higher threshold is used in infants16 � 4 mEq/L).1

In our case, the sAG level was borderline (sAG � 16 mEq/L); therefore, bothormal and increased anion gap metabolic acidosis etiologies were approached. For an

ncreased sAG etiology, there were no clinical or laboratory findings of sepsis, theetabolic acidosis persisted long after the initial dehydration had resolved, and serum

actate was normal, thus excluding lactic acidosis. Although there was a known consan-uinity, there was no family history of inborn errors of metabolism and physical exami-ation showed no dysmorphic features. Laboratory evaluation to exclude amino or organiccidemias was negative. The absence of ketonuria, and normal blood glucose levels,xcluded ketoacidosis. Because increased sAG etiologies were excluded, normal AGetabolic acidosis etiologies were approached. As the diarrhea seemed to resolve, renal

ubular acidosis (RTA) was the major working diagnosis at this stage.

ID Microvillus inclusion diseaseAS periodic acid-Schiff

RTA Renal tubular acidosissAG Serum anion gap

From the Departments of Pediatrics (A.B.,R.S.), NICU (A.R.), Department of Pathol-ogy (A.A.), Bnai Zion Medical Center; Elec-tron Microscopy Unit, Faculty of Medicine,Technion - Israel Institute of Technology(T.I., I.M.), Haifa, Israel.

Reprint requests: Dr Ron Shaoul, Head,Pediatric Day Care Unit, Department ofPediatrics, Bnai Zion Medical Center, 47Golomb St. POB 4940, Haifa 31048, Israel.E-mail: [email protected].

J Pediatr 2007;150:198-201



DI Protracted diarrhea of infancy uAG Urine anion gap

98

10.1016/j.jpeds.2006.11.041

R

mr(ptiolbphltcltTIpttt

tHHicfanc(dpu(Ractt

D

onvcbstdc

scs

cdtdfoibat

pttbncpbipvdiwlattcdbs(icrcae

ba(tibtm

A

enal Tubular AcidosisRTA is characterized by normal sAG hyperchloremic

etabolic acidosis resulting from either impaired HCO3�

eabsorption or impaired H� excretion.2 In proximal RTAtype 2) the threshold for bicarbonate reabsorption in theroximal tubule is lower than normal (15-16 mEq/L ratherhan 22-23 mEq/L). Because 85% of the filtered bicarbonates normally reabsorbed in the proximal tubule, high doses ofral bicarbonate are needed to normalize the HCO3

� serumevel in patients with proximal RTA. In these patients, serumicarbonate levels will not reach the normal range, despiterovision of very large amounts of bicarbonate. On the otherand, even without administration of bicarbonate, the serum

evel will not drop below 15 to 16 mEq/L (the proximalubule threshold for HCO3

� reabsorption). In the presentase, the serum HCO3

�level was substantially below thatevel (�10 mEq/L). After intensive bicarbonate supplementa-ion, the serum level was within the normal range (23 mEq/L).hese findings made the diagnosis of proximal RTA unlikely.

n addition, proximal RTA is rarely isolated; it is usually aart of Fanconi’s syndrome, which causes other proximalubule reabsorption abnormalities (e.g. glycosuria, phospha-uria, and aminoaciduria). These abnormalities were absent inhis case.

In distal RTA (type 1), the distal tubule cells are unableo secrete H� to the tubule lumen because of abnormal

�/Na� exchanger. Normally, the distal tubule cells secrete� ions coupled with NH3 to form NH4

�, which is excretedn the urine as NH4�/Cl�. This distal tubule urine acidifi-ation mechanism allows alleviation of metabolic acidosis, aunction that can be assessed by the calculation of the urinenion gap (uAG), calculated as: (Na� � K�) � Cl�. Inormal subjects, in conditions of metabolic acidosis, the in-reased excretion of NH4�/Cl� results in negative uAGexcess of Cl�) and a low urine pH. In contrast, patients withistal RTA present with alkaline urine pH, and zero orositive uAG. In our case, during metabolic acidosis, therine was acidified (urine pH: 5.0), and the uAG was negative�21.8 mEq/L), which contradicts the diagnosis of distalTA.2 In addition, serum potassium levels are frequently

bnormal in RTA.2 Therefore, the laboratory findings in ourase did not match the characteristic presentations the RTAypes. This, along with the persistent diarrhea, made a gas-rointestinal tract etiology highly probable.

iarrheaInfectious diarrhea, the most common identified cause

f diarrhea in the first week of life,3 was ruled out based onegative stool samples for bacteria, parasites, and viruses. Theolume of diarrhea reached 150 to 200 mL/kg/day, apparentlyeased with fasting (60 mL/kg/day), and recommenced onreast-feeding; thus, an osmotic diarrhea was presumed. Stoolample analysis can provide supportive findings for identifyinghe mechanism of diarrhea; osmotic versus secretory. Osmoticiarrhea is characterized by lower pH (�5.0), lower sodium
oncentration (�70 mEq/L), and positive reducing substances.4 w
Newborn Infant with Protracted Diarrhea and Metabolic Acidosis

In the present case, stool sample examination revealed aodium content of 61 mEq/L, potassium of 50 mEq/L,hloride of 57 mEq/L, pH of 6, and mildly positive reducingubstances.

Two congenital carbohydrate malabsorption disordersan present with severe osmotic diarrhea within the first feways of life: congenital lactase deficiency and glucose-galac-ose malabsorption.5 Both are rare autosomal recessive disor-ers. Because the diarrhea did not resolve on hydrolyzedormula (Neocate®) containing glucose polymer, an etiologyf disacchridase deficiency was unlikely. Subsequently, thenfant was fed a carbohydrate-free formula (RCF, Ross Car-ohydrate Free®) with no resolution of the diarrhea or met-bolic acidosis. This ruled out glucose-galactose malabsorp-ion, and any other carbohydrate malabsorption.

At this stage, the baby was nil per os, receiving totalarenteral nutrition. In contrast to the previous fasting trials,he diarrhea persisted. This raised the suspicion that some ofhe watery diarrheal stools were previously mistaken for urineecause of their clear nature without any content. Rectalasogastric (NG) tube insertion revealed a large amount oflear watery diarrhea. Therefore, the working diagnosis wasrotracted diarrhea of infancy (PDI). The causes of PDI cane classified into two categories based on the findings ofntestinal biopsy: PDI with normal villi (without villus atro-hy), and PDI with villus atrophy.3 In the category of normalillous PDI, congenital ion transport defects cause secretoryiarrhea presenting at birth. The most common congenitalon transport defect is the chloride-bicarbonate exchanger, inhich HCO3� is not secreted into the gastrointestinal tract

umen, leading to alkalosis. Obviously, the severe metaboliccidosis in our case did not support this diagnosis. In addi-ion, the fecal chloride concentration, which did not exceedhe fecal sodium level, did not support the diagnosis ofongenital chloride diarrhea. Congenital sodium diarrhea, aisorder of the intestinal H�/Na�-exchanger, is characterizedy hyponatremia, alkaline diarrhea, and high concentration oftool sodium. In the other congenital ion transport defectssodium co-transporters), abnormal reabsorption of sodium isnvolved, which usually results in hyponatremia, unlike ourase.3 Therefore, an ion transporter defect was unlikely. Othereported causes of protracted diarrhea with normal villi in-lude heparan sulphate deficiency,6 glycosylation defect,7 andmutation in the neurogenin-3, a protein that is required for

ndocrine-cell development in the pancreas and intestine.8

Protracted diarrhea of infancy with villous atrophy cane a result of infectious or post-infectious enteropathy, andutoimmune or allergic enteropathy, the IPEX syndromei.e., Immune dysregulation, Polyendocrinopathy, Enteropa-hy, X-linkage). In these disorders, an immune process isnvolved, and histologically the villus atrophy is characterizedy T-cell infiltration. Based on the report by Cuenod et al,9

he etiology of villous atrophy may be classified as immune-ediated or developmental. PDI with villous atrophy but
ithout an inflammatory infiltrate is caused by two primary
199

sastftmmttmctscfo

tue

ssefi

roei

tbf

gtd

a

2

tructural enteropathies: microvillus inclusion disease (MID)nd tufting enteropathy. MID is a severe enteropathy, theecond most common cause (following infectious) of pro-racted diarrhea starting at the first few days of life, although,ew may present later up to 2 months of age,10 whereas,ufting enteropathy usually presents later, in the first fewonths of life. MID is characterized by an abnormal accu-ulation of periodic acid-Schiff (PAS) positive material in

he apical cytoplasm of epithelial cells, an increase in “secre-ory granules” by electron microscopy, and the presence oficrovillous-lined vacuoles, or inclusions, within the apical

ytoplasm of surface epithelial cells.11-13 It is hypothesizedhat the PAS stain abnormality and the electron microscopyecretory granules represent the same phenomenon.14 Theharacteristic feature of tufting enteropathy is the presence ofocal epithelial “tufts” composed of closely packed enterocytesn light microscopy.3

The next diagnostic step therefore was to perform in-estinal biopsies, which were retrieved from the duodenum bypper endoscopy. Mucosal samples examined by light and
lectron microscopy were compatible with MID. Figure 1 i
00 Bar et al

hows CD10 (1A), PAS (1B), and alkaline phosphatase (1C)tains compatible with MID. Figure 2 shows the duodenallectron micrograph from this patient showing the typicalndings of MID.

Because all patients with MID have poor prognosis andemain dependent on parenteral nutrition, the only treatmentption for them is small bowel transplantation.15 Ruemmelet al have recently reported a survival rate of 86% afterntestinal transplantation in seven patients with MID.15

AMBULATORY COURSEThe baby, who is now 18 months old, is thriving on

otal parenteral nutrition and is dependent on 50 mEq/L oficarbonate day. An attempt of oral feedings with hydrolyzedormula failed.

The mother has recently delivered a preterm (36 weeks)irl, who developed watery diarrhea and metabolic acidosis onhe 3rd day of life. Subsequent duodenal and rectal biopsiesemonstrated the characteristic findings of MID.

It has been proposed that the primary molecular defect

igure 1. Unlike the sharp linear brush border stain of the normaltestine,18 this figure shows the typical prominent cytoplasmic

taining of surface enterocytes (arrow heads) seen in MID. A,taining for CD10,18 a membrane-associated neutral peptidase,idely used in lymphoma phenotyping. B, PAS stain showing an

pical cytoplasmatic PAS positive material.18,19 C, staining forlkaline phosphatase showing the same findings.20,21

Finsswa

n MID involves membrane trafficking of immature/differen-


tarupaltig

Wi

12223p4N5GR46Cd7Nc28n9Cc1o1a1N16L1aG1pp1oG1o1dP1w2i

FlaiSc(O

A

iating enterocytes.14 Similar to other reports of occurrencemong siblings,10,16 our cases strongly suggest an autosomalecessive inheritance. In addition to siblings’ data, a cluster ofnrelated cases of MID has been reported in the Navajoopulation (Arizona, US).17 This clustering of cases suggests“founder effect,” in which rare genes in the general popu-

ation are more frequent in a genetically isolated populationhat has shared ancestor genes. No specific genes have beendentified or have been proposed as participants in the patho-

igure 2. Electron micrograph of duodenal enterocyte of patient 1. Theuminal face shows a few small, distorted microvilli (small arrows) as wells completely denuded areas. Beneath such an area, a round typicalnclusion is seen, containing inward-pointing microvilli (large arrow).ome of the microvilli are symmetrically arranged, other are transversallyut. Beyond the inclusion, small round or vermiform bodies can bee seenarrowheads), frequently seen in enterocytes of patients with MID.riginal magnification � 8000.

enesis of MID.2a

Newborn Infant with Protracted Diarrhea and Metabolic Acidosis

e thank Dr Erella Livne and Mrs Pessia Schenzer for their helpn the alkaline phosphatase stain.

REFERENCES. Schwaderer AL, Schwartz GJ. Back to basics: acidosis and alkalosis. Pediatr Rev004;25:350-7.. Rodriguez SJ. Renal tubular acidosis: the clinical entity. J Am Soc Nephrol002;13:2160-70.. Sherman PM, Mitchell DJ, Cutz E. Neonatal enteropathies: defining the causes ofrotracted diarrhea of infancy. J Pediatr Gastroenterol Nutr 2004;38:16-26.. Ghishan FK. Chronic diarrhea. In: Behrman RE, Kliegman RM, Jenson HB, eds.elson Textbook of Pediatrics. 17th ed. Philadelphia: Elsevier; 2004:1276-81.

. Naim HY, Zimmer KP. Congenital disease of dysfunction and absorption 1.enetically determined disaccharidase deficiency. In: Walker WA, Goulet O, KleinmanE, Sherman PM, Shneider BL, Sanderson IR, eds. Pediatric Gastrointestinal Disease.th ed. Hamilton, Ontario, Canada: BC Decker; 2004:880-97.. Murch SH, Winyard PJ, Koletzko S, Wehner B, Cheema HA, Risdon RA, et al.ongenital enterocyte heparan sulphate deficiency with massive albumin loss, secretoryiarrhoea, and malnutrition. Lancet 1996;347:1299-301.. Mention K, Michaud L, Dobbelaere D, Guimber D, Gottrand F, Turck D.eonatal severe intractable diarrhoea as the presenting manifestation of an unclassified

ongenital disorder of glycosylation (CDG-x). (Arch Dis Child Fetal Neonatal Ed001;85:F217-F219.. Wang J, Cortina G, Wu SV, Tran R, Cho JH, Tsai MJ, et al. Mutant neuroge-in-3 in congenital malabsorptive diarrhea. N Engl J Med 2006;355:270-80.. Cuenod B, Brousse N, Goulet O, De PS, Mougenot JF, Ricour C, et al.lassification of intractable diarrhea in infancy using clinical and immunohistological

riteria. Gastroenterology 1990;99:1037-43.0. Phillips AD, Schmitz J. Familial microvillous atrophy: a clinicopathological surveyf 23 cases. J Pediatr Gastroenterol Nutr 1992;14:380-96.1. Phillips AD, Jenkins P, Raafat F, Walker-Smith JA. Congenital microvilloustrophy: specific diagnostic features. Arch Dis Child 1985;60:135-40.2. Phillips AD. Intractable diarrhoea due to congenital enterocyte defects. Annestle 2006;61:15-23.

3. Rund CR, Carmichael B, Harris AA, Schaffner V. Small bowel biopsy in a-week-old infant. Microvillous inclusion disease of the small intestine. Arch Patholab Med 2006;130:e19-e21.4. Phillips AD, Szafranski M, Man LY, Wall WJ. Periodic acid-Schiff stainingbnormality in microvillous atrophy: photometric and ultrastructural studies. J Pediatrastroenterol Nutr 2000;30:34-42.

5. Ruemmele FM, Jan D, Lacaille F, Cezard JP, Canioni D, Phillips AD, et al. Newerspectives for children with microvillous inclusion disease: early small bowel trans-lantation. Transplantation 2004;77:1024-8.6. Davidson GP, Cutz E, Hamilton JR, Gall DG. Familial enteropathy: a syndromef protracted diarrhea from birth, failure to thrive, and hypoplastic villus atrophy.astroenterology 1978;75:783-90.

7. Pohl JF, Shub MD, Trevelline EE, Ingebo K, Silber G, Rayhorn N, et al. A clusterf microvillous inclusion disease in the Navajo population. J Pediatr 1999;134:103-6.8. Groisman GM, Amar M, Livne E. CD10: a valuable tool for the light microscopiciagnosis of microvillous inclusion disease (familial microvillous atrophy). Am J Surgathol 2002; 26:902-7.9. Weeks DA, Zuppan CW, Malott RL, Mierau GW. Microvillous inclusion diseaseith abundant vermiform, electron-lucent vesicles. Ultrastruct Pathol 2003;27:337-40.0. Reinshagen K, Naim HY, Zimmer KP. Autophagocytosis of the apical membranen microvillus inclusion disease. Gut 2002;51:514-21.
1. Lake BD. Microvillus inclusion disease: specific diagnostic features shown bylkaline phosphatase histochemistry. J Clin Pathol 1988;41:880-2.
201

ocd

Gma

swa(Gio

vwrc

pctw

ppf

a(m

CG

2

CLINICAL AND LABORATORYOBSERVATIONS

Gaucher Disease: Progressive Mesenteric and Mediastinal LymphadenopathyDespite Enzyme Therapy

T. ANDREW BURROW, MD, MITCHELL B. COHEN, MD, RONALD BOKULIC, MD, GAIL DEUTSCH, MD, ARABINDA CHOUDHARY, MD,RICHARD A. FALCONE, JR, MD, AND GREGORY A. GRABOWSKI, MD

A 5-year-old male with Gaucher’s disease type 3 developed progressive mesenteric and mediastinal lymphadenopathyver 12 months, despite enzyme replacement therapy, contributing to the development of a protein-losing enteropathy. Theseomplications are unique, indicating poorly accessible, differentially responsive compartments in patients with Gaucher’sisease who are receiving enzyme therapy. (J Pediatr 2007;150:202-6)

aucher’s disease is an autosomal recessively inherited disorder resulting from mutations in the gene encoding thelysosomal enzyme, acid �-glucosidase (GCase, glucocerebrosidase). The consequent defective activity of GCase resultsin accumulation of glucosylceramide (glucocerebroside) in cells of monocyte/macrophage lineage. These lipid-laden

acrophages, Gaucher cells, reside within the liver, spleen, bone marrow, and lymph nodes, leading to hepatosplenomegaly,nemia, thrombocytopenia, destructive bone disease, and lymphadenopathy, respectively.

Three phenotypes of Gaucher’s disease have been described, based on the absence, or presence and severity, of neurologicaligns. Gaucher’s disease type 1, the non-neuronopathic variant that accounts for �90% of all cases of the disease in the Westernorld, is restricted to the visceral organs. Gaucher’s disease types 2 and 3 are considered neuronopathic variants of the disease

nd are distinguished by the age at onset and rate of progression of neuronopathic signs. Gaucher’s disease type 2 has early onset�3–6 months), severe, rapidly progressing neurological deterioration and visceral disease, followed by death at 1 to 2 years.aucher’s disease type 3 has later onset of neurological disease, including characteristic eye movement abnormalities (saccadic

nitiation defects) and varying degrees of visceral disease. The neuronopathic diseases are thought to result from CNS storagef the deacylated analogue of glucosylceramide, glucosylsphingosine,1 and subsequent neuronal death.

Regular, periodic intravenous infusions of imiglucerase (Cerezyme) have proven safe and effective in ameliorating theisceral manifestations of Gaucher’s disease.2 Excellent hematologic responses and reductions in hepatosplenomegaly occurithin 0.5 to 2 years of such enzyme therapy. However, the diminished or lack of response of lung or CNS involvement,

espectively, to enzyme therapy has highlighted the presence of restricted or inaccessibleompartments to the intravenously administered drug.

A 5-year-old Caucasian male with Gaucher’s disease type 3 is reported withrogressive mesenteric lymphadenopathy as a result of massive sequestration of Gaucherells within lymph nodes despite enzyme replacement therapy. The progressive medias-inal and mesenteric lymphadenopathy and the related severe protein-losing enteropathyith its associated complications are unique and indicate another restricted compartment.

CASE REPORTThis previously reported Caucasian male3 was the 7-lb, 11-oz product of a term

regnancy to a 34-year-old gravida III, para II mother and a 35-year-old father. Theregnancy was complicated by preeclampsia, and a cesarean section was performed forailure to progress. At birth the infant was 19.75 inches in length (25th percentile).

Hepatosplenomegaly was noted at 9 months. Evaluation showed elevated serumminotransferase levels, anemia, and thrombocytopenia. Magnetic resonance imagingMRI) of the abdomen confirmed hepatomegaly (581 mL; �2.5� normal) and spleno-egaly (294 mL; �16� normal). Significantly, no lymphadenopathy was apparent. A

T Computed tomography MRI Magnetic Resonance Imaging

From the Divisions of Human Genetics,Gastroenterology, Hepatology and Nutri-tion, and Pulmonology, Departments ofPathology, Radiology, and Pediatric Surgery,Cincinnati Children’s Hospital Medical Cen-ter, Cincinnati, Ohio; and the Departmentsof Pediatrics, Pathology, Radiology, andSurgery, University of Cincinnati College ofMedicine, Cincinnati, Ohio.

Submitted for publication Feb 23, 2006; lastrevision received Jul 27, 2006; acceptedOct 9, 2006.

Reprint requests: Dr Gregory A. Grabowski,Director, Division of Human Genetics, Cin-cinnati Children’s Hospital Medical Center,3333 Burnet Avenue, MLC 4006, Cincinnati,OH 45229. E-mail: [email protected].



Case Glucocerebrosidase

02

10.1016/j.jpeds.2006.10.062

lor

cDt1Aasst

t[iwm5dnn2((p

mbnita

wwmml

plamapcs

anrsmopa

bcmveiuofi

Fm e grea

G

iver biopsy showed typical Gaucher cells, but these were notbserved on the bone marrow biopsy. Brain MRI and skeletaladiographs were normal.

Gaucher’s disease was documented by severely de-reased acid �-glucosidase activity in peripheral leukocytes.NA sequencing of GBA showed two alleles; one encoding

hree mutations (D409H, L444P, and A456P) in exons 9 and0, and the other allele encoded a K79N mutation in exon 4.3

t that time, the patient was developmentally delayed in grossnd fine motor skills and speech, but without neurologicaligns. However, at 2 years of age, he was noted to haveaccadic initiation failure and hyperreflexia, predominately inhe lower extremities.

Enzyme therapy (60 U/kg/2 weeks) with mannose-erminated recombinant glucocerebrosidase (imigluceraseCerezyme], Genzyme Corporation, Cambridge, Mass) wasnitiated shortly after the diagnosis. Marked improvementas evident at 4.5 years of age by decreased hepatospleno-egaly (2.5� normal to 1.4� normal, and 16� normal to

.7� normal, respectively), and the biomarkers, chitotriosi-ase (from 39,845 to 11,031 nmol/hour/mL; normal 4–76mol/hour/mL), acid phosphatase (from 39.4 to 19.1 U/L;ormal 3.1–7 U/L), and angiotensin converting enzyme (from50 to 160 U/L; normal 13–100 U/L) levels. The hematocritfrom 31% to 39.5%; normal 35%–45%) and platelet countsfrom 78 to 271 K/�L; normal 135–466 K/�L) also im-roved. Imiglucerase-specific antibodies were not detected.

At 1.5 years of age, the patient was evaluated for inter-ittent wheezing and dyspnea that clinically improved after

ronchodilator and inhaled corticosteroid therapy but thatever required oral corticosteroid therapy. Chest computer-

zed tomography (CT) evaluation showed significant medias-inal lymphadenopathy, focal “ground glass” opacities, and

igure 1. Axial CT image of the chest with contrast (5 years of age). A, Tiliary nodularity and alveolar infiltrates. B, Mediastinal nodes (arrow) wer

lveolar infiltrative disease. No calcifications or bronchiectasis l

aucher’s Disease: Progressive Mesenteric and Mediastinal Lymphadeno

ere observed. Diagnostic testing for pulmonary infectionas unrevealing. Bronchoscopic examination revealed a nor-al airway anatomy without compression of the large oredium airways. Bronchoalveolar lavage revealed 5% to 10%

ipid-laden macrophages, consistent with Gaucher cells.Repeat chest CT at 5 years of age (Figure 1) revealed

rogression of lung disease, demonstrated by miliary nodu-arity and diffuse alveolar infiltrates, in addition to mediastinalnd hilar lymphadenopathy. Polysomnography demonstratedild hypoventilation without signs of obstructive or central

pnea, without desaturation. Spirometry and plethsymogra-hy were not obtained because the patient was too young toooperate. The etiology of his pulmonary disease was pre-umed to be a result of Gaucher’s disease and asthma.

At 3 years of age, intermittent bouts of diarrhea andbdominal bloating led to abdominal MRI that showed large,odular, soft tissue masses posterior and anterior to the leftenal vein. By 4 years of age, these masses encased the aorta,uperior mesenteric artery, inferior vena cava, renal vein, andesentery, resulting in a mass effect on the vasculature but no

bstruction to circulation. The mesenteric masses were pal-able on physical examination, but only minor lymphadenop-thy was present elsewhere.

The mesenteric masses further progressed in sizey 5 years of age (Figure 2). Alpha-fetoprotein, �-humanhorionic gonadotropin, homovanillic acid, and vanillicandelic acid were within normal ranges. Laparotomy re-

ealed extensive adenopathy throughout the root of the mes-ntery (Figure 3). The diffuse nature of the adenopathy andts location throughout the mesentery made a surgical solutionntenable. Multiple biopsies of the mesenteric masses dem-nstrated nodular histiocytic lesions surrounded by a densebrous capsule and rims of lymphoid tissue, consistent with

larging pulmonary nodule is demonstrated (arrowhead) as well as diffusetly enlarged.

he en

ymph node replacement (Figure 4A,B). The lesions had

pathy Despite Enzyme Therapy 203

dsatsetst(ow

w

ssbo(lCn

tgtn

Fa

Fa

2

ense aggregates of CD-68 (a macrophage marker) positive,ingle and multinucleated cells with prominent reactive changesnd occasional mitotic activity. The centers of the lesions con-ained proteinaceous eosinophilic material and large histiocytes,ome resembling Gaucher cells (Figure 4C). Intracellular andxtracellular lipid, cholesterol clefts, and dystrophic calcifica-ion were visible throughout the biopsied materials. Ultra-tructural examination revealed cells with membrane-boundubular-appearing structures characteristic of Gaucher cellsFigure 4D). There were no signs of malignancy. No micro-rganisms were visualized. Gram and acid-fast bacillus stainsere negative. The liver biopsy was essentially normal.

Cytomegalovirus and Epstein-Barr virus infections

igure 2. Axial CT images of the abdomen with contrast (5 years of age).orta (a), inferior vena cava (i), and mesentery with calcification (arrow).

igure 3. Small intestine at the time of biopsy. A, Massive mesenteric lymppearance of the enlarged nodes within the mesentery (arrow).

ere excluded by polymerase chain reaction or serological m

04 Burrow et al

tudies. Actinomyces israelii was cultured from only one ofeveral biopsied mesenteric lymph nodes. Stool viral andacterial cultures were negative, but antigen studies on twoccasions were positive for Clostridium difficile. Long-term8 weeks) intravenous penicillin was followed by oral penicil-in (1 year) and metronidazole therapy to treat A israelii and

difficile infections, respectively. However, the lymphade-opathy and abdominal distention were unchanged.

Hypoalbuminemia was noted at 5 years of age and sincehat time, serum albumin levels have ranged from 1.7 to 2.6/dL (normal 3.4–5.2 g/dL). Normal hepatic synthetic func-ion, renal evaluation demonstrating absent proteinuria, andormal serum immunoglobin levels with the exception of a

, Images demonstrate an extensively enlarged mass (outlined) encasing the

de enlargement (arrow) is shown. B, The pale yellow to white typical

A, B

ph no

ildly diminished IgG level, combined with large amounts of


sNcstpo

zctocalbp

dnwptw

Gnmetw

FamUO

G

tool �-1-antitrypsin, indicated a protein-losing enteropathy.umerous serious complications have resulted, including as-

ites and peripheral edema. More recently, a severe, progres-ive, and difficult-to-control thrombophilia, with significanthrombus formation in the right internal jugular, brachioce-halic, and subclavian veins developed, likely a complicationf the protein-losing enteropathy.4

At 5.5 years of age, the frequency of recombinant en-yme infusions was increased to 60 U/kg/week. Despite thishange and symptomatic management, the patient continueso exhibit marked abdominal distention without a reductionf the intra-abdominal lymphadenopathy. In addition, heontinues to exhibit progression of mediastinal lymphadenop-thy and pulmonary disease. An enteral, medium chain trig-yceride–containing diet has been instituted, which has sta-ilized the hypoalbuminemia, but the other sequelae of the

igure 4. Histology of biopsied mesenteric masses/lymph nodes. A, A dennd central portions (CN) are magnified in (B) and (C), respectively. B, Reany of which resembled Gaucher cells (C). C, These Gaucher cells contailtrastructural examination revealed degenerating cells containing lipid dropriginal magnification �40 (A); �200 (B); �400 (C); �12,000 (D).

rotein-losing enteropathy continue to progress. o

aucher’s Disease: Progressive Mesenteric and Mediastinal Lymphadeno

DISCUSSIONSince the inception of enzyme therapy for Gaucher’s

isease, selected tissues including lung, CNS, and lymphodes have had diminished/absent responses comparedith other tissues.5,6 This indicates that certain organs areoorly accessible to intravenously administered enzymeherapy products; however, the mechanistic bases are notell understood.

The present case is the second report of a child withaucher’s disease exhibiting massive mesenteric lymphade-

opathy. In comparison, Lim and colleagues reported a 13-onth-old female with Gaucher’s disease type 3 who, despite

nzyme replacement therapy, developed significant mesen-eric lymphadenopathy that had been stable in size for 3 yearshen reported.7 The present case is much more severe, dem-

rous capsule and rim of lymphoid tissue (arrowheads). The peripheral (P)ent of major parts of the lymph node with engorged large storage cells,

ne to coarsely fibrillar Periodic Acid Schiff positive material (arrows). D,rrowheads) and abundant fibrillar inclusions typical for Gaucher’s disease.

se fibplacemned filets (a

nstrating progression of the mesenteric masses over 2 years,

pathy Despite Enzyme Therapy 205

cernc

laiaollshtr

asapitecm

dsusmsppr

1VY2Ed23vt4p5od6fi7f8

2

oincident with similar progression of mediastinal lymphad-nopathy, protein-losing enteropathy, thrombophilia, and pa-enchymal and interstitial lung disease, despite relatively mildeurological manifestations. No other cause, other than Gau-her’s disease, has been discovered to explain these events.

The clinical course in this patient further illustrates thatymph nodes, including mesenteric and mediastinal nodes,nd the lungs are relatively sequestered or poorly accessible tontravenous enzyme therapy. The protein-losing enteropathynd related consequences appear to be a result of lymphaticbstruction, resulting in inadequate drainage of the intestinalymphatic tissue and increased hydrostatic pressure in theacteals, particularly in the face of normal liver biopsy and liverynthetic function studies. The C difficile or A israelii mightave contributed to the development of the lymphadenopa-hy, but it preceded these infections and did not decrease oresolve with adequate therapy for either potential infection.

Increased intensity of enzyme therapy (60 U/kg/week)ppears ineffective for the current patient. Alternatives ofubstrate reduction therapy and bone marrow transplantationre being considered, but each has significant potential com-lications for this patient. The diarrhea induced by miglustatn most Gaucher’s patients (�90%)8 would be unwelcome inhis patient. Stem-cell transplant would be complicated by thexisting lung involvement. Moreover, the apparent poor vas-ularization of the enlarged nodes may inhibit access by donoracrophages following transplant.

With the advent of enzyme therapy, the Gaucher’sts

06 Burrow et al

isease phenotypes have been transformed into an evolvingpectrum of continuing disease involvement. Consequently,nanticipated and unexpected results of partial treatment ofpecific, poorly accessible organs complicate disease manage-ent. Such unusual cases highlight the need for comprehen-

ive therapies in affected patients, as well as the presence ofoorly accessible/differentially responsive compartments inatients with Gaucher’s disease and other lysosomal diseaseseceiving enzyme therapy or other therapeutic modalities.

REFERENCES. Beutler E, Grabowski GA. Gaucher Disease. In: Scriver CR, Beaudet AL, Sly WS,alle D, eds. The Metabolic and Molecular Bases of Inherited Diseases. 8th ed. Nework: McGraw-Hill; 2001:3635-68.. Weinreb NJ, Charrow J, Andersson HC, Kaplan P, Kolodny EH, Mistry P, et al.ffectiveness of enzyme replacement therapy in 1028 patients with type 1 Gaucherisease after 2 to 5 years of treatment: a report from the Gaucher Registry. Am J Med002;113:112-9.. Zhao H, Bailey LA, Elsas LJ II, Grinzaid KA, Grabowski GA. Gaucher disease: inivo evidence for allele dose leading to neuronopathic and nonneuronopathic pheno-ypes. Am J Med Genet A 2003;116:52-6.. Muntean W, Rossipal E. Loss of inhibitors of the blood coagulation system inrotein-losing enteropathy [author’s transl]. Klin Padiatr 1979;191:20-3.. Altarescu G, Hill S, Wiggs E, Jeffries N, Kreps C, Parker CC, et al. The efficacyf enzyme replacement therapy in patients with chronic neuronopathic Gaucher’sisease. J Pediatr 2001;138:539-47.. Grabowski GA, Leslie N, Wenstrup R. Enzyme therapy for Gaucher disease: therst 5 years. Blood Rev 1998;12:115-33.. Lim AK, Vellodi A, McHugh K. Mesenteric mass in a young girl—an unusual siteor Gaucher’s disease. Pediatr Radiol 2002;32:674-6.. Cox T, Lachmann R, Hollak C, Aerts J, van Weely S, Hrebicek M, et al. Novel oral
reatment of Gaucher’s disease with N-butyldeoxynojirimycin (OGT 918) to decreaseubstrate biosynthesis. Lancet 2000;355:1481-5.

ct

TpT

C

iwhteai

C

orU(td(

C

sawimd

Rt

Recipient Twin Limb Ischemia with Postnatal Onset

ROLAND SPENCER BROADBENT, MB, CHB

After the occurrence of 3 local cases of limb ischemia in newborn twins, we reviewed the literature to investigate thisombination systematically. This review reveals a distinct condition: postnatal onset limb ischemia affecting recipient twins inwin-twin transfusion syndrome. (J Pediatr 2007;150:207-9)

win-twin transfusion syndrome (TTTS) occurs during fetal life when vascular communications between monochorionictwins feed more blood to one twin, the recipient, than to the other, the donor. Perinatal limb ischemia in twins has notbeen recognized as a distinct entity, apart from a small series of antenatal-onset cases.1 The 3 cases reported here, all

ostnatal cases in recipient twins, prompted a review of the literature where postnatal-onset limb ischemia occurred in a twin.he hypothesis was that like the index cases, limb ischemia in the published cases would occur in TTTS recipients.

PATIENTS AND METHODS

ase 1This female recipient twin was born at 25 weeks gestation, after severe polyhydramnios in this twin and oligohydramnios

n the other fetus. She weighed 700 g, and her monochorionic co-twin weighed 510 g. A low umbilical artery catheter (UAC)as inserted and infused with heparinized saline 1 U/mL. Intravenous indomethacin 0.2 mg/kg was given at 7 hours. At 24ours, the right leg became discolored. The catheter was removed, after which the leg reperfused normally. Another catheter washen inserted into the same umbilical artery, and again the leg became ischemic. Leg pulses were absent to palpation and Dopplerxamination. The catheter was withdrawn until the tip was in the umbilical artery within the anterior abdominal wall. Anrteriogram (Figure) demonstrated complete occlusion of the right external iliac artery. Removing the catheter produced nomprovement. Arterial flow reappeared on day 3, and the limb recovered fully.

ase 2Signs of TTTS developed at 19 weeks in this monochorionic pregnancy. Severe polyhydramnios in the larger twin and

ligohydramnios (stuck twin) in the other twin persisted until the membranes ruptured after amnioreduction at 24 weeks. Theecipient twin weighed 610 g, 200 g more than her cotwin. A low UAC was inserted and infused with heparinized saline 1/mL. Mild hypotension was treated with low-dose intravenous dopamine. Three doses of intravenous indomethacin,

0.1 mg/kg, 0.2 mg/kg, and 0.2 mg/kg) were given in the first 30 hours. At 36 hours of age, the left leg became ischemic, andhe arterial catheter was removed. Doppler ultrasound imaging revealed absent arterial flow in the left external iliac artery andistally. Despite intravenous thrombolytic therapy with tissue plasminogen activatort-PA), extensive gangrene developed, and an above-knee amputation was required.

ase 3This infant was the larger of monochorionic twins. Antenatal scans at 24 weeks

howed severe polyhydramnios in this and stuck twin in the other. Labor began aftermnioreduction at 25 weeks, and delivery was by caesarian section. The recipient twineighed 680 g. A low UAC was inserted and infused with heparinized saline 1 U/mL. No

ndomethacin was given. On the second day, the right leg became white from theid-thigh to the toes, resolving after 5 minutes. Several hours later, the right leg became

usky; the UAC was withdrawn, whereupon the leg returned to its normal color.

TLI Recipient twin limb ischemia TTTS Twin-twin transfusion syndrome

From the Department of Women’s andChildren’s Health, Faculty of Medicine, Uni-versity of Otago, Dunedin, New Zealand.

Submitted for publication Jan 21, 2006; lastrevision received Sep 16, 2006; acceptedNov 6, 2006.

Reprint requests: Roland S. Broadbent,Department of Women’s and Children’sHealth, Faculty of Medicine, University ofOtago, Great King Street, Dunedin 9001,New Zealand. E-mail: [email protected].



-PA Tissue plasminogen activator UAC Umbilical artery catheter
10.1016/j.jpeds.2006.11.033
207

M

ocP“

“wfdtoiwhcw

lirchlawcp

TAa

Ftr(Toab

T

PCt ave bee*

2

ethodsWe searched the literature for twins with neonatal-

nset limb ischemia to ascertain the characteristics of thisombination. Peer-reviewed articles were found by searchingubmed and Google Scholar, using “twin” along with “fetus,”newborn,” “limb,” “arm,” “leg,” “necrosis,” “ischemia,” and

igure. Arteriogram from case 1, performed through a 3.5 F catheter inhe right umbilical artery. Filling defects indicate mural thrombus in theight hypogastric artery (vertical arrows) and at the aortic bifurcationhorizontal arrow). The oblique arrow shows the left external iliac artery.he oval area shows the right internal iliac artery filled by retrograde flowf contrast. There is complete occlusion of the right external iliac arterynd no supply to the right leg. Note the lack of contrast blush in the

able. Limb Ischemia in Twins

Case SourceGestation(weeks) Sex

Birthweight

Twin Cot

1. Case 1 25 F/F 700 52. Case 2 24 F/F 610 43. Case 3 25 F/F 680 64. Arad3 33 M/M 1910 145. Arad3 30 M/M 1015 76. Breschan4 24 M/M 640 57. Cincotta5 268. Kataria6 29 M/M 10809. Kennedy7 28 820

10. Letts2 33 F/F 1680 1111. Letts2 28 F 940 1112. Letts2 25 M/M 650 313. Rayner8 34 M 210014. Stringel9 31 F 168015. Zenker10 31 F/F 1520 10

AL, peripheral arterial line; Hb, hemoglobin; Hct, hematocrit.ases 1, 2, and 3 are the cases reported here. Details of the other cases have been extrac

o onset of the ischemia. “Indomethacin” indicates that indomethacin was known to hPartial exchange transfusion in the first 24 hours to reduce the Hct to this level.

pones and soft tissues on the right.

08 Broadbent

gangrene” in various combinations. The most recent searchas done in July 2006. Authors were contacted to obtain

urther information. Cases were accepted if there was evi-ence of limb ischemia with onset in the newborn period in awin. TTTS was accepted where this was stated in the article,r where there was polyhydramnios around the larger (recip-ent) and oligohydramnios around the smaller (donor) twin,ith nothing to indicate dichorionicity. The recipient has theigher hemoglobin level of the 2. Severe polycythemia wasonsidered to be present if so reported or if hematocritas � 70 or hemoglobin was � 220 g/L.

RESULTSThe literature search revealed 13 cases of neonatal onset

imb ischemia in 1 of twins. The sex of both twins was knownn 8 cases, all like-sexed except for the twin with meningitiseported by Letts et al.2 Sepsis appears to be the cause of thatase of ischemia; in the interest of clarity, that case is excludedere. The Table reports 15 cases, the remaining 12 from the

iterature and the 3 reported here. TTTS occurred in 9 cases,ll recipient twins. In the other 6 cases there was nothing thatould exclude either TTTS or recipient status. Severe poly-

ythemia occurred in 5 cases. At the onset of ischemia, severeolycythemia was present in 2 cases and absent in 8 cases.

DISCUSSIONCarr et al1 reported 6 prenatal cases of limb ischemia in

TTS and noted that some, but not all, had polycythemia.ll were recipient twins, although the authors did not draw

ttention to this fact. The present report demonstrates that

Indomethacin TTTS/affected twin Co-twin

Postnatal Yes/recipient LivebornPostnatal Yes/recipient Died at birthNil Yes/recipient LivebornAntenatal Yes/recipient LivebornAntenatal Yes/recipient Liveborn

LivebornYes/recipient Liveborn

StillbornLiveborn

Yes/recipient Died at birthLiveborn

Yes/recipient Stillborn

Yes/recipient Liveborn

m the literature with additional data supplied by the authors. Hb or Hct at onset refersn administered before the onset of ischemia.

(g)

win

101035100035

002070

20

ted fro

ostnatal-onset limb ischemia in twins also occurs in TTTS.


Bh

RtlwwwceJw

1smc

ioi

ii

1t2O3m4fb5t6S7cJ8c9a1

T

R

oth in utero and ex utero, it is the recipient that is affected;ence the term recipient twin limb ischemia (RTLI).

Cases 8, 11, 13, and 15 show that postnatal-onsetTLI does occur without an arterial catheter in place, al-

hough an arterial catheter was present in 10 cases. Where theocation of the catheter was known (in cases 1, 2, 3, and 6), itas same side as the ischemia. In case 1, the ischemia resolvedhen the initial catheter was removed but recurred when itas replaced. In case 3, the ischemia resolved as soon as the

atheter was withdrawn. These findings support the hypoth-sis that UACs played a part in these ischemic episodes.udging from case 6, the period of risk extends into the secondeek.

Intra-arterial thrombus was reported in cases 8, 9, 12,3, and 15. The filling defects in the arteriogram (Figure) alsouggest intra-arterial thrombus in case 1. Thrombolytic treat-ent with t-PA was unsuccessful in case 2 but successful in

ases 9 and 15.Severe polycythemia is a risk factor for neonatal limb

schemia,2 but it is not the primary etiology in RTLI; itccurred at onset in only 2 of 10 cases. Indomethacin admin-

able. Continued

Hb or Hct at birth Hb or Hct atonsetAffected twin Cotwin

52.0% 46.0% 54.2%153 g/L 125 g/L155 g/L 143 g/L 134 g/L71% 24%

174 g/L 154 g/L 126 g/LSevere polycythemia

55%40.6%180 g/L 134 g/L260 g/L Very pale 228 g/L150 g/L 121 g/L221 g/L 221 g/L

225 g/L 55%*

stration preceded onset of the ischemia in 4 cases, but this ist1

ecipient Twin Limb Ischemia with Postnatal Onset

nsufficient information to indicate whether indomethacinncreases the risk of limb ischemia.

REFERENCES. Carr SR, Luks F, Tracy T, Plevyak M. Antenatal necrotic injury in severewin-to-twin transfusion syndrome: a case and review. Fetal Diagn Ther 2004;19:370-2.. Letts M, Blastorah B, al-Azzam S. Neonatal gangrene of the extremities. J Pediatrrthop 1997;17:397-401.

. Arad I, Bar-Oz B, Amit Y, Ergaz Z, Peleg O. Neonatal limb ischemia followingaternal indomethacin treatment in twin pregnancies. J Perinat Med 1995;23:487-91.

. Breschan C, Kraschl R, Jost R, Marhofer P, Likar R. Axillary brachial plexus blockor treatment of severe forearm ischemia after arterial cannulation in an extremely lowirth-weight infant. Paediatr Anaesth 2004;14:681-4.. Cincotta R, Oldham J, Sampson A. Antepartum and postpartum complications ofwin-twin transfusion. Aust N Z J Obstet Gynaecol 1996;36:303-8.. Kataria R, Gupta DK, Mitra DK. Colonic mucormycosis in the neonate. Pediatrurg Int 1995;10:271-73.. Kennedy LA, Drummond WH, Knight ME, Millsaps MM, Williams JL. Suc-essful treatment of neonatal aortic thrombosis with tissue plasminogen activator.Pediatr 1990;116:798-801.. Rayner CR, Lloyd DJ, Ward K. Neonatal upper limb ischaemia: the case foronservative management. J Hand Surg [Br] 1984;9:57-8.. Stringel G, Mercer S, Richler M, McMurray B. Catheterization of the umbilicalrtery in neonates: surgical implications. Can J Surg 1985;28:143-6.0. Zenker M, Ries M, Vetter V, Rauch R, Harms D. Non–catheter-related aortic

C Age of onset Limb Outcome

s 24 hours R lower limb No necrosiss 36 hours L lower limb Necrosiss 48 hours R lower limb No necrosiss 1 day L lower limb Necrosiss 2 hours L lower limb No necrosisL 12 days R upper limb Necrosis

1 day L lower limb Necrosis9 days L lower limb Necrosis

s 3 days Lower limbs Necrosiss 48 hours R lower limb Necrosis

24 hours R lower limb Necrosiss 4 days Lower limbs Necrosis

Few minutes L upper limb Necrosiss 3 days R lower limb Necrosis

36 hours Lower limbs No necrosis

UA

YeYeYeYeYePA

NoYeYeNoYeNoYeNo

hrombosis and resistance to activated protein C in a premature newborn. Acta Paediatr999;88:1035-8.

209

o(

CG

vwcw

rlm

pp

P

berwn

fftr

p

AHI

2

Congenital Patent Ductus Venosus: An Association with the HyperIgE Syndrome

KEREN SAGIV-FRIEDGUT, MD, MICHAELA WITZLING, MD, ILAN DALAL, MD, CHANA VINKLER, MD, ELI SOMEH, MD, AND

ARIE LEVINE, MD

We describe an association between congenital patent ductus venosus and hyper immunoglobulin E syndrome in a pairf siblings. The possibility that this is a separate entity or a genetically linked association is discussed.J Pediatr 2007;150:210-2)

ongenital portosystemic shunts (PSS) are rare malformations involving the vasculature leading into and out of the liver.The clinical consequences of a shunt from the portal vein to the systemic veins are portal hypertension and hepaticencephalopathy. Congenital shunts have been occasionally associated with other malformations that include heterotaxia,

oldenhar’s syndrome, biliary atresia, mental retardation, and genitourinary malformations.1

Congenital patent ductus venosus (PDV) is a rare disorder involving a shunt from the fetal umbilical vein to the inferiorena cava. Spontaneous closure of the ductus venosus is the rule in normal infancy, usually immediately or within the first feweeks after birth.2 There is no evidence linking hypercoagulable states and the development of PDV at present. Familial

ongenital PDV has been reported only 3 times to date, in 2 case reports involving 2 sets of 3 siblings each and in 1 set of twinsithout underlying diseases or syndromes.3-5

Hyper immunoglobulin E (IgE) syndrome (HIES) is a rare primary immunodeficiency syndrome characterized byecurrent staphylococcal infections, pneumonia, eczema, prominent jaw with coarse facies, and markedly elevated serum IgEevels, usually �2000 IU/L. Most cases are sporadic; however, families with autosomal dominant (AD) and autosomal recessive

odes of inheritance have been described.6

An association between HIES and vascular liver anomalies has not been described. We present 2 siblings with identicalortosystemic hepatic shunt–PDV and HIES and discuss whether these 2 entities are separate or share a common geneticathway.

CASE REPORT

atient 1A 5-year-old male previously diagnosed with HIES was admitted to our hospital

ecause of a staphylococcal hepatic abscess and peritonitis. The diagnosis of HIES wasstablished on the basis of clinical features, including recurrent skin abscesses, eczemas,ecurrent oral thrush, prominent teeth, and upper jaw with coarse facies. Immunologicorkup revealed IgE levels usually above 4000 units/mL (normal � 230). Other immu-oglobulin levels were within the normal range.

Absolute eosinophil counts were in excess of 2000/mL (normal � 800). Neutrophilunction testing did not detect chemotactic or bacteriocidic abnormalities; superoxideormation was normal. There was no response to a Candida sp. skin test, and the responseo the diphteria tetanus (DT) skin test was normal. Lymphocytic subpopulation andesponse to mitogen was normal.

Past history revealed 4 normal siblings and a younger sister described below asatient 2. The patient’s parents are of non-consanguineous Arab origin. His father has a

D Autosomal dominantIES Hyper immunoglobulin E syndrome

PDV Patent ductus venosusPSS Portosystemic shunt.

From the Pediatric Gastroenterology Unit,Department of Radiology, the Allergy andImmunology Unit and The Genetics Insti-tute, Wolfson Medical Center, Holon, andTel Aviv University, Sackler School of Med-icine, Tel Aviv, Israel.

Submitted for publication May 24, 2006;last revision received Jul 25, 2006; acceptedOct 16, 2006.

Reprint requests: Arie Levine, PediatricGastroenterology Unit, or Ilan Dalal, Pedi-atric Infectious/Allergy/Immunology Unit,the E Wolfson Medical Center, Depart-ment of Pediatrics, Halokhamim Street 62,PO Box 5, 58100 Holon, Israel. E-mail:[email protected], [email protected].



gE Immunoglobulin E

10

10.1016/j.jpeds.2006.10.052

hluv

tlcrtf

tfp4voswelpdPp

P

Ha2pPcsac

utuwadsNlpond

a

ealtm

oc(imwa

uGaocc

dsehds

oRsdztc

Fpi

C

ypereosinophilic syndrome with recurrent episodes of cellu-itis and cutaneous fungal infection. The father’s abdominalltrasound scan demonstrated a normal liver with normal liverasculature.

Physical examination revealed microcephaly, prominenteeth and upper jaw as mentioned, and prominent ears. Theiver was not palpable; the spleen was palpated 2 cm below theostal margin. A neurologic and developmental assessmentevealed that the boy had mild mental retardation with at-ention deficit disorder with hyperactivity. No neurologicocal findings or asterixis were demonstrated.

Relevant laboratory test results include alanine aminoransferase of 39 units/L (norm � 41), aspartate aminotrans-erase of 52 units/L (norm � 38), mildly increased alkalinehosphatase of 744 units/L (norm 145-320), and albumin of.1 g/dL (norm 3.8-5.4). The ammonia was constantly ele-ated above 200 units/L (normal range 9-33 U/L). The resultf echocardiography was interpreted as normal. A liver biopsypecimen from an area of nodularity disclosed congestionithout evidence of inflammation and without fibrosis. The

lectroencephalography result was normal. During the post-aparotomy assessment, contrast-enhanced abdominal com-uterized tomography and ultrasound Doppler examinationemonstrated a shunt from the left portal vein through aDV to the infradiaphragmatic inferior vena cava. The rightortal vein was hypoplastic (Figure).

atient 2Briefly, patient 2 is the 9-year-old sister of patient 1.

er history was unremarkable except for a periodontal abscessnd Thalassemia minor. Because of an enlarged liver palpated

cm below costal margins and palpable spleen (edge) onhysical examination and a brother known to have congenitalDV and HIES, she was evaluated for the latter 2. Her facialharacteristics show prominent teeth and upper jaw veryimilar to her brother’s. Neurologic and developmentalssessment was normal. She, like her sibling, had never hadyanosis.

Her blood test revealed elevated ammonia of 130nits/L, normal alanine amino transferase, aspartate amino-ransferase, and alkaline phosphatase. Her total IgE is 1896nits/L (normal � 230), and the absolute eosinophil countas 390/mL. Liver ultrasonography with Doppler scanning

nd contrast enhanced abdominal computerized tomographyemonstrated similar findings to those of her brother—a largehunt through a PDV with a hypoplastic right portal vein.either patient had evidence of a hypercoagulable state by

aboratory evaluation. Cytogenic analysis was performed oneripheral blood lymphocytes of patient 1 and patient 2 by usef G-bands by trypsin using Giemsa (GTG) banding tech-iques. No major deletion or translocation of the 4q regionid was detected.

DISCUSSIONThe rarity of coexisting PDV and HIES in 2 siblings

nd the fact that their father also has elevated IgE with i

ongenital Patent Ductus Venosus: An Association with the Hyper IgE S

osinophilia suggests that this is not a coincidence. Thisssociation raises several possibilities: a single genetic defecteading to both syndromes, 2 different syndromes occurring inhe same individuals, or that the presence of 1 syndromeight lead to the other.

PDV is considered a rare sporadic phenomenon. Thenly prior evidence for a genetic cause as a possible causeomes from 2 previous reports, each containing 3 siblingsdescribed in the introduction). A second clue is that congen-tal PSS is not rare in dogs, suggesting that a genetic mutation

ay cause PSS.7 The fact that 2 siblings share the same shuntithout evidence of a hypercoagulable state lends credence togenetic hypothesis for PDV.

HIES on the other hand, has been clearly linked to annderlying genetically inherited abnormality in some patients.rimbacher et al8 described 19 kindreds with 57 individuals

ffected with HIES in an AD mode of inheritance. In somef the patients they found linkage to 4q21-4q21.1 locus. Aandidate gene is not known. The 2 siblings and their fatherould fit a hyper IgE–associated AD trait.

One possible explanation of the association we haveescribed is that this is a previously unpublished geneticyndrome. Another possible connection between these 2 rarentities could be 2 rare genetic traits in the same family. Inighly consanguineous populations, such as exist in the Mid-le East, it is not uncommon to find 2 genetic traits in theame family.

Another possible genetic cause may involve a mutationr deletion to a common locus involving more than genes.egulation of fetal vasculature and closure of intrauterine

hunts is often mediated by prostaglandins. The prostaglan-in D2 synthase (the hematopoietic type among the 3 isoen-ymes) is located downstream from the hyper IgE locus onhe 4q21-q22 locus (omim 602598). Prostaglandin D2 is aofactor in the development regulation of the fetal circulation

igure. Abdominal CT. Coronal reconstruction. Minimal intensifyingrojection (MIP). Case 1. White arrow indicates portal vein; black arrowndicates PDV.

n animal models.9-12

yndrome 211

tfhrllvna

taPs

1C2D3ps

4d5d2627IM8G19t1RA1n51P9e1ce

2

We did not find a translocation or a major deletion inhis location by chromosomal analysis of the 2 siblings and theather. Although a mutation scenario is appealing, we do notave any evidence that mutation in the PGD synthase iselated to PDV. Lastly, presence of an underlying immuno-ogic deficiency may predispose individuals toward dysregu-ation of the fetal vasculature. Although mast cells may secreteascular endothelial growth factor in response to IgE, there iso prior evidence to date to suggest that situations withbnormal IgE affect the fetal circulation.13

In conclusion, we have described an association be-ween PDV and HIES in 2 siblings. This association may benovel syndrome. The finding of a fourth set of siblings withDV strongly suggests that genetic mutations may be respon-ible for some cases of this entity.

REFERENCES. Carmen G, María M, Carlos M, Purificación M, Gabino G, Enrique GH.ongenital Hepatic Shunt. Radiographics 2004;24:755-72.

. Meyer VW, Lind J. The ductus venosus and the mechanism of its closure. Archis Child 1966;41:597-605.

. Uchino T, Endo F, Ikeda S, Shiraki K, Sera Y, Matsuda I. Three brothers withrogressive hepatic dysfunction and severe hepatic steatosis due to patent ductus veno-
us. Gastroenterology 1996;111:1964-8. r
12 Sagiv et al

. Jacob S, Farr G, De Vun D, Takiff H. Hepatic manifestations of familial patentuctus venosus in adults. Gut 1999;45:442-5.. Gillespie MJ, Golden A, Sivarajan VB, Rome JJ. Transcatheter closure of patentuctus venosus with the Amplatzer vascular plug in twin brothers. Pediatr Cardiol006;27:142-5.. Grimbacher B, Holland SM, Puck JM. Hyper-IgE syndromes. Immunol Rev005;203:244-50.. van Straten G, Leegwater PA, de Vries M, van den Broom WE, Rothuizen J.nherited congenital extrahepatic portosystemic shunts in Cairn terriers. J Vet Intern

ed 2005;19:321-4.. Grimbacher B, Schaffer AA, Holland SM, Davis J, Gallin JI, Malech HL, et al.enetic linkage of hyper-IgE syndrome to chromosome 4. Am J Hum Genet

999;65:735-44.. Perreault T, Coe JY, Olley PM, Coceani F. Pulmonary vascular effects of pros-aglandin D2 in newborn pig. Am J Physiol 1990;258(Pt 2):H1292-9.0. Dumont I, Hardy P, Peri KG, Hou X, Molotchnikoff S, Varma DR, et al.egulation of endothelial nitric oxide synthase by PGD(2) in the developing choroid.m J Physiol Heart Circ Physiol 2000;278:H60-66.

1. Warren JB, Loi RK, Wilson AJ. PGD2 is an intermediate in agonist-stimulateditric oxide release in rabbit skin microcirculation. Am J Physiol 1994;266(Pt 2):H1846-3.2. Spatz M, Stanimirovic D, Uematsu S, Roberts LJ 2nd, Bembry J, McCarron RM.rostaglandin D2 and endothelin-1 induce the production of prostaglandin F2 alpha,alpha, 11 beta- Prostaglandin F2, Prostaglandin E2, and thromboxan in capillaryndothelium of human brain. Prostaglandin Leukot Essent Fatty Acids 1993;49:789-93.3. Boesiger J, Tsai M, Maurer M, Yamaguchi M, Brown LF, Claffey KP, et al. Mastells secrete vascular permeability factor/vascular endothelial cell growth factor andxhibit enhanced release after immunoglobulin E dependent up-regulation of fc epsilon
eceptor I expression. J Exp Med 1998;188:1135-45.
Fifty Years Ago in The Journal of PediatricsCLASSIFICATION AND ETIOLOGICAL FACTORS IN MENTAL RETARDATION

Yanet H. J Pediatr 1957;50:226-30

In 1956, Yanet classified the underlying causes of mental retardation for more than 2500 admissions to the SouthburyTraining School for Mentally Retarded in Connecticut. Only 40% of these children were mildly retarded, as compared withthe 75% prevalence of mild mental retardation in the larger community; this reflected the selection bias of which children wereinstitutionalized. Ninety-one percent of the causes were categorized as prenatal, 3% perinatal, and 6% postnatal. The prenatalcauses were divided into those that were definitely genetic (only 5% had recognized single-gene disorders) and those that were“not definitely genetically determined.” Oddly, “mongolism” (Down syndrome) was classified in the latter group; the influenceof maternal age was thought to reflect some yet unidentified environmental influence, and Jerome Lejeune would not reportthe chromosomal duplication for trisomy 21 until 1958. A long-standing miscount in the total number of human chromosomeshad contributed to the delay in making this identification. Down syndrome nevertheless accounted for 80% of the 10% of theprenatal group that included prenatal infections, kernicterus, and cretinism. Of the other 81% with prenatal causes, almost half(40%) represented “familial” or “subcultural” retardation; there were both environmental deprivation as well as “physiological”contributions in the form of the derivative genetic pool for this group. A large number of children were placed in the geneticgroup with minimal specific findings (eg, isolated microcephaly), so that many of those would today probably be classified asidiopathic. The 3% incidence for perinatal cause seems low when considering the history of malpractice lawsuits over thefollowing half century. Although fetal alcohol syndrome, fragile X syndrome, and many other specific causes were still in thefuture, the shape of the distribution of causes has nevertheless remained remarkably constant.

Pasquale J. Accardo, MDDepartment of Pediatrics

Virginia Commonwealth University-Medical College of Virginia

Richmond, VA10.1016/j.jpeds.2006.09.008


LETTERS

Erythromycin for treatment of feeding intolerancein preterm infants

To the Editor:Nuntnarumit et al1 reported that intermediate doses of

oral erythromycin (EM) were effective and safe for the treat-ment of feeding intolerance in preterm infants. Their protocolused an intermediate oral dose instead of intravenous low-dose EM. Inclusion criteria were relatively loose. Infants wereenrolled for treatment at 6 to 8 postnatal days, and eveninfants in the control group reached full feeding at a medianage of 13 days (range 9-15). Therefore the infants had onlymild feeding intolerance. In such a population of prematureinfants, a degree of feeding intolerance occurs within the firstweek of life and can possibly improve on stimulation with asmall volume of breast milk, a so-called gut stimulation pro-tocol.2 We doubt if these infants required EM treatment or iftheir protocol would be as effective in premature infants withsevere feeding intolerance.

We have accumulated experience with EM treatmentfor feeding intolerance in premature infants since we began touse it in 1998.3 We now use an intravenous loading dose ofEM (30 mg/kg/d) for 2 days, followed by an oral maintenancedose (6 to 10 mg/kg/d) when enteral feeding can be started.EM is continued until full feeding (120 ml/kg/d) is wellestablished. The mean time to full feeding is 11 days. We useEM only in premature infants with feeding intolerance lastingbeyond 2 weeks of age.

One complication of EM treatment is hypertrophicpyloric stenosis (HPS). The authors described that, unlikeusing the antimicrobial dose of EM, no infant had HPS.Several points merit discussion. First, the study that reportedinfants with HPS induced by the antimicrobial dose of EMwere near- or full-term infants.4 Not understanding how thepyloric muscle of the premature infant responds to EM, onecan not deduce that an intermediate-dose EM will not induceHPS in premature infants. Second, there is a strong linkbetween systemic EM use in infants and subsequent HPS,with the highest risk in the first 2 weeks of age. We should bevery cautious in using EM treatment during this time frame.Third, premature infants with HPS do not exhibit typicalsymptoms, and it is well known that onset of vomiting pre-sents after the corrected age for a “term baby” instead of thethird postnatal week, as do full-term infants.5,6 Fourth, py-loric length and muscle thickness evaluated by ultrasonogra-phy in premature infants with HPS are less than the estab-lished pyloric muscle index for diagnosis of hypertrophy.5,7-9

The article failed to define diagnostic criteria of HPS forpremature infants. Fifth, some infants with HPS may not beidentified. Based on rates of 2.65% in infants treated with EMup to 2 weeks of age and 0.25% if untreated, 42 infants wouldneed EM treatment to cause a HPS case.4

We suggest a further randomized controlled trial to clarifyefficacy and safety of oral EM in severe feeding intolerance.

Bai-Horng Su, MDHung-Chih Lin, MD

Department of PediatricsChina Medical University Hospital

Taichung, Taiwan10.1016/j.jpeds.2006.10.064

REFERENCES1. Nuntnarumit P, Kiatchoosakun P, Tantiprapa W, Boonkasidecha S. Efficacy of oralerythromycin for treatment of feeding intolerance in preterm infants. J Pediatr2006;148:600-5.2. Kliegman RM. The relationship of neonatal feeding practices and the pathogenesisand prevention of necrotizing enterocolitis. Pediatrics 2003;111:671-2.3. Su BH, Lin HC, Peng CT, Tsai CH. Effect of erythromycin on feeding intolerancein very low birth weight infants: a preliminary observation. Acta Paediatrica Sinica1998;39:324-6.4. Mahon BE, Rosenman MB, Kleiman MB. Maternal and infant use of erythromycinand other macrolide antibiotics as risk factors for infantile hypertrophic pyloric stenosis.J Pediatr 2001;139:380-4.5. Janik JS, Wayne ER, Janik JP. Pyloric stenosis in premature infants. Arch PediatrAdolesc Med 1996;150:223-4.6. Kumar RK. Pyloric stenosis in premature infants. Aust Fam Physician 1998;27:989.7. Cosman BC, Sudekum AE, Oakes DD, deVries PA. Pyloric stenosis in a prematureinfant. J Pediatr Surg 1992;27:1534-6.8. Bisset RAL, Gupta SC. Hypertrophic pyloric stenosis: ultrasonic appearances in asmall baby. Pediatr Radiol 1988;18:405.9. Haider N, Spicer R, Grier D. Ultrasound diagnosis of infantile hypertrophic pyloricstenosis: determinants of pyloric length and the effect of prematurity. Clin Radiol2002;57:136-9.

To the Editor:We read with interest the article by Nuntnarumit et al1

showing that oral erythromycin is effective and safe for treat-ment of feeding intolerance in preterm infants. We would liketo highlight three issues:

1) We are intrigued that even though erythromycin halved theduration of time to enteral feeds in the intervention group, itdid not make any difference to the duration of hospital stay.

2) It is inappropriate to reassure readers by commenting thatthere was no significant difference in adverse effects,namely sepsis, especially when the study was not ade-quately powered to address this outcome.

3) Our population of babies who suffer from feeding intol-erance is more immature (�28 weeks), unlike the moremature babies in the study, and hence we question theapplicability of these results to a wider population.

We believe a large, randomized controlled trial is re-quired to find the long-needed answer of what role erythro-mycin has to play in resolving this problem.

Sandra Mascarenhas, MRCPCHMithilesh Lal, MRCPCH

James Cook University HospitalMiddlesborough, United Kingdom

10.1016/j.jpeds.2006.09.026

e30 The Journal of Pediatrics • February 2007

REFERENCE1. Nuntnarumit P, Kiatchoosakun P, Tantiprapa W, Boonkasidecha S. Efficacy oforal erythromycin for treatment of feeding intolerance in preterm infants. J Pediatr2006;148:600-5.

Reply

To the Editor:We appreciate the comments regarding our article and

would like to thank Drs. Su and Lin for their relevant obser-vations and concerns. All comments are well taken and shouldbe kept in mind in the interpretation of results of erythromy-cin (EM) trials in preterm infants. We agree that the infantsin our study probably had mild feeding intolerance, but theresponse to EM was dramatic in treated infants as demon-strated by significantly reduced gastric residual and number offeeds withheld within 24 to 48 hours after initiation of thetreatment.1 In cases with more severe feeding intolerance, ahigher dose and increased duration of EM may be required.2

To our knowledge, no serious complications were re-ported in more than 184 newborn infants who received EMfor prophylaxis or treatment of feeding intolerance either inobservational or randomized control trials.3 Certainly, furtherstudies of dose-response relationship and correlation withblood levels of EM are needed to determine the therapeuticdosage, as well as serious side effects such as hypertrophicpyloric stenosis in this particular group.

We also appreciate the questions and concerns of Drs.Mascarenhas and Lal. The following are responses:

1. The length of hospital stay was not significantly differentbetween the two groups (median 46 days in the groupreceiving EM vs 60 days in the group receiving placebo,P � .07). The small sample size was not sufficient to detecta difference. In this study, infants in the group receivingEM reached full enteral feeding approximately 1 weekearlier than in the control group. However, there are manyfactors that affect the hospital course of premature infants,which could explain the lack of difference in length of stay.

2. We do not recommend routine use of EM. Definitely, largerclinical trials are needed to confirm its safety and efficacy. Wehave mentioned in the discussion that the calculated samplesize in this study has sufficient power to evaluate the efficacyof EM, but not the incidence of side effects.

3. Approximately one fourth of infants in the group receivingEM in our study were below 28 weeks’ gestation. Webelieve the very preterm infants might have different re-sponses than more mature infants.

Pracha Nuntnarumit, MDDepartment of Pediatrics

Faculty of Medicine, Ramathibodi HospitalMahidol University, Bangkok, Thailand

Pakaphan Kiatchoosakun, MDDepartment of Pediatrics

Faculty of Medicine, Srinagarind HospitalKhon Kaen University, Khon Kaen, Thailand

Wacharee Tantiprapa, MDDepartment of Pediatrics

Faculty of MedicineChiang Mai University, Chiang Mai, Thailand

Suppawat Boonkasidecha, MDDepartment of Pediatrics

Faculty of MedicineSrinakarinwirot University, Bangkok, Thailand

10.1016/j.jpeds.2006.10.065

REFERENCES1. Nuntnarumit P, Kiatchoosakun P, Tantiprapa W, Boonkasidecha S. Efficacy of oralerythromycin for treatment of feeding intolerance in preterm infants. J Pediatr2006;148:600-5.2. Ng PC, So KW, Fung KS, Lee CH, Fok TF, Wong E, et al. Randomisedcontrolled study of oral erythromycin for treatment of gastrointestinal dysmotility inpreterm infants. Arch Dis Child Fetal Neonatal Ed 2001;84:177-82.3. Patole S, Rao S, Doherty D. Erythromycin as a prokinetic agent in pretermneonates: a systematic review. Arch Dis Child Fetal Neonatal Ed 2005:90:301-6.

Cardiovascular support in the preterm: Treatmentsin search of indications

To the Editor:I read with interest the recent editorial1 commenting on

the findings of a pilot study by Paradisis et al2 on the use ofmilrinone for low systemic flow in the very low birth weight(VLBW) neonate during the immediate postnatal period.Barrington and Dempsey1 discuss the management optionspracticed by most neonatologists to treat cardiovascular com-promise in this patient population and point out the lack ofavailable evidence supporting the efficacy of these treatmentstrategies. The authors conclude that there is no publishedevidence that the use of dopamine (or that of epinephrine ordobutamine) to treat cardiovascular compromise in theVLBW neonate during the immediate postnatal period de-creases mortality and improves outcomes, and that it is es-sential that we now design and perform the right trials todetermine whether the use of present treatment modalities ishelping these patients rather than harming them. These con-clusions correctly sum up the major issues that have plaguedthe field of neonatal hemodynamics for many years. However,although the description of these major concerns is thought-ful, certain aspects of the discussion on developmental car-diovascular physiology and the effects of vasopressors andinotropes in this patient population warrant a more compre-hensive analysis of the frequently conflicting findings.

First, the major debate regarding neonatal cardiovascu-lar compromise and its management centers around the ques-tion of whether systemic blood pressure adequately represents thestatus of systemic, and thus cerebral, blood flow (CBF) in theVLBW neonate with its unique hemodynamic profile in theimmediate postnatal period.3 However, because the lowerlimits of the normal gestational- and postnatal-age dependentblood pressure range in this patient population are not knownand because there are no data available on the effectiveness ofthe current management, no definite answer can be given to

Letters e31

this question. Instead, Barrington and Dempsey propose thatbecause “there is a very poor relationship between bloodpressure and systemic blood flow[4] and because treatments ofhypotension may elevate blood pressure but decrease perfu-sion,” measurements of flow rather than those of blood pres-sure should guide management of neonatal circulatory com-promise in the VLBW neonate. Although assessment ofsystemic blood flow is very important in VLBW neonates,especially during the early transitional period, the notion thatblood pressure changes in response to treatment are useless inthe assessment of systemic and CBF has to be challenged.

As stated by the authors, findings of studies usingsuperior vena cava (SVC) blood flow measurements as asurrogate to CBF suggest that cardiac output (systemic bloodflow) rather than blood pressure is the most important factordetermining CBF3 and demonstrate that decreases in sys-temic blood flow in the VLBW neonate during the firstpostnatal days are indeed associated with central nervoussystem injury and poor neurodevelopmental outcome.5 Fur-thermore, an earlier study using near infrared spectroscopy(NIRS) also could not find a significant relationship betweenblood pressure and CBF.6 However, other investigators usingNIRS have found that hypotensive preterm infants on thefirst postnatal day have low CBF,7,8 and that increases inblood pressure in response to dopamine7,8 or epinephrine8

administration are associated with increases in CBF. In ad-dition, available data indicate that a number of VLBW neo-nates present with a pressure passive cerebral circulation.9

Finally, more recent findings using NIRS and/or functionalechocardiography indicate that CBF is clearly affected by bothblood pressure and cardiac output in the VLBW neonate duringthe immediate transitional period.10-12 Thus, although long-term outcome data are not available, following the bloodpressure response to vasopressor/inotrope administration ispractical, as it can provide indirect information on organblood flow changes (see below). Obviously, the best approachwould be to continuously monitor both blood pressure andsystemic blood flow in these patients. Unfortunately, the toolsavailable to assess systemic blood flow in the VLBW neonateat the bedside (functional echocardiography and NIRS) havesignificant limitations.3,9,10,12 In addition, they have primarilybeen employed for research purposes, and it is unknownwhether their routine clinical use would affect outcome.

Second, the statement that “dopamine appears to ele-vate blood pressure predominantly by vasoconstriction at theexpense of systemic blood flow” is based on a study that usedfixed doses of dopamine at 10 and 20 mcg/kg/minute insteadof titration of the medication to achieve the desired hemo-dynamic response.13 As mentioned earlier, a recent random-ized controlled trial comparing the effects of dopamine withepinephrine on blood pressure and CBF in VLBW neonatesduring the first postnatal day found that, when titrated toachieve an “optimal” blood pressure response, both medica-tions successfully increased blood pressure and CBF at low-to-moderate doses.8 If, at the applied dose range, either dopa-mine or epinephrine had increased blood pressure by

primarily causing a generalized vasoconstriction, CBF wouldlikely have not increased. The authors of this study8 con-cluded that their “findings provide additional support for theuse of mean blood pressure changes as a consistent outcomemeasure of systemic organ blood flow” in the VLBW neonateduring the first postnatal day. Findings of another observa-tional study lend further support to this notion.7

Third, as for the inotropic effects of dopamine, severalstudies have demonstrated that, if carefully titrated, moderateto moderately high doses of the drug increase blood pressurewhile cardiac output is either increased14-17 or unchanged.18

However, as stated in the editorial, some patients may indeedrespond to dopamine with an increase in blood pressure anda decrease in cardiac output. It is of note though that thestudy17 cited in the editorial to support the notion that“dopamine may increase blood pressure but decrease perfu-sion” suffers from significant shortcomings including its smallpatient population and its use of left cardiac output to assesssystemic blood flow in patients with a large patent ductusarteriosus (PDA). As for its findings,17 out of the 14 pretermneonates enrolled in this study, all responded to mediumdoses of dopamine with increases in blood pressure, whereasleft cardiac output increased in 10 patients by 50 mL/kg/minute and decreased in the remaining 4 patients by onlyaround 20 mL/kg/minute. Again, as all of the patients had aPDA, their left cardiac output cannot be used to assess thechanges in systemic blood flow. Taking all the available in-formation into consideration, I do support the notion thatdopamine, particularly at high doses, may increase bloodpressure at the expense of systemic perfusion in certain pa-tients. However, this type of hemodynamic response is notthe rule, especially if the drug is carefully titrated.19

In summary, the main conclusions of the editorial aboutthe lack of evidence of the effectiveness of the management ofneonatal hypotension with vasopressor/inotropes and theneed for well-designed randomized controlled trials evaluat-ing medium- and long-term outcomes of VLBW neonatestreated for hypotension and/or low systemic blood flow are ofgreat importance. However, the notions that blood pressure ismore or less a useless hemodynamic parameter for the assess-ment of the cardiovascular status of the VLBW neonate in theimmediate postnatal period and that dopamine is a purevasopressor without positive inotropic effects are not sup-ported by the available findings in the literature.

Istvan Seri, MD, PhDChildren’s Hospital Los Angeles and

The Women’s and Children’s Hospitalof the LAC�USC Medical Center

Keck School of Medicineof the University of Southern California

Los Angeles, CA10.1016/j.jpeds.2006.09.023

REFERENCES1. Barrington KJ, Dempsey EM. Cardiovascular support in the preterm: treatmentsin search of indications. J Pediatr 2006;148:289-91.

e32 The Journal of Pediatrics • February 2007

2. Paradisis M, Evans N, Kluckow M, Osborn D, McLachlan AJ. Pilot study ofmilrinone for low systemic blood flow in very preterm infants. J Pediatr2006;148:306-13.3. Kluckow M, Evans N. Superior vena cava flow in newborn infants: a novel markerof systemic blood flow. Arch Dis Child Fetal Neonatal Ed 2000;82:F182-F187.4. Kluckow M, Evans N. Relationship between blood pressure and cardiac output inpreterm infants requiring mechanical ventilation. J Pediatr 1996;129:506-12.5. Hunter RW, Evans N, Rieger I, Kluckow M. Low superior vena cava flow andneurodevelopment at 3 years in very preterm infants. J Pediatr 2004;145:588-92.6. Tyszczuk L, Meek J, Elwell C, Wyatt JS. Cerebral blood flow is independent ofmean arterial blood pressure in preterm infants undergoing intensive care. Pediatrics1998;102:337-41.7. Munro MJ, Walker AM, Barfield CP. Hypotensive extremely low birth weightinfants have reduced cerebral blood flow. Pediatrics 2004;114:1591-96.8. Pellicer A, Valverde E, Elorza MD, Madero R, Gaya F, Quero J, Cabanas F.Cardiovascular support fro low birth weight infants and cerebral hemodynamics: A random-ized, blinded clinical trial. Pediatrics 2005;115:1501-12.9. Tsuji M, Saul JP, du Plessis A, Eichenwald E, Sobh J, Crocker R, VolpeJJ. Cerebral intravascular oxygenation correlates with mean arterial pressure in criticallyill premature infants. Pediatrics 2000;106:625-32.10. Kissack CM, Garr R, Wardle SP, Weindling AM. Cerebral fractional oxygenextraction in very low birth weight infants is high when there is low left ventricularoutput and hypocarbia but is unaffected by hypotension. Pediatr Res 2004;55:400-5.11. Victor S, Marson AG, Appleton RE, Beirne M, Weindling AM. Relationshipbetween blood pressure, cerebral electrical activity, cerebral fractional oxygen extractionand peripheral blood flow in very low birth weight newborn infants. Pediatr Res2006;59:314-9.12. West CR, Groves AM, Williams CE, Harding JE, Skinner JR, Kuschel CA, et al.Early low cardiac output is associated with compromised electroencephalographic ac-tivity in very preterm infants. Pediatr res 2006;59:610-5.13. Osborn D, Evans N, Kluckow M. Randomized trial of dobutamine versus dopa-mine in preterm infants with low systemic blood flow. J Pediatr 2002;140:183-91.14. Padbury JF, Agata Y, Baylen BG, Ludlow JK, Polk DH, Goldblatt E, et al.Dopamine pharmacokinetics in critically ill newborn infants. J Pediatr 1986;110:293-8.15. Lopez SL, Leighton JO, Walther FJ. Supranormal cardiac output in the dopa-mine- and dobutamine-dependent preterm infant. Pediatr Cardiol 1997;18:292-6.16. Lundstrom K, Pryds O, Greisen G. The hemodynamic effects of fopamine andvolume expansion in sick preterm infants. E Hum Dev 2000;57:157-63.17. Zhang J, Penny DJ, Kim NS, Yu VYH, Smolich JJ. Mechanism of blood pressureincrease induced by dopamine in hypotensive preterm neonates. Arch Dis Child FetalNeonatal Ed 1999;81:F99-F104.18. Roze JC, Tohier C, Maingureneau C, Lefevre M, Mouzard A. Response todobutamine and dopamine in the hypotensive very preterm infant. Arch Dis Child1993;69:59-63.19. Seri I. Cardiovascular, renal and endocrine actions of dopamine in neonates andchildren. J Pediatr 1995;126:333-44.

Reply

To the Editor:We thank Dr Seri for his thoughtful consideration of

our editorial, and we certainly recognize his expertise andmajor contributions in this area. However, we wish to furtherclarify some of the issues raised. Blood flow through anyvascular bed is determined by the perfusion pressure and thevascular resistance. Thus, cerebral blood flow (CBF), forexample, is determined by cerebral vascular resistance and thecerebral perfusion pressure (generally mean arterial pressureminus either central venous or intracranial pressure, which-ever is the greater). Published data do not demonstrate thatCBF is partly determined by cardiac output but do show acorrelation between the two because CBF is a major compo-nent of systemic perfusion1 and therefore of venous return tothe heart: thus, cardiac output is partly determined by CBF.

Any drug that causes cerebral vasoconstriction has a riskof reducing CBF, even if the perfusion pressure is increased,and such a response to dopamine (an extremely potent �-ad-renergic agonist) has been demonstrated in some experimen-tal animal preparations.2

It has been clear for many years that the unique struc-tural and functional limitations of the neonatal myocardiummake it extremely susceptible to increases in afterload, whichcommonly lead to a decrease in cardiac output.3 Unfortu-nately, the uncontrolled studies quoted by Dr Seri all have thesame limitations that he notes for Zhang et al that leftventricular output (LVO) has been the primary flow measuredand that sample sizes have tended to be small (eg, LVOreported in 4 infants).4 The largest of those studies in factdemonstrated exactly what we were stressing—that there is nodirect relation between blood pressure and cardiac output.5

That study did not have any measurements taken before theinstitution of therapy for “hypotension” (defined as a meanarterial pressure less than 30 mm Hg, regardless of gestationalor postnatal age) and certainly did not demonstrate that“careful titration” of dopamine doses will produce an increasein systemic perfusion. Similarly, the study by Lundstrom etal,6 the only controlled study of dopamine in the newborninfant that has shown an increase in LVO, did not show anyincrease in CBF despite the elevation in blood pressure,therefore confirming that cerebral vasoconstriction may occurwith dopamine infusion. Three other controlled trials of do-pamine in the newborn infant that have included measure-ments of “cardiac output” have all shown that dopamine,when carefully titrated to achieve the desired increase in bloodpressure, caused a decrease in LVO7-9; only one measuredright ventricular output9 and showed a decrease in rightventricular output when dopamine was “carefully titrated” toachieve the desired blood pressure response.

It is clear from this discussion that there are many unan-swered questions. We maintain that many low birth weightinfants with low numerical values for blood pressure do notrequire treatment. The challenge is to determine how to selectthose infants who may benefit from intervention and then todetermine which interventions may improve their outcomes.

Keith J. Barrington, MDPediatrics and Obstetrics & Gynaecology

McGill University NICURoyal Victoria Hospital

Montreal, Quebec, Canada10.1016/j.jpeds.2006.09.032

REFERENCES1. Kluckow M, Evans N. Superior vena cava flow in newborn infants: a novel markerof systemic blood flow. Arch Dis Child Fetal Neonatal Ed 2000;82:F182-7.2. Gleason CA, Robinson R, Harris A, Traystman RJ. Cerebrovascular effects of IVdopamine infusions in preterm and near-term fetal sheep. J Invest Med 1998;46:138A.3. Hawkins J, Van Hare GF, Schmidt KG, Rudolph AM. Effects of increasingafterload on left ventricular output in fetal lambs. Circ Res 1989;65:127-34.4. Padbury JF, Agata Y, Baylen BG, Ludlow JK, Polk DH, Goldblatt E, et al.Dopamine pharmacokinetics in critically ill newborn infants. J Pediatr 1986;110:293-8.5. Lopez SL, Leighton JO, Walther FJ. Supranormal cardiac output in the dopa-mine- and dobutamine-dependent preterm infant. Pediatr Cardiol 1997;18:292-6.6. Lundstrom K, Pryds O, Greisen G. The haemodynamic effects of dopamine andvolume expansion in sick preterm infants. Early Hum Dev 2000;57:157-63.7. Roze JC, Tohier C, Maingueneau C, Lefevre M, Mouzard A. Response to dobutamineand dopamine in the hypotensive very preterm infant. Arch Dis Child 1993;69:59-63.8. Chatterjee A, Bussey ME, Leuschen MP, Latson LA, Hole EB, Christian KK, et al. Thepharmacodynamics of inotropic drugs in premature neonates. Pediatr Res 1993;33:206A.9. Phillipos EZ, Barrington KJ, Robertson MA. Dopamine versus epinephrine forinotropic support in the neonate: a randomized blinded trial. Pediatr Res 1996;39:238A.

Letters e33

Documents

The Journal of Pediatrics - lib.ajaums.ac.irlib.ajaums.ac.ir/booklist/94457.pdf · Tridapalli, Fabrizio Sandri, Maddalena Menarini, Emanuela Ferretti and Giacomo Faldella . 23. Comparative