LETTERS - Diabetes Care · 2004-01-09 · Diabetes T he 2000 Dietary Guidelines for Americans and the dietary guide-lines for the control of diabetes in Mexico include recommendations

OBSERVATIONS

RelationshipBetween PeriodontalDisease and DiabeticRetinopathy

R ecently, various studies have re-ported that periodontal disease ad-versely affects diabetes (1). The

control of periodontal disease in elderlyindividuals has been reported to improvethe control of blood glucose (2). Severeperiodontal disease is associated with el-evated blood lipopolysaccharide levels asa result of periodontogenic bacteria,which induce higher levels of interleu-kin-6 (IL-6) and tumor necrosis factor-�(TNF-�) (3,4). Control of periodontaldisease is now considered not only a den-tal problem but also an issue affecting thepatient’s overall quality of life. Proinflam-matory cytokines such as IL-6 have beenshown to be involved in the pathogenesisof diabetic retinopathy (DR) (5), while therelationship between diabetic retinopathyand periodontal disease remains unclear.We investigated whether periodontal dis-ease is correlated with diabetic retinopathy.

The study was based on a prospectivereview of 73 eyes in 73 consecutive dia-betic patients. The mean duration of dia-betes was 14.3 � 7.1 years (range 2–33),and the mean HbA1c was 7.5 � 1.6%(5.2–13.7). IL-6 and TNF-� levels in thevitreous fluid samples from 32 eyes ob-tained during vitrectomy and in pairedplasma samples were measured by en-zyme-linked immunosorbent assay. Non-diabetic patients included 10 withmacular hole and 2 with epiretinal mem-brane. Institutional ethics committee ap-proval was obtained, and all participantsgave informed consent. The severity of di-abetic retinopathy was quantified accord-ing to the modified Early TreatmentDiabetic Retinopathy Study (ETDRS) ret-inopathy severity scale (6). The severity ofperiodontal disease was quantified ac-cording to bone loss and then graded andevaluated (7). Patients with periodontaldisease were classified as positive or neg-ative based on median values. Diabeticpatients were classified as having nonpro-

liferative or proliferative diabetic retinop-athy. Data are presented as means � SD.The Mann-Whitney U test was used tocompare IL-6 and TNF-� levels. To deter-mine the relationship between the sever-ity of periodontal disease and ETDRS,retinopathy severity, or angiogenic fac-tors, as well as between X and Y parame-ters, Spearman’s rank-order correlationcoefficient and logistic regression modelwere applied.

The severity of periodontal diseasewas significantly correlated with the se-verity of diabetic retinopathy (P �0.0012), and the risk of proliferative dia-betic retinopathy was significantly higherin the presence of periodontal disease(odds ratio � 2.80, P � 0.036). Therewas no significant relationship betweenthe severity of periodontal disease andHbA1c or duration of diabetes (P � 0.098and 0.295, respectively). There was a sig-nificant relationship between the severityof diabetic retinopathy and duration ofdiabetes (P � 0.002). The vitreous fluidlevel of IL-6 (mean 154.2 � 164.6 pg/ml[range 0.993–597.0]) was significantly el-evated in patients with diabetic retinopa-thy compared with that in nondiabeticpatients (mean 1.34 � 0.91 pg/ml [0.6–3.68]) (P � 0.0001). Furthermore, thevitreous fluid level of IL-6 was signifi-cantly correlated with the severity of peri-odontal disease (P � 0.012). There wasno significant relationship between thevitreous fluid level of IL-6 and HbA1c orduration of diabetes (P � 0.293 and0.705, respectively). In contrast, the vit-reous fluid level of TNF-� was not signif-icantly correlated with the severity ofperiodontal disease. The IL-6 concentra-tion in vitreous fluid (mean 154.2 �164.6 pg/ml [0.993–597.0]) was signifi-cantly higher than that in plasma (mean1.89 � 3.47 pg/ml [0.156–18.8]) (P �0.0001).

There was a significant relationshipbetween periodontal disease and severityof diabetic retinopathy, but it was unclearwhether periodontal disease directly af-fects the progression of diabetic retinopa-thy because this was a cross-sectionalstudy. Further prospective studies, in-cluding evaluation of systemic factors, arenecessary.

HIDETAKA NOMA, MD1

IKUO SAKAMOTO, PHD1

HIDEKI MOCHIZUKI, MD1

HIDETOSHI TSUKAMOTO, PHD2

ATSUSHI MINAMOTO, PHD1

HIDEHARU FUNATSU, PHD3

HIDETOSHI YAMASHITA, PHD4

SHIGEO NAKAMURA, PHD5

KEN KIRIYAMA, PHD4

HIDEMI KURIHARA, PHD5

HIROMU K. MISHIMA, PHD1

From the 1Department of Ophthalmology and Vi-sual Science, Hiroshima University Graduate Schoolof Biomedical Sciences, Hiroshima, Japan; the 2De-partment of Ophthalmology, Hiroshima PrefecturalHospital, Hiroshima, Japan; the 3Department ofOphthalmology, Diabetes Center, Tokyo Women’sMedical University, Tokyo, Japan; the 4Departmentof Ophthalmology and Visual Science, YamagataUniversity School of Medicine, Yamagata, Japan;and the 5Department of Periodontal Medicine, Hi-roshima University Graduate School of BiomedicalSciences, Hiroshima, Japan.

Address correspondence to Hiromu K. Mishima,PhD, Hiroshima University Graduate School of Bio-medical Sciences, Department of Ophthalmologyand Visual Science, 1-2-3 Kasumi, Minami-ku, Hi-roshima 734-8551, Japan. E-mail: [email protected].

© 2004 by the American Diabetes Association.

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Loe H: Periodontal disease: the sixth com-

plication of diabetes mellitus. DiabetesCare 16:329–334, 1993

2. Grossi SG, Skrepcinski FB, DeCaro T,Robertson DC, Ho AW, Dunford RG,Genco RJ: Treatment of periodontal dis-ease in diabetics reduces glycated hemo-globin. J Periodontol 68:713–719, 1997

3. Grossi SG, Genco RJ: Periodontal diseaseand diabetes mellitus: a two-way relation-ship. Ann Periodontol 3:51–61, 1998

4. Soolari AS, Champagne C, Punzi JS, AmarS, Van Dyke TE: Serum modulation ofneutrophil response to Porphyromonasgingivalis LPS in periodontal disease. J IntAcad Periodontol 1:101–109, 1999

5. Funatsu H, Yamashita H, Shimizu E,Kojima R, Hori S: Relationship betweenvascular endothelial growth factor and in-terleukin-6 in diabetic retinopathy. Retina21:469–477, 2001

6. The Early Treatment Diabetic Retinopa-thy Study Research Group: Fundus pho-tographic risk factors for progression ofdiabetic retinopathy: ETDRS report num-ber 12. Ophthalmology 98:823–833, 1991

7. Jeffcoat MK, Wang IC, Reddy MS: Radio-graphic diagnosis in periodontics. Peri-odontol 2000 7:54–68, 1995

L E T T E R S

DIABETES CARE, VOLUME 27, NUMBER 2, FEBRUARY 2004 615

Effect of DiabetesInterventionPrograms onPhysical ActivityAmong MigrantMexican WomenWith Type 2Diabetes

T he 2000 Dietary Guidelines forAmericans and the dietary guide-lines for the control of diabetes in

Mexico include recommendations thatadults participate in at least 30 min ofmoderate physical activity, preferablydaily (1,2). However, to our knowledge,there have been no studies that explorethe effect of diabetes intervention pro-grams on physical activity among migrantMexican women with type 2 diabetes.

All women from seven diabetes edu-cation groups from three different Mexi-can institutions located in Tijuana wereinvited to answer a previously validatedquestionnaire on physical activity. Of 111questionnaires, 100 were adequately an-swered. The mean age was 53 � 12 years.The majority of the sample was migrantsfrom other Mexican states, and 40% wereclassified as overweight and 31% asobese. Of the women, 62, 45, and 15%reported �20, �30, and �60 min ofphysical activity per day, respectively.Seventy-three percent reported �80 minof weekly physical activity. Daily outdooractivity of participants was 39 � 4.2 min(mean � SE), and daily indoor activitywas 5.72 � 0.27 h. Total light activity(�3.0 metabolic equivalents [METs]) was5.28 � 0.24 h/day, total moderate activity(3–6 METs) was 55 � 14 min/day, andtotal vigorous activity (�6.0 METs) was4.2 � 0.6 min/day. The average physicalactivity level was 1.54 � 0.03. The mainindoor activities were cooking (11h/week), dishwashing and clothes wash-ing (3.2 h/week), cleaning (3.1 h/week),and shopping (1.9 h/week), and the mainoutdoor physical activities were walking(3.1 h/week), semiactive exercise andstretching (1.26 h/week), running (0.23h/week), and bicycle riding (0.18h/week). The main resting activity wassleeping (49.16 h/week), followed bywatching television (11.3 h/week), rest-ing in bed (2 h/week), driving or sitting in

a car (1.5 h/week), and sitting at home(1.38 h/week).

This study shows that the majority ofMexicans with diabetes who are willing toparticipate in diabetes education groupsat the primary health care clinics engagein �20 min of daily physical activity,which therefore reinforces the need forpromoting culturally based interventions(3,4). These results are better than the na-tional data for adults with type 2 diabetesin Mexico and the U.S. (5,6); however,the groups we studied were especiallymotivated subjects interested in obtainingbetter metabolic control through diabeteseducation groups. On the other hand, thepopulation from this study, which has alow socioeconomic status, usually con-fronts major environmental or economicbarriers that prevent access to safe recre-ational areas or fitness facilities. Thus,even with economic constraints and inad-equate environmental access to physicalactivity, promoting physical activity in aculturally based intervention is a worth-while strategy that should be supported.Although further studies in large popula-tions are still required to evaluate the ef-fectiveness at a larger scale, at the primarycare level of Mexican institutions, stron-ger emphasis should be placed on pro-moting physical activity.

MONTSERRAT BACARDı-GASCON, MD

PERLA ROSALES-GARAY, MD

ARTURO JIMENEZ-CRUZ, MD, PHD

From the Nutrition Program, Medical School, Uni-versidad Autonoma de Baja California, Tijuana,Mexico.

Address correspondence to Montserrat Bacardı-Gascon, MD, Universidad Autonoma de Baja Cali-fornia, Medical School, Nutrition, Av. Tecnologico14418, Mesa de Otay, Baja California, Tijuana22390, Mexico. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. US Department of Agriculture/Depart-

ment of Health and Human Services:Home and garden bulletin no. 232. In2000 Dietary Guidelines for Americans. 5thed. Washington, DC, U.S. Department ofAgriculture, 2000

2. Secretaria de Salud: Norma Oficial Mexi-cana Para la Prevencion, Tratamiento yControl de la Diabetes Mellitus en la Aten-cion Primaria a la Salud. Mexico, DF, Mex-ico, Secretaria de Salud, 2000 (NOM-015-SSA2-1994)

3. Jimenez-Cruz A, Bacardı-Gascon M, Ro-sales-Garay P, Herrera-Espinoza J, Willis

OW: A culturally sensitive tool for Mexi-can people with diabetes: “La Manzana dela Salud”. Rev Biomed 14:51–59, 2003

4. Clark DO: Physical activity efficacy andeffectiveness among older adults and mi-norities. Diabetes Care 20:1176–1182

5. Aguilar-Salinas CA, Velazquez-MonroyO, Gomez-Perez FJ, Gonzalez-Chavez A,Esqueda AL, Molina-Cuevas V, Rull-Rodrigo JA, Tapia Conyer R, the EncuestaNacional de Salud 2000 Group: Charac-teristics of patients with type 2 diabetes inMexico: results from a large population-based nationwide survey. Diabetes Care26:2021–2026, 2003

6. Nelson KM, Reiber G, Boyko EJ: Diet andexercise among adults with type 2 diabe-tes. Diabetes Care 25:1722–1728, 2002

Alternative SiteTesting at theEarlobe Tip

Reliability of glucose measurementsand pain perception

There is growing interest in alterna-tive less painful sites for capillaryblood glucose (CBG) monitoring.

The earlobe tip is a potential site (1) thathas been occasionally used by nurseswhen fingertip testing is refused or diffi-cult. We investigated the clinical valueand accuracy of earlobe CBG measure-ments as an alternative to fingertip andforearm testing.

A total of 50 patients with type 2 di-abetes (aged 42–82 years, 28% with neu-ropathy) were enrolled in the study. Theduration of diabetes was 9.5 � 8.3 years(means � SD). Testing sites (lateral aspectof the fingertip, earlobe tip, and flexorsurface of the forearm) were rubbed andcleaned before lancing was performed bya physician using Lifescan Unistick-2 lan-cets (Lifescan, Milpitas, CA). The forearmwas lanced using the Microlet-Vaculancedevice (Bayer, Tarrytown, NY). The orderof site testing was randomized. Pain wasimmediately assessed after the first at-tempt using a 100-mm graphic visualscale (2). CBG was measured with anAccu-Check Advantage glucose meter(Roche, Indianapolis, IN).

First-attempt sampling success rateswere 88% (fingertip), 74% (earlobe), and

Letters

616 DIABETES CARE, VOLUME 27, NUMBER 2, FEBRUARY 2004

62% (forearm). After earlobe pricking,bleeding lasted for �60 s in all subjects(�30 s in 90%). Earlobe pain (medianscore 5.5 mm) was less uncomfortablethan fingertip pain (17 mm, P � 0.01,Wilcoxon test) but not statistically differ-ent from forearm (8.0 mm). Although ear-lobe pain was more tolerable, a limitationin this study is that it did not evaluate painperception after repeated testing on mul-tiple days, when skin soreness might be-come relevant.

CBG measurements from the earlobedeviated from the fingertip by 9.5 � 1.0%(mean � SE). The correlation coefficientbetween these two sites was 0.97 (P �0.01). Of earlobe measurements, 97.8%were within clinically acceptable zonesA�B by error grid analysis (3). When ear-lobe was compared with forearm, CBGdeviation was 10 � 1.4%, correlation co-efficient 0.96 (P � 0.01), and 95.3% ofmeasurements were within zones A�B byerror grid analysis.

In contrast to other sites that requirevacuum-assisted lancing devices or so-phisticated glucose meters, our observa-tions demonstrate that CBG monitoring atthe earlobes is attainable with regularstandard lancets and that earlobe CBGconcentrations correlate well and deviateminimally from either fingertip or fore-arm values. Earlobe testing was also clin-ically accurate by the error grid analysismethod, which takes into account the ef-fects on the clinical decision that wouldhave been made if the CBG had been mea-sured in the reference site. These datasupport the notion that earlobe CBG con-centrations can be used in substitution offingertip values. However, one limitationin our data series is that glucose values�70 mg/dl were not observed. Therefore,earlobe testing should be avoided if hypo-glycemia is suspected.

For many patients, the earlobe prick-ing technique may be less convenient be-cause it requires a second person to lancethe skin and collect the blood. This cer-tainly limits the applicability of the tech-nique in the outpat ient se t t ing .Nevertheless, earlobe testing seems to bea useful alternative that could minimizecosts and discomfort in patients assistedby a relative or nurse in hospitals andnursing homes.

FREDERICO G.S. TOLEDO, MD1,2

ANDREW TAYLOR, MD2

From the 1Department of Medicine, Division of En-docrinology and Metabolism, University of Pitts-burgh Medical Center, Pittsburgh, Pennsylvania;and the 2Department of Internal Medicine, Divisionof Endocrinology and Metabolism, University of Mi-ami School of Medicine and Miami VA Medical Cen-ter, Miami, Florida.

Address correspondence to Frederico G.S. To-ledo, MD, University of Pittsburgh Medical Center,Department of Medicine, Division of Endocrinologyand Metabolism, E1140 Biomedical Science Tower,200 Lothrop St., Pittsburgh, PA 15261. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Carley SD, Libetta C, Flavin B, Butler J,

Tong N, Sammy I: An open trial to reducethe pain of blood glucose testing: ear ver-sus thumb. BMJ 321:20, 2000

2. Huskisson EC: Measurement of pain. Lan-cet 9:1127–1131, 1974

3. Clarke WL, Cox D, Gonder-Frederick LA,Carter W, Pohl SL: Evaluating clinical ac-curacy of systems for self-monitoring ofblood glucose. Diabetes Care 10:622–628,1987

Needle Reuse andTip Damage

Q uestions about needle reuse areraised in recommendations by theAmerican Diabetes Association (1)

and by practitioners and manufacturingcompanies (2). As an initial approach toexamining these questions, we investi-gated whether multiple insertions of nee-dles through the rubber stoppers oninsulin vials would cause damage to nee-dle tips (3). New needles were selected,attached to a precision manipulator, andinserted into insulin vials attached to atransducer that measured the force neces-sary to insert and remove a needlethrough the stopper. If needles were dam-aged by multiple insertions, an increase inforce required to penetrate the rubberstopper would be expected.

In five experiments, we found thatthere was no significant difference in themean force required to penetrate the rub-ber stopper of an insulin vial between 17initial (41.5 g) and 17 subsequent (42.0 g)insertions. This suggests that needle tipsare not dulled or damaged by multipleinsertions to a degree that more forcewould be required to penetrate a vial

stopper. Light micrographs and scanningelectron micrographs supported the con-clusion that little or no damage to needletips occurred as a result of multiple inser-tions. We observed no hooks, bending, orother needle tip damage, although therewas some evidence for deterioration of thesilicon lubricant coating of needles.

At the same time, our examination ofnew, unused needles revealed imperfec-tions when observed at high magnifica-tion. Our observations make it clear thatdamage attributed to reuse must be dis-tinguished from inherent imperfectionsassociated with the manufacturingprocess.

Clinical and manufacturer recom-mendations that discourage reuse of nee-dles because of assumed dulling,bending, and/or fragmenting of needletips are at variance with our findings, inwhich repeated penetration of insulin vialstoppers did not damage needle tips (2).Although we found no evidence of needledamage when penetrating the rubberstopper on insulin vials, there is need toascertain the effect of cutaneous tissuepenetration on needle tips. The reuse ofneedles by a significant number of pa-tients with diabetes as a matter of conve-nience, or out of concern for cost and/orthe creation of nonbiodegradable waste,indicates that further examination of nee-dle reuse in vial stoppers and cutaneoustissue is warranted.

DOUGLAS KLINE, PHD1

TERRY KUHN, PHD2

From the 1Department of Biological Sciences, KentState University, Kent, Ohio; and the 2Division ofUndergraduate Studies, Kent State University, Kent,Ohio.

Address correspondence to Terry Kuhn, KentState University, Division of Undergraduate Studies,P.O. Box 5190, Kent, OH 44242-0001. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. American Diabetes Association: American

Diabetes Association Complete Guide to Di-abetes. 3rd ed. American Diabetes Associ-ation, Alexandria, VA, 2002, p. 106, 480

2. Effect of multiple insulin vial insertions onneedle tips [article online], 2003. Availablefrom http://www.diabetesneedlereuse.org.Accessed 8 December 2003

3. The Editors: Questions and answers (Let-ter). Diabetes Self Manag 17:94–95, 2000

Letters


The DiabetesTreatmentSatisfactionQuestionnaire

A cross-cultural South Africanperspective

R eliable and valid multicultural in-struments are important in multi-cultural societies that are typical of

modern cities, and clinicians, using psy-chosocial assessments, need to ensurethat their diagnostic and screening toolsare appropriate. This study was con-ducted with 176 diabetic outpatientsfrom two culturally distinct groups (95Bantu-speaking and 81 Afrikaans-speaking subjects) to 1) ascertain the un-derlying dimensions of treatmentsatisfaction as measured by the DiabetesTreatment Satisfaction Questionnaire(DTSQ status) (1), 2) determine the reli-ability (internal consistency) of the mea-sures, and 3) investigate the effects ofobjective (HbA1c results) and subjectivemetabolic control, health (2), and well-being (3) on satisfaction with diabetestreatment.

Principal components analysis wasconducted on the 8-item DTSQ (1). Allcommunality estimates exceeded the cri-terion of 0.30 (4) for both Bantu-speakingand Afrikaans-speaking patients (range0.62–0.79 and 0.55–0.76, respectively).Two factors explained 71% of the vari-ance for Bantu-speaking patients and68% of the variance for Afrikaans-speaking patients. The first factor con-sisted of the six treatment satisfactionitems, and the second factor consisted ofthe two subjective metabolic controlitems. Reliability (internal consistency)coefficients were excellent (5) and verysimilar for both groups (�0.80 on allmeasures).

Treatment satisfaction was associatedwith fewer incidents of hyperglycemia(r � �0.58, P � 0.01) and hypoglycemia(r � �0.32, P � 0.01), higher generalwell-being (r � 0.56, P � 0.01), and bet-ter health (r � 0.44, P � 0.01) for Bantu-speaking patients. For Afrikaans-speaking patients, greater treatmentsatisfaction was associated with fewer in-cidents of hyperglycemia (r � �0.29, P �0.01), higher general well-being (r �

0.54, P � 0.01), and better health (r �0.50, P � 0.01). Language, sex, age, andemployment status were not related totreatment satisfaction or general well-being (P � 0.05), confirming the con-struct validity of the measures. HbA1c

results were not significantly related totreatment satisfaction, subjective meta-bolic control, general well-being, or gen-eral health for either group (P � 0.05).

For Bantu-speaking patients, fewerincidents of hyperglycemia significantlypredicted 33% of the variance (P �0.001) in treatment satisfaction; an addi-tional 11% of the variance (P � 0.001)was explained by general well-being. ForAfrikaans-speaking patients, generalwell-being predicted 29% of the variance(P � 0.001) in treatment satisfaction; anadditional 7% of the variance (P � 0.001)was explained by general health.

In conclusion, the study demon-strated that the underlying dimensions ofthe DTSQ for both groups were treatmentsatisfaction and hyper- and hypoglyce-mia, all measures had excellent reliability(5), and well-being is an important pre-dictor of treatment satisfaction for bothgroups of patients. These findings wereconsistent with those reported in the U.K.and Sweden (6–7) and support the ideathat the DTSQ can be used in multicul-tural settings.

Acknowledgments— We thank Novo Nor-disk (South Africa) for funding the interview-ers’ salaries. We also thank Professor ClareBradley for permission to use her measuresand Rosalind Plowright for constructive com-ments on the application of the DTSQ in mul-ticultural settings.

MARGARET SANDRA WESTAWAY, PHD1,2

JOHN R. SEAGER, PHD1,3

From the 1South Africa Medical Research Council,Health and Development Research Group, Pretoria,Gauteng, South Africa; the 2School of Health Sys-tems and Public Health, University of Pretoria, Pre-toria, Gauteng, South Africa, and the 3Faculty ofCommunity and Health Sciences, University of theWestern Cape, Cape Town, Western Cape, SouthAfrica.

Address correspondence to Prof. Margaret San-dra Westaway, SA Medical Research Council, Healthand Development Department, Private Bag X385,Pretoria, Gauteng, 0001 South Africa. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Bradley C: Diabetes treatment satisfaction

questionnaire. In Handbook of Psychologyand Diabetes: A Guide to Psychological Mea-surement in Diabetes Research and Practice.Bradley C, Ed. Chur, Switzerland, Har-wood Academic, 1994, p. 111–132

2. Stewart AL, Hays RD, Ware JE: The MOSShort-Form General Health Survey: reli-ability and validity in a patient popula-tion. Med Care 26:724–735, 1988

3. Bradley C: The well-being questionnaire.In Handbook of Psychology and Diabetes: AGuide to Psychological Measurement in Dia-betes Research and Practice. Bradley C, Ed.Chur, Switzerland, Harwood Academic,1994, p. 89–109

4. Child D: The Essentials of Factor Analysis.London, Holt, Rinehart & Winston,1970, p. 33–34

5. Arias E, de Vos S: Using housing items toindicate socio-economic status: LatinAmerica. Soc Indic Res 38:53–80, 1996

6. Petterson T, Young B, Lee P, Newton P,Hollis S, Dornan T: Well-being and treat-ment satisfaction in older people with di-abetes. Diabetes Care 21:930–935, 1998

7. Wredling R, Stalhammar J, Adamson U,Berne C, Larsson Y, Ostman J: Well-beingand treatment satisfaction in adults withdiabetes: a Swedish population-basedstudy. QualLife Res 4:515–522, 1995

Depression,Diabetes, andGlycemic Control inPima Indians

F ew studies have addressed the rela-tionship of depression and diabetesin ethnic minority groups, especially

Native Americans (1). We examined therelationship between depression and dia-betes in a community-based sample of541 Pima Indians aged �18 years (192with and 349 without diabetes) examinedfrom September 2002 through February2003.

Depression was defined by five ormore depressive symptoms lasting �2weeks, as assessed with PRIME-MD(Mood Module in the Primary Care Eval-uation of Mental Disorders) (2). Diabeteswas defined by a glucose tolerance test(fasting plasma glucose �7.0 mmol/l or2-h plasma glucose �11.1 mmol/l) orprevious clinical diagnosis.

Letters


The prevalence of depression was16.3% (18.7% in women and 12.6% inmen, P � 0.06). In both sexes, the prev-alence of depression was higher in dia-betic individuals (men 17.2 vs. 10.9%,women 20.2 vs. 17.6%), although thesedifferences were not statistically signifi-cant (for total sample: age- and sex-adjusted odds ratio 1.3 [95% CI 0.7–2.1]). In diabetic individuals, HbA1c washigher by 1.2% in those with depression(9.3 vs. 8.1%, P � 0.01), although de-pression was not related to HbA1c in non-diabetic individuals (5.2 vs. 5.3%, P �0.2). This association remained signifi-cant in a multivariate linear regressionmodel that included age, sex, duration ofdiabetes, and BMI (HbA1c higher by 1.1%in depressed persons, P � 0.01). Fastingplasma glucose was also higher, but notsignificantly so, in depressed diabetic in-dividuals (10.2 vs. 9.5 mmol/l, P � 0.3).

Although studies of depression in Na-tive-American communities are limited,our findings are consistent with previoussuggestions that depression is severaltimes more prevalent among NativeAmericans than in the general U.S. popu-lation (3). Our finding that the prevalenceof depression was somewhat higher in di-abetic individuals is also consistent withprevious studies (1,4–6). Our study lacksprecision to estimate the association ofdepression with diabetes because of therelatively small sample size (541, as com-pared with 21,513 to 1.3 million in otherrecent reports [4–6]). The high preva-lence of depression in our study suggeststhat certain social, cultural, or economicfactors may overshadow the influence ofdiabetes on depression in this population.

The higher HbA1c in depressed dia-betic individuals is consistent with previ-ous findings in other populations (7).Treatment of depression reportedly im-proves glycemic control in diabetic pa-tients, although the long-term effects arenot known (8,9). This study adds to thesparse literature on depression and diabe-tes in ethnic minority groups. Identifica-tion and treatment of depression may bean important aspect of treating diabetes inNative Americans.

PUNEET K. SINGH, BA

HELEN C. LOOKER, MBBS

ROBERT L. HANSON, MD, MPH

JONATHAN KRAKOFF, MD

PETER H. BENNETT, MB, FRCP

WILLIAM C. KNOWLER, MD, DRPH

From the National Institute of Diabetes and Diges-tive and Kidney Diseases, National Institutes ofHealth, Phoenix, Arizona.

Address correspondence to Helen C. Looker, Na-tional Institutes of Health, 1550 E. Indian SchoolRd., Phoenix, AZ 85014. E-mail: [email protected].


Acknowledgments— We thank Dr. RichardRubin, Dr. Patrick Lustman, and Dr. Maryde Groot for advice; Dr. Diane Montella andPriscilla Foote, MSW, of Gila River HealthCare; and the members of the Gila River IndianCommunity for their participation.

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Anderson RJ, Freedland KE, Clouse RE,

Lustman PJ: The prevalence of comorbiddepression in adults with diabetes: ameta-analysis. Diabetes Care 24:1069–1078, 2001

2. Spitzer RL, Williams JB, Kroenke K, Lin-zer M, deGruy FV III, Hahn SR, Brody D,Johnson JG: Utility of a new procedure fordiagnosing mental disorders in primarycare: the PRIME-MD 1000 study. JAMA272:1749–1756, 1994

3. US Department of Health and HumanServices: Mental health. A supplement toMental Health: A Report of the Surgeon Gen-eral: Culture, Race, and Ethnicity. Rock-ville, MD, U.S. Department of Health andHuman Services, 2001

4. Egede LE, Zheng D, Simpson K: Comor-bid depression is associated with in-creased health care use and expendituresin individuals with diabetes. Diabetes Care25:464–470, 2002

5. Nichols GA, Brown JB: Unadjusted andadjusted prevalence of diagnosed depres-sion in type 2 diabetes. Diabetes Care 26:744–749, 2003

6. Finkelstein EA, Bray JW, Chen H, LarsonMJ, Miller K, Tompkins C, Keme A,Manderscheid R: Prevalence and costs ofmajor depression among elderly claim-ants with diabetes. Diabetes Care 26:415–420, 2003

7. Lustman PJ, Anderson RJ, Freedland KE,de Groot M, Carney RM, Clouse RE: De-pression and poor glycemic control: ameta-analytic review of the literature. Di-abetes Care 23:934–942, 2000

8. Lustman PJ, Freedland KE, Griffith LS,Clouse RE: Fluoxetine for depressionin diabetes: a randomized double-blindplacebo-controlled trial. Diabetes Care 23:618–623, 2000

9. Lustman PJ, Griffith LS, Freedland KE,Kissel SS, Clouse RE: Cognitive behaviortherapy for depression in type 2 diabetesmellitus: a randomized, controlled trial.Ann Intern Med 129:613–621, 1998

Improper InsulinCompliance MayLead toHepatomegaly andElevated HepaticEnzymes in Type 1Diabetic Patients

W e have encountered hepatomeg-aly and pronounced elevation ofliver enzymes AST and ALT in

four patients with type 1 diabetes. Thesepatients shared similar clinical features.They were all female (aged 11–14 years)with poor glycemic control. All had fre-quent hyperglycemia and intermittenthypoglycemia related to their history ofpoor compliance. Most of them had mul-tiple hospital admissions for severe hy-perglycemia and/or diabetic ketoacidosis.In addition to their high daily doses ofinsulin (1.3–2.2 units � kg�1 � day�1),most were receiving extra doses of insulinto correct their frequent hyperglycemia.A1C levels were all higher than normal(ranging from 9.2 to 14.5%). Their initialAST and ALT levels were at least 30- and14-fold higher than the normal limits, re-spectively, but the other liver functiontests, such as alkaline phosphatase, pro-thrombin/partial prothrombintime, andtotal bilirubin, were normal except forone patient who had a minimal increase inalkaline phosphatase and total billirubin.The degree of hepatomegaly did not cor-relate with the liver enzyme levels, nor didit correlate with glycemic control orHbA1c levels.

Upon admission to the hospital,proper insulin dosing was established.Three of the four patients were able tolower their insulin dose to 0.9–1.2 units �kg�1 � day�1 and achieve normal glyce-mic control. The AST and ALT levels werequickly decreased in just a few days afterthe patients obtained better glycemic con-trol during hospitalization. Except for onepatient, who was admitted for diabetic ke-toacidosis, the patients had no apparentsymptoms of liver disease before the admis-sion. Their hepatomegaly was an inciden-tal finding. Other than poorly controlleddiabetes, the investigations did not revealany other causes for hepatomegaly andincreased liver enzymes. The normal cre-atine phosphokinase level and negative

Letters


myoglobinuria from one patient ruled outthe possibility of rhabdomyolysis. Theliver biopsy obtained in one patient re-vealed abundant glycogen deposits inhepatocytes that were consistent with theabdominal computed tomography find-ing of “fatty” appearance of those enlargedlivers in all four patients. There were alsosome features that these patients did notshare. One patient with the most pro-found hepatomegaly had significant delayin growth and puberty consistent withMauriac syndrome as previously de-scribed (1), whereas the other three pa-tients had normal growth and puberty.One patient, who was found to have hep-atomegaly and elevated hepatic enzymesduring one of her admissions to the hos-pital for diabetic ketoacidosis, had somenonspecific gastrointestinal symptomsthat might have been related to her dia-betic ketoacidosis rather than the hepaticdisorder. Although all of our cases weregirls, similar cases have been identified inboys (2).

Hepatomegaly and elevated hepaticenzymes, reported in both adult and pe-diatric patients with type 1 diabetes (2,3),could be relatively common but may beunder-recognized or misidentified as themore common nonalcoholic steatohepa-titis (NASH) because of similar clinicalfeatures. NASH is commonly seen inobese type 2 diabetic patients with insulinresistance. The hepatic enzyme elevationis slow to resolve. Our patients, however,were all nonobese type 1 diabetic pa-tients. Their pronounced elevation of he-patic enzymes was resolved in just a fewdays once they achieved reasonable glyce-mic control at insulin dosages that werelower than what they were prescribed athome. Though the mild hepatomegalyand abnormal liver enzymes were be-lieved to be associated with liver steaticchange in NASH, whether the pro-nounced elevation of the liver enzymeswas directly caused by liver glycogen de-posit is not known. The pathogenesis forthese problems has not been well studied.Nevertheless, both the increased hepaticenzymes and glycogen deposits may berelated to poor glycemic control. Most ofour patients received relatively high dosesof insulin at home. We question the pos-sible role of insulin over-treatment thatmight contribute to the pathogenesis ofhepatomegaly because insulin is clearly apromoting agent for glycogenesis. Similarcases of hepatomegaly and elevated

hepatic enzymes have been reportedin children and adolescents who werechronically over-treated with insulin(4,5).

We therefore advocate the high vigi-lance in promoting patient compliance toinsulin dosing rather than simply increas-ing insulin dosage in response to hyper-glycemia. The swift reduction of hepaticenzymes in our cases after achieving rea-sonable glycemic control suggests thatliver biopsy and other extensive work-upmay be unnecessary in managing similarpatients.

Y. MILES YU, MD

CAMPBELL P. HOWARD, MD

From the Section of Endocrinology, Children’sMercy Hospital, University of Missouri at KansasCity School of Medicine, Kansas City, Missouri.

Address correspondence to Miles Yu, Section ofEndocrinology, Children’s Mercy Hospital, 2401Gillham Rd., Kansas City, MO 64108. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Lee RGG, Bode HH: Stunted growth and

hepatomegaly in diabetes mellitus. J Peds91:82–84, 1977

2. Chatila R, West AB: Hepatomagely andabnormal liver tests due to glycogenesis inadults with diabetes. Medicine 75:327–333, 1996

3. Olssen R, Wesslau C, William-Olsen T,Zettergren L: Elevated aminotransferasesand alkaline phosphatases in unstable di-abetes mellitus without ketoacidosis orhypoglycemia. J Clin Gastroenterol 11:541–545, 1989

4. Asherov J, Mimouni M, Varsano I, LubinE, Laron Z: Hepatomegaly due to self-in-duced hyperinsulinism. Arch Dis Child 54:148–149, 1979

5. Rosenbloom AL, Giordano BP: Chronicovertreatment with insulin in childrenand adolescents. Am J Dis Child 131:881–885, 1977

Acute Hyperglycemia

Implications for contrast-inducednephropathy during cardiaccatheterization

A cute hyperglycemia exacerbatesischemic injury of the brain andheart (1,2). Renal contrast agents

are nephrotoxic, largely due to acute isch-

emia secondary to renal artery vasocon-striction (3). Historically, diabeticpatients have been identified as a high-risk group for the development of con-trast-induced nephropathy followingcardiac catheterization; however, themechanism for this increased risk is un-clear (4). The purpose of this study was todetermine whether acute hyperglycemiais an independent risk factor for the de-velopment of contrast-induced nephrop-athy after cardiac catheterizationprocedures.

A prospective, observational studywas performed on all patients with diabe-tes (insulin dependent and independent)or any patient with a baseline serum cre-atinine �1.2 mg/dl receiving a cardiaccatheterization procedures in a universi-ty-affiliated cardiac catheterization facil-ity between June 2001 and January 2002.Patients with a diagnosis of acute renalfailure or patients on dialysis were ex-cluded. Patients were divided into twogroups, hyperglycemic (AHG) (serumglucose �150 mg/dl) and nonhypergly-cemic (NHG) (serum glucose �150 mg/dl), at the time of cardiac catheterizationprocedures. Contrast-induced nephropa-thy was defined as an increase in serumcreatinine �0.3 mg/dl or �25% abovethe patient’s baseline, determined 3–5days following cardiac catheterizationprocedures.

The mean age, baseline creatinine,presence or absence of diabetes, hydra-tion status, type and dose of contrastagent received, and use of specific medi-cations, including acetylcysteine, werenot different between groups. The per-centage of inpatients was greater in theAHG group (74%) than in the NHG(26%), P � 0.049. Ventricular function,as measured by left ventricular end-diastolic pressure, was the same betweengroups (AHG � 17 � 10 mmHg vs.NHG � 13 � 3 mmHg, P � 0.21), andleft ventricular ejection fraction was sig-nificantly lower in the AHG group(AHG � 45 � 13% vs. NHG � 59 �14%, P � 0.023). A total of 38 patientswere studied, including 33 diabetic sub-jects (87%). One-half of the study group(19 patients) was found to have hypergly-cemia at the time of their cardiac catheter-ization procedure. Mean serum glucosewas 217 � 78 mg/dl for the AHG groupvs. 124 � 15 mg/dl for the NHG group,P � 0.001. The incidence of contrast-induced nephropathy for the entire study

Letters


population was 24% (9 of 38). The inci-dence of contrast-induced nephropathyin the AHG group was 42% (8 of 19) andwas significantly greater than that for theNHG group, 5.3% (1 of 19), P � 0.01.

Acute hyperglycemia is a potential in-dependent risk factor for the develop-ment of contrast-induced nephropathy indiabetic patients undergoing cardiaccatheterization procedures. The glucosemolecule has been shown to be a potentialcytotoxin in the context of hyperglycemia(5). Acute hyperglycemia in patients withor without diabetes can detract from clin-ical outcomes in cardiovascular disease(1). The mechanism by which acute hy-perglycemia worsens ischemic myocar-dial injury is currently under study.Conceivably, hyperglycemia may exacer-bate acute renal ischemia associated withadministration of radiographic contrastagents. The observational design of thisstudy limits the relationship betweenacute hyperglycemia and contrast-induced nephropathy to that of a tempo-ral association and does not addresscausality. Confounding variables, such asthe slightly worse left ventricular ejectionfraction in the AHG group, may have con-tributed to the development of contrast-induced nephropathy; however, thehyperglycemia in the AHG group mayhave contributed to poorer ventricularfunction. The relationship between acutehyperglycemia and contrast-induced ne-phropathy reported here will require arandomized controlled clinical trial fordefinitive characterization. This reportsuggests that a temporal association existsbetween acute hyperglycemia and con-trast-induced nephropathy at the time ofcardiac catheterization procedures in dia-betic patients with mild renal dysfunc-tion, and this topic bears further study.

DIANE B. TURCOT, MD1

FRANCIS J. KIERNAN, MD, FACC1

RAYMOND G. MCKAY, MD, FACC1

NEIL J. GREY, MD, FACP2

WILLIAM BODEN, MD, FACC1

GEORGE A. PERDRIZET, MD, PHD, FACS3

From the 1Division of Cardiology, Hartford Hospi-tal, University of Connecticut School of Medicine,Hartford, Connecticut; the 2Division of Endocrinol-ogy, Hartford Hospital, University of ConnecticutSchool of Medicine, Hartford, Connecticut; and the3Division of Trauma, Hartford Hospital, Universityof Connecticut School of Medicine, Hartford, Con-necticut.

Address correspondence to George A. Perdrizet,MD, PhD, FACS, Hartford Hospital, Division of

EMS/Trauma, 80 Seymour St., P.O. Box 5037, Hart-ford, CT 06102-5037. E-mai l : gperdr [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Malmberg K, for the DIGAMI (Diabetes

Mellitus, Insulin Glucose Infusion inAcute Myocardial Infarction) StudyGroup: Prospective randomized study ofintensive insulin treatment on long-termsurvival after acute myocardial infarctionin patients with diabetes mellitus. BMJ314:1512–1515, 1997

2. Weir CJ, Murray GD, Dyker AG, Lees KR:Is hyperglycemia a predictor of poor out-come after acute stroke? BMJ 314:1303–1306, 1997

3. Brezis M, Rosen S: Hypoxia of the renalmedulla: its implication for disease.N Engl J Med 332:647–655, 1995

4. Parfrey PS, Griffiths SM, Barrett BJ, PaulMD, Genge M, Withers J, Farid N, Mc-Manamon PJ: Contrast material-inducedrenal failure in patients with diabetes mel-litus, renal insufficiency, or both: a pro-spective controlled study. N Engl J Med320:143–149, 1989

5. Porte D Jr, Schwartz MW: Diabetes com-plications: why is glucose potentiallytoxic? Science 272:699–700, 1996

Influence of thePolymorphismsTpr64Arg in the �3-Adrenergic ReceptorGene and Pro12Alain the PPAR�2 Geneon MetabolicSyndrome–RelatedPhenotypes in anIndigenousPopulation of theBrazilian Amazon

M etabolic syndrome is a cluster ofrisk factors for type 2 diabetes andcardiovascular disease. Multiple

mechanisms, including genetic factors,may contribute to this condition. TheTrp64Arg variant in the �3-adrenergic re-ceptor has been associated with featuresof the metabolic syndrome (1). A rela-tively common gene variant, Pro12Ala ofthe peroxisome proliferator–activated re-ceptor-2 (PPAR2) has been previously

studied for association with obesity andtype 2 diabetes (2,3).

The Parkataje Indians, from the Bra-zilian Amazon region, remained largelyisolated. Recently, they underwent arapid and intensive process of accultura-tion, with important changes in their life-style. Accompanying these changes, anincreasing prevalence of obesity and otherfeatures of the metabolic syndrome havebeen observed. This study examines therelevance of the Trp64Arg mutation in the�3-adrenergic receptor gene and thePro12Ala mutation in the PPAR2 gene asa susceptibility factor to features of themetabolic syndrome in this population.

Participants were individuals aged�20 years; those with admixture andpregnant women were excluded. Thestudy population comprised 85 (52 menand 33 women) individuals (mean age41 � 14.9 years). The degree of related-ness among the individuals was deter-mined, and 37 nuclear families from sixpedigrees were verified. Polygamy, in-cluding polyandry, occurs in this popula-tion. BMI, waist-to-hip ratio, systolic anddiastolic blood pressures, and serum li-poproteins were studied. Fasting and 2-hblood samples were drawn for glucoseand insulin measurements. Changes inbody weight were analyzed in 80 individ-uals for a 3-year period. Genotypes weredetermined by PCR/restriction fragment–length polymorphism, as previously de-scribed (1,2).

A principal component analyses fromthe correlation matrix of the variablesmeasured was performed. Statistical anal-yses (ANOVA and family-based associa-tion test) were done with the first twoprincipal components because the mea-sured variables are all related to the met-abo l i c syndrome. The pr inc ipa lcomponent, therefore, reflects the vari-ance common to these variables andavoids corrections for multiple indepen-dent tests.

Obesity rates were higher in womenthan in men (27.2 vs. 3.8% at baseline,P � 0.006 and 45.2 vs. 16.3% at 3-yearfollow-up, P � 0.01), and for both sexes,there was an increase in these rates duringthe follow-up period (12.94 vs. 27.5%,P � 0.03). Diabetes was diagnosed in oneindividual, impaired glucose tolerance inanother, and the remaining were classi-fied as normal glucose tolerant accordingto World Health Organization criteria.

Frequencies of the �3-adrenergic re-

Letters


ceptor Arg and the PPAR2 Ala variantswere 0.33 and 0.31, respectively. Thesefrequencies are in the Hardy-Weinbergequilibrium. The �3-adrenergic receptorArg allele frequency (0.33) is much higherthan those reported in other populations,except for Pima Indians (4). Similarly, thePPAR2 Ala allele was more prevalent inthe Parkateje Indians than in the otherpopulations, whose frequency rangesfrom 0.12 among Caucasians to 0.01 inChinese (5). ANOVA (with Welch’s cor-rection) showed that the first principalcomponent was heterogeneous amongthe genotypic classes of the PPAR2 lo-cus; the AlaAla genotype was differentfrom the others (F � 3.51, P � 0.035).The �3-adrenergic receptor locus showedno differences among the genotypes. TheFBAT analyses showed that the PPAR2locus presented a significant segregationdistortion with the recessive model (P �0.032) but not with the additive or dom-inant models.

Among the Parkateje Indians, thePro12Ala variant in the PPAR2 gene, butnot the Trp64Arg variant in the �3-adrenergic receptor, was associated withfeatures of the metabolic syndrome.

JOAO PAULO B. VIEIRA-FILHO, MD, PHD1

ANDRE F. REIS, MD, PHD1

TERESA S. KASAMATSU, BSC1

EDELWEISS F. TAVARES, MD, PHD1

LAERCIO J. FRANCO, MD, PHD2

SERGIO R. MATIOLI, PHD3

REGINA S. MOISES, MD, PHD1

From the 1Division of Endocrinology, Departmentof Medicine, Federal University of Sao Paulo, SaoPaulo, Brazil; the 2Department of Social Medicine,Faculty of Medicine of Ribeirao Preto-University ofSao Paulo, Sao Paulo, Brazil; and the 3Department ofBiology, University of Sao Paulo, Sao Paulo, Brazil.

Address correspondence to Regina S. Moises,MD, PhD, Universidade Federal de Sao Paulo, Es-cola Paulista de Medicina, Disciplina de Endocrino-logia, Rua Botucatu, 740-2o,° andar, 04034-970 SaoPau lo , SP , Braz i l . E -ma i l : rmoi se [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Widen E, Lehto M, Kanninen T, Walston

J, Shuldiner AR, Groop LC: Association ofa polymorphism in the B3-adrenergic-re-ceptor gene with features of the insulinresistance syndrome in Finns. N EnglJ Med 333:348–351, 1995

2. Ek J, Urhammer SA, Sorensen TIA,Andersen T, Auwerx J, Pedersen O: Ho-mozygosity of the Pro12Ala variant of the

peroxisome proliferation-activated recep-tor-2 (PPAR-2): divergent modulatingeffects on body mass index in obese andlean Caucasian men. Diabetologia 42:892–895, 1999

3. Altshuler D, Hirschhorn JN, KlannermarkM, Lindgren CM, Vohl MC, Nemesh J,Lane CR, Schaffner SF, Bolk S, Brewer C,Tuomi T, Gaudet D, Hudson TJ, Daly M,Groop L, Lander ES: The common PPARgamma Pro12Ala polymorphism is asso-ciated with decreased risk of type 2 dia-betes. Nat Genet 26:76–80, 2000

4. Walston J, Silver K, Bogardus C, KnowlerWC, Celli FS, Austin S, Manning B, Stros-berg AD, Stern MP, Raben N, Sorkin JD,Roth J, Shuldiner AR: Time of onset ofnon-insulin-dependent diabetes mellitusand genetic variation in the B3-adrener-gic-receptor gene. N Engl J Med 333:347–347, 1995

5. Yen CJ, Beamer BA, Negri C, Silver K,Brown KA, Yamall DP, Burns DK, Roth J,Shuldiner AR: Molecular scanning of thehuman peroxisome proliferator activatedreceptor gamma gene in diabetic Cauca-sians: identification of a Pro12Ala PPARgamma 2 missense mutation. Biochem Bio-phys Res Commun 241:270–274, 1997

Elevated SerumFerritinConcentrations in aGlucose-ImpairedPopulation and inNormal GlucoseTolerant First-Degree Relatives inFamilial Type 2Diabetic Pedigrees

Two large epidemiological studieshave recently reported a strong as-sociation between elevated serum

ferritin concentration and increased riskfor diabetes (1,2). Moreover, other stud-ies have revealed the relationship amongexcess ferritin, coronary heart disease,and insulin resistance and have thereforerenewed interest in ferritin as a risk factorfor diabetes.

This study further investigates the as-sociation between ferritin metabolismand different status of glucose tolerance,including 121 type 2 diabetic subjects, 86impaired glucose tolerant (IGT) subjects,58 normal glucose tolerant (NGT) first-

degree relatives in type 2 diabetic pedi-grees, and 85 healthy control subjects. Allpatients underwent an oral glucose toler-ance test (OGTT) and insulin release testsafter 8 h of fasting, and blood levels offerritin, HbA1c, glucose, insulin, C-peptide, and lipids were measured. Se-rum ferritin levels were measured withthe radioimmunoassay kit (Beijing NorthInstitute of Biological Technology). Nor-mal ranges for ferritin concentration are12–245 ng/ml for adult men and 5–130 ng/ml for women. We defined ele-vated concentrations of ferritin as �295ng/ml for men and �155 ng/ml forwomen.

Levels of fasting and postprandialplasma glucose in the NGT group wereremarkably higher than in the healthycontrol subjects. Fasting insulin concen-trations in the NGT group were alsohigher than those of the other groups,while postprandial insulin concentrationsincreased significantly when comparedwith healthy control subjects. Ferritinconcentrations were the highest in type 2diabetic subjects, followed by the IGTgroup, the NGT group, and the healthycontrol group (412.88 � 155.58,354.19 � 173.03, 231.31 � 130.32 [P �0.05 compared with healthy control sub-jects], and 164.69 � 110.54 ng/ml, re-spectively). In the type 2 diabetic group,the newly diagnosed patients had higherferritin concentrations than previously di-agnosed (461.72 � 132.41 vs. 354.19 �173.03 ng/ml, P � 0.05).

We also compared concentrations ofserum ferritin in men and women for eachgroup. In general, concentrations of fer-ritin in men were higher than in women(P � 0.05) except for in the healthy con-trol group. In male subjects, ferritin con-centrations of both newly and previouslydiagnosed type 2 diabetic, IGT, NGT, andhealthy control groups showed the sametrend as the whole group (494.30 �142.6, 425.01 � 136.77, 390.07 �125.09, 284.74 � 112.04 [P � 0.001compared with the healthy control sub-jects], and 197.93 � 110.41 ng/ml, re-spectively). However, in female subjects,ferr i t in concentra t ions in newly(425.65 � 137.5 ng/ml) and previously(295.37 � 150.98 ng/ml) diagnosed type2 diabetes and IGT (330.72 � 131.03 ng/ml) were higher than the NGT (174.06 �123.45 ng/ml) and healthy contol(137.28 � 89.63 ng/ml) groups (P �0.001). No significant difference was

Letters


found between female NGT and femalehealthy control subjects. Moreover, innewly diagnosed type 2 diabetes, the con-centrations of ferritin were significantlyhigher than in the previously diagnosedtype 2 diabetic and IGT patients.

Using multiple regression analysis,we found an association between ferritinconcentration and BMI, waist-to-hip ra-tio, systolic blood pressure, diastolicblood pressure. HbA1c, FPG, 2-h plasmapostprandial glucose, triglycerides, andtotal cholesterol were positively related toferritin concentrations, while HDL cho-lesterol levels were inversely related toferritin concentrations.

In recent years, the issue of the poten-tial pathology of serum ferritin in type 2diabetes has gained remarkable interest(3). In this study, we found that serumferritin concentrations were remarkablyincreased in type 2 diabetes, especially innewly diagnosed patients. Subjects withhigher concentrations of ferritin conse-quently had higher HbA1c, glucose, andinsulin concentrations. These results fur-ther proved a positive association be-tween type 2 diabetes and high plasmaferritin concentrations.

The exact mechanism through whichelevated ferritin promotes the develop-ment of type 2 diabetes is unknown.Some investigations argued that abnor-malities in ferritin metabolism might be aprimary cause of type 2 diabetes (4–6). Inour study, ferritin concentration in IGTsubjects, the high-risk population for type2 diabetes, already significantly increasedwhen compared with normal control sub-jects, implying that hyperferritinemia oc-curs before elevation of plasma glucoseconcentrations. This observation was fur-ther substantiated by evidence that NGTfirst-degree relatives in the type 2 diabeticpedigrees had higher ferritin concentra-tions than normal control subjects.

YAN REN, MD, PHD1

HAOMING TIAN, MD1

XIUJUN LI, MD1

JINZHONG LIANG, MD1

GUIZHI ZHAO,2

From the 1Division of Endocrinology, Departmentof Internal Medicine, West China Hospital, SichuanUniversity; and the 2Laboratory of Endocrinologyand Metabolism, West China Hospital, Sichuan Uni-versity, Chengdu, China.

Address correspondence to Haoming Tian, MD,Division of Endocrinology, Department of InternalMedicine, West China Hospital of Sichuan Univer-

sity, 37 Guoxue Lane, Chengdu, Sichuan China610041. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Tuomainen TP, Nyyssonen K, Salonen R,

Tervahauta A, Korpela H, Lakka T,Kaplan GA, Salonen JT: Body iron storesare associated with serum insulin andblood glucose concentration: populationstudy in 1,013 eastern Finnish men. Dia-betes Care 20:426–428, 1997

2. Ford ES, Cogswell ME: Diabetes and serumferritin concentration among U.S. adults.Diabetes Care 22:1978–1983, 1999

3. Salonen JT, Tuomainen TP, Nyyssonen K,Lakka HM, Punnonen K: Relation be-tween iron stores and non-insulin depen-dent diabetes in men: case-control study(Abstract). BMJ 317:727, 1998

4. Moczulski DK, Grzeszczak W, Gawlik B:Role of hemochromatosis C282Y and H63Dmutations in HFE gene in development oftype 2 diabetes and diabetic nephropathy.Diabetes Care 24:1187–1191, 2001

5. Van Lerberghe S, Hermans MP, Dahan K,Buysschaert M: Clinical expression andinsulin sensitivity in type 2 diabetic pa-tients with heterozygous for haemochro-matosis. Diabetes Metab 28:33–38, 2002

6. Salonen JT, Tuomainen TP, Kontula K:Role of C282Y mutation of in haemochro-matosis gene in development of type 2diabetes in healthy men: prospective co-hort study. BMJ 24:1706–1707, 2000

SimultaneousBilateral Facial Palsyin a Diabetic Patient

Unilateral facial paralysis is a rela-tively common condition with anincidence of 20 –25 per 100,000

population. However, simultaneous bilat-eral facial palsy (facial diplegia) is an ex-tremely rare clinical entity and occurs in0.3–2% of facial paralysis patients (1).The annual incidence is approximately 1per 5 million (2).

A 78-year-old diabetic patient pre-sented to the emergency room of our hos-pital with dysarthry and bilateralsymmetrical facial weakness. He was un-able to show his teeth, close his eyelids, ordilate his cheeks. From the neurologic ex-amination, there were no other importantfindings, except for a minor instabilityduring walking. The patient did not referhead injury or febrile viral infection in therecent past. We made the presumptive di-

agnosis of bilateral peripheral facial paral-ysis. Five weeks after his admission to ourhospital, he made a full recovery. We haveto note that glucocorticoids were not ad-ministered to him.

His full blood count, erythrocyte sed-imentation rate, liver function tests, tu-mor markers, thyroid hormones, serumprotein immunoelectrophoresis, serumACE levels, C-reactive protein, and rapidprotein reagent (RPR) were all within nor-mal limits. HbA1c was 7.0%, and the au-toantibody screen was negative. Purifiedprotein derivative was 5 mm. Serologicaltests for varied infectious agents, includ-ing herpes simplex virus (HSV)-I and -II,Varicella-Zoster virus (VZV), Epstein-Barr, Coxsackie, HIV-I and -II, cytomeg-alovirus (CMV), and hepatitis B viruses,as well as Mycoplasma and Borrelia Burg-dorferi, were all negative.

Lumbar puncture revealed a normalpressure. Glucose, protein, and whiteblood count of the cerebrospinal fluid(CSF) were all within normal limits. Fur-thermore, stains and cultures for micro-organisms were negative, as were tests forviruses (HSV and HSV-II, VZV, amdCMV), Borrelia Burgdorferi, and syphilis(venereal disease reaction level [VDRL]test).

Magnetic resonance imaging (MRI) ofthe brain and computed tomography(CT) scans of the head, thorax, and abdo-men were all normal.

Facial diplegia may have diverse eti-ologies and may prove to be a diagnosticdilemma. The most common causes arebilateral Bell’s palsy, Lyme disease, Guil-lain-Barre syndrome, sarcoidosis, Moebi-ous syndrome, leukemia, viral infections,syphilis, basilar skull fractures, and pon-tine gliomas.

The most common infectious cause offacial diplegia is Lyme disease, caused byBorrelia Burgdorferi (3). Regarding thecase presented, the IgG antibodies againstthis agent in serum, as well as in CSF, werenegative. Other rare infectious causes in-clude syphilis and Mycoplasma (4). How-ever, VDRL tests in CSF and RPR in serumwere negative, while antibody titer againstMycoplasma was negative.

Guillan-Barre syndrome is thought tobe a postinfectious inflammatory polyra-diculoneuritis. Up to 50% of the fatalcases have bilateral facial paralysis (5).The diagnosis is made on lumbar punc-ture (with a typically elevated protein inthe absence of a raised number of cells)

Letters


and peripheral areflexia. Our patient hadneither peripheral muscle weakness norareflexia, and the CSF examination wasnormal.

Basilar skull fractures and pontine gli-omas were excluded by means of bothbrain CT and MRI. Because there was nohilar adenopathy on chest CT and be-cause serum ACE levels were normal, sar-coidosis was rejected.

Bilateral Bell’s palsy does not seem tobe a plausible diagnosis because our pa-tient had neither a preceding viral infec-tion nor the characteristic symptoms ofthis condition (facial numbness or pain,change in taste, numbness of the tongue,hyperacusis, etc.).

Diabetes has previously been associ-ated with facial diplegia (4,6,7). Accord-ing to Adour, Wingerd, and Doty (7),diabetes was present in 28.4% of 67 pa-tients with recurrent or bilateral facialpalsy. A plausible explanation could bethat diabetic patients are more prone tonerve degeneration. In another series of43 patients with bilateral seventh nervepalsy, there was one case associated withdiabetes (4). Thus, having excluded allthe other possible causes of this disorderafter extensive evaluation, we could as-sume that the most likely cause of facialdiplegia in the case presented is diabetes.

In conclusion, bilateral facial paraly-sis may be due to a life-threatening condi-tion and, therefore, the practitioner shouldbe aware of the diagnostic possibilitiesthat cause this extremely rare condition. Areview of the literature reveals that diabe-tes is associated with facial diplegia andshould always be included in the differ-ential diagnosis of this condition.

ALEXANDER KAMARATOS, MD, PHD1

STELIOS KOKKORIS, MD1

JOHN PROTOPSALTIS, MD, PHD1

DIMITRIOS AGORGIANITIS, MD2

HARIS KOUMPOULIS, MD2

JOHN LENTZAS, MD1

ANDREAS MELIDONIS, MD, PHD3

GREGORY GIANNOULIS, MD, PHD1

From the 1Second Department of Internal Medicine,Tzanio Hospital, Piraeus, Greece; the 2NeurologyDepartment, Tzanio Hospital, Piraeus, Greece; andthe 3Diabetologic Center, Tzanio Hospital, Piraeus,Greece.

Address correspondence to Alexander Kamara-tos, MD, PhD, Tzanio General Hospital of Piraeus,Second Department of Internal Medicine,Tzani andAfentouli 1, Piraeus 185 36, Greece. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Stahl N, Ferit T: Recurrent bilateral pe-

ripheral facial palsy. J Laryngol Otol 103:117–119, 1989

2. George MK, Pahor AL: A cause for bilat-eral facial palsy. Ear Nose Throat J 70:492–3, 1991

3. Clark JR, Carlson RD, Pachner AR: Facialparalysis in Lyme disease. Laryngoscope95:1341–1345, 1985

4. Keane JR: Bilateral seventh nerve palsy:analysis of 43 cases and review of the lit-erature. Neurology 44:1198–1202, 1994

5. Arias G, Nogues J, Manos M, Amilibia E,Dicenta M: Bilateral facial nerve palsy:four case reports. ORL J OtorhinolaryngolRelat Spec 60:227–229, 1998

6. Hattori T, Schlagenhauff RE: Bilateral fa-cial palsy: occurrence with diabetes mel-litus. N Y State J Med 77:1492–1494, 1977

7. Adour K, Wingerd J, Doty HE: Prevalenceof concurrent diabetes mellitus and idio-pathic facial paralysis (Bell’s palsy). Dia-betes 24:449–451, 1975

Reduced Fear ofHypoglycemia inSuccessful IsletTransplantation

The recent dramatic improvement inclinical outcomes in islet transplan-tation in type 1 diabetes with the Ed-

monton Protocol has led to considerableexcitement in the field of diabetes (1,2).The unprecedented 1-year success ratesprovide considerable evidence of the clin-ical effectiveness of the procedure (2,3).However, the benefits of freeing or reduc-ing insulin requirements for these pa-tients must be weighed against the risks ofthe procedure itself, as well as the life-long immunosuppression. Before makingthis treatment available to a larger num-ber of people with type 1 diabetes, mea-sures of quality of care and of clinicaleffectiveness must be incorporated tofully evaluate the benefit of this treatment.

Episodes of severe hypoglycemia, acommon occurrence in patients with la-bile type 1 diabetes and hypoglycemia un-awareness, result in considerable fear andanxiety (4,5). When these concerns be-come an overwhelming burden for pa-tients with type 1 diabetes, islettransplantation with the Edmonton Pro-tocol is a potential solution (1–3). To de-termine the potential impact of islet

transplantation on self-reported health-related quality of life (HRQL) outcomes,we compared islet-transplanted patientswith pretransplant patients on measuresof fear of hypoglycemia and anxiety.

Patients were asked to self-complete abattery of measures, including the Hypo-glycemia Fear Survey (HFS) (4,5) and theHealth Utilities Index Mark two (HUI2)(6). The HFS contains 23 questions thatassess patients’ concerns and worriesabout hypoglycemia and the behaviors inwhich patients may engage to avoid lowblood glucose. The emotion attribute ofthe HUI2 can be used as an index of anx-iety (6). Our standard protocol for admin-istration of HRQL questionnaires occursat baseline (pretransplant); midtransplant(i.e., between the first and second); 1, 3,6, and 12 months posttransplant; and an-nually thereafter. Because islet-transplantedpatients may have completed multiplesurveys during follow-up, we initiallyused only the last available HRQL assess-ment. Surveys were completed by 81 (46pretransplant and 35 islet-transplanted)patients. Among the islet-transplantedpatients, questionnaires were completeda median of 11.9 months (range 1–36)after transplant. Scores between the twogroups of patients were compared usingnonparametric statistical tests.

Fear of hypoglycemia was signifi-cantly lower in islet-transplanted (median5.0) compared with pretransplant (me-dian 47.0) patients for the HFS total score(P � 0.001). The magnitude of the differ-ence in HUI2 emotion scores betweenpretransplant and islet-transplanted pa-tients would be considered clinically im-portant (6) (1.00 vs. 0.86, respectively),although the difference was not staticallysignificant (P � 0.96). Among all islet-transplanted patients, the small number(n � 3) without C-peptide secretion andrequiring exogenous insulin had substan-tially more fear about hypoglycemia (P �0.041) and reported more anxiety on theHUI2 emotion attribute (P � 0.023) thanislet-transplanted patients with successfultransplants.

Because anxiety pre- and posttrans-plant could be related to the procedureitself, we also compared HFS and HUI2emotion scores between pretransplantand islet-transplanted patients in the im-mediate posttransplant period; for thesecomparisons, we used all available HRQLassessments at 1 and 3 months posttrans-plant. We found that fear of hypoglyce-

Letters


mia was lower, with a median HFS totalof 30.0 for islet-transplanted patients(n � 20) at 1 month and 6.5 (n � 18) at 3months, both of which were significantlylower (P � 0.01) than pretransplant.Conversely, the HUI2 emotion scorewas not significantly different frompretransplant at either 1 or 3 monthsposttransplant.

These initial evaluations of self-reported HRQL outcomes of in islet trans-plant recipients demonstrate that clinicalsuccess is associated with substantial re-duction in emotional burden through re-duced fear of hypoglycemia. Generalanxiety in islet-transplanted patients is re-duced overall, which seems to be relatedto the freedom from requirement of exog-enous insulin rather than to recoveringfrom the transplant procedure itself. Al-though the interpretation of our initialdata is interesting and informative, sev-eral limitations and questions remain.These initial data were collected cross-sectionally and on a relatively small butgrowing sample of islet-transplanted pa-tients; even with the small sample sizes,the observed differences were statisticallysignificant. Longitudinal assessments tomeasure within-person change over timeare required to fully assess the impact onHRQL.

JEFFREY A. JOHNSON, PHD1,2

MARIA KOTOVYCH, MA2

EDMOND A. RYAN, MD, FRCPC3,4

A.M. JAMES SHAPIRO, MBBS, BMEDSCI,

PHD, FRCSC4,5

From the 1Department of Public Health Sciences,University of Alberta, Edmonton, Alberta, Canada;the 2Institute of Health Economics, Edmonton, Al-berta, Canada; the 3Division of Endocrinology andMetabolism, Department of Medicine, University ofAlberta, Edmonton, Alberta, Canada; the 4ClinicalIslet Transplant Program, University of Alberta, Ed-monton, Alberta, Canada; and the 5Department ofSurgery, University of Alberta, Edmonton, Alberta,Canada.

Address correspondence to Jeffrey A. Johnson,PhD, Institute of Health Economics, 1200-10405Jasper Ave., Edmonton, Alberta, Canada T5J 3N4.E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Shapiro AMJ, Lakey JRT, Ryan EA, Kor-

butt GS, Toth E, Warnock GL, KnetemanNM, Rajotte RV: Islet transplantation inseven patients with type 1 diabetes melli-tus using a glucocorticoid-free immuno-suppressive regimen. N Engl J Med 343:

230–238, 20002. Ryan EA, Lakey JR, Rajotte RV, Korbutt

GS, Kin T, Imes S, Rabinovitch A, ElliottJF, Bigam D, Kneteman NM, WarnockGL, Larsen I, Shapiro AM: Clinical out-comes and insulin secretion after islettransplantation with the Edmonton Pro-tocol. Diabetes 50:710–719, 2001

3. Ryan EA, Lakey JRT, Paty BW, Imes S,Korbutt GS, Kneteman NM, Bigam D, Ra-jotte EV, Shapiro AMJ: Successful islettransplantation: continued insulin reserveprovides long-term glycemic control. Di-abetes 51:2148–2157, 2002

4. Cox D, Irvine A, Gonder-Frederick L,Nowacek G, Butterfield J: Fear of hypo-glycemia: quantification, validation, andutilization. Diabetes Care 10:617–621,1987

5. Irvine A, Cox D, Gonder-Frederick L:Thefear of hypoglycemia scale. In Hand-book of Psychology and Diabetes. Bradley C,Ed. Amsterdam, Hardwood Academic,1994, p. 133–155

6. Feeny DH, Torrance GW, Furlong WJ:Health utilities index. In Quality of Life andPharmacoeconomics in Clinical Trials. 2nded. Spilker B, Ed. Philadelphia, Lippin-cott-Raven, 1996, p. 239–251

A Case ofLipoatrophy WithLispro InsulinWithout InsulinPump Therapy

L ocalized lipoatrophy occurring inthe subcutaneous insulin injectionarea in diabetic patients was a phe-

nomenon practically forgotten after theintroduction of human insulin in medicalpractice. In recent years, there have beenvery few publications in relation to thismatter.

Three cases of patients with type 1diabetes who presented with subcutane-ous localized lipoatrophy areas and whowere in treatment with Lispro insulinwere recently reported (1,2). The threepatients used a continuous subcutaneousinsulin infusion (CSII) system; therefore,the authors posed the doubt of whethersuch an administration system locallyplayed a determinant role in the occurrenceof subcutaneous localized lipoatrophy.

We present a case of localized lipoat-rophy associated with treatment with

Lispro insulin administered in a multipledose regimen that disregards the role ofCSII as a necessary factor for its genesis.

Our patient is a 35-year-old womandiagnosed in January of 1992 at 22 yearsof age. From the start, she was treatedwith recombinant DNA human insulin(Humulin Regular and Humulin NPH;Lilly) in a regimen of three daily doses.She always exhibited a good degree ofmetabolic control, with HbA1c between 6and 7%. Seven years after diagnosis, shebegan to exhibit episodes of hypoglyce-mia not perceived with accompanyingneuroglycopenia, which persisted in spiteof several changes of her prior insulin reg-imen. For that reason, in November of2000 it was decided that she wouldchange to LisPro insulin administered be-fore breakfast, lunch, snack, and dinner,and NPH insulin administered beforebreakfast and dinner. With the new regi-men, metabolic control remained similarto the previous control and the episodesof neuroglycopenia persisted. Anti-insulin antibody (IAA) levels were mea-sured and were high (49.6%, referencevalue �8.5%). In October of 2002, 23months after beginning with LisPro, thepatient consulted the physician becauseshe had a circumscribed localized lipoat-rophy area of 3 cm in diameter on theanterior aspect of the right thigh, one ofher normal injection areas. Six monthslater, a period in which injection in saidarea was avoided, the lesion remained un-changed, but an incipient localized lipoat-rophy area could be observed in the samearea of the contralateral thigh. For thisreason, it was decided to change fromLispro to Aspart insulin.

Six months after said change of insu-lin, which was when this letter was sent,neither progression nor improvement ofthe localized lipoatrophy lesions had beenobserved. IAA levels were 30.5%, slightlylower than the previous levels.

The development of localized lipoat-rophy in the insulin injection area is apractically exclusive complication of type1 diabetic patients, although cases havebeen reported in patients with type 2 di-abetes (3). From the etiopathogenic pointof view, it is considered an immunologicalphenomenon. Although this has not beensufficiently clarified, a strong associationbetween the lesions and high IAA plasmalevels and the presence of insulin and im-munoglobulin G deposits in subcutane-ous tissue of the affected areas (4) have

Letters


been reported. The consequences of thisimmunological activation are the local in-hibition of adipocyte differentiation,probably mediated by the local hyperpro-duction of tumor necrosis factor-� (5).

In affected patients, the pharmacoki-netic variations of insulin due to high IAAlevels and the erratic absorption of thedrug when it is injected in the areas af-fected with localized lipoatrophy imply aglycemic variability making it very diffi-cult to achieve suitable metabolic control.

Although the immunogenic profile ofthe patients treated with Lispro insulinand recombinant human insulin are com-parable (6), the recent occurrence of de-scriptions of localized lipoatrophyassociated to this analogue can decreasethe therapeutic alternatives of this com-plication when, especially in recent years,in the few published cases of human in-sulin–induced localized lipoatrophy theattempted solution to the problem was tochange to Lispro. It is possible that the useof CSII may favor the occurrence of local-ized lipoatrophy but, as can be seen in thecase we present, it is not a factor sine quanon for its development. Curiously, severecases of human insulin–induced localizedlipoatrophy have been previously re-ported that responded satisfactorily afterthe introduction of CSII (7).

The association of localized lipoatro-phy and Lispro insulin without the con-course of CSII has not been reportedpreviously. We therefore believe it is in-teresting to disclose our case and toencourage publishing for other diabetolo-gists who have observed similar cases forthe purpose of clarifying its pathogenesisand therapeutic approach.

ALFONSO ARRANZ, MD

VICTOR ANDIA, MD

ANTONIO LOPEZ-GUZMAN, MD

From the Endocrinology Unit, Hospital Nuestra Se-nora de Sonsoles, Avila, Spain.

Address correspondence to Dr. Alfonso Arranz,Endocrinology Unit, Ntra. Sra. De Sonsoles, Car-retera de Madrid, km 109, Avila, Spain 05071. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Griffin ME, Feder A, Tamborlane WV: Li-

poatrophy associated with lispro insulinin insulin pump therapy (Letter). DiabetesCare 24:174, 2001

2. Ampudia-Blasco FJ, Hasbum B, CarmenaR: A new case of lipoatrophy with lispro

insulin in insulin pump therapy (Letter).Diabetes Care 26:953–954, 2003

3. Mu L, Goldman JM: Human recombinantDNA insulin-induced lipoatrophy in pa-tient with type 2 diabetes mellitus. EndocrPract 6:151–152, 2000

4. Reeves WG, Allen BR, Tattersall RB: Insu-lin induced lipoatrophy evidence for animmune pathogenesis. BMJ 1:1500–1506, 1980

5. Atlan-Gepner C, Bondgrand P, FarnarierC, Xerri L, Choux R, Gauthier JF, Brue T,Vague P, Grob JJ, Vialettes B: Insulin-in-duced lipoatrophy in type 1 diabetes: apossible tumor necrosis factor-�–medi-ated dedifferentiation of adipocytes. Dia-betes Care 19:1283–1285, 1996

6. Fineberg SE, Huang J, Brunelle R, GulliyaKS, Anderson JH: Effect of long-term ex-posure to insulin Lispro on the inductionof antibody response in patients with type1 or type 2 diabetes. Diabetes Care 26:89–96, 2003

7. Chantelau E, Reuter M, Schotes S, StarkeAA: Severe lipoatrophy with human insu-lin: successfully treated by CSII (Letter).Diabet Med 10:580–581, 1993

Diabetes In ANonpancreatectomizedChild WithNesidioblastosis

P ersistent hyperinsulinemic hypogly-cemia of infancy (PHHI) (OnlineMendelian Inheritance in Man

[OMIM] 256450), formerly known as ne-sidioblastosis, is a glucose metabolismdisorder characterized by profound hy-poglycemia and inappropriate secretionof insulin (1). Affected children run therisk of severe neurological damage unlessimmediate and adequate steps are taken(2). Treatment with diazoxide and/or so-matostatin analogue is the first line oftherapy. However, it not always effective,especially in familial cases, which may ne-cessitate an alternative intervention suchas pancreatectomy (3).

Several studies have suggested thatpartial pancreatectomy endangers futureislet cell function (4,5). The incidence ofdiabetes increases with age and correlateswith the extent of surgical resection (6,7).However, there was no report of occur-rence of overt diabetes in medicallytreated patients (8). In this report, we de-

scribed an adolescent female with neona-tal nesidioblastosis who developeddiabetes after medical treatment with dia-zoxide/octreotide. To our knowledge, thisis the first nesidioblasosis case subjectwho developed diabetes following medi-cal therapy.

A 14-year-old Saudi female presentedwith severe persistent hypoglycemiaduring the first few days of life. She wasdiagnosed with hyperinsulinemic hypo-glycemia of infancy based on her intrave-nous glucose requirement of �14 mg �kg�1 � min�1, an insulin-to-glucose ratioof �0.3 (her insulin level was 98 �U/ml ata serum glucose of 32 mg/dl), negativeurinary ketones, a 30-min glucose incre-ment of �30 mg/dl in response to intra-muscular 0.5 mg glucagon, and normalblood spot acylcarnitine profile deter-mined by tandem mass spectrometry. Shealso had a normal growth hormone levelof �20 mU/l and a normal cortisol level of�500 nmol/l during hypoglycemia. Shewas treated initially with frequent feedingsupplemented with complex carbohy-drates (polycose/corn starch) and thenstarted on diazoxide 15 mg � kg�1 � day�1

divided three times a day, which kept hereuglycemic with occasional hypoglyce-mic episodes. In 1992, octreotide was firstintroduced in our hospital as an adjunc-tive therapy to diazoxide. She was startedon 25 �g � kg�1 � day�1 of subcutaneousoctreotide divided four times a day. Sheresponded to medical treatment with nohypoglycemic episodes. She was contin-ued on diazoxide and octreatide until theage of 10 years, when she became eugly-cemic and these two medications werestopped. At the age of 14, she developedhyperglycemia associated with weightgain. Her blood glucose ranged from 200to 300 mg/dl, and her weight was 75 kg(�95%). She had an insulin level of 10�U/ml and C-peptide level of 0.16 nmol/lat a serum glucose level of 350 mg/dl. An-tiglutamic acid decarboxylase, insulin,and islet cell antibodies were negative.She responded to metformin 250 mgtwice a day with a serum mean glucoselevel of 109 mg/dl and HbA1c of 7.5%.

The long-term outcome of PHHI isnot well documented. Previous reportssuggested that subtotal or near total pan-createctomy in infants will endanger thefuture islet function (4–8). Long-termfollow-up studies in medically treated pa-tients with diazoxide or octreotideshowed that some of these patients re-

Letters


sponded to medical therapy and becameeuglycemic (9–11). Some patients wereweaned off medical therapy and contin-ued to be euglycemic; however, none ofthem became hyperglycemic or diabetic.Leibowitz et al. (8) followed six conserva-tively treated patients with PHHI. Intra-venous glucose tolerance was performedin all patients and showed a blunted in-sulin response in two with no overt hy-perglycemia. Histologically, Kassem et al.(12) showed that �-cell proliferation andapoptosis, which normally occurrs in thenormal developing human pancreas, alsooccurs in the PHHI pancreas with a higherfrequency of apoptosis. They suggestedthat this phenomenon will result in aslow, progressive, and complete loss of�-cell mass. This histological report andthe development of diabetes in our non-pancreatectomized PHHI patient maysuggest that patients with PHHI will nat-urally develop diabetes whether theywere treated medically or surgically oreven if they are left untreated. This hy-pothesis was further raised when the nat-ural history of this disease was discussedin knockout mouse models. Transgenicmice engineered to express a dominant-negative form of Kir6.2 or mice with ATP-sensitive K� channel deficiency devel-oped hyperinsulinemic hypoglycemiafollowed by hypoinsulinemic hyperglyce-mia. Diabetes in these transgenic micewas thought to be due to sustained unreg-ulated Ca influx and premature �-cell ap-optosis (burn-out phenomenon) (13,14).Seino et al. (15,16) reported another pos-sible predisposing factor to hyperglyce-mia in PHHI patients. They showed thathyperglycemia in Kir6.2 knockout micewas more evident with age and increasingweight. They suggested that the Kir6.2knockout mouse provides a model of type2 diabetes, and that both the genetic de-fect in glucose-induced insulin secretionand the acquired insulin resistance due toenvironmental factors are necessary to de-velop diabetes in the Kir6.2 knockoutmouse.

We hypothesized that diabetes wasinduced by weight gain and obesity in ourpatient. She responded to metformin,which may suggest that her diabetes isdue to insulin resistance induced by bothweight gain and insulin insufficiency.Simple type 2 diabetes is still a possibility,although there was no history of diabetesin the family. This patient could be thehuman example of the Kir6.2 knockout

mouse model. We recommend, based onthis human clinical evidence, weight con-trol in aged PHHI patients to decease theincidence of diabetes.

BASSAM S. BIN-ABBAS, MD

ABDULLAH A. AL-ASHWAL, MD

From the Department of Pediatrics, King Faisal Spe-cialist Hospital and Research Center, Riyadh, SaudiArabia.

Address correspondence to Bassam Saleh Bin-Abbas, MD Consultant, Section of Pediatric Endo-crinology Department of Pediatrics, MBC 58 KingFaisal Specialist Hospital and Research Center, P.O.Box 3354 Riyadh 11211 Saudi Arabia. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Aynsley-Green A: Nesidioblastosis of the

pancreas in infancy. Dev Med Child Neurol23:372–379, 1981

2. Schwitzgebel VM, Gitelman SE: Neonatalhyperinsulinism. Clin Perinatol 25:1015–1038, 1998

3. Shilyansky J, Fisher S, Cutz E, Perlman K,Filler RM: Is 95% pancreatectomy theprocedure of choice for the treatment ofpersistent hyperinsulinemic hypoglyce-mia of the neonate? J Pediatr Surg 32:342–346, 1997

4. Dunger DB, Burns C, Ghale GK, MullerDP, Spitz L, Grant DB: Pancreatic exo-crine and endocrine function after subto-tal pancreatectomy for nesidioblastosis.J Pediatr Surg 23:112–115, 1988

5. Mahachoklertwattana P, Suprasongsin C,Teeraratkul S, Preeyasombat C: Persistenthyperinsulinemic hypoglycemia of infan-cy: long-term outcome following subtotalpancreatectomy. J Pediatr EndocrinolMetabol 13:37–44, 2000

6. De Lonlay- Debeney P, Poggi-Travert F,Fournet JC, Sempoux C, Vici CD,Brunelle F, Touati G, Rahier J, Junien C,Nihoul-Fekete C, Robert JJ, SaudubrayJM: Clinical features of 52 neonates withhyperinsulism. N Engl J Med 340:1169–1175, 1999

7. Dacou-Voutetakis C, Psychou F, Maniati-Christidis M: Persistent hyperinsulinemichypoglycemia of infancy: long term re-sults. J Pediatr Endocrinol Metabol 11:131–141, 1998

8. Leibowitz G, Glaser B, Higazi AA,Salameh M, Cerasi E, Landau H: Hyper-insulinemic hypoglycemia of infancy (ne-sidioblastosis) in clinical remission: highincidence of diabetes mellitus and persis-tent beta-cell dysfunction at long term fol-low-up. J Endocrinol Metabol 80:386–392,1995

9. Tuoati G, Poggi-Travert F, Ogier deBaulny H, Rahier J, Brunelle F, Nihoul-

Fekete C, Czernichow P, Saudubray JM:Long-term treatment of persistent hyper-insulinemic hypoglycemia of infancy withdiazoxide: a retrospective review of 77cases and analysis of efficacy-predictingcriteria. Eur J Pediatr 157:628–633, 1998

10. Glaser B, Landaw H: Long-term treatmentwith somatostatin analogue SMS 201–995: alternative to pancreatectomy in per-sistent hyperinsulinemic hypoglycemia ofinfancy. Digestion 45:27–35, 1990

11. Darendeliler F, Bundak R, Bas F, Saka N,Gunoz H: Long term diazoxide treatmentin persistent hyperinsulinemic hypogly-cemia of infancy: a patient report. J PediatrEndocrinolo Metabol 10:79–81, 1997

12. Kassem SA, Ariel I, Thornton PS, Scheim-berg I, Glaser B: �-Cell proliferation andapoptosis in the developing normal hu-man pancreas and in the hyperinsulin-ism of infancy. Diabetes 49:1325–1333,2000

13. Miki T, Tashiro F, Iwanaga T, NagashimaK, Yoshitomi H, Aihara H, Nitta Y, GonoiT, Inagaki N, Miyazaki JI, Seino S: Abnor-malities of pancreatic islets by targeted ex-pression of a dominant-negative KATPchannel. Proc Natl Acad Sci U S A94:11969–11973, 1997

14. Miki T, Nagashima K, Tashiro F, KotakeK, Yoshitomi H, Tamamoto A, Gonoi T,Iwanaga T, Miyazaki JI, Seino S: Defect ininsulin secretion and enhanced insulinaction in KATP deficient mice. Proc NatlAcad Sci U S A 95:10402–10406, 1998

15. Seino S, Iwanaga T, Nagashima K, Miki T:Diverse roles of KATP channels learnedfrom Kir6.2 genetically engineered mice.Diabetes 49:311–318, 2000

16. Winarto A, Miki T, Seino S, Iwanaga T:Morphological changes in pancreatic is-lets of KATP channel-deficient mice: theinvolvement of KATP channels in the sur-vival of insulin cells and the maintenanceof islet architecture. Arch Histol Cytol 64:59–67, 2001

Oral GlucoseTolerance TestEvaluation WithForearm andFingertip GlucoseMeasurements inPregnant Women

I t is known that glucose levels in capil-lary blood in the fingertip after a liquidglucose load are constantly higher

when compared with venous blood mea-surements (1). Recently alternative sites

Letters


for capillary blood drawing (e.g., fore-arm) have been proposed (2) that are lesspainful compared with fingertip. Datahave shown that there was no significantdifference between the capillary blooddrawn from forearm and fingertip in dia-betic patients with glucose values in awide range (3). Nevertheless, some datahave shown that glucose results from al-ternative sites and fingertip were notidentical. This difference was more pro-nounced when there was a rapid increaseor decrease of blood glucose values (4). Itseemed that significant differences ap-peared when glucose values declined at amean rate �2 mg � dl�1 � min�1 (5), butnot at a lower rate (6). All the above-mentioned reports compared capillaryblood drawn from either the forearm orfingertip, but so far, it appears that nodirect comparison has been made be-tween venous plasma blood and capillaryforearm blood.

Thus, the purpose of this investiga-tion is to study the pattern of capillaryforearm blood and that of capillary finger-tip blood glucose using the same glucom-eter (FreeStyle; Therasense) and tocompare both with venous blood labora-tory measurements during a 100-g oralglucose tolerance test (OGTT) in preg-nant women. A total of 47 pregnantwomen (age 31 � 3 years, BMI 24 � 3kg/m2, and gestational age 24–28 weeks)underwent a 100-g OGTT. Half of thesewomen (n � 23) had simultaneous glu-cose samples drawn from the forearm af-ter rubbing (7) using FreeStyle in 0�, 60�,120�, and 180�, whereas the other half(n � 24) underwent the same procedurewith blood drawn from the fingertip. Thetwo groups were matched for age, BMI,and gestational age. Glucose difference inpercentage (GDP) was calculated for bothgroups separately. Mean GDP betweenfinger glucose and venous glucose sam-ples was significantly higher at 60�(14.6 � 20.4%), 120� (25.2 � 34.7%),and 180� (26.4 � 26.7%) than at 0�(�3.1 � 14.1%) (P � 0.01). Mean GDPbetween forearm glucose and venous glu-cose samples was significantly higher at120� (16.3 � 21.5%) and 180� (16.3 �21.5%) than at 0� (�2.5 � 16.3%) (P �0.01). On the contrary, mean GDP at 60�(6.7 � 20.9%) was not found signifi-cantly different.

These findings confirmed the alreadyreported observation that up to 3 h after aliquid glucose load, capillary finger glu-

cose levels are constantly higher (15–26%) than venous glucose levels. On thecontrary, forearm glucose levels werecloser to venous plasma glucose levels:There was no significant difference be-tween them after 1 h, whereas a signifi-cant increase of 16% appeared at 2 and3 h. These findings are in accordance withthe concept of slower glucose kinetics atthe forearm than the fingertip due tolesser arteriovenous anastomoses (4). Tobe sure, this physiological differenceneeds to be taken into consideration inthe detection of hypoglycemia in diabeticpatients. However, it is precisely thisphysiological difference that supports thesuggestion that capillary forearm glucosemeasurements using a portable glucosemeter may be useful for the 50-g chal-lenge test for gestational diabetes screen-ing in an outpatient environment.

CHARALAMPOS STAVRIANOS, MD

ELENI ANASTASIOU, MD

From the First Endocrine Section, Diabetes Center,Alexandra Hospital, Athens, Greece.

Address correspondence to Eleni Anastasiou,MD, Alexandra Hospital, First Endocrine Section,Diabetes Center, 80 V. Sofias Ave., Athens 11528,Greece. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Kuwa K, Nakayama T, Hoshino T, Tomi-

naga M: Relationships of glucose concen-trations in capillary whole blood, venouswhole blood and venous plasma. ClinChim Acta 307:187–192, 2001

2. Ellison JM, Stegmann JM, Colner SL, Mi-chael RH, Sharma MK, Ervin KR, HorwitzDL: Rapid changes in postprandial bloodglucose produce concentration differ-ences at finger, forearm, and thigh sam-pling sites. Diabetes Care 25:961–964,2002

3. Lock JP, Szuts EZ, Malomo KJ, Anagnos-topoulos A: Whole-blood glucose testingat alternate sites: glucose values and he-matocrit of capillary blood drawn fromfingertip and forearm. Diabetes Care 25:337–341, 2002

4. Jungheim K, Koschinsky T: Glucose mon-itoring at the arm: risky delays of hypo-glycemia and hyperglycemia detection.Diabetes Care 25:956–960, 2002

5. Jungheim K, Koschinsky T: Response tothe letter by Pfutzner and Forst (Letter).Diabetes Care 25:639–640, 2002

6. Pfutzner A, Forst T: Response to Jung-heim and Koschinsky (Letter). DiabetesCare 25:638–639, 2002

7. McGarraugh G: Response to Jungheim

and Koschinsky: glucose monitoring atthe arm (Letter). Diabetes Care 24:1304–1306, 2001

Silent HypoglycemiaPresenting AsDysesthesias

H ypoglycemia is not often in the dif-ferential diagnosis for dysesthesiasbut should be considered when in-

volved in the care of diabetic patients.Such symptoms may herald silent hypo-glycemia and resultant nerve injury, as il-lustrated in the following case.

A 26-year-old female with type 1 di-abetes presented with a 2-month historyof numbness and tingling in her handsand feet upon waking in the morning.Symptoms began when her treatment wasaltered from NPH 50 units q A.M. to NPH35 and Regular 3 q A.M. and NPH 8 andRegular 5 at dinner. The patient moni-tored her glucose more than four timeseach day and reported three to four glu-cose values a week that were �60 mg/dlwithout symptoms. Her morning glucoselevels averaged 60 mg/dl. The symptomswere more pronounced in her hands thanfeet and resolved within minutes. Onexam, she showed no objective sensoryloss, possessed good muscle tone, bulk,and strength, had intact reflexes (2�) bi-laterally, and had no focal neurologicalsigns. HbA1c was 6.8%.

Symptoms were attributed to periph-eral neuropathy secondary to hypoglyce-mia. Her insulin regimen was adjusted toNPH 35 and Regular 3 q A.M., NPH 4 qHS, and Regular 5 before dinner for glu-cose �200 mg/dl. One month later, shereported the disappearance of the symp-toms and a reduction in the frequency ofvalues �60 mg/dl to once a week.

Hypoglycemia has been proposed toinduce nerve injury by several mecha-nisms. Lack of substrate leads to a reduc-tion in axonal transport, causing anaccumulation of intraneural metabolitesand neuronal injury (1). Hypoglycemiacan induce a reduction in blood flow,leading to neural hypoxia (2–4). Thesemechanisms may all play a role in nerveinjury; disturbance in neural blood flowmay be the initial manifestation of hypo-

Letters


glycemia, while prolonged hypoglycemiamay induce axonal damage (2).

Peripheral neural injury has been re-ported in patients with hypoglycemiadue to insulinomas (5). These patientsdisplayed paresthesias and/or musclewasting and weakness. After tumor resec-tion, patients showed resolution of sen-sory symptoms, while muscle wastingpersisted.

We propose that practitioners con-sider undetected hypoglycemia as a pos-sible cause of paresthesias in diabeticsubjects. Frequent episodes of hypoglyce-mia can hinder patients’ efforts to achievenormoglycemia. Early measures takento reduce such episodes will promotenormoglycemia.

NOLAWIT TESFAYE, BS

ELIZABETH R. SEAQUIST, MD

From the Division of Endocrinology and Diabetes,Department of Medicine, University of MinnesotaMedical School, Minneapolis, Minnesota.

Address correspondence to Elizabeth R. Seaquist,MD, MMC 101, 420 Delaware St. SE, University ofMinnesota, Minneapolis, MN 55455. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Sidenius P, Jakobsen J: Anterograde fast

component of axonal transport during in-sulin-induced hypoglycemia in nondia-betic and diabetic rats. Diabetes 36:853–858, 1987

2. Ohshima J, Nukada H: Hypoglycaemicneuropathy: microvascular changes dueto recurrent hypoglycaemic episodes inrat sciatic nerve. Brain Res 947:84–89,2002

3. Kihara M, Zollman PJ, Smithson IL, La-gerlund TD, Low PA: Hypoxic effect ofexogenous insulin on normal and diabeticperipheral nerve. Am J Physiol 266:E980–E985, 1994

4. Hilsted J, Bonde-Petersen F, NorgaardMB, Greniman M, Christensen NJ, Parv-ing HH, Suzuki M: Haemodynamicchanges in insulin-induced hypoglycae-mia in normal man. Diabetologia 26:328–332, 1984

5. Jaspan JB, Wollman RL, Bernstein L,Rubenstein AH: Hypoglycemic peripheralneuropathy in association with insuli-noma: implication of glucopenia ratherthan hyperinsulinism. Medicine 61:33–44, 1982

Exercise IncreasesAdiponectin Levelsand InsulinSensitivity inHumans

Adiponectin is an abundant circulat-ing adipocytokine with anti-inflammatory properties (1) linked

to cardiovascular disease, type 2 diabetes,and obesity (2–5). Numerous reports (3–5), including the present one, confirmplasma adiponectin levels to be inverselyrelated to insulin resistance. Longer term,a rise in adiponectin has been shown tooccur in response to weight loss and gli-tazone therapy, but not after chronic ex-ercise training. However, understandingof the shorter-term regulation of adi-ponectin in particular remains unclear. Asan extension to a previously reported ex-ercise intervention in sedentary males byour group (6), we have now examined theeffects of this training intervention on adi-ponectin levels in overweight males. Wedemonstrate that the short-term exercisetraining increased circulating adiponectinlevels with accompanied improved insu-lin sensitivity.

Twenty-six overweight males partici-pated in an exercise program, as previ-ously described (6). Full data wereavailable on 19 subjects who completedthe entire program. At baseline and pos-texercise intervention, all subjects wereassessed for anthropometric measures(dual-energy X-ray absorptiometry, mag-netic resonance imaging, and BMI), insu-lin sensitivity (insulin clamp), andindirect calorimetry (for fat oxidationrates), and overnight fasting plasma sam-ples were collected for adiponectin levels.Briefly, exercise consisted of aerobic exer-cise (brisk walking mixed with light jog-ging) 4–5 days per week for 40 min persession (55–70% VO2max) over 10weeks (6). Plasma adiponectin was deter-mined using a radioimmunoassay kit(Linco Research, St. Charles, MO). Two-tailed paired Student’s t tests were usedfor comparisons between time points be-fore and after exercise, and associationsbetween continuous variables were inves-tigated using simple regression analyses.Analyses were performed using StatViewsoftware (version 4.5; Abacus Concepts,Berkeley, CA).

The subjects’ mean age was 37.1 �1.3 years, and before exercise mean BMIwas 30.7 � 0.7 kg/m2 and VO2max was48.4 � 0.8 ml � kg fat-free mass (FFM)�1

� min�1. Correlations between glucose in-fusion rate (GIR), a measure of insulinsensitivity, and indexes of adiposity in thesedentary males were highly significant(all P � 0.0001). Fasting plasma adi-ponectin levels were strongly inversely re-lated to insulin resistance in thesesubjects (r � �0.52, P � 0.0007) and tototal fat (r � �0.39, P � 0.015), centralsubcutaneous fat (r � �0.37, P � 0.02),and visceral fat mass (r � �0.32, P �0.05). Two to three bouts of moderatelyintense aerobic exercise performed within1 week of baseline assessments resultedin a mean 23% increase in GIR (35.0 �2.7 vs. 43.0 � 2.8 �mol � min�1 � kgFFM�1, P � 0.0001) and a mean 37%increase in basal fat oxidation rate(1.05 � 0.14 vs. 1.44 � 0.08 g � day�1 �kg FFM�1). These effects were main-tained after 10 weeks of exercise training(42.4 � 3.1 �mol � min�1 � kg FFM�1 and1.35 � 0.07 g � day�1 � kg FFM�1, respec-tively). Body weight was unchanged aftertwo to three bouts of exercise (93.5 � 1.9vs. baseline 93.4 � 1.8 kg) and was notsignificantly reduced at 10 weeks (92.6 �1.9 kg, P � 0.08) in this cohort.

Adiponectin levels rose by 260% aftertwo to three bouts of exercise (1 week)(7.0 � 0.7 vs. 18.2 � 1.9 �g/ml, P �0.0001) despite unchanged body weightand remained elevated (16.4 � 1.9 �g/ml, P � 0.0001) after 10 weeks. However,individual changes in adiponectin levelsafter two to three bouts (1 week) andafter 10 weeks of exercise were not corre-lated with the respective changes in insu-lin sensitivity or fat oxidation rate. Ourresults contrast with Hulver et al. (7)where adiponectin is unaltered with exer-cise training despite enhanced insulin ac-tion. However, we assessed the acuteeffect of exercise after two to three boutsof exercise (6), whereas they took their“basal” samples 6 weeks after a rampingexercise period before the 6-month en-durance exercise training program (7).Our data indicate that elevated adiponec-tin levels are first apparent after 1 week(two to three bouts) of moderately intenseexercise. We suggest that it is likely thatthis short-term moderate exercise train-ing can modify regulation of adiponectin,and this could be postulated to provideanother mechanism by which exercise re-

Letters


duces atherogenic risk, at least in over-weight males.

Acknowledgments— We acknowledge theassistance of the nursing staff of the ClinicalResearch Facility, laboratory technicians of theDiabetes Research Group, Dr. Judith Freundand technicians of the Nuclear Medicine De-partment, St Vincent’s Hospital Sydney, andvolunteers who participated in this study.

ADAMANDIA D. KRIKETOS, PHD

SENG KHEE GAN, MBBS, FRACP

ANN M. POYNTEN, MBBS, FRACP

STUART M. FURLER, PHD

DONALD J. CHISHOLM, MBBS, FRACP

LESLEY V. CAMPBELL, MBBS, FRACP

From the Diabetes and Obesity Research Program,Garvan Institute of Medical Research, Sydney, Aus-tralia.

Address correspondence to Dr. Adamandia D.Kriketos, Diabetes and Obesity Research Program,Garvan Institute of Medical Research, 384 VictoriaSt., Sydney, NSW 2010, Australia. E-mail:[email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Okamoto Y, Arita Y, Nishida M, Muraga-

chi M, Ouchi N, Takahashi M, Igura T,Inui Y, Kihara S, Nakamura T, YamashitaS, Miyagawa J, Funahashi T, MatsuzawawY: An adipocyte-derived protein, adi-ponectin, adheres to injured vascularwalls. Horm Metab Res 32:47–50, 2001

2. Arita Y, Kihara S, Ouchi N, Takahashi M,Maeda K, Miyagawa J, Hotta K, Shimo-mura I, Nakamura T, Miyaoka K,Kuriyama H, Nishida M, Yamashita S,Okubo K, Matsubara K, Muraguchi M,Ohmoto Y, Funahashi T, Matsuzawa Y:Paradoxical decrease of an adipose-spe-cific protein, adiponectin, in obesity. Bio-chem Biophys Res Commun 257:79–83,1999

3. Hotta K, Funahashi T, Arita Y, TakahashiM, Motsuda M, Okamoto Y, Iwahashi H,Kuriyama H, Ouchi N, Maeda K, NishidaM, Kihara S, Sakai N, Nakajima T, Hase-gawa K, Muraguchi M, Ohmoto Y, Naka-mura T, Yamashita S, Hanafusa T,Matsuzawa Y: Plasma concentrations ofa novel, adipose-specific protein, adipo-nectin, in type 2 diabetic patients. Arte-rioscler Thromb Vasc Biol 20:1595–1599,2000

4. Hotta K, Funahashi T, Bodkin NL, Ort-meyer HK, Arita Y, Hansen BC, Matsu-zawa Y: Circulating concentrations of theadipocyte protein adiponectin are de-creased in parallel with reduced insulinsensitivity during the progression to type2 diabetes in rhesus monkeys. Diabetes 50:

1126–1133, 20015. Weyer C, Funahashi T, Tanaka S, Hotta K,

Matsuzawa Y, Pratley RE: Hypoadi-ponectinemia in obesity and type 2 diabe-tes: close association with insulinresistance and hyperinsulinemia. J Endo-crinol Metab 86:1930–1935, 2001

6. Gan SK, Kriketos AD, Ellis BA, ThompsonCH, Kraegen EW, Chisholm DJ: Changesin aerobic capacity and visceral fat but notmyocyte lipid levels predict increased in-sulin action after exercise in overweightand obese men. Diabetes Care 26:1706–1713, 2003

7. Hulver MW, Zheng D, Tanner CJ,Houmard JA, Kraus WE, Slentz CA,Sinha MK, Pories WJ, MacDonald KG,Dohm GL: Adiponectin is not alteredwith exercise training despite enhancedinsulin action. Am J Physiol 283:E861–E865, 2002

High Glucose LevelsInduce an Increase inMembraneAntioxidants, inTerms of Vitamin Eand Coenzyme Q10,in Children andAdolescents WithType 1 Diabetes

O xidative stress is defined as an im-balance between prooxidants andantioxidants in favor of the former

(1), and diabetic patients are considered arisk group for increased oxidative stress(2,3). Studies regarding oxidant/antioxidant balance in type 1 diabeticchildren and adolescents have given con-flicting results (4 –7). The aim of thisstudy was to determine whether serumhydroperoxides (reactive oxygen metabo-lites [ROMs]) as oxidative markers andplasma �-tocopherol (vitamin E) and co-enzyme Q10 as indexes of antioxidant ca-pacity could be related to metaboliccontrol in 75 unselected children, adoles-cents, and young adults with type 1 dia-betes. ROMs are the first markers ofoxidation and one of the most reliable in-dicators of oxidative stress. Vitamin E isan important chain-breaking antioxidant

factor controlling LDL oxidation. Coen-zyme Q10 is an electron carrier–protontranslocator in the respiratory chain andis an antioxidant factor by directly scav-enging radicals or indirectly by regenerat-ing vitamin E. ROMs were assayed usingthe kit d-ROMs test (Diacron), which isbased on the Fenton reaction (8). VitaminE was determined by reversed-phasehigh-performance liquid chromatogra-phy. Coenzyme Q10 was also determinedby reversed-phase high-performance liq-uid chromatography, according to themethod of Grossi et al. (9). Statistical sig-nificance was assessed using Student’s ttest and Pearson correlation index fornormally distributed data and usingMann-Whitney and Spearman rank cor-relation for nonnormally distributed data.All results that were nominally significantat P � 0.05 are indicated. Diabetic pa-tients did not have different ROMs, vita-min E, and coenzyme Q10 levels fromage-matched control subjects. Significantpositive correlations were found betweenthe following parameters: vitamin E andcoenzyme Q10, coenzyme Q10 andHbA1c, and vitamin E and HbA1c. No cor-relation was observed between ROM lev-els and coenzyme Q10, vitamin E, orHbA1c values. Vitamin E and coenzymeQ10 values were higher in patients (n �37) with poor control (HbA1c �8%) thanin those (n � 38) with good control(HbA1c �8%) (vitamin E, 25.2 � 9.5 vs.20.9 � 4.6, P � 0.044; coenzyme Q10,1.12 � 0.56 vs. 0.82 � 0.33, P � 0.012,respectively). The patients with retinal orrenal complications (n � 19) comparedwith those without had higher values ofvitamin E (25.8 � 7.1 vs. 20.9 � 4.8, P �0.009).

Therefore, in our patients vitamin Elevels increased in all of the situationswhere an increase of oxidative stress wasputative, i.e., in the presence of poor met-abolic control and complications. This re-sult is in disagreement with most of thedata of the literature (5,7,10,11), but inagreement with a few studies (12,13).However, a further confirmation of thisresult is indirectly provided by our find-ings regarding coenzyme Q10. In fact, co-enzyme Q10 levels, like vitamin E levels,are also higher in poorly controlled thanin well-controlled patients and are posi-tively correlated with HbA1c values. Thisfinding is not surprising because thesetwo antioxidants have strict physiologicalinterrelationships and are positively inter-

Letters


correlated in both the patients and controlsubjects. On the other hand, it has alreadybeen demonstrated that high-glucoseconditions produce an overexpression ofintracellular antioxidant enzymes in hu-man endothelial cells in culture (14) or inskin fibroblasts from diabetic patients(15) and that the decreased susceptibilityto oxidative stress in diabetic rats is asso-ciated with an increase in mitochondrialglutathione and coenzyme Q contents(16). This effect seems to represent anadaptive response to increased oxidativestress. In very young patients, this re-sponse is high enough to neutralize theincrease in reactive oxygen species. Infact, we found this unchanged in theblood of our patients.

SILVANA SALARDI, MD1

STEFANO ZUCCHINI, MD1

DANIELA ELLERI, MD1

GABRIELE GROSSI, MC2

ALBERTO M. BARGOSSI, AM, MD2

STEFANO GUALANDI, PHD1

ROBERTA SANTONI, MD1

ALESSANDRO CICOGNANI, MD1

EMANUELE CACCIARI, MD1

From the 1Department of Pediatrics, University ofBologna, Bologna, Italy; and the 2Central Laboratoryof “S.Orsola-Malpighi” Hospital, Bologna, Italy.

Address correspondence to Prof. Silvana Salardi,Department of Pediatrics, Via Massarenti 11, 40138Bologna, Italy. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Betteridge DJ: What is oxidative stress?

Metabolism 49:3–8, 20002. Baynes JW: Role of oxidative stress in de-

velopment of complications in diabetes.Diabetes 40:405–412, 1991

3. Hartnett ME, Stratton RD, Browne RW,Rosner BA, Lanham RJ, Armstrong D: Se-rum markers of oxidative stress and sever-ity of diabetic retinopathy. Diabetes Care23:234–240, 2000

4. Asayama K, Uchida N, Nakane T, Hayash-ibe H, Dobashi K, Amemiya S, Kato K,Nakazawa S: Antioxidants in the serum ofchildren with insulin-dependent diabetesmellitus. Free Radic Biol Med 15:597–602,1993

5. Dominguez C, Ruiz E, Gussinye M,Carrascosa A: Oxidative stress at onsetand in early stages of type 1 diabetes inchildren and adolescents. Diabetes Care21:1736–1742, 1998

6. Varvarovska J, Racek J, Stozicky F, Soucek

J, Trefil L, Pomahacova R: Parameters of ox-idative stress in children with type 1 diabe-tes mellitus and their relatives. J DiabetesComplications 17:7–10, 2003

7. Willems D, Dorchy H, Dufrasne D: Serumantioxidant status and oxidized LDL inwell-controlled young type 1 diabetic pa-tients with and without subclinical com-plications. Atherosclerosis 137 (Suppl.):S61–S64, 1998

8. Alberti A, Bolognini L, Macciantelli D,Caratelli M: The radical cation of N, N-Diethyl-para-phenylediamine: a possibleindicator of oxidative stress in biologicalsamples. Res Chem Intermed 26:253–267,2000

9. Grossi G, Bargossi AM, Fiorella PL, PiazziS, Battino M, Bianchi GP: Improved high-performance liquid chromatographicmethod for the determination of coen-zyme Q10 in plasma. J Chromatogr 593:217–226, 1992

10. Krempf M, Ranganathan S, Ritz P, Morin M,Charbonnel B: Plasma vitamin A and E intype 1 (insulin-dependent) and type 2 (noninsulin-dependent) adult diabetic patients.Int J Vitam Nutr Res 61:38–42, 1991

11. Cinaz P, Hasanoglu A, Bideci A, BiberogluG: Plasma and erythrocyte vitamin E lev-els in children with insulin dependent di-abetes mellitus. J Pediatr Endocrinol Metab12:193–196, 1999

12. Hozumi M, Murata T, Morinobu T,Manago M, Kuno T, Tokuda M, KonishiK, Mingci Z, Tamai H: Plasma beta-caro-tene, retinol, and alpha-tocopherol levelsin relation to glycemic control of childrenwith insulin-dependent diabetes mellitus.J Nutr Sci Vitaminol 44:1–9, 1998

13. Campoy C, Baena RM, Blanca E, Lopez-Sabater C, Fernandez-Garcia JM, MirandaMT, Molina-Font JA, Bayes R: Effects ofmetabolic control on vitamin E nutri-tional status in children with type 1 dia-betes mellitus. Clin Nutr 22:81–86, 2003

14. Ceriello A, dello Russo P, Amstad P,Cerutti P: High glucose induces antioxi-dant enzymes in human endothelial cellsin culture: evidence linking hyperglyce-mia and oxidative stress. Diabetes 45:471–477, 1996

15. Ceriello A, Morocutti A, Mercuri F, Qua-gliaro L, Moro M, Damante G, Viberti GC:Defective intracellular antioxidant en-zyme production in type 1 diabetic pa-tients with nephropathy. Diabetes49:2170–2177, 2000

16. Palmeira CM, Santos DL, Seica R, MorenoAJ, Santos MS: Enhanced mitochondrialtesticular antioxidant capacity in Goto-Kakizaki diabetic rats: role of coenzyme QAm J Physiol Cell Physiol 281:C1023–C1028, 2001

Increased OxidativeStress Is AssociatedWith Serum Levels ofTriglyceride, InsulinResistance, andHyperinsulinemia inJapaneseMetabolically Obese,Normal-Weight Men

M etabolically obese, normal-weight(MONW) subjects (BMI �25 kg/m2) are characterized by an excess

(�100 cm2 by abdominal computed to-mography scanning) visceral fat area(VFA), insulin resistance, and hyperinsu-linemia (1,2). The criteria for MONWsubjects and the insulin resistance syn-drome are very similar, and the patho-physiological events occurring in MONWsubjects have recently been the focus ofmany investigators (1–4).

Several studies have reported the as-sociation of oxidative stress with insulinresistance and hyperinsulinemia in obesesubjects (5,6). However, the degree of ox-idative stress and its correlation with in-sulin resistance and insulin secretion havenot yet been evaluated in MONWsubjects.

The present study comprised 18 Jap-anese MONW (aged 34.7 � 1.7 years,BMI 23.9 � 0.3 kg/m2, and VFA 146.3 �5.8 cm2 [means � SE]) and 18 age-matched normal (BMI �25 kg/m2 andVFA �100 cm2) men (aged 33.8 � 1.4years, BMI 21.9 � 0.5 kg/m2, and VFA59.3 � 5.3 cm2).

According to the American DiabetesAssociation’s diagnostic criteria, all sub-jects had normal glucose tolerance basedon the 75-g oral glucose tolerance test(OGTT) (7).

The plasma levels of free 8-epi-prostaglandin F2� (8-epi-PGF2�) weremeasured as marker of oxidative stress us-ing a commercially available enzyme im-munoassay kit (Cayman Chemical, AnnArbor, MI).

8-epi-PGF2� plasma levels inMONW men (40.4 � 6.2 pg/ml; P �0.01) were significantly increased com-pared with normal subjects (8.5 � 1.5pg/ml). The glucose infusion rates (indexof insulin resistance during the euglyce-mic-hyperinsulinemic clamp study) in

Letters


MONW subjects (53.9 � 3.2 �mol � kg�1

� min�1; P � 0.02) were significantly de-creased compared with normal subjects(65.0 � 2.5 �mol � kg�1 � min�1). Fastingserum levels of insulin (49.1 � 4.1pmol/l; P � 0.01), insulin area under thecurve (AUC) during the 75-g OGTT(44721.7 � 3811.3 pmol/l; P � 0.02),and serum levels of triglycerides (1.6 �0.1 mmol/l; P � 0.01) were significantlyincreased in MONW subjects comparedwith normal subjects (fasting insulin lev-els 29.9 � 2.9 pmol/l, insulin AUC31341.7 � 3388.9 pmol/l, and serum lev-els of triglyceride 0.9 � 0.1 mmol/l).

The 8-epi-PGF2� plasma levels weresignificantly correlated with the glucoseinfusion rate (r � �0.513, P � 0.05),VFA (r � 0.868, P � 0.01), serum levelsof triglyceride (r � 0.658, P � 0.02), fast-ing serum levels of insulin (r � 0.502,P � 0.05), and the insulin AUC (r �0.655, P � 0.01) only in MONWsubjects.

Bakker et al. (8) have previously re-ported that elevated concentration of cy-tosolic long-chain acyl-CoA, which isassociated with increased cytosolic tri-glyceride stores, induces mitochondrialoxygen free radical production due to in-tramitochondrial ADP deficiency. There-fore, increased trigylceride content innonadipose tissue together with increasedserum levels of trigylcerides may play animportant role in the production of oxi-dative stress in Japanese MONW subjects.

8-epi-PGF2� plasma levels were sig-nificantly correlated with insulin resis-tance in MONW men. This relationshipwas also observed in obese men (6). Al-though correlation does not prove causa-tion, these findings suggest that oxidativestress may contribute to the developmentof insulin resistance in obese and MONWmen.

The results of the present study are inagreement with a previous study (9,10)that showed increased cytosolic long-chain acyl-CoA and oxidative stress lowerglucose-induced insulin secretion frompancreatic �-cells. On the other hand, ithas been reported that hyperinsulinemiareduces oxidative stress production(11,12). Hyperinsulinemia may also havea protective role against increased oxida-tive stress in MONW men.

AKIRA KATSUKI, MD1

YASUHIRO SUMIDA, MD1

HIDEKI URAKAWA, MD1

ESTEBAN C. GABAZZA, MD1

SHUICHI MURASHIMA, MD2

KANAME NAKATANI, MD3

YUTAKA YANO, MD1

YUKIHIKO ADACHI, MD1

From the 1Third Department of Internal Medicine,Mie University School of Medicine, Mie, Japan; the2Department of Radiology, Mie University School ofMedicine, Mie, Japan; and the 3Department of Lab-oratory Medicine, Mie University School of Medi-cine, Mie, Japan.

Address correspondence to Y. Sumida, MD,Third Department of Internal Medicine, Mie Univer-sity School of Medicine, 2-174 Edobashi, Tsu, Mie514-8507, Japan. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Ruderman N, Chisholm D, Pi-Sunyer X,

Schneider S: The metabolically obese,normal-weight individual revisited. Dia-betes 47:699–713, 1998

2. Katsuki A, Sumida Y, Urakawa H, Gaba-zza EC, Murashima S, Maruyama N, Mo-rioka K, Nakatani K, Yano Y, Adachi Y:Increased visceral fat and serum levels oftriglyceride are associated with insulin re-sistance in Japanese metabolically obese,normal-weight subjects with normal glu-cose tolerance. Diabetes Care 26:2341–2344, 2003

3. Davidson MB: Is treatment of insulin re-sistance beneficial independent of glyce-mia? Diabetes Care 26:3184–3186, 2003

4. Lorenzo C, Okoloise M, Williams K, SternMP, Haffner SM: The metabolic syndromeas predictor of type 2 diabetes: the SanAntonio Heart Study. Diabetes Care 26:3153–3159, 2003

5. Davi G, Guagnano MT, Ciabattoni G, BasiliS, Falco A, Marinopiccoli M, Nutini M,Sensi S, Patrono C: Platelet activation inobese women: role of inflammation and ox-idant stress. JAMA 288:2008–2014, 2002

6. Urakawa H, Katsuki A, Sumida Y, Gaba-zza EC, Murashima S, Morioka K, Ma-ruyama N, Kitagawa N, Tanaka T, Hori Y,Nakatani K, Yano Y, Adachi Y: Oxidativestress is associated with adiposity and in-sulin resistance in men. J Clin EndocrinolMetab 88:4673–4676, 2003

7. The Expert Committee on the Diagnosisand Classification of Diabetes Mellitus: Re-port of the Expert Committee on the Diag-nosis and Classification of DiabetesMellitus. Diabetes Care 20:1183–1197,1997

8. Bakker SJL, IJzermen RG, Teerlink T,Westerhoff HV, Gans ROB, Heine RJ: Cy-tosolic triglycerides and oxidative stress incentral obesity: the missing link betweenexcessive atherosclerosis, endothelial dys-function, and �-cell failure? Atherosclero-

sis 148:17–21, 20009. Prentki M, Corkey BE: Are the �-cell sig-

naling molecules malonyl-CoA and cyto-solic long-chain acyl-CoA implicated inmultiple tissue defects of obesity andNIDDM. Diabetes 45:273–283, 1996

10. Milburn JL Jr, Hirose H, Lee YH, Naga-sawa Y, Ogawa A, Ohneda M, Beltrandel-Rio H, Newgard CB, Johnson JH, UngerRH: Pancreatic �-cell in obesity: evidencefor induction of functional, morphologic,and metabolic abnormalities by increasedlong chain fatty acids. J Biol Chem 270:1295–1299, 1995

11. Dandona P, Aljada A, Mohanty P, GhanimH, Hamouda W, Assian E, Ahmad S: In-sulin inhibits intranuclear nuclear factor B and stimulates I B in mononuclearcells in obese subjects: evidence for an an-ti-inflammatory effect? J Clin EndocrinolMetab 86:3257–3265, 2001

12. Kyselova P, Zourek M, Rusavy Z, Trefil L,Racek J: Hyperinsulinemia and oxidativestress. Physiol Res 51:591–595, 2002

Efficacy ofConversion FromBedtime NPH InsulinInjection to Once- orTwice-DailyInjections of InsulinGlargine in Type 1Diabetic PatientsUsing Basal/BolusTherapy

The efficacy of glycemic control intype 1 diabetic patients with eitheronce- or twice-daily glargine insulin

injection was evaluated in this long-term,prospective, nonrandomized study.Eighty-two type 1 diabetic patients werefollowed over 12–15 months after con-version from a single bedtime NPH insu-lin injection to a single bedtime insulinglargine injection. These patients wereswitched with the availability of glargineinsulin to reduce frequency and severityof nocturnal hypoglycemia and to im-prove fasting glucose levels. This group oftype 1 diabetic patients was switched toglargine insulin in place of twice-dailyNPH. These patients continued theirsame bolus therapy with either insulin lis-pro or aspart and underwent frequent(three to five times daily) home glucose

Letters


monitoring. This study showed the ex-pected fewer nocturnal hypoglycemicevents, but the primary outcome was animprovement in glycemic control basedon the HbA1c values. Patients HbA1c val-ues were determined every 8 weeks, andinsulin doses were titrated, with glargineadjusted based on morning fasting glu-cose values. If the HbA1c remained abovegoal, the intensity of home glucose mon-itoring was increased and bolus therapywas adjusted accordingly. In one-quarterof patients, the lunch bolus titration re-sulted in mid-afternoon hypoglycemia,and when reduced, the patients had ele-vated presupper glucose values. Patientswho had an increase in their HbA1cand/or persistent elevation of presupperglucose despite titration of both bolus in-sulin and glargine insulin were thenplaced on twice-daily glargine injections.A spilt dose of glargine was given onlyafter titration of glargine insulin resultedin morning hypoglycemia and/or persis-tent elevation of the afternoon blood glu-coses that could not be corrected withbolus titration.

Sixty-two subjects were usingglargine insulin once daily, and the re-maining 20 (24.2%) subjects requiredtwice-daily therapy. The 24.2% of pa-tients on split glargine were convertedfrom once-daily glargine injections afteran average of 289 � 203 days (median259). At that time, their HbA1c had dete-riorated from an initial value of 7.9 � 1.5to 8.1 � 1.4% (P � 0.16) and titrationwas limited by the symptoms outlinedabove. Subjects on split glargine did notdiffer from those subjects using once-daily glargine injections in regard to theirage (P � 0.21), duration of diabetes (P �0.21), baseline HbA1c (P � 0.91), pres-ence of detectable C-peptide (P � 0.78),or the presence of microvascular compli-cations: retinopathy (P � 0.37), nephrop-athy (P � 0.44), neuropathy (P � 0.30),or macrovascular complications (P �0.88).

In the single-daily injection patients,the HbA1c improved significantly (from7.8 to 7.3%, P � 0.01) after 476 � 178days on glargine insulin. The splitglargine injection subjects also had an im-provement in the HbA1c from 7.9 to 7.4%(P � 0.03) over a 3- to 6-month period.The ending HbA1c between groups wasnot significant (P � 0.80). The decreasefrom the mean starting HbA1c was identi-cal between groups; however, a more sig-

nificant drop in the HbA1c from the timeof spilt to the end of the study (8.1 to7.4%) did reach statistical significance(P � 0.001). To achieve this improvedglycemic control in these patients, 70%more glargine was required (44 � 26 vs.26 � 13 units, P � 0.008).

In conclusion, in this prospective,nonblinded, nonrandomized, prospec-tive study, one-quarter of type 1 diabeticpatients required twice-daily glargine in-sulin injections to achieve acceptable gly-cemic control. The reason for this was thatmore glargine insulin could be safelyand/or effectively used when split in thispopulation. Regardless, type 1 diabeticpatients on glargine insulin improvedtheir glycemic control, as measured by theHbA1c values. Patients who do notachieve control after a titration periodshould receive split daily doses to achieveglycemic control.

This is the first study to demonstratean improvement in HbA1c in type 1 dia-betic subjects using basal-bolus therapyafter conversion from basal NPH insulinto basal glargine insulin. Previous studieswere not continued beyond 6 months,and the protocols restricted glargine toonce-daily dosing (1–3). Using our studydata to project the outcome, if all patientsremained on single daily dosing, then theaverage HbA1c reduction would havebeen 0.3%, which is similar to other stud-ies, and would not have been statisticallysignificant.

ERIC S. ALBRIGHT, MD

RENEE DESMOND, DVM, PHD

DAVID S.H. BELL, MB

From the Department of Medicine, University of Al-abama at Birmingham, Birmingham, Alabama.

Address correspondence to David S.H. Bell, MD,University of Alabama at Birmingham, Departmentof Medicine, 1808 Seventh Ave. S., BDB 802, Bir-mingham, AL 35294. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Raskin P, Klaff L, Bergenstal R, Halle JP,

Donley D, Mecca T: A 16-week compari-son of the novel insulin analog insulinglargine (HOE 901) and NPH human in-sulin used with insulin lispro in patientswith type 1 diabetes. Diabetes Care 23:1666–1671, 2000

2. Rosenstock J, Park G, Zimmerman J, theUS Insulin Glargine (HOE 901) Type 1Diabetes Investigator Group: Basal insulin

glargine (HOE 901) versus NPH insulin inpatients with type 1 diabetes on multipledaily insulin regimens. Diabetes Care 23:1137–1142, 2000

3. Pieber TR, Eugene-Jolchine I, Derobert E,the European Study Group of HOE 901 inType 1 Diabetes: Efficacy and safety ofHOE 901 versus NPH insulin in patientswith type 1 diabetes. Diabetes Care 23:157–162, 2000

COMMENTS ANDRESPONSES

MemoryImpairmentsAssociated WithPostprandialHyperglycemia andGlycemic Control

Comment on Greenwood et al.

I t was with interest that we read thestudy by Greenwood et al. (1), whichinvestigated the impact of postprandial

hyperglycemia on memory function intype 2 diabetic patients and demonstratedimpaired memory function after carbohy-drate ingestion. As they thoroughly dis-cussed, the impact of glycemic controland transient hyperglycemia has been un-der investigation since the mid-1980s(2,3), with study results that are hetero-geneous and not very conclusive. In fact,data from a study at our diabetes center(4) comprising 53 type 2 diabetic patientssuggest that glycemic control has no in-fluence on cognitive functioning, includ-ing memory (Auditory Verbal LearningTest), whereas patients with diabeticcomplications show lower performance.

One reason for the heterogeneity ofresults probably stems from the lack ofconsensus on which instruments to usefor cognitive function assessment (5) andthe usually small sample sizes. In this re-spect, unfortunately, Greenwood et al.did not use the standard versions of thetests for memory assessment but insteadused instruments that were constructed ofparallel forms and had obviously under-gone profound changes, like omittingitems. The precise nature of the test that

Letters


was applied is not described in their pre-vious study either (6). Although theythoroughly addressed parallels of the ver-sions used, the possible interference ef-fects of several verbal memory tests usedin a row are not discussed.

Taken together with the small samplesize, the large interindividual variability ofperformance within the groups, and hencethe fact that the adequacy of regressionanalysis is disputable, in our point of viewthe conclusions of Greenwood et al. aredaring. Surely, neuropsychological ef-fects of transient hyperglycemic excur-sions are worth being studied further, butconcluding that ingestion of one-half ba-gel and grape juice leads to acute memoryimpairment seems, in our opinion, toofar-reaching.

THOMAS KUBIAK, PHD1

NORBERT HERMANNS, PHD1

MICHAEL PREIER, MA2

BERNHARD KULZER, MA3

THOMAS HAAK, MD1,3

From the 1Research Institute of the Diabetes Acad-emy, Diabetes Center Mergentheim, Bad Mergen-theim, Germany; the 2Department of Neuropsy-chology, Rehabilitation Clinic Staffelstein, BadStaffelstein, Germany; the 3Diabetes Clinic, DiabetesCenter Mergentheim, Mergentheim, Germany.

Address correspondence to Dr. Thomas Kubiak,PhD, Forschungsinstitut der Diabetes AkademieMergentheim (FIDAM), Diabetes Zentrum Mergen-theim, Th.-Klotzbuecher-Str. 12, D-97980 Bad Mer-gentheim, Germany. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Greenwood CE, Kaplan RJ, Hebbleth-

waite S, Jenkins DJA: Carbohydrate-in-duced memory impairment in adults withtype 2 diabetes. Diabetes Care 26:1961–1966, 2003

2. Holmes CS: Metabolic control and audi-tory information processing at altered glu-cose levels in insulin-dependent diabetes.Brain Cogn 6:161–174, 1987

3. Holmes CS: Neuropsychological and Behav-ioral Aspects of Diabetes. New York,Springer, 1990

4. Hewer W, Mussell M, Rist F, Kulzer B,Bergis K: Short-term effects of improvedglycemic control on cognitive function inpatients with type 2 diabetes. Gerontology49:86–92, 2003

5. Strachan MW, Frier BM, Deary IJ: Cogni-tive assessment in diabetes: the need forconsensus (Editorial). Diabet Med 14:421–422, 1997

6. Kaplan RJ, Greenwood CE, Winocur G,

Wolever TMS: Cognitive performance isassociated with glucose regulation inhealthy elderly persons and can be en-hanced with glucose and dietary carbohy-drates. Am J Clin Nutr 72:825–836, 2000

MemoryImpairmentsAssociated WithPostprandialHyperglycemia andGlycemic Control

Response to Kubiak et al.

W e thank Kubiak et al. (1) for theirthoughtful appraisal of our studyrelating to memory response in

adults with type 2 diabetes following car-bohydrate ingestion (2). We agree en-tirely with their comment that theunderlying origins of memory impair-ment in this population is poorly under-stood and that lack of consensus onstandardized neuropsychologic testingprocedures may, in part, be contributingto this confusion (3). Clearly, as Kubiak etal. point out, a major contributor to thevariance in cognitive performance ob-served in this population is the high prev-alence of other risk factors for cognitivedecline, including cardiovascular disease,hypertension, and depression (4,5), mak-ing it challenging to isolate the potentialcontribution of diabetes per se.

We sought to explore cognitive func-tion in adults with type 2 diabetes byperturbing the system through the ad-ministration of glucose, which is a treat-ment commonly used in studies ofcognitive aging to explore the system’splasticity in the face of underlying age-related deficits (6,7). A major advantageof applying this approach to the type 2diabetic population is that changes ob-served in response to the challenge wereunlikely to be directly attributable to vas-cular complications. Our data providedevidence for cognitive deficits, primarilyrelated to declarative memory function,following the ingestion of 50 g of glucosein the form of rapidly absorbed carbohy-drate foods (bagel and juice). We then ar-gued that this impairment was consistentwith observations in healthy senioradults, in whom moderate elevations in

blood glucose resulted in memory en-hancement and more extreme increases inblood glucose were associated with defi-cits (what is often referred to as an invert-ed-U dose-response relationship) (8).Based on this argument, we concludedthat adults with type 2 diabetes likely re-sponded to a glucose challenge in a man-ner comparable with that of older adults,with the caveat that they were more likelyto attain levels of hyperglycemia associ-ated with cognitive impairment giventheir underlying disease. Clearly, this is aconclusion requiring further verification.

One issue of concern raised by Ku-biak et al. is that we include alternate ver-sions, developed by us and others (9), ofstandardized neuropsychologic tests, al-though we apply these versions usingstandardized methodology. Our within-individual design, i.e., requiring multipletesting of subjects, necessitates their use.While we do not provide the precise de-tails used to develop these alternate ver-sions, we previously directed readers tothose publications that we relied on to doso. Importantly, we have never stated orimplied that these alternate versionsshould be used clinically from a diagnos-tic perspective; rather they are only usedas experimental research tools. We agreethat the use of these alternate versions po-tentially adds unwanted variance to ourmeasures, but disagree that this detractsfrom the results obtained. Rather, the ad-ditional variance contributed by the use ofalternate test versions makes it more, notless, difficult to observe change followingglucose ingestion or associations betweensubject characteristics and performancelevels. We address test version variabilitythrough the random allocation of differ-ent test versions both between and withinsubjects to uniformly distribute this addi-tional variance (as much as possible)throughout the data, thereby minimizingpotential bias associated with their use.

Another concern expressed was thelimited number (n � 19) of subjects inour study, which is in essence an exten-sion of those concerns related to test ver-sion variability. While not commented onin the original publication, this samplesize was based on power analyses drawingon our results in healthy senior adults re-ceiving glucose in the form of carbohy-drate foods and including the samealternate versions of the neuropsycho-logic tests (9). Thus we believe our studyto be statistically sound. Nevertheless, as

Letters


with all studies, the extension of the re-sults to the broader population is compli-cated by the fact that subjects willing toparticipate in experimental proceduresare somewhat unique and, in this sense,differ from the more heterogeneous pop-ulation typically observed in clinical prac-tice—a factor of importance in all studiesdrawing on human volunteers.

Finally, Kubiak et al. comment onconcerns related to interference whenmultiple tests probing declarative mem-ory are used. This factor is not discussedin this work but is addressed by us previ-ously (9) in studies conducted on healthysenior adults. This comment raises multi-ple issues of interest. The first is that theexact nature of the declarative memorydeficits observed in adults with type 2 di-abetes remains largely unexplored.Clearly, multiple components of cogni-tive function are recruited and contributeto performance on end measures of de-layed verbal recall; yet the precise deficit,potentially including interference and in-hibitory control, remains largely unex-plored. Admittedly, we observed deficitsin our study on the second, not the first,verbal recall test used, thus raising thepossibility that interference is an impor-tant contributor and one requiring furtherexploration. Yet this does not detract fromthe fact that performance on this secondtest was poorer when subjects were testedfollowing carbohydrate ingestion com-pared with when they were tested follow-ing placebo (water) ingestion. The secondissue relating to the comment by Kubiaket al. is the degree to which one controlsfor external factors influencing cognitiveperformance. Clearly, our data suggestthat the fed/fasted state of the individualmay be an important contributor to ob-served variance. Similarly, time of day oftesting is another recognized contributorto within-individual variance in cognitivefunction and shifts in peak performancetimes occur in states, such as aging,wherein disruptions to circadian sleeprhythms are apparent (10). This is clearlya pattern disruption to which adults withtype 2 diabetes may be especially vulner-able. Yet rarely do authors address whenduring the day testing occurred andwhether a fixed time of day was used, aswe did in our studies. All of these factorsare likely important and contribute to the“unexplained” variance in regressionmodels. Some will view this as excitingopportunities for new exploration,

whereas others will see this as reason todiscount “explained” variance in func-tion.

Certainly, it is essential that as wewade through conflicting data regardingthe origins of cognitive deficits in adultswith type 2 diabetes, be they primarily orsecondarily associated with the high prev-alence of other risk factors in this popu-lation, that we not lose sight of the sharedand common interest, which is helpingthose with type 2 diabetes prevent or min-imize their risk of cognitive dysfunction.Kubiak et al. state that we were daring toconclude that transitory food-inducedhyperglycemic episodes could be associ-ated with acute cognitive deficits in thispopulation. We continue to stand by ourconclusions but recognize that furtherstudies will either support or refute ourconclusions.

CAROL E. GREENWOOD, PHD1,2,3

RANDALL J. KAPLAN, PHD1,2

STACEY HEBBLETHWAITE, BSC1

DAVID J.A. JENKINS, MD, PHD1,4,5

From the 1Department of Nutritional Sciences, Fac-ulty of Medicine, University of Toronto, Toronto,Ontario, Canada; the 2Department of Food and Nu-trition Services, Baycrest Centre for Geriatric Care,Toronto, Ontario, Canada; the 3Kunin-LunenfeldApplied Research Unit, Baycrest Centre for GeriatricCare, Toronto, Ontario, Canada; the 4Clinical Nu-trition and Risk Modification Centre, St. Michael’sHospital, Toronto, Ontario, Canada; and the 5Divi-sion of Endocrinology and Metabolism, St. Michael’sHospital, Toronto, Ontario, Canada.

Address correspondence to Carol Greenwood,Department of Nutritional Sciences, Faculty ofMedicine, University of Toronto, Toronto, Ontario,Canada M5S 3E2. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Kubiak T, Hermanns N, Preier M, Kulzer

B, Haak T: Memory impairments associ-ated with postprandial hyperglycemiaand glycemic control: comment onGreenwood et al. (Letter). Diabetes Care27:633–634, 2004

2. Greenwood CE, Kaplan RJ, Hebbleth-waite S, Jenkins DJA: Carbohydrate-in-duced memory impairment in adults withtype 2 diabetes. Diabetes Care 26:1961–1966, 2003

3. Strachan MW, Frier BM, Deary IJ: Cogni-tive assessment in diabetes: the need forconsensus (Editorial). Diabet Med 14:421–422, 1997

4. Strachan MW, Deary IJ, Ewing FM, FrierBM: Is type II diabetes associated with anincreased risk of cognitive dysfunction? A

critical review of published studies. Dia-betes Care 20:438–445, 1997

5. Stewart R, Liolitsa D: Type 2 diabetes mel-litus, cognitive impairment and dementia.Diabet Med 16:93–112, 1999

6. Gold P: Glucose modulation of memorystorage processing. Behav Neural Biol 45:342–349, 1986

7. Messier C, Tsiakas M, Gagnon M, Desro-chers A, Awad N: Effect of age and glu-coregulation on cognitive performance.Neurobiol Aging 24:985–1003, 2003

8. Parsons MW, Gold PE: Glucose enhance-ment of memory in elderly humans: aninverted-U dose-response curve. Neuro-biol Aging 13:401–404, 1992

9. Kaplan RJ, Greenwood CE, Winocur G,Wolever TMS: Cognitive performance isassociated with glucose regulation inhealthy elderly persons and can be en-hanced with glucose and dietary carbohy-drates. Am J Clin Nutr 72:825–836, 2000

10. Li KZ, Hasher L, Jonas D, Rahhal TA, MayCP: Distractibility, circadian arousal, andaging: a boundary condition? Psychol Ag-ing 13:574–583, 1998

Off-Loading in Trialsin NeuropathicDiabetic FootUlceration

No, it’s not time for a paradigmshift

P rofessor Boulton and Dr. Armstrong(1) argued recently that “all futuretrials of therapy should use a nonre-

movable off-loading device.” In doing so,they betray a failure to understand howthe structure of trials must be determinedby their purpose: those designed to deter-mine the efficacy (“Can it work in idealcircumstances?”) may differ from thosedesigned to determine effectiveness(“Does it work in practice?”). Two factorsthat underlie the capacity of a controlledtrial to demonstrate efficacy are 1) the ef-fect, or lack of it, of the intervention and2) the effect of the control. Boulton andArmstrong concluded (with no evidence)that the failure of Veves et al. (2) to dem-onstrate any benefit of Promogran was“likely” to be the result of a failure to stan-dardize off-loading techniques. Anotherinterpretation is that the product is com-paratively ineffective in routine practice.

The classical total contact cast (TCC)does not have a dressing window, and sohow can it be used in trials of dressings

Letters


and applications designed to be changedmore often than the off-loading device? ATCC, but not their modified walker, canbe modified by incorporating a dressingwindow, but dressing windows have theirproblems. If too small, they limit the abil-ity to clean and dress the wound properly.If too large, they limit the effectiveness ofoff-loading by allowing the ulcerated areato prolapse.

Crucially, however, Boulton andArmstrong fail to satisfactorily address thequestions of acceptability and safety.They acknowledge that TCCs have ad-verse effects and suggest that these may beovercome with their modified walker, butadmit that relevant trials have not beencompleted. In truth, many people findnonremovable devices unacceptable,with reasons that include secondary ul-ceration of the index foot, abrasions onthe contralateral foot, unsteadiness (espe-cially in the elderly, those with posturalhypotension or impaired propriocep-tion), and falls from tripping, not to men-tion the ease—or lack of it—with whichpatients can shower or take a bath. Trialsof nonremovable off-loading devices maybe critically biased by population selection.

In conclusion, we emphasize ourenormous respect for the work under-taken by Boulton and Armstrong butthink that their arguments are simply notjustified. The recent Cochrane review ofoff-loading (3) concluded that “there isvery limited evidence of the effectivenessof total contact casts” and highlighted thefact that there has been no comparisonundertaken between TCC and Scotchcast(or equivalent) removable boots, whichare widely used in many countries. TheTCC is an option, but not sine qua non ineither clinical practice or future trials.

WILLIAM J. JEFFCOATE, MRCP1

FRANCES L. GAME, MRCP, MRCPATH1

PATRICIA E. PRICE, PHD2

From the 1Department of Diabetes and Endocrinol-ogy, Foot Ulcer Trials Unit, City Hospital, Notting-ham, U.K.; and the 2Wound Healing Research Unit,Cardiff Medicenter, Heath Park, Cardiff, U.K.

Address correspondence to Dr. W.J. Jeffcoate,Foot Ulcer Trials Unit, Department of Diabetes andEndocrinology, City Hospital, Nottingham NG51PB, U.K. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Boulton AJM, Armstrong DG: Trials in

neuropathic diabetic foot ulceration: time

for a paradigm shift? (Editorial). DiabetesCare 26:2689–2690, 2003

2. Veves A, Sheehan P, Pham HT: A random-ized controlled trial of promogran vs stan-dard treatment in the management in themanagement of diabetic foot ulcers. ArchSurg 137:822–827, 2002

3. Spencer S: Pressure relieving interven-tions for preventing and treating diabeticfoot ulcers. Cochrane Database Syst Rev3:CD002302, 2000

Off-Loading in Trialsof NeuropathicDiabetic FootUlceration

Further evidence of the need for aparadigm shift

W e are happy that our recent edi-torial (1) stimulated interestfrom other experts in the dia-

betic foot. However, we find ourselves indisagreement with the letter by Jeffcoateand Game in this issue of Diabetes Care (2)in several respects. They suggest that oureditorial betrays “a failure to understandhow the structure of trials must be deter-mined.” Surely, this cannot be the case.Any trial assessing dressings, drugs, orconstructs should be designed to providethe maximum opportunity for the prod-uct to demonstrate efficacy by removingall possible confounding variables. As wehave recently demonstrated (3), those pa-tients provided with removable cast walk-ers only wear their device for 28% ofactivity daily, so we proposed that futuretrials should therefore standardize off-loading, preferably using a nonremovabledevice. As off-loading in the trial of pro-mogram (4) was “left to the individualcenter,” we stand by our assessment that alikely explanation of the failure to dem-onstrate efficacy was related to a failure tostandardize off-loading.

Having demonstrated the efficacy ofany new product, it then behooves us totranslate the results into clinical practice.Here we agree with Jeffcoate and Gamethat not all patients can tolerate casts;however, our experience to date suggeststhat the instant total contact cast (TCC) isbetter tolerated by patients than the TCC(5). (The instant TCC is a removable castwalker rendered nonremovable by wrap-ping it with cast material.) Further studieson this will be published in 2004. Rather

than stating that many patients cannottolerate nonremovable devices, surely re-search should be directed at improvingthe design of such casts to make themmore safe and acceptable. We suggest thatthe failure to develop satisfactory off-loading in recent years is responsible forthe poor results of trials of potential newtherapies for plantar ulcers.

Jeffcoate and Game then assert thatTCCs do not have dressing windows. Co-incidentally, in the very next issue of Di-abetes Care, Ha Van et al. (6) describe aTCC incorporating just such a window.

Further support for our position ap-pears in several articles published in themonths since the appearance of our edi-torial. In addition to describing the incor-poration of a dressing window into anonremovable cast, Ha Van et al. reportedthat only 10% of patients complied withthe removable off-loading device in theirstudies, suggesting that the 28% reportedin our study (3) was likely realistic, if notoptimistic. Secondly, Caravaggi et al. (7)demonstrated that trials of new dressingscould be successfully executed using anonremovable cast. Finally, Piaggesi et al.(8) provide pivotal histological evidencestrongly demonstrating the importance ofadequate off-loading. It is now clear whyso many trials have failed to demonstrateefficacy in recent years. Hence, we reiter-ate the need for a paradigm shift in thedesign of future clinical trials of putativetherapies for plantar neuropathic ulcers.

ANDREW J.M. BOULTON, MD, FRCP1

DAVID G. ARMSTRONG, MSC, DPM2

From the 1Diabetes Research Unit, University of Mi-ami, Miami, Florida; and the 2Department of Sur-gery, Tucson Veterans Affairs Medical Center,Tucson, Arizona.

Address correspondence to Andrew J.M. Boul-ton, MD, FRCP, Diabetes Research Unit, Universityof Miami, P.O. Box 016960 (D110), Miami, FL33101. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Boulton AJM, Armstrong DG: Trials in

neuropathic diabetic foot ulcers: time fora paradigm shift? (Editorial). DiabetesCare 26:2689–2690, 2003

2. Jeffcoate WJ, Game FL: Off-loading in tri-als in neuropathic diabetic foot ulcer-ation: no, it’s not time for a paradigm shift(Letter). Diabetes Care 27:635–636, 2004

3. Armstrong DG, Lavery LA, Kimbriel HR,Nixon BP, Boulton AJ: Activity patterns ofpatients with diabetic foot ulceration: pa-

Letters


tients with active ulceration may not ad-here to a standard pressure off-loadingregimen. Diabetes Care 26:2595–2597,2003

4. Veves A, Sheehan P, Pham HT: A random-ized controlled trial of promogran vs stan-dard treatment in the management ofdiabetic foot ulcers. Arch Surg 137:822–827, 2002

5. Armstrong DG, Short B, Espensen EH,Abu-Rumman PL, Nixon BP, Boulton AJ:Technique for fabrication of an “instanttotal-contact cast” for treatment of neuro-pathic diabetic foot ulcers. J Am PodiatrMed Assoc 92:405–408, 2002

6. Ha Van G, Siney H, Hartmann-HeurtierA, Jacqueminet S, Greau F, Grimaldi A:Nonremovable, windowed, fiberglass castboot in the treatment of diabetic plantarulcers: efficacy, safety, and compliance.Diabetes Care 26:2848–2852, 2003

7. Caravaggi C, DeGiglio R, Pritelli C, Som-maria M, Dalla Noce S, Faglia E, ManteroM, Clerici G, Fratino P, Morabito A:HYAFF 11–based autologous dermal andepidermal grafts in the treatment of non-infected diabetic plantar and dorsal footulcers: a prospective, multicenter, con-trolled, randomized clinical trial. DiabetesCare 26:2853–2859, 2003

8. Piaggesi A, Viacava P, Rizzo L, NaccaratoG, Baccetti F, Romanelli M, Zampa V, Del-Prato S: Semiquantitative analysis of thehistopathological features of the neuro-pathic foot ulcer: effect of pressure relief.Diabetes Care 26:3123–3128, 2003

C-Reactive Proteinand Glycemic Controlin Adults WithDiabetes

Response to King et al.

K ing et al. (1) recently suggested anassociation between glycemic con-trol and systemic inflammation, i.e.,

between HbA1c levels and highly sensitiveC-reactive protein (hsCRP) levels, basedon data from 1,018 participants in theThird National Health and Nutrition Ex-amination Survey. This report promptedus to search for a similar association in ourclinical practice.

Since hsCRP levels can be lowered bystatins, thiazolidinediones (TZDs), andanti-inflammatory drugs, we first lookedat 64 C-peptide–negative type 1 diabeticpatients whose only medication was insu-

lin and found no association (r � 0.0748,P � 0.28) between HbA1c levels andhsCRP. With this negative association, weinvestigated 108 C-peptide–positive type2 diabetic patients, all of whom were on astatin, an aspirin, and a TZD, to seewhether there was an association betweenhsCRP and HbA1c in this homogenousgroup on maximal hsCRP-lowering ther-apy. The association was again negative,with an r value of 0.0424 and a P value of0.78.

Why then did King et al. find an as-sociation of HbA1c with hsCRP and wedid not? We believe that King et al.’s as-sociation was with insulin resistance andnot hyperglycemia. An association of in-sulin resistance and hsCRP has been welldocumented, and theoretically at least,the greater the insulin resistance theworse the glycemic control and, con-versely, the higher the glucose the greaterthe insulin resistance (glucotoxicity). Inour group of type 2 diabetic patients whowere all on a TZD, insulin resistanceshould be maximally treated so that if hy-perglycemia did affect the hsCRP, its ef-fects would not be confounded by theeffects of insulin resistance.

That insulin resistance was not a fac-tor in the King et al. study could be con-cluded from the inclusion of fastinginsulin levels in the regression model.When diabetic subjects are treated withinsulin, insulin secretagogues, or insulinsensitizers, the effectiveness of a fastingserum insulin level as a marker for insulinresistance is negated and the conclusionthat insulin resistance was eliminated as afactor nullified.

To resolve this problem of differingconclusions from an epidemiologicalcross-sectional study and a retrospectivecross-sectional clinical study, a prospec-tive longitudinal study should be per-formed. An ideal study would be of type 1diabetic patients at onset who are clini-cally free of infection, with measurementsof hsCRP being performed before insulintherapy and 2 months later when they arewell controlled in the honeymoon period.This is of clinical importance because ifhsCRP levels are elevated due to hyper-glycemia, then hsCRP levels should onlybe measured when glycemia is controlledto avoid unnecessary prescribing.

DAVID S.H. BELL, MB, FACE

ROBERT W. HARDY, PHD

RENEE DESMOND, PHD

From the Department of Medicine, University of Al-abama Medical School, Birmingham, Alabama.

Address correspondence to David S.H. Bell, MD,University of Alabama at Birmingham, Departmentof Medicine, 1808 Seventh Ave. S., BDB 802, Bir-mingham, AL 35294. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. King DE, Mainous AG, Buchanan TA,

Pearson WS: C-reactive protein and gly-cemic control in adults with diabetes. Di-abetes Care 26:1535–1539, 2003

2. Ford ES: Body mass index, diabetes, andC-reactive protein among U.S. adults. Di-abetes Care 22:1971–1977, 1999

C-Reactive Proteinand Glycemic Controlin Adults WithDiabetes

Response to Bell, Hardy, andDesmond

W e thank Bell, Hardy, and Des-mond (1) for their comments re-garding our recent article (2),

and we appreciate the opportunity to re-spond to the issues they have raised.Based on their analysis of two groups ofpatients in their practice, Bell, Hardy, andDesmond question whether there is an as-sociation between CRP and glycemic con-trol. Several possible explanations existfor the difference in our findings. First, weused a nationally representative popula-tion-derived database that may be morediverse than the one used by them. Sec-ond, we specifically excluded people onanti-inflammatory and cholesterol-lowering medications, precisely becausethe use of such individuals is likely to con-found the relationship between C-reac-tive protein (CRP) and HbA1c (insulin-sensitizing drugs were not widelyavailable at the time of the study [1988–1994]). Another reason for the differencein our findings could be our ability to ac-count for several other factors that mightconfound or mask the relationship, in-cluding age, race, sex, BMI, smoking,length of time with diabetes, and fastinginsulin levels. Further supporting our

Letters


findings, other researchers have found asimilar association between CRP andHbA1c in nondiabetic individuals (3).

Bell, Hardy, and Desmond correctlynote the limitation of fasting insulin levelas a measure of insulin resistance. How-ever, their conclusion that CRP is relatedto insulin resistance rather than glycemiamay also be premature, since the term in-sulin resistance is a very general one thatincludes several possible underlyingmechanisms. Our report did not addressspecific mechanisms for the associationwe found, but instead called for more re-search to further delineate the nature ofthe association. We agree with Bell,Hardy, and Desmond that more definitive

prospective and interventional studies areneeded to investigate the association be-tween CRP and glycemia, as we urged inour article.

DANA E. KING, MD1

ARCH G. MAINOUS, III, PHD1

THOMAS A. BUCHANAN, MD2

WILLIAM S. PEARSON, MHA1

From the 1Department of Family Medicine, MedicalUniversity of South Carolina, Charleston, SouthCarolina; and the 2General Clinical Research Center,University of Southern California School of Medi-cine, Los Angeles, California.

Address correspondence to Dana E. King, MD,Medical University of South Carolina, Departmentof Family Medicine, 295 Calhoun St., Charleston,SC 29425. E-mail: [email protected].


● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

References1. Bell DSH, Hardy RW, Desmond R: C-re-

active protein and glycemic control inadults with diabetes (Letter). DiabetesCare 27:637, 2004

2. King DE, Mainous AG, Buchanan TA,Pearson WS: C-reactive protein and gly-cemic control in adults with diabetes. Di-abetes Care 26:1535–1539, 2003

3. Wu T, Dorn JP, Donahue RP, Sempos CT,Trevisan M: Associations of serum C-re-active protein with fasting insulin, glu-cose, and glycosylated hemoglobin: theThird National Health and Nutrition Ex-amination Survey, 1988–1994. Am J Epi-demiol 155:65–71, 2002

Letters


Documents

LETTERS - Diabetes Care · 2004-01-09 · Diabetes T he 2000 Dietary Guidelines for Americans and the dietary guide-lines for the control of diabetes in Mexico include recommendations