NEOREVIEWS 1.11

DOI: 10.1542/neo.12-1-e2 2011;12;e2-e7 NeoReviews

Dara Brodsky and Lori R. Newman Educational Perspectives: A Systematic Approach to Curriculum Development

http://neoreviews.aappublications.org/cgi/content/full/neoreviews;12/1/e2located on the World Wide Web at:

The online version of this article, along with updated information and services, is

Online ISSN: 1526-9906. Illinois, 60007. Copyright © 2011 by the American Academy of Pediatrics. All rights reserved. by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village,it has been published continuously since 2000. NeoReviews is owned, published, and trademarked NeoReviews is the official journal of the American Academy of Pediatrics. A monthly publication,

. Provided by Loyola Univ Med Ctr on January 5, 2011 http://neoreviews.aappublications.orgDownloaded from

http://neoreviews.aappublications.org/cgi/content/full/neoreviews;12/1/e2

http://http://neoreviews.aappublications.org

Author Disclosure

Drs Brodsky and Ms Newman have

disclosed no financial relationships

relevant to this article. This

commentary does not contain a

discussion of an unapproved/

investigative use of a commercial

product/device.

A Systematic Approach to CurriculumDevelopmentDara Brodsky, MD,* Lori R. Newman, MEd†

AbstractThis review presents a systematic approach to curriculum development,divided into five steps: 1) perform a needs assessment, 2) define the goalsand learning objectives, 3) identify resources, 4) develop educationalstrategies and implement the curriculum, and 5) evaluate and modify thecurriculum. Although the curriculum developmental stages are presentedin five ordered steps, curriculum development is actually a dynamic,interactive process, in which development of one step naturally affectsother steps. Learners are central to this process, and with each step, theinstructor needs to be mindful of the learners’ needs and prior experiences,using a variety of educational strategies to reach trainees with differentlearning styles.

Objectives After completing this article, readers should be able to:

1. Identify key steps in developing a curriculum.2. Distinguish between goals and learning objectives.3. Select learner-centered teaching strategies.4. Recognize the importance of evaluating a curriculum.

IntroductionMedical school curricula receive sub-stantial attention and are largely stan-dardized across the country, butpostgraduate curricula in most hos-pitals are inconsistent, depending onthe ability, interest level, and avail-ability of faculty. By using a system-atic approach to curriculum develop-ment and standardizing teachingpractices, clinician educators can en-sure that their learners master theknowledge and skills necessary to at-tain the next level of training. Thisarticle describes one such systematicapproach: a five-step, easy-to-use

framework for curriculum develop-ment (modified from Kern and asso-ciates). The learner is central to thisprocess, and with each step, the in-structor should be mindful of thelearner’s needs and prior experiences,using a variety of educational strate-gies to reach trainees with differentlearning styles.

Definition of CurriculumA curriculum is any planned educa-tional experience. Clinicians tend tothink of curricula as teaching guide-lines for clinical rotations or entiretraining programs, but the term alsoincorporates planned teaching ses-sions at the bedside or in an outpa-tient clinic. Using this broader defi-nition, most neonatology fellows andneonatologists have experience withcurriculum development activities.

Within this definition of curricu-

*Assistant Professor in Pediatrics, Harvard MedicalSchool, Beth Israel Deaconess Medical Center,Boston, MA.†Director, Office of Faculty Education, Center forEducation, Shapiro Institute for Education atHarvard Medical School and Beth Israel DeaconessMedical Center, Boston, MA.

educational perspectives

e2 NeoReviews Vol.12 No.1 January 2011. Provided by Loyola Univ Med Ctr on January 5, 2011 http://neoreviews.aappublications.orgDownloaded from


lum are four categories. The officialcurriculum is generally a written setof information and skills that learnersmust master by the end of their train-ing experience. The operational cur-riculum is that portion of the officialcurriculum and any additional con-tent that actually is taught. The nullcurriculum category describes theknowledge and skills that are ne-glected or intentionally not taught.Finally, the hidden curriculum cate-gory is the set of values, attitudes,and beliefs embedded in the trainingprogram’s cultural milieu, which isconveyed through verbal and non-verbal cues.

Step 1: Needs AssessmentThe first step in curriculum develop-ment is for the instructor to assess thelearners’ educational needs, whichwill help determine the breadth anddepth of the curriculum. A needs as-sessment is critical for effective cur-riculum development, and althoughit is a simple concept to understand,it can be challenging to execute. Cur-riculum developers must determinewhat the learners already know, whatthey need to know to move to thenext training level, and what infor-mation and skills need to be taughtto fill knowledge gaps and meet edu-cational benchmarks. Methods forassessing the learners’ current knowl-edge and determining their prior ex-periences range from formal writtentests to informal surveys. Curriculumdevelopers should consider the localand national expectations of theirtraining programs to help determinethe educational standards for theirtrainees. Once curriculum develop-ers have identified what the traineesknow and need to learn, they canfocus on the information they needto teach.

Developers should select a needsassessment method based on the typeof curriculum they are planning and

the amount of face-to-face contacttime they will have with the learners.For example, a clinician educatordeveloping a yearlong curriculummight need to use multiple assess-ment methods, such as a knowledgepretest, focus group input, and awritten survey. On the other hand, ifa clinician is planning a single bedsideteaching session, one method, suchas informally asking the trainees whatthey want to learn, should suffice.

To demonstrate how this curricu-lar design process works, we use aclinical scenario throughout this re-view:

An attending in the neonatal in-tensive care unit will be working withthree pediatric residents for a 2-weekrotation. A few weeks before the rota-tion, the Department Chair tells theattending that she has heard a lot ofcriticism from the residents about thecurriculum this year. The traineeshave complained that they are notlearning anything they really need toknow. The Chair asks the attending todesign a new curriculum for the resi-dents.

The attending is already familiarwith the Accreditation Council forGraduate Medical Education’s pedi-atric residency program requirements.Because the rotation is fast approach-ing and there is no time for formaltesting, the attending meets with thechief resident, a group of pediatric res-idents, and other neonatologists to de-termine what the residents typicallyknow at baseline. Based on this infor-mation, the attending determines thatan important subject that the residentsneed to know, but do not consistentlylearn, is how to manage a hospitalizedinfant with respiratory distress.

Step 2: Goals and LearningObjectivesThe second step of curriculum devel-opment is to define the learning goalsand objectives. A learning goal com-

municates the educational aim andpurpose of the instruction. It alsoidentifies the learner group and thescope of the curricular content. Us-ing the scenario, the attending de-cides that his primary learning goalis: To have the residents develop theknowledge and skills necessary to carefor a hospitalized infant with respira-tory distress. In this example, “resi-dents” encompass the learner group,“develop the knowledge and skillsnecessary” is the purpose of the in-struction, and the “hospitalized in-fant with respiratory distress” definesthe scope.

After defining the learning goal,the attending needs to determinethe learning objectives. The learningobjectives describe what the learnerswill be able to know, show, or doat the end of the teaching ses-sion(s). The objectives can be cog-nitive (knowledge-based), psycho-motor (skill and performance-based), and affective (attitudinal).Objectives answer five questions:

● Who?● Will do?● How much?● Of what?● By when?

Thus, a learning objective for thescenario might be: The pediatric resi-dents will intubate five infants whohave respiratory distress by the end oftheir 2-week rotation. This statementanswers the five questions:

● Who?�The pediatric residents● Will do?�will intubate● How much?�five● Of what?�infants who have respi-

ratory distress● By when?�by the end of the 2-week

rotation

Ideally, the learning objectivesshould incorporate the acronym“SMART”: Specific, Measurable,Attainable, Relevant, and Targeted


NeoReviews Vol.12 No.1 January 2011 e3. Provided by Loyola Univ Med Ctr on January 5, 2011 http://neoreviews.aappublications.orgDownloaded from


to the learner. The example learningobjective fits these criteria. The fol-lowing two learning objectives, onthe other hand, are flawed:

1. The pediatric residents will un-derstand the causes of respiratory dis-tress in an infant by the end of therotation. This learning objective isnot measurable because there is noobservable indication of how to mea-sure the residents’ understanding.Because there are many causes of re-spiratory distress in infants, this ob-jective needs to be more specific anddemonstrable. The objective mightbe changed to: The pediatric residentswill summarize the five most commoncauses of respiratory distress in a hospi-talized infant by the end of the rota-tion. This is measurable and demon-strable; at the end of the rotation, theinstructor can assess the residents’acquired knowledge by having themprovide a written or verbal summaryof the causes of respiratory distress.This objective is also specific, aimedat describing the five most commoncauses of respiratory distress in a hos-pitalized infant.

2. The pediatric residents will suc-cessfully intubate an infant with alarge neck mass obstructing the airwayby the end of the rotation. Althoughthis objective is measurable (the res-ident will be able to do it or not) andspecific, it is neither realistic norlikely attainable. An improved objec-tive would be: The pediatric residentswill successfully intubate three infantsby the end of the rotation.

To ensure that the learners areactively engaged with the curricu-lum, find the content meaningful,and can apply and retain new knowl-edge and skills, curriculum designersshould use Bloom’s taxonomy as aguide when creating learning objec-tives. This multi-tiered model classi-fies educational objectives accordingto six cognitive levels of increasingcomplexity: knowledge, comprehen-

sion, application, analysis, synthesis,and evaluation. To engage the learn-ers in higher-order thinking anddeep understanding of concepts,mastery of the lower levels is re-quired. Figure 1 provides examplesof learning objectives for the scenariothat target each cognitive domain.

Step 3: ResourceIdentificationWhen developing a curriculum, in-structors need to identify the people,time, facilities, materials, and fund-ing necessary to build, implement,and sustain the curriculum. Al-though this seems obvious, if curric-ulum designers do not focus on thisstep in advance, they may find thattheir curriculum is complete butlacks the critical resources to imple-ment it. For example, to teach resi-dents how to intubate effectively,the curriculum should involve sometype of simulation-based training.Depending on the resources avail-able, intubation training can rangefrom use of a mannequin head to ahigh-fidelity mannequin that has acardiovascular monitor.

To obtain the necessary resources,curriculum developers need to iden-

tify stakeholders and gain their sup-port. Stakeholders are directly af-fected by a curriculum, such as theDepartment Chair in the scenario.Additional stakeholders include cli-nicians who need to commit theirown time to help teach the curricu-lum. Before approaching the stake-holder, the instructor should knowhow to respond to the often unspo-ken question, “What’s in it for me?”In response, the curriculum devel-oper can reveal a well-developedneeds assessment, well-crafted edu-cational objectives, and possibly anevaluation plan to show how the cur-riculum will be an excellent return onthe stakeholder’s investment.

Step 4: Development ofEducational Strategies andImplementation of theCurriculumEducational strategies are the teach-ing methods and activities used toengage learners actively and enablethem to meet the learning objectives.These methods include:

● Readings from textbooks or jour-nal articles

● Lectures

Figure 1. Bloom’s taxonomy. This multi-tiered model classifies educational objectivesaccording to six cognitive levels of increasing complexity: knowledge, comprehension,application, analysis, synthesis, and evaluation. To the right of the model are examplesof learning objectives that target each cognitive domain.




● Small group or case-based discus-sions

● Simulations with patient-actor,family member-actor, or manne-quins

● Bedside teaching sessions

In addition to connecting theteaching method with the specificcurricular content, curriculum de-signers also must think about thelearner when selecting educationalstrategies. Some trainees prefer tolearn by reading; others may learnbest through discussions or activepractice. To appeal to all types oflearners, curricula should include avariety of teaching methods. In thescenario, the attending might usereadings and small group discussionsto help the learners summarize thefive most common causes of respiratorydistress in a hospitalized infant anduse simulation and bedside teachingto have the residents apply knowledgeof pharyngeal anatomy and intubatean infant.

Each type of educational strategyhas its own strengths and weak-nesses. For example, readings are lowcost, but learners need to be self-motivated to complete their assign-ments. Although lectures offer agreat opportunity to teach to a largenumber of learners simultaneously,this approach can be passive, withminimal interaction between thelearners and teachers. Small-groupdiscussions are ideal for problem-solving and teaching clinicaldecision-making, but success de-pends on the learners’ interest, expe-rience, and knowledge. Realism is atremendous benefit of simulation,but the more realistic the simulation,the higher the cost. Finally, bedsideteaching fosters learner motivationand responsibility but requires inten-sive faculty supervision.

After determining the educationalstrategies and curriculum content,

the instructor can implement thecurriculum. During this step, it iseasy to fall into a teacher-centeredapproach with an overly rigid adher-ence to the preplanned curriculum.It is essential at this point for thecurriculum designer to maintain hisor her commitment to learner-centered learning. Teachers need toguide or facilitate learning andshould consider the following tech-niques:

● Avoid information overload and al-low time for group discussion

● Ask higher-order questions, com-pelling learners to analyze and eval-uate information

● Encourage learner-to-learner dia-logue

● Be flexible and allow the learners’interests, experiences, and needs toguide the direction of teaching

Step 5: Evaluation andModificationEvaluation and modification com-pletes the curriculum developmentmodel. The process of evaluation al-lows the teacher to ask and answerthe critical question, “Was the curric-ulum successful in achieving thelearning goals and objectives?” Eval-uation of a curriculum consists oftwo interconnected domains: learner

evaluation and program evaluation(Fig. 2). Learner evaluation asks thequestion, “Did the learners get it?”and determines whether learners canprovide evidence that they under-stand what they were taught, candemonstrate new skills, and will beable to apply new learning when car-ing for their next patient. Programevaluation asks the question, “Doesthe curriculum work?” This is an-swered by quantitatively measuringthe achievements of all learners, col-lecting feedback about the quality ofthe teaching, and assessing the im-pact of the curriculum on future clin-ical care.

There are many methods to eval-uate the learner. Some examples in-clude:

● Knows: Written examinations, pre-and posttesting

● Shows: Observation by a supervisor,simulation, objective structuredclinical examination, standardizedpatients

● Does: Medical record audit, multi-source feedback, learner portfolio,clinical evaluation exercise

In the scenario, if a learning ob-jective is to have the resident summa-rize the five most common causes ofrespiratory distress in a hospitalized

Figure 2. The questions that encompass the two interconnected domains of curricu-lum evaluation: learner evaluation and program evaluation. Adapted from Tess AV.Introduction to curriculum development and instructional design. In: Principles ofMedical Education. Boston, MA: Beth Israel Deaconess Medical Center; 2009.




infant, the evaluation might includea written or verbal knowledge-basedexamination. If another learning ob-jective is to have the resident applyknowledge of pharyngeal anatomy andintubate an infant, the use of simu-lation or observation of a resident’sintubation procedure is an excellentmethod of learner evaluation.

Program evaluation measures thesuccess of the curriculum by analyz-ing the quantity, quality, and impactof the curriculum. Quantitative eval-uation of the curriculum can includeresults of knowledge-based tests orthe number of learning objectivesthat each resident attained. For ex-ample, in the scenario, measuring theresident’s success rate of intubationbefore and after the curriculum willhelp evaluate the program’s effec-tiveness. Feedback from learners, at-tendings, and clinical team memberson the structure and content of thecurriculum can help to assess the pro-gram qualitatively. In the scenario,improved quality of the program may

be shown by an increase in positiveresident evaluations about the curric-ulum meeting their learning needscompared with precurriculum evalu-ations. To determine if the curricu-lum had an impact on clinical care,the attending in the scenario can col-lect data to see if there are fewerresident-associated complicationsfrom intubation after the curriculumwas introduced.

After compiling the findings fromlearner and program evaluations, theattending can modify and improvethe curriculum. By analyzing thestrengths and weaknesses of the cur-riculum and determining the needsof the next set of learners, curriculumdevelopers can engage in continuousquality improvement as they recon-sider and recalibrate each step of themodel for the next clinical rotation.

ConclusionThe diagram in Figure 3 depicts thesystematic approach to curriculumdevelopment discussed in this article.

The learner is central to this process,reflected within the star model. Dur-ing each step of curriculum develop-ment, clinical teachers need to bemindful of their learners’ needs andprior experiences and the educationalstrategies to engage them. Curricu-lum development is a dynamic, inter-active process, in which developmentof one step naturally affects othersteps. As clinicians move forward intheir own curriculum developmentprocess, we encourage them to in-corporate learner-centered instruc-tion into their everyday interactionswith trainees, to reflect on what as-pects of a curriculum works well andwhat needs to be improved, to bewilling to modify the curriculum tomeet the needs of the learners, and tohave fun teaching.

Suggested ReadingBillings D, Halstead JA. Teaching in Nurs-

ing: A Guide for Faculty. St. Louis, MO:Saunders, Elsevier Science Health Sci-ence; 1998

Bloom B. Taxonomy of Educational Objec-tives. Handbook I: Cognitive Domain.New York, NY: David McKay Com-pany, Inc; 1956

Figure 3. Systematic approach to curriculum development.

American Board of PediatricsNeonatal-Perinatal MedicineContent Specifications• Understand the

strengths andweaknesses ofvarious teachingmethods (eg,lecture, small group discussion,bedside teaching, simulation).

• Understand that individuals may learnmore effectively with certain teachingmethods (eg, reading, hearing, doing)than with others.

• Understand the role of needsassessment in educational planning.

• Distinguish between goals andlearning objectives.

• Identify components of well-formulated learning objectives.




Brodsky D, Huang G. Principles of Teachingand Learning. Computer-Based Mod-ule. Boston, MA: 2010. Available at:http://tinyurl.com/brodsky1

Brodsky D, Martin C. Principles of teachingand learning. In: Neonatology Review.2nd ed. www.lulu.com; 2010

Chandran L, Gusic M, Baldwin C, et al.Evaluating the performance of medicaleducators: a novel analysis tool to dem-onstrate the quality and impact of edu-cational activities. Acad Med. 2009;84:54–66

Eisner E. The Educational Imagination: Onthe Design and Evaluation of School Pro-grams. 3rd ed. New York, NY: Macmil-lan College Publishing; 1994

Green ML. Identifying, appraising, and im-plementing medical education curricula:a guide for medical educators. Ann In-tern Med. 2001;135:889–896

Kern DE, Thomas PA, Howard DM, Bass

EB. Curriculum Development for Medi-cal Education: A Six-step Approach. Bal-timore, MD: Johns Hopkins UniversityPress; 1998

Lake FR, Hamdorf JM. Teaching on therun tips 6: determining competence.Med J Aust. 2004;181:502–503

Ludmerer KM. Time to Heal: AmericanMedical Education from the Turn of theCentury to the Era of Managed Care.Oxford, England: Oxford UniversityPress, Inc; 1999

Mager RF. Preparing Instructional Objec-tives. Belmont, CA: David S. Lake Pub-lishers; 1984

Miller GE. The assessment of clinical skills/competence/performance. Acad Med.1990;65:S63–S67

Quirk ME. How to Learn and Teach in Med-ical School: A Learner-centered Ap-proach. Springfield, IL: Charles CThomas Publishers; 1994

Rodney JW, Peyton JW. Teaching & Learn-ing in Medical Practice. Rickmansworth,UK: Manticore Europe Ltd; 1998

Prideaux D. ABC of learning and teaching:curriculum design. BMJ. 2003;326:268–270

Sheets KJ, Anderson WA, Alguire PC. Cur-riculum development and evaluation inmedical education. J Gen Intern Med.1992;7:538–543

Thomas PA, Kern DE. Internet resourcesfor curriculum development in medicaleducation: an annotated bibliography.J Gen Intern Med. 2004;19:599–605

University of British Columbia Faculty ofMedicine. Teaching Skills for Commu-nity-based Preceptors. Accessed October2010 at: http://www.med.ubc.ca/faculty_staff/faculty_development/educational_material/teaching_skills_booklet.htm





Dara Brodsky and Lori R. Newman Educational Perspectives: A Systematic Approach to Curriculum Development

& ServicesUpdated Information

s;12/1/e2http://neoreviews.aappublications.org/cgi/content/full/neoreviewincluding high-resolution figures, can be found at:

Permissions & Licensing

http://neoreviews.aappublications.org/misc/Permissions.shtmltables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://neoreviews.aappublications.org/misc/reprints.shtml

Information about ordering reprints can be found online:



http://neoreviews.aappublications.org/misc/Permissions.shtml

http://neoreviews.aappublications.org/misc/reprints.shtml



Robin K. Ohls Core Concepts: The Biology of Hemoglobin







Core Concepts: The Biology of HemoglobinRobin K. Ohls, MD*

Author Disclosure

Dr Ohls has disclosed

no financial

relationships relevant

to this article. This

commentary does not

contain a discussion

of an unapproved/

investigative use of a

commercial

product/device.

AbstractA consistent and organized transition from embryonic to fetal to adult hemoglobin(Hgb) occurs during human fetal development. Hgb concentrations gradually in-crease, averaging 18 g/dL (180 g/L) by 40 weeks’ gestation. The ability to deliveroxygen to tissues in the fetus and neonate is primarily determined by the percentage offetal versus adult Hgb and the concentration of 2,3 diphosphoglycerate (2,3-DPG).Studies continue to evaluate the relationship between Hgb concentrations and oxygendelivery in neonates to determine what Hgb concentrations best meet the needs of awide variety of clinical situations from the critically ill extremely low-birthweightinfant to the stable growing preterm infant. Biochemical interactions between nitricoxide (NO) and Hgb beyond the production of methemoglobin do occur and may bea source of deliverable NO to the microcirculation under hypoxic conditions.

Objectives After completing the article, readers should be able to:

1. Describe the development of globin gene synthesis and Hgb formation.2. Explain fetal to neonatal transition of Hgb F to Hgb A.3. Review the development and function of 2,3-DPG.4. Describe the relationship between NO and Hgb.

Hemoglobin Concentration During DevelopmentRed blood cell indices such as Hgb, hematocrit, mean cell volume, and red cell distributionwidth change during gestation and continue to change through the first postnatal year. (1)Hgb concentrations gradually rise during gestation. At 10 weeks’ gestation, the averageconcentration is approximately 9 g/dL (90 g/L); (1)(2) by the start of the third trimester,values in the developing fetus reach 11 to 12 g/dL (110 to 120 g/L); and by 30 weeks,Hgb concentrations are 13 to 14 g/dL (130 to 140 g/L). (1) Jopling and associates (3)have identified reference ranges based on approximately 25,000 preterm and term infants.From 22 to 40 weeks’ gestation, there is a consistent increase in Hgb of 0.21 g/dL(2.1 g/L) per week (Fig. 1). (3) In this large cross-sectional study, no sex differences werenoted in Hgb concentrations, but some studies have reported a slight difference in Hgbconcentrations between white and African American preterm infants. (4) At delivery, a 1-to 2-g/dL (10- to 20-g/L) rise in Hgb may result from transfusion of placental blood intothe infant. In term and late preterm infants, Hgb concentrations increase by approximately4% at 4 hours of postnatal age, resulting from a decrease in plasma volume. (5) By 8 to12 hours of age, Hgb concentrations achieve a relatively constant level. In contrast, in anevaluation of more than 20,000 preterm infants (29 to 34 weeks’ gestation), a decrease ofapproximately 6% was measured at 4 hours of age. (3) This decrease might be due to a lackof placental transfusion because the umbilical cord in preterm infants is rapidly clamped toexpedite resuscitation.

Red blood cell production decreases significantly after birth, primarily as a result of theincreased availability of oxygen, which greatly reduces erythropoietin (Epo) productionand endogenous erythropoiesis. By the end of the first postnatal week, Hgb concentrationsbegin to decline (3) and continue to decrease over the next several weeks as a result ofdecreased erythrocyte production, a shortened erythrocyte life span, and an increase in

*Professor of Pediatrics, University of New Mexico, Albuquerque, NM.

Article core concepts



blood volume related to growth (Fig. 2). Term infantsreach their Hgb nadir at approximately 8 weeks, withan average Hgb concentration of 11.2 g/dL (112 g/L).(6) Hgb concentrations subsequently rise so that by6 months, the concentration averages 12.1 g/dL(121 g/L). (7)

The decline in Hgb in very low-birthweight infants isgreater than that in term infants, in part because ofphlebotomy losses and in part because of the suppressiveimpact of transfusions on endogenous erythropoiesis.Such infants reach the Hgb nadir of 8 g/dL (80 g/L)at 4 to 8 weeks of age. (8) Figure 2 demonstratesrelationships among birthweight, chronologic age, andHgb in term and preterm infants. (3)

Maternal conditions can influence fetal Hgb concen-trations. Infants born small for gestational age can havehigher Hgb concentrations due to placental insufficiencyand secondary polycythemia. (9)(10) Infants of dia-betic mothers, infants of smoking mothers, and infantsborn at higher altitudes also tend to have higher Hgbconcentrations at birth (11)(12)(13)(14) In all of theseconditions, accelerated erythropoiesis is believed to bepart of a compensating mechanism designed to raiseoxygen-carrying capacity to maintain an adequate oxy-gen supply to the fetus. In the case of the fetus of amother who has diabetes, increased metabolic demandsof the fetus (as evidenced by a positive correlation be-tween maternal Hgb A1c and neonatal Hgb) may ac-

count for the higher oxygen needsand the compensatory increase inerythropoiesis and Hgb concentra-tion. (15)

Hemoglobin SynthesisDuring fetal erythropoiesis, an or-derly evolution of the productionof various Hgbs occurs. Eight glo-bin genes direct the synthesis of sixdifferent polypeptide chains, desig-nated alpha (�), beta (�), gamma(�), delta (�), epsilon (�), and zeta(�). These globin chains combinein the developing erythroblast toform seven different Hgb tetramers:Gower 1 (�2-�2), Gower 2 (�2-�2),Portland (�2-�2), fetal hemoglobin(Hgb F: �2-�2), and two types ofadult hemoglobin: �2-�2, known asHgb A, and �2-�2, known as HgbA2 (Table 1).

Globin GenesThe globin genes are organized into two clusters(Fig. 3). The �-like genes are located along a 20-kb distalsegment of the short arm of chromosome 16. The clustercontains three functional genes (�1, �2, and �2), threepseudogenes (evolutionary remnants of genes that arenot expressed because of inactivating mutations thatprevent production of a functional globin protein), andone gene of undetermined function (a globin-like genewithout inactivating mutations). The �-like gene clusteris located along a 60-kb segment of the short arm ofchromosome 11, and it contains five functional genes(�, �, A�, G�, and �) and one pseudogene. Within eachcomplex, the genes are all in the same 5� to 3� orienta-tion, and they are arranged in the order in which they areexpressed during development. (16)

Timing of Globin Chain SynthesisGlobin chain production has been determined at eachstage of development, from initial yolk sac (primitive) tohepatic (definitive) and marrow erythropoiesis. It is notclear why or how primitive erythroid progenitors pro-grammed to produce embryonic Hgb transition to de-finitive progenitors programmed to produce Hgb F.Because quantification of globin gene expression usingreal time polymerase chain reaction methods reflectsproduction by a heterogeneous source of erythrocytes,production of a specific Hgb is usually reported as a

Figure 1. Reference ranges for blood hemoglobin concentrations at birth in 24,416patients at 22 to 42 weeks’ gestation. The solid line represents the mean value and thedashed lines represent the 5% to 95% reference range. Reprinted with permission fromJopling. (3)

core concepts hemoglobin



percentage of the total Hgb measured. Studies measur-ing Hgb production by erythroid colonies in cultureshow that individual cells in a colony produce predomi-nantly one type of Hgb. (17)

During the fourth to fifth week of gestation, �, �, and� chains are the primary globin chains produced (Fig. 4).

During the sixth to seventh week ofgestation, primitive erythroblastscontinue to produce �, �, and �

chains, while definitive erythrocytesproduce �, �, G�, and A� chains. Bythe seventh to eighth week, �- and�-chain synthesis is no longer de-tectable, and the primary globinchains produced are �, G�, and A�.�-chain production is just barelydetectable at this time and gradu-ally increases, comprising up to 10%of total non–�-chain production by10 weeks of gestation. (18) Geneticdisorders associated with �-chainsynthetic or structural abnormali-ties may be detected in utero assoon as �-chain production occursbut are often not clinically apparentuntil after birth.

From 10 to 33 weeks of gesta-tion, the primary globin chains syn-thesized are �, G�, A�, and �. As-sessment of the output of the twolinked �-globin genes by mRNAanalysis suggests that they are ex-pressed in the ratio (�2:�1) rang-ing from 1.5 to 3.0:1 throughoutfetal life, and this ratio continuesthrough normal adulthood. Therelative rates of G�-chain and A�-chain production are also constantthroughout fetal life at a G�:A� ratioof approximately 3:1. (19) Duringthe transition from 32 to 36 weeksof gestation, the relative rate of�-chain synthesis increases and thatof �-chain production declines, soat birth, �-chain synthesis makesup approximately 50% of non–�-chain synthesis. There is consider-able variation among infants, how-ever, with many infants showingprolonged dependence on Hgb F.After birth, the level of �-chain pro-

duction increases, while the level of �-chain productionsteadily declines, so by the end of the first year, �-chainsynthesis reaches the low concentration that is character-istic of adult life (�2%). Over the first few months afterbirth, the G�:A� ratio changes from 3:1 to 2:3, althoughthis ratio varies in adults. (20)(21)(22)

Figure 2. Reference ranges for blood hemoglobin concentrations in 39,559 patientsduring the 28 days after birth in late preterm and term infants 35 to 42 weeks’ gestation(panel A) and in preterm infants 29 to 34 weeks’ gestation (panel B). The solid linerepresents the mean value and the dashed lines represent the 5% to 95% reference range.Reprinted with permission from Jopling. (3)




Delta-chain production has been observed as early as32 weeks’ gestation. Delta-gene production lags behind�-gene production, so the adult �/� synthesis ratio is notreached until 4 to 6 months after birth.

Hemoglobin Production During DevelopmentDevelopmental changes in the production of Hgb can beseen in Figure 5. Before the onset of other chain forma-tion, unpaired globin chains may form tetramers, result-

ing in the presence of �4. (23) Soon thereafter, �- and�-chain production begins, and Hgb Gower 1 (�2-�2),Gower 2 (�2-�2), and Portland (�2-�2) are formed. (24)By 5 to 6 weeks’ gestation, Hgb Gower 1 and Gower 2constitute 42% and 24% of the total Hgb, respectively,with Hgb F (�2-�2) making up the remainder. By 14 to16 weeks, Hgb F constitutes 50% of the total Hgb, andby 20 weeks, it forms more than 90% of the Hgb. (25)Small quantities of Hgb A (�2-�2) are found beginningat 6 to 8 weeks’ gestation. The increase in �-chainproduction occurring between 12 and 20 weeks’ gesta-tion accounts for the sudden rise in Hgb A found atthe end of the first trimester of pregnancy. Tetramers of�-chains (�4 or Hgb Barts) and �-chains (�4 or Hgb H)can be found in conditions in which �-chain synthesis isimpaired or absent, such as �-thalassemia syndromes.

Hgb F is easily distinguished immunologically and bio-chemically from Hgb A. The primary differentiating physi-ologic characteristic of Hgb F is its decreased interactionwith 2,3-DPG (also known as 2, 3 bisphosphoglycerate).2,3-DPG binds to deoxyhemoglobin in a cavity betweenthe �-chains and stabilizes the deoxy form of Hgb, resultingin reduced Hgb-oxygen affinity. 2,3-DPG binds less effec-tively to �-globin chains because of the differing amino acid

sequence in the non–�-chain. Con-sequently, 2,3-DPG does not reducethe oxygen affinity of Hgb F as muchas that of Hgb A.

Fetal and adult Hgb also differ insolubility. Hgb F is more soluble instrong phosphate buffers than Hgb Aand is oxidized to methemoglobinmore easily than Hgb A. Hgb F has aconsiderably greater affinity for oxy-gen as a result of differences in bind-ing to 2,3-DPG mentioned previ-ously. Hgb F is resistant to acidelution, which allows differentiationof cells containing Hgb F from cellscontaining Hgb A. (26) This prop-erty forms the basis of differentiatingfetal from maternal red cells using theKleihauer Betke stain.

G�-chains represent 70% to 80%of the total �-chains in the blood ofthe fetus and newborn. The percentof �-chains made up of G� falls toabout 40% by 5 months of age. Thisunique difference in G�-chain pro-duction found in the fetus helps todistinguish fetal hematopoiesis from

Figure 3. Organization of the globin genes. Transcription takes place from the 5� to the 3�end; for both chromosomes, the genes are arranged in order of their developmental activation.The upper part of the figure represents the beta-like globin genes on the short arm ofchromosome 11, and the lower part of the figure represents the alpha-like genes on the distalshort arm of chromosome 16. Regions of the gene that code for primary globin proteins areshown as shaded ovals. Regions that code for pseudogenes (y-nonexpressed remnants thathave a number of inactivating mutations that prevent transcription and translation intofunctional globin protein) are shown as open ovals. �-1 is a globin-like gene withoutinactivating mutations. The locus control region (LCR) is shown as a hatched segment.

Table 1. Globin Chain Compositionof Common Hemoglobin

Type of Hemoglobin Composition

Embryonic Hemoglobin:Gower 1 �2-�2Gower 2 �2-�2Portland �2-�2

Fetal Hemoglobin:F �2-�2

Adult Hemoglobin:A �2-�2A2 �2-�2




that found in later life. Under stress, the older infant andadult increase production of Hgb F. Increased Hgb Fproduction often occurs in leukemic states and in otherconditions. (27)(28) The delay in the switch of Hgb F toHgb A has been noted in conditions of maternal hypoxia, ininfants who are small for gestational age, (29) and in infants

of diabetic mothers. (30)(31) Ele-vated concentrations of Hgb F mayhave protective effects in some dis-ease states. For example, a high con-centration of fetal Hgb F in patientswho have sickle cell disease may be apredictor of increased adult life ex-pectancy. (32)(33)

On a body weight basis, red bloodcell production during the lattermonths of gestation is significantlygreater compared with that in adultlife. Immediately after birth, erythro-poiesis is considerably reduced, asso-ciated with a steady and linear declinein �-chain synthesis during the pe-riod of reduced neonatal erythropoi-esis. Newly synthesized red blood cellsappearing in the circulation whenerythropoiesis resumes contain pre-dominantly Hgb A. The postpartumdecline of Hgb F production and of

the intercellular distribution of fetal and adult Hgb overthe first postnatal months has been extensively examined.Immediately after birth, there is a brief rise in Hgb F,followed by a steady decline. Studies of the intercellulardistribution of Hgb F (using an acid-elution technique)have shown that the distribution of Hgb F is heterogeneous

over the first few months after birth.At 3 months, the distribution ofHgb F becomes bimodal, with pop-ulations of cells that contain acid-resistant Hgb F and populations ofcells containing Hgb A. These obser-vations have suggested that HgbF-containing cells are replaced by apopulation of cells containing Hgb Aduring the early postnatal period.

Studies show that the type ofglobin chains produced at differentstages of development are not closelyrelated to the site of erythropoiesis.It appears that �- and �-chains aresynthesized in both primitive anddefinitive cell lines. Moreover, dur-ing the later stages of fetal develop-ment, the switch from �- to �-chainproduction occurs synchronouslythroughout the liver and bonemarrow. The transition from �- to�-chain synthesis is most closely re-

Figure 4. Production of globin chains during the fetal and neonatal period.

Figure 5. Hemoglobin production during the fetal and neonatal period.




lated to postconceptional age and not chronologic age.(27) Thus, infants born preterm continue to synthesizesignificant amounts of �-globin (and fetal Hgb) until40 weeks’ gestation.

2,3-DPG MetabolismThe affinity of Hgb for oxygen can be decreased byinteraction with certain organic phosphates, such as2,3-DPG and adenosine triphosphate. (34) The highlycharged anion 2,3-DPG binds to deoxyhemoglobin butnot to oxyhemoglobin. Deoxyhemoglobin F does notpossess as great an affinity for 2,3-DPG as does deoxy-hemoglobin A and, therefore, cannot bind 2,3-DPG tothe same degree as Hgb A. Thus, the fetal leftward-shifted Hgb oxygen dissociation curve (Fig. 6) is noteasily modulated in the presence of 2,3-DPG.

The P50 (partial pressure of oxygen at which half ofHgb is saturated) of fetal blood is 19 to 21 mm Hg, some6 to 8 mm Hg lower than that of adult blood. As Hgb Fconcentration declines after birth, however, there is amarked rightward shift in the postnatal Hgb oxygenequilibrium curve. The percentage of Hgb A and the redcell 2,3-DPG content play the greatest roles in alteringthe position of the Hgb oxygen dissociation curve. As aresult, preterm infants who have a greater proportion ofHgb A but less 2,3-DPG (which occurs following packedred blood cell transfusion) may have a similar P50 as thosewho have increased quantities of Hgb F.

Certain factors are known to alter the affinity of Hgb foroxygen (Table 2). The most important of these are the HgbF concentration and the red cell 2,3-DPG content. Theconcentration of red blood cell 2,3-DPG gradually in-creases with gestation. At term, the concentration is similarto that of adults. By the end of the first postnatal week, the2,3-DPG concentrations are considerably higher than theyare at birth. After the first week, red blood cell 2,3-DPGconcentrations remain relatively unchanged for the next6 months. In term infants, the Hgb-oxygen dissociationcurve gradually shifts to the right, and by 4 to 6 months ofage, the P50 values approximate those of the adult.

The situation is somewhat different in preterm in-fants. Because Hgb F synthesis is still active, increases inP50 seen in term infants as a result of the switch fromHgb F to Hgb A do not occur as rapidly. The red bloodcell 2,3-DPG concentrations also are slightly lower inpreterm infants. (35) These concentrations are increasedwith the use of human recombinant Epo, which shifts theoxygen dissociation curve to the right. (36)(37)

Nitric Oxide-hemoglobin InteractionsNO plays a significant role in vasoactive regulation.Under baseline conditions, NO is produced by endo-thelial NO synthase and diffuses into surrounding smoothmuscle cells, activating soluble guanylyl cyclase to producecyclic guanosine 5�-monophosphate, and regulates vasculartone. NO reacts with oxyhemoglobin to form methemo-

globin, which is reversed by erythro-cytic methemoglobin reductase. Asecond reaction also can occur, inwhich NO reacts with deoxyhemo-globin to form nitrosyl hemoglobin(NO-Hgb). There is some evidencethat erythrocytes containing NO-Hgb may be able to release NO intothe circulation, thus causing vasodi-latation in the microvasculature. Athird reaction has been studied thatinvolves the binding of NO to the�-chain cysteine amino acid to formS-nitrosyl-hemoglobin (SNO-Hgb).It has been postulated that nitriteions within erythrocytes can be re-duced to NO by deoxyhemoglobin,so NO is generated as erythrocytesenter hypoxic regions. All of thesepotential mechanisms result in NOcontrolling blood flow via hypoxicvasodilatation. These mechanismsare especially important in preterm

Figure 6. Hemoglobin-oxygen dissociation curve. The curve representing fetal hemoglo-bin is on the left, and the curve representing adult hemoglobin is on the right. The P50 isshown as a hatched line for each curve.




infants because vulnerable vascular beds such as the splanch-nic system are at risk for hypoxic injury via shunting ofblood to the brain and heart. Blood that is collected fortransfusion loses its SNO-Hgb within hours, so transfusedpacked red blood cells (PRBCs) lack NO. TransfusingPRBCs that cannot deliver NO to the microvasculaturemay, in fact, reduce local oxygen delivery to tissues throughvasoconstriction. Studies are in progress to evaluate therenitrosylization of Hgb before transfusion.

Oxygen TransportThe mechanisms controlling oxygen transport in uteroand immediately postpartum are complex. During pre-natal life, the fetal arterial oxygen tension (PO2) is ap-proximately 30 mm Hg, and the venous PO2 is approxi-mately 15 mm Hg (Fig. 6). These low PO2s contribute tothe development of relative polycythemia in the fetus.After birth, numerous factors affect oxygenation, includ-ing the inspired gas mixture, pulmonary function, thearterial oxygen dissociation curve, and the ability toextract oxygen at the tissue level. (38) It has been spec-ulated that the actual amount of oxygen released to

tissues may be greater in utero because of the character-istics of the Hgb-oxygen dissociation curve.

If pulmonary function is normal, the PO2 of pulmo-nary blood rises from the 40 mm Hg of pulmonaryarterial blood to the 100 mm Hg of pulmonary venousblood. Because of the shape of the Hgb-oxygen dissoci-ation curve, these PO2 values permit 95% saturation ofHgb by oxygen. Further increases in PO2 produce littleadditional increase in saturation. In the healthy adult,approximately 50% of Hgb is saturated with oxygenwhen the PO2 falls to 27 mm Hg (P50�27 mm Hg). Insituations in which the Hgb-oxygen dissociation curvehas shifted to the right, the affinity of Hgb for oxygen isreduced. Thus, at any given PO2, more oxygen is releasedto tissues. Conversely, if the curve is shifted to the left,the affinity of Hgb for oxygen is increased. Therefore, atany given PO2, less oxygen is released to the tissues.

A precise relationship does not exist between the declinein Hgb F and the decrease in oxygen affinity of a neonate’sblood. (38) Rather, changes in P50 reflect the interactionbetween red blood cell 2,3-DPG, the increase in Hgb Athat occurs after birth, and the decline in Hgb F. Althoughoxygen-carrying capacity (Hgb concentration � oxygensaturation � 1.36 mL oxygen/g of Hgb) decreases overthe first few postnatal months as Hgb concentration de-clines, the amount of oxygen delivery can remain similar oreven increase. (35) For example, a preterm infant born witha Hgb concentration of 15 g/dL (150 g/L) delivers 1 mLof oxygen to the tissues for every 100 mL of circulatingblood (based on a P50 of 19 and a central venous PO2 of40 mm Hg). As the percent of Hgb A increases over time,the P50 shifts to the right. The infant can now deliver2.1 mL of oxygen per 100 mL of blood, despite a decreasein total Hgb to 8 g/dL (80 g/L) (based on a P50 of 24 mmHg and a central venous PO2 of 40 mm Hg). (35)

After intrauterine transfusion, infants have oxygen-unloading properties characteristic of adult blood. Despitethe decrease in oxygen affinity that accompanies intrauter-ine transfusion, no deleterious effects of this procedure withrespect to oxygen uptake by the fetus have been docu-mented. (39) The physiologic significance of manipulatingthe Hgb-oxygen affinity of extremely preterm infants viaPRBC transfusions continues to be studied. It is importantto understand an infant’s ability to deliver oxygen to tissueswhen determining whether to administer an erythrocytetransfusion. The decision to transfuse should not be basedon Hgb concentration alone. Transfusions significantly af-fect an infant’s endogenous erythropoiesis: for infants whoundergo exchange transfusion or multiple transfusions,both Epo concentrations and reticulocyte counts are lowerat any given Hgb concentration. (8)(40) The search con-

Table 2. Factors AffectingHemoglobin-Oxygen AffinityIncreased P50, increased red blood cell 2,3-DPG:

● Adaptation to high altitude● Hypoxemia associated with chronic pulmonary disease● Hypoxemia associated with cyanotic heart disease● Anemia● Decreased red blood cell mass● Hyperthyroidism● Red cell pyruvate kinase deficiency

Increased P50, no consistent alteration in red blood cellDPG:

● Abnormal hemoglobins (Kansas, Seattle,Hammersmith, Tacoma, E)

● Vigorous exercise

Decreased P50, decreased red blood cell 2,3-DPG:

● Septic shock● Severe acidosis● Following massive transfusions of stored blood● Neonatal respiratory distress syndrome

Decreased P50, no consistent alteration in red blood cellDPG:

● Abnormal hemoglobins (Kempsey, Chesapeake,Capetown, Yakima, Rainier)

2,3-DPG�2,3 diphosphoglycerate, P50�partial pressure of oxygen atwhich half of hemoglobin is saturated




tinues for a specific marker that reflects the need for im-proved oxygen delivery to tissues (via red blood cell trans-fusion). Currently, no ideal marker that is simple, requireslittle or no blood, is reproducible, and can be applied topreterm infants exists for clinical use in neonates.

SummaryThe organized transition from embryonic to fetal to adultHgb has been extensively studied, providing significantunderstanding of the molecular basis of Hgb development.Studies continue to evaluate the relationship between Hgbconcentrations and oxygen delivery in neonates to bestdetermine what Hgb concentrations best meet the needs ofa wide variety of clinical situations from the critically illextremely low-birthweight infant to the stable growingpreterm infant. Studies are underway to explore the mech-anisms linking Hgb metabolism and the transfer of NO byerythrocytes, and these studies have the potential to addgreatly to the body of evidence regarding transfusion guide-lines in various neonatal populations.

ACKNOWLEDGMENTS. I wish to thank Rebecca Mo-ran, MD, and Andrea Duncan, MD, for their thoughtfulreview and comments and Ann Chavez for her assistancein completing the manuscript.

References1. Forestier F, Daffos F, Catherine N, Renard M, Andreux JP.Developmental hematopoiesis in normal human fetal blood. Blood.1991;77:23602. Bratteby L. Studies on the erythro-kinetics in infancy: red cellvolume of newborn infants in relation to gestational age. ActaPaediatr Scand. 1968;57:1323. Jopling J, Henry E, Wiedmeier SE, Christensen RD. Referenceranges for hematocrit and blood hemoglobin concentration duringthe neonatal period: data from a multihospital health care system.Pediatrics. 2009;123:e333–e3374. Alur P, Devapatla SS, Super DM, et al. Impact of race andgestational age on red blood cell indices in very low birth weightinfants. Pediatrics. 2000;106:306–3105. Usher R, Shephard M, Lind J. The blood volume of a newborninfant and placental transfusion. Acta Paediatr Scand. 1963;52:4976. Oski FA. The erythrocyte and its disorders. In: Oski FA, Nathan

DG, eds. Hematology of Infancy and Childhood. Philadelphia, PA:WB Saunders; 1993:18–437. Kling PJ, Schmidt RL, Roberts RA, Widness JA. Serum eryth-ropoietin levels during infancy: associations with erythropoiesis.J Pediatr. 1996;128:7918. Stockman JA 3rd, Garcia JF, Oski FA. The anemia of prematu-rity: factors governing the erythropoietin response. N Engl J Med.1977;296:6479. Humbert JR, Abelson H, Hathaway WE, Battaglia FC. Poly-cythemia in small for gestational age infants. J Pediatr. 1969;75:81210. Hakanson DO, Oh W. Hyperviscosity in the small for gesta-tional age infant. Pediatr Res. 1977;11:472A11. Moore LG, Newberry MA, Freeby GM, Crnic LS. Increasedincidence of neonatal hyperbilirubinemia at 3100 m in Colorado.Am J Dis Child. 1984;138:15812. Bureau MA, Shapcott D, Berthiaumey, et al. Maternal ciga-rette smoking and fetal oxygen transport: a study of P50, 2,3-diphosphoglycerate, total hemoglobin, hematocrit, and type F he-moglobin in fetal blood. Pediatrics. 1983;2:2213. Matoth Y, Zaizove R, Varsano I. Postnatal changes in some redcell parameters. Acta Paediatr Scand. 1971;60:31714. Gonzales GF, Steenland K, Tapia V. Maternal hemoglobinlevel and fetal outcome at low and high altitudes. Am J Physiol RegulIntegr Comp Physiol. 2009;297:R1477–R148515. Nelson SM, Freeman DJ, Sattar N, Lindsay RS. Erythrocytosisin offspring of mothers with type 1 diabetes—are factors other thaninsulin critical determinants? Diabet Med. 2009;26:887–89216. Weatherall DJ, Wood WG, Jones RW, Clegg JB. The develop-mental genetics of human hemoglobin. In: Stamatoyannopoulos G,Nienhuis AW, eds. Experimental Approaches for the Study of Hemo-globin Switching. New York, NY: Alan R Liss; 1985:3–2517. Papayannopoulou T, Kurachi S, Brice M, Nakamoto B, Stama-toyannopoulos G. Asynchronous synthesis of HbF and HbA duringerythroblast maturation. II. Studies of G gamma, A gamma, andbeta chain synthesis in individual erythroid clones from neonataland adult BFU-E cultures. Blood. 1981;57:53118. Kleihauer E. The hemoglobins. In: Stave U, ed. Physiology of thePerinatal Period. Vol 1. New York, NY: Appleton-Century-Crofts;1970:25519. Masala B, Manca L, Formato M, Pilo G. A study of the switchof fetal hemoglobin and newborn erythrocytes fractionated bydensity gradient. Hemoglobin. 1983;7:56720. Schroeder WA, Shelton JR, Shelton JB, Apell G, Huisman TH,Bouver NG. Worldwide occurrence of nonallelic genes for the gamma-chain of human foetal haemoglobin in newborns. Nature. 1972;240:27321. Schroeder WA, Huisman THJ, Brown AK, et al. Postnatalchanges in chemical heterogeneity of human fetal hemoglobin.Pediatr Res. 1971;5:47322. Schroeder WA, Huisman THJ. Human gamma-chains: struc-tural features. In: Stamatoyannopoulos G, Nienhuis AW, eds. Cel-lular and Molecular Regulation of Hemoglobin Switching. NewYork, NY: Grune & Stratton; 1979:29–4523. Szelengi JG, Holland SR. Studies on the structure of humanembryonic hemoglobin. Acta Biochim Biophys Acad Sci Hung. 1969;4:4724. Hecht F, Motulsky AG, Lemire RJ, Shepard TE. Predomi-nance of hemoglobin Gower 1 in early human embryonic develop-ment. Science. 1966;152:9125. Heizman THJ, Schroeder WA, Brown AK, Hyman CB, Or-tega JA, Sukumaran PK. Further studies of the postnatal change inchemical heterogeneity of human fetal hemoglobin in several ab-normal conditions. Pediatr Res. 1975;9:1

American Board of Pediatrics Neonatal-PerinatalMedicine Content Specifications• Know the biochemical characteristics of

fetal hemoglobin.• Know the developmental biology of

hemoglobin types.• Know normal erythropoiesis in the fetus

and neonate.




26. Kleihauer E, Braun H, Betke K. Demonstration of fetal hemoglo-bin in the erythrocytes of a blood smear. Klin Wochenschr. 1957;35:63727. Bard H, Lachance C, Widness JA, Gagnon C. The reactivationof fetal hemoglobin synthesis during anemia of prematurity. PediarRes. 1994;36:25328. Bard H, Fouron JC, Gagnon C, Gagnon J. Hypoxemia andincreased fetal hemoglobin synthesis. J Pediatr. 1994;124:94129. Bard H. The effect of placental insufficiency on fetal hemoglobinand adult hemoglobin synthesis. Am J Obstet Gynecol. 1974;120:6730. Bard H, Prosmanne J. Relative rates of fetal hemoglobin andadult hemoglobin synthesis in the cord blood of infants of insulin-dependent diabetic mothers. Pediatrics. 1985;75:114331. Perrine SP, Greene MF, Faller DV. Delay in fetal hemoglobinswitch in infants of diabetic mothers. N Engl J Med. 1985;312:33432. Platt OS, Brambilla DJ, Rosse WF, et al. Mortality in sickle celldisease. Life expectancy and risk factors for early death. N EnglJ Med. 1994;330:163933. Trompeter S, Roberts I. Haemoglobin F modulation in child-hood sickle cell disease. Br J Haematol. 2009;144:308–31634. Benesch R, Benesch RE, Yu CL. Reciprocal binding of oxygen

and diphosphoglycerate by human hemoglobin. Proc Natl Acad SciU S A. 1968;59:52635. Delivoria-Papadopoulos M, Roncevic NP, Oski FA. Postnatalchanges in oxygen transport of term, preterm and sick infants: the roleof red cell 2,3 diphosphoglycerate in adult hemoglobin. Pediatr Res.1971;5:23536. Soubasi V, Kremenopoulos G, Tsantal C, Savopoulou P, Mus-safiris C, Dimitriou M. Use of erythropoietin and its effects on bloodlactate and 2, 3-diphosphoglycerate in premature neonates. Biol Neo-nate. 2000;78:28137. Debska-Slizien A, Owczarzak A, Lysiak-Szydlowska W, RutkowskiB. Erythrocyte metabolism during renal anemia treatment with recombi-nant human erythropoietin. Int J Artif Organs. 2004;27:935–94238. Stockman JA. Anemia of prematurity: current concepts in theissue of when to transfuse. Pediatr Clin North Am. 1986;33:11139. Novy MJ, Frigoletto FD, Easterday CL, Umansky I, Nelson NM.Changes in umbilical-cord blood oxygen affinity after intrauterinetransfusions for erythroblastosis. N Engl J Med. 1971;285:58940. Oski FA, Stockman JA. Anaemia in early infancy. Br J Haema-tol. 1974;27:195




NeoReviews Quiz

9. Red blood cell production decreases after birth, primarily as a result of increased availability of oxygen,which greatly reduces erythropoietin production and endogenous erythropoiesis. A decrease in hemoglobinconcentration follows this reduced red blood cell production. Of the following, the hemoglobin nadir inhealthy term infants is reached at a postnatal age closest to:

A. 8 weeks.B. 12 weeks.C. 16 weeks.D. 20 weeks.E. 24 weeks.

10. Hemoglobin consists of heme, an iron-containing protoporphyrin, and globin, a polypeptide. Eight globingenes direct the synthesis of six different polypeptide chains, designated as alpha (�), beta (�), gamma(�), delta (�), epsilon (�), and zeta (�). These globin chains combine in the developing erythroblast toform seven different hemoglobin tetramers: Gower 1 (�2-�2), Gower 2 (�2-�2), Portland (�2-�2), fetalhemoglobin (Hgb F: �2-�2), adult hemoglobin (Hgb A: �2-�2), and adult hemoglobin A2 (Hgb A2: �2-�2).Of the following, the most prevalent hemoglobin tetramer in the fetus at 18 weeks of gestational age is:

A. Gower 1.B. Hgb A.C. Hgb A2.D. Hgb F.E. Portland.

11. A term infant is born with severe anemia. A Kleihauer Betke test is performed on maternal blood todetermine whether fetomaternal hemorrhage is the cause. Of the following, the property of fetalhemoglobin that best differentiates fetal from maternal red blood cells using the Kleihauer Betke test isthat the fetal hemoglobin, relative to adult hemoglobin, is/has:

A. Decreased interaction with 2,3-diphosphoglycerate.B. Greater affinity for oxygen.C. Greater solubility in strong phosphate buffer.D. Readily oxidized to methemoglobin.E. Resistant to acid elution.





Robin K. Ohls Core Concepts: The Biology of Hemoglobin













Maureen E. Sims Legal Briefs: Neonatal Anemia at Birth







Author Disclosure

Dr Sims has disclosed no financial

relationships relevant to this

discussion. This commentary does not

contain a discussion of an

unapproved/investigative use of a

commercial product/device.

Neonatal Anemia at BirthMaureen E. Sims, MD*

A Subacute FetomaternalHemorrhageA 2,980-g 35-weeks’ gestation fe-male infant is delivered to a 30-year-old woman whose pregnancy wascomplicated by chronic hyperten-sion. Four days before delivery shehad a clinic visit, and fetal ultra-sonography performed at that timewas read as normal. Two days later(2 days before delivery), the womancalled the clinic because she beganexperiencing chills, a “crampy feel-ing,” and loose stools. She reportedgood fetal movement.

Several hours later, she visitedthe emergency department becauseof additional symptoms of nausea,myalgia, back pain, and vomiting.During this visit, her temperaturewas 38.9°C. The long-term variabil-ity on a nonstress test (NST) was re-ported as fair, but one spontaneousdeceleration was noted. The womanreported that fetal movements de-creased from earlier in the day. (Themonitoring strip was missing andunavailable for review at the time oflitigation.) Complete blood countfindings were unremarkable, and uri-nalysis showed 0 to 3 red blood cellsand a large amount of hemoglobin.The plaintiff experts said the physi-cian caring for the woman shouldhave considered a transfusion reac-tion from the fetus because of the he-moglobinuria, the woman’s symp-toms, and the report of decreasedfetal movement. A more detailedevaluation should have been per-formed, including a biophysical pro-file (BPP) and a Kleihauer-Betketest (KB). The box on the paperworkfor “fetal movement” was left blank.

The woman said in her depositionthat she shared with the nurses thatfetal movement was reduced. Theplaintiff experts said this was the fe-tus’ attempt to conserve energy be-cause of the anemia. The defense ex-perts said that the reduction in fetalmovement meant that the damagealready had occurred. The motherwas discharged.

Later that day, the woman visitedher obstetrician and had a repeatNST. The obstetrician read the NSTas reactive, although he stated on hisdeposition that it took more than60 minutes to become “somewhatreactive.” The woman reported de-creased fetal movements during thisvisit. The plaintiff obstetric expertssaid the NST was nonreactive andnot only was the misinterpretation ofthe NST below the standard of care,but an additional violation was notto have pursued additional testingsuch as a BPP and a KB (in light ofthe hemoglobinuria on the previousday).

The following day, the woman re-turned to her obstetrician because ofcontinued decreased fetal move-ment. The NST was repeated andwas interpreted by the treating obste-trician as nonreassuring. The firsthour showed a flat baseline with aspontaneous deceleration. By thenext hour, a sinusoidal pattern ap-peared. A BPP showed 2/10 (2 foramniotic fluid and 0 for tone, reac-tive heart rate, movement, andbreathing). The obstetrician wasconcerned about a concealed abrup-tion or a fetomaternal hemorrhage.During the next hour, a KB showed1% fetal cells and estimated a bloodloss from the fetus of 50 mL. Onehour later, a cesarean section was per-

*Professor of Pediatrics, UCLA School of Medicine,Los Angeles, CA.

legal briefs



formed. The treating obstetrician(defendant) said he saw no reasonfor an emergency cesarean section.The plaintiff obstetric experts saidthat much time was wasted and thecesarean section should have beenperformed much sooner. Further-more, the lengthy evaluation kept thefetus in a hostile environment un-necessarily longer. Arterial and ve-nous cord gases showed pH values of7.25 and 7.38, respectively. The de-fense pointed out that the essentialcriteria to diagnose intrapartumhypoxia were not met because the ar-terial cord blood gases were normal.(1) The plaintiff experts questionedthe motives behind the conclusions inthe American College of Obstetri-cians and Gynecologists (ACOG)statements. The plaintiff experts fur-ther explained that the placentacleared the lactic acid produced bythe fetal cells, which were undergo-ing anaerobic metabolism due toanemic hypoxia, and that the fetallactic acid was cleared by the moth-er’s liver.

A profoundly pale infant was bornwith Apgar scores of 6 at 1 minuteand 8 at 5 minutes. The defense ex-perts pointed out that the baby was ingood clinical condition, as evidencedby the good Apgar scores, and quotedthe ACOG consensus criteria. (1)The plaintiff experts pointed out thatin a subacute condition in whichfluid adjustments are made, the Ap-gar scores can be good because thecardiovascular system is still intact.Blow-by oxygen was provided andthe infant was transported to thenewborn intensive care unit. On ad-mission, the infant’s temperature was34.9°C, heart rate was 150 beats/min, blood pressure was 57/34 mmHg (with a mean of 39 mm Hg), andoxygen saturation was 100% on 25%oxygen via hood. The defense expertsmaintained that the baby’s normalblood pressure, saturation, and heart

rate supported her normalcy at birth.The plaintiff experts explained thatthe baby was not hypovolemic at birthbecause of the subacute nature of thetransfusion that allowed fluid shiftsin utero. They further explained thata saturation of 100% was not reflec-tive of the level of tissue oxygenation.The inadequate numbers of red bloodcells produced tissue hypoxia.

Umbilical lines were inserted andcompleted by 37 minutes of age, andan arterial blood gas showed a pH of6.99, PCO2 of 36 mm Hg, PO2 of50 mm Hg, and base deficit of�20 mEq/L. The plaintiff expertspointed out that, considering the pro-found paleness of this infant, a moreappropriate approach would havebeen to insert an umbilical venousline emergently, obtain a venous gasand hematocrit (Hct), and immedi-ately start a partial exchange withemergency-release O-negative blood.The Hct was 11% (0.11), hemoglo-bin was 3.7 g/dL (37 g/L), andnucleated red blood cells (nRBCs)were 55/100 white blood cells. Theplaintiff experts pointed out that theestimated blood volume of this babywas 253 mL (85% of her birth-weight). She must have lost a total ofapproximately 80% of her blood vol-ume for her Hct to be reduced to 11%(0.11). The nRBC value was rela-tively low because of the subacute on-set of the anemia. The anemia wasnot chronic because the nRBC wouldhave been much higher and the babydid not have hydrops. The baby wascrossed-matched for a transfusion.Her blood was A Rh-positive and themother’s was O Rh-negative. Theplaintiff experts pointed out that theblood group incompatibility betweenthe mother and the fetus created he-molysis of the fetal cells in the moth-er’s circulation as the explanationfor the underestimate of the fetal-maternal transfusion. (2) The di-rect Coombs test result was negative.

One hour later, a partial exchangetransfusion was begun in which 80mL was transfused and 70 mL wasremoved. The plaintiff experts saidthat a sooner partial exchange anduse of a higher volume would havebeen ideal, but the action taken didnot fall below standard of care orcontribute substantially to the ulti-mate outcome. A posttransfusion Hctwas 20% (0.20). Repeat arterialblood gas showed a pH of 7.07,PCO2 of 23 mm Hg, PO2 of 69 mmHg, and base deficit of �21 mEq/L.A 45-mL blood transfusion was per-formed 1 hour later that increasedthe Hct to 34% (0.34).

On postnatal day 1, the infant de-veloped seizures and oliguria. Elec-troencephalography showed mark-edly abnormal results. On postnatalday 2, magnetic resonance imaging(MRI) showed diffuse bilateral corti-cal laminar necrosis and loss of gray-white differentiation. The infant’screatinine values were elevated andpeaked on day 2 at 1.6 mg/dL(141.4 �mol/L), but her urine out-put eventually normalized. Her ala-nine aminotransferase peaked on day2 at 3,421 U/L, but liver functionwas never clinically compromised.The plaintiff experts pointed out thatdespite the fluid adjustments thatprevented fetal death or neonatalshock, a toll was paid at the cellularlevel. Eventually, the shift from oxy-gen being used to nonoxygen path-ways to support energy metabolismbecame depleted. The liver and kid-neys recovered; the brain did not. Theplaintiff experts explained that thebaby showed she could not tolerate thechange from intrauterine to extra-uterine life. In utero, very little ef-fort is expended by the fetus for bloodgas exchange. However, when thebaby is born, the left heart pumpsagainst higher pressures than theplacenta and the baby needs tobreathe on its own, both of which

legal briefs



greatly increase the metabolic de-mand and oxygen consumption. Al-though the baby was stable at birth,the hypoxic anemia rapidly created ahuge ongoing base deficit until theanemia was corrected. The conse-quences of this profound anemia inutero were manifested after birth inthe kidney, liver, and brain.

The infant subsequently devel-oped dystonic and spastic cerebralpalsy and microcephaly. Her seizurescontinued and were difficult to con-trol. She had developmental delays inmotor, speech, and social behavior.Her kidneys performed normally.

The neonatologist was not sued.Although the neonatologist did notprovide optimal intervention interms of correcting the profoundanemia, the plaintiff experts be-lieved that the damage occurred inutero. The obstetrician and the fam-ily settled out of court.

Acute HemorrhageImmediately Before DeliveryA 3,370-g 40-weeks’ gestation whitemale infant was delivered by vacuum-assisted vaginal delivery. The preg-nancy was uncomplicated until1 hour before delivery, when thickmeconium was passed and a subse-quent fetal bradycardia to the 50soccurred for several minutes, fol-lowed by a brief period of shoulderdystocia. A profoundly pale, limp,and lifeless infant was handed tothe pediatrician. The infant wasintubated immediately and givenpositive-pressure ventilation. Nalox-one was administered 2 minutes afterbirth, two doses of endotracheal epi-nephrine were administered at 5 and8 minutes, atropine was administeredat 7 minutes, and a dose of lidocainefollowed. The plaintiff experts werehighly critical of the resuscitation.They stated that Neonatal Resuscita-tion Program (NRP) guidelineswere not followed. (3) The defense

experts stated that the NRP does notset standards but serves merely as aguideline.

Cardiac compressions were givenintermittently when the heart ratedrifted to less than 60 beats/min,but most of the time during the first20 minutes after birth, the heart ratewas more than 100 beats/min, al-though the infant’s color remainedpale and his pulses were weak. Four-teen minutes after birth the obstetri-cian discovered on delivering the pla-centa that the umbilical cord wasprofusely bleeding near its velamen-tous insertion site. During the obste-trician’s deposition, he stated that heimmediately informed the pediatri-cian and the nursing staff of this andincluded the finding in his deliverynote. The finding was validated bythe pathologist who examined theplacenta. During the deposition, thepediatrician denied hearing aboutthis. Nursing staff could not remem-ber and did not document anythingabout it. The plaintiff experts saidthat even without the knowledge ofthe cord bleeding after birth, a palebaby in cardiovascular collapseneeds volume and that volume re-placement must be blood. Further-more, the plaintiff experts main-tained that the color of a baby is animportant finding. Pallor, in con-trast to a gray or dusky color, at birthpoints to anemia rather than purehypoxemia. Pale mucous membranescan be evaluated as an additional orsubstitute tool for neonates of pig-mented ethnic backgrounds, but thisinfant was white and all cliniciansadmitted that the patient appearedextremely white. In pure hypoxemicsituations, as the heart rate improveswith oxygenation and ventilation, sodoes the color. In severe hypoxemicanemias, the heart rate may improve(as it did in this situation), but thecolor remains pale. Even in the faceof an improved heart rate, the tissues

are still in need of better oxygen-ation from improved oxygen deliv-ery from the red cells.

Ten minutes later, the pediatri-cian ordered blood to be cross-matched. The plaintiff contendedthat blood loss should have been sus-pected immediately because of the in-fant’s pallor. Moreover, the infantwas in shock, which necessitated im-mediate blood administration. Aperipheral intravenous line wasplaced at 28 minutes, followed im-mediately by an infusion of 30 mL ofnormal saline. An umbilical venousline was placed at 31 minutes. Theplaintiff experts maintained thataccess should have been attemptedimmediately. The umbilical vein isthe most quickly accessible direct in-travenous route in the newborn. (3)In addition to access, a blood gasfrom the umbilical vein should havebeen sent while waiting foremergency-release O-negative blood.A baseline Hct should have been ob-tained, although the value couldhave been misleading because itwould be fairly normal before equil-ibration occurs.

The values for the cord gaseswere available at 32 minutes. Thearterial cord gas had a pH of 7.05,PCO2 of 76 mm Hg, PO2 of 35 mmHg, bicarbonate of 20 mEq/L(20 mmol/L), and base deficit of�12 mEq/L; the venous cord gashad a pH of 7.12, PCO2 of 60 mmHg, PO2 of 45 mm Hg, bicarbonateof 21 mEq/L (21 mmol/L), andbase deficit of �9 mEq/L. Apgarscores were 1, 2, 4, 1, and 2 at 1, 5,10, 15, and 20 minutes, respectively.Another two boluses of normal salinewere administered by 45 minutes ofage. The plaintiff experts explainedthat normal saline was not the ap-propriate volume expander in theface of profound anemia, but it wasa good choice as long as emergencyblood was being processed. The de-

legal briefs



fense experts maintained that it wastoo risky to administer blood that wasnot cross-matched. They furthermaintained that it was impossible toacquire emergency blood for the de-livery room or a newborn intensivecare unit in 5 to 10 minutes. Theplaintiff pointed out that the hospi-tal needed to have had a processstreamlined and even suggested tothe defense attorney that drills shouldhave been done at that center, espe-cially because the need is not thatcommon but crucial when it doesoccur.

At approximately 1 hour of age,an Hct measured 12% (0.12). An ar-terial blood gas at this time showed apH of 6.5, PCo2 of 158 mm Hg, PO2

of 47 mm Hg, and base deficit of�32 mEq/L. The plaintiff expertspointed out that although the oxygenconcentration in the blood was satis-factory, the oxygen concentration de-livered to the tissues because of thelack of red blood cells was profoundlyreduced. The transport team arrivedat the same time that the blood bankissued the cross-matched blood, anda transfusion was begun duringtransport.

At 1 day of age at the referralhospital, the patient developed sei-zures and electroencephalographyreadings were very abnormal. MRIshowed cortical necrosis in the occip-ital, temporal, and parietal lobes andabnormal signals in the thalamic andbasal ganglia areas. The baby hadanuria for 3 days, and his creatininepeaked at 8.5 mg/dL (751.4 �mol/L). Renal ultrasonography suggestedpapillary necrosis. The renal functionremained abnormal throughout thehospitalization. The plaintiff expertspointed out that hypoxia may pro-duce acute tubular necrosis that usu-ally is transitory. However, hypoxicanemia in the fetus or newborn hasbeen associated with chronic renal

insufficiency. (4) The infant’s liverfunction tests were elevated into the1,000s and his direct bilirubinpeaked at 14.9 mg/dL (254.8�mol/L), which was believed to bedue to a postnecrotic hypoxic eventresulting from the anemia. The liverfunction recovered.

On long-term follow-up evalua-tion, the child has cerebral palsy, re-nal insufficiency, a seizure disorder,microcephaly, and significant devel-opmental delays.

The obstetrician was not sued.The family and the pediatrician set-tled out of court.

DiscussionNeonatal anemia at birth can be clas-sified according to the cause. (5)Anemia from hemorrhage is the mostcommon cause. Mechanical traumaduring delivery, uterine rupture orplacental abruption, rupture of theumbilical cord, and fetal-fetal or feto-maternal hemorrhage are the mostcommon causes of hemorrhagic ane-mia at birth. A less common causeof anemia at birth is hemolysis, thecauses of which include an immuneresponse between mother and baby(Rh or ABO incompatibility) andred blood cell membrane defects suchas spherocytosis. Anemia at birthfrom hypoplasia of erythrocyte pro-duction is very rare, especially pre-senting at birth. The inherited causes(Diamond-Blackfan) must be differ-entiated from infection with parvo-virus B19, which can cause a purered blood cell aplasia at birth. Whenthe anemia is chronic and severe, hy-drops develops. In chronic anemia,the fetus develops a microcytic, hy-pochromic anemia. After birth, themean corpuscular volume and meancorpuscular hemoglobin are low.

In the first case discussed in thisarticle, the fetomaternal hemorrhagewas subacute, which was supported

by only a moderate elevation of thenRBCs, the absence of hydrops, andthe normal blood pressure at birth.The second baby with the rupturedumbilical cord was born in hypovole-mic shock because of the acute na-ture of the hemorrhage. An appreci-ation of pallor at birth and the needto move quickly with blood is vitalfor appropriate intervention from thedamaging effects of hypoxic anemiathat have long-term devastating im-pacts on the brain and kidneys.

References1. The American College of Obstetriciansand Gynecologists’ Task Force on NeonatalEncephalopathy and Cerebral Palsy, theAmerican College of Obstetricians and Gy-necologists, the American Academy of Pedi-atrics. Neonatal Encephalopathy and Cere-bral Palsy: Defining the Pathogenesis andPathophysiology. Washington, DC: Ameri-can College of Obstetricians and Gynecolo-gists; 2003:1–852. Glasser L, West JH, Hagood RM. In-compatible fetomaternal transfusion withmaternal intravascular lysis. Transfusion.1970;10:322–3253. Neonatal Resuscitation Textbook. ElkGrove Village, IL; Dallas, TX: AmericanAcademy of Pediatrics; American Heart As-sociation; 20064. Ringer SA. Acute renal failure in theneonate. NeoReviews. 2010;11:e243–e2515. Widness JA. Pathophysiology of anemiaduring the neonatal period, including ane-mia of prematurity. NeoReviews. 2008;9:e520–e525

American Board of PediatricsNeonatal-Perinatal MedicineContent Specifications• Know the causes of

and diagnosticapproach to aninfant who isanemic at birth.

• Know the causes and pathophysiologyof acute fetal and neonatal blood loss.

• Know the clinical and laboratoryfindings and management of acutefetal and neonatal blood loss.

legal briefs




Maureen E. Sims Legal Briefs: Neonatal Anemia at Birth













Roger F. Soll Meta-analysis in Neonatal Perinatal Medicine







Meta-analysis in Neonatal Perinatal MedicineRoger F. Soll, MD*

Author Disclosure

Dr Soll has disclosed

that he is president

of the Vermont

Oxford Network and

editor of the

Cochrane Neonatal

Review. This

commentary does not


of an unapproved/


commercial

product/device.

AbstractSystematic overviews provide a comprehensive and thorough review of the availabledata from clinical trials. When these reviews include meta-analyses, clinicians cansynthesize the results of related studies and gain greater precision in their estimates ofthe effects of therapy. Even when inconclusive, meta-analyses allow for the explorationof differences between studies and may point toward promising areas of futureresearch (or steer clinicians away from further nonproductive areas). In the field ofneonatal-perinatal medicine, systematic overviews have provided the basis of severalmajor changes in guidelines with measurable impact on neonatal outcome.


1. Describe the process of conducting a systematic review.2. Delineate biases inherent in meta-analyses.3. Review how heterogeneity affects meta-analyses.

IntroductionIt is almost impossible to keep up with the scientific evidence on which clinicians base theirpractice and policies. The simple exercise of searching a common bibliographic databasesuch as PubMed, using the word “neonate” and limiting the search to “all infants” and“randomized, controlled trials” retrieves more than 7,000 citations. No clinician couldpossibly keep up with this mass of literature. When multiple trials are available, cliniciansare confronted with a situation more like taking a Rorschach test than like interpreting theevidence; what we see more reflects our biases than the evidence at hand.

A systematic approach to reviewing the literature is necessary to address this difficulty.Systematic reviews identify, appraise, and synthesize research-based evidence and present itin an accessible format. (1) Systematic overviews apply specific research strategies to datafrom all relevant studies. If appropriate, these reviews can include meta-analyses, aquantitative statistical method used to combine the results of similar randomized, con-trolled trials (RCTs) to produce typical estimates of effect size. By combining informationfrom all relevant studies, meta-analysis can provide more precise estimates of the effect ofhealth care than those derived from the individual studies included within the review.(2)(3) They also facilitate evaluation of the consistency of evidence across studies and theexploration of differences between studies.

The following review discusses the approach to creating a systematic overview andhow to interpret the results of meta-analyses. A more detailed and technical discus-sion of creating a systematic review and meta-analysis is given in “The CochraneHandbook.” (4)

How is a Systematic Review Conducted?Systematic reviews are essentially “studies of studies.” They use rigorous scientific methodssimilar to those used in any clinical trial. The specific objective of the review, framed as ananswerable question, provides the backbone of the analysis. Like a clinical trial, a protocolfor a systematic review should be developed that clearly states the objectives of the review,the population and intervention of interest, and the methods used at each stage of thereview.

*H.W. Wallace Professor of Neonatology; Coordinating Editor, Cochrane Neonatal Review Group, Burlington, VT.

Article research



The next major step is selection of studies for inclu-sion. As noted previously, the criteria for study inclusionmust be specifically defined in the protocol. This includesthe specifics regarding the intervention, the population,and the outcomes assessed. If appropriate, exclusioncriteria must be explicitly stated.

Searching for trials for inclusion is much simpler in theage of computer databases. Long gone are the days ofsitting in the medical library pouring over volumes ofIndex Medicus. Today there are multiple resources foraccess to the general medical literature and some re-sources unique to the field of neonatal-perinatal medi-cine. Searching the medical literature is now widelyavailable through the internet, including several biblio-graphic databases. These include the Cochrane Library,MEDLINE, EMBASE, and CINAHL.

Once studies are located, rigorous evaluation ofwhether they meet the criteria for inclusion is necessary.If included, the study must be assessed for the validity ofthe methodology in study design, conduct, analysis, andreporting. For the purpose of Cochrane reviews, onlyrandomized or quasi-randomized, controlled trials areincluded. (5)

Outcome data must be extracted and tabulated foreach included trial. There are two distinct stages to theanalysis. First, a clinically relevant standard statistic iscalculated for each study to describe the observed inter-vention effect. For example, the standard statistic maybe the relative risk (RR), the RR reduction, the riskdifference if the data are dichotomous, or a differencebetween means if the data are continuous. Second, anestimate of the summary (pooled) intervention effect iscalculated as a weighted average of the interventioneffects estimated in the individual studies. The statisticalmethods for pooling results are similar to the statisticalmethods used in analyzing the data for multicenter trials.Pooling the results of similar RCTs increases the statisti-cal power lacking in individual smaller trials and enablesthe clinician to have greater security in accepting orrejecting treatment differences demonstrated by thetrials. (2)

Once the analysis has been completed, it is importantto assess the importance of the evidence. Clinical trialsmay use a variety of statistical techniques in reporting theresults. A “statistically significant” reported differencedoes not make the finding clinically relevant. (6) Toassess whether the results of a trial are clinically relevantrequires calculation of some simple statistics from studyfindings (Table). The relative risk reduction (RRR) indi-cates the relative, but not absolute, reduction in the event

rate. The absolute risk reduction (ARR) indicates theabsolute reduction in the event rate. If the overall inci-dence of the event is low, the ARR also will be low, evenif there is a relatively large difference in the relative risk.Understanding both the RR and ARR is essential tomaking any clinical judgment. Also of use is the numberneeded to treat (NNT), which is calculated by dividing1 by the ARR. For example, if the ARR is 20%, thenthe NNT is 5. In other words, five infants would need tobe treated to prevent one theoretical event. In addition,confidence intervals (CIs) should be reported with eachof these statistics. A 95% CI reflects 95% certainty that thetrue value of the measure lies within the bounds of theinterval.

A meta-analysis can be subject to many of its ownbiases. Any attempt at pooling results from various stud-ies not only incorporates the biases of the primary studiesbut adds further bias attributable to study selection andthe inevitable heterogeneity of the selected studies. (7)Publication bias, the tendency for investigators to choosepositive studies for publication, skews the medical litera-ture toward favorable reports of treatment. Unless theauthors of the meta-analysis scrupulously research allavailable resources, these studies are not located and themeta-analysis could report a false-positive finding. Thisproblem is compounded further by the greater chancethat such a false-positive finding will be published. Meta-analysis can offer false-negative conclusions because ofinappropriate study selection. If the studies selected can-not be grouped (heterogenous), the positive effects ob-served with one specific treatment or in one specific

Table. Key Terminology forEstimating the Size of theTreatment Effect

OutcomeRisk of OutcomePositive Negative

Treated (Y) a b Y�a/(a�b)Control (X) c d X�c/(c�d)

Relative Risk (RR) is the risk of the outcome in the treated group (Y)compared to the risk in the control group. RR�Y/XRelative Risk Reduction (RRR) is the percent reduction in risk in thetreated group (Y) compared to the control group (X). RRR�1-Y/X�100% or 1-RR�100%Absolute Risk Reduction (ARR) is the difference in risk between thecontrol group (X) and the treatment group (Y). ARR�X-YNumber Needed to Treat (NNT) is the number of patients that mustbe treated over a given period of time to prevent one adverse outcome.NNT�1/(X-Y) or 1/ARR

research meta-analysis



population may be lost. To minimize bias, the authors ofthe meta–analysis and the readers of the review mustdemand the same methodologic quality from these anal-yses that they would from individual RCTs. It is essentialthat all meta-analyses include a prospectively designedprotocol, a comprehensive and extensive search strategy,strict criteria for inclusion of studies, standard definitionsof outcomes, and standard statistical techniques.

What is Heterogeneity?Invariably, studies brought together in a systematic re-view have differences. Variability among studies in asystematic review is termed “heterogeneity.” Clinicalheterogeneity refers to variability in the participants,interventions, and outcome included in the studies.Methodologic heterogeneity refers to variability in thestudy design. Clinical or methodologic heterogeneitymay contribute to measurable statistical heterogeneity.Methods have been developed for quantifying inconsis-tency across studies that move the focus away fromtesting whether heterogeneity is present (which is almostinevitable) to assessing the impactof heterogeneity on the meta-analysis. (4) The I-squared (I2) sta-tistic can be used to evaluatewhether substantial heterogeneityis present and may influence theinterpretation of the analysis. Ifnoted, heterogeneity can be ex-plored using subgroup analyses.

Sample Meta-analysis:Early Corticosteroidsfor the Prevention ofChronic Lung DiseaseTo understand a meta-analysis, itmay be useful to examine a system-atic review and meta-analysis thathas influenced practice in neonatalperinatal medicine, such as earlycorticosteroids for the preventionof chronic lung disease in preterminfants. (8) This review examinesthe relative benefits and adverse ef-fects of postnatal corticosteroidsadministered within the first 7 daysof birth to preterm infants at risk ofdeveloping chronic lung disease. Thereview can be found on the NationalInstitute of Child Health and Hu-

man Development web site (http://www.nichd.nih.gov/cochrane/hallida3/hallida.htm) or in the Cochrane Library.

Chronic lung disease remains a significant problemamong very low-birthweight infants. It carries with it bothcosts, in terms of longer hospital stay, and risks, in terms oflater development. (9) Corticosteroids can reduce lunginflammation in newborns who have chronic lung disease,but there are major short- and long-term adverse effects.

For the purposes of this review, the authors includedonly RCTs of postnatal corticosteroid therapy. The“study participants” were defined as preterm infants be-lieved to be at risk of developing chronic lung diseasewho were enrolled within the first 7 days of birth. Infantswere not required to be on ventilator support, which is animportant distinction. The authors could have limitedthe population to very low-birthweight or extreme low-birthweight infants and restricted the review only tothose receiving ventilator support. The authors chose tocast a wide net for studies, but this could potentially leadto clinical heterogeneity.

The authors chose to study intravenous or oral corti-

Figure 1. Early (<8 days) postnatal corticosteroids for preventing chronic lung disease inpreterm infants. Effect on chronic lung disease or death. Reprinted with permission fromHalliday et al. (8)




costeroids (including dexamethasone and hydrocorti-sone). Trials of inhaled corticosteroids were not in-cluded. The outcome measures included the importantshort-term measures related to neonatal intensive care(eg, mortality, chronic lung disease) as well as longer-term outcomes, including development at 18 to 24months. The search strategy was specifically detailed inthe review. In addition, a formal plan for data collectionand analysis was proposed.

The authors identified 28 trials that qualified forinclusion in the review. This discussion is limited to theprimary outcome measures of chronic lung disease ordeath and cerebral palsy. Figures 1 and 2 are the “forestplots” of the relative risk reported in each study and a“summary” statistic showing the “pooled” relative riskestimated by the meta-analysis. In the review of all stud-ies (using either dexamethasone or hydrocortisone),chronic lung disease (defined as oxygen requirement at36 weeks adjusted age) and death was significantly re-duced (typical RR, 0.89; 95% CI, 0.84 to 0.95 andtypical risk difference, 0.06; 95% CI, �0.09 to �0.02).Twenty-one studies were included in the analysis, in-volving more than 3,300 infants. For this outcome, theI2 is reported as 43%, which represents a moderatedegree of heterogeneity. What might be the source forthis heterogeneity? The primary analysis includes a sub-group analysis of dexamethasone and hydrocortisone.

For the studies of hydrocortisone,little heterogeneity is noted (I2 24%).For the studies exclusively usingdexamethasone, a moderate degreeof heterogeneity persists (I2 49%).Close inspection of these studiesshows many differences in the dexa-methasone studies, including patientpopulation, length of treatment, tim-ing of treatment, and dosage.

Although the impact of cortico-steroids on chronic lung diseaseseems promising, the results regard-ing cerebral palsy raise grave con-cerns. Cerebral palsy was increasedwith the use of corticosteroids (typi-cal RR, 1.45; 95% CI, 1.06, 1.98 andtypical risk difference, 0.03; 95% CI,0.00, 0.06 for 12 studies and 1,452infants). Again, there was moderateheterogeneity (I2 66%) in the overallanalysis that was particularly notablein the studies that used dexametha-sone (I2 34%). The meta-analysis

provides a framework to examine individual studies com-pared with the other included studies. Of interest, Shinwelland associates (10) reported the greatest risk of cerebralpalsy and yet represented some of the least exposure todexamethasone, suggesting to investigators that perhapstiming of treatment is critical.

The understanding of this meta-analysis led to strongstatements from the Committee on Fetus and Newborn ofthe American Academy of Pediatrics. (11) These statementsrecommend severe restriction in the use of postnatal corti-costeroids to prevent and treat chronic lung disease, notingthat “outside the context of a randomized, controlled trial,the use of corticosteroids should be limited to exceptionalclinical circumstances (eg, an infant on maximal ventilatorsupport and oxygen support).”

ConclusionsSystematic overviews and meta-analysis are critical toolsin efforts to practice evidence-based medicine. Althoughflawed by the deficiencies and limitations of the includedstudies as well as by biases created by the analysis itself,meta-analysis provides a framework to inspect individualtrials as well as a method to gain more precise estimates ofeffects by analyzing the trials in aggregate. Analyses suchas the review of early postnatal corticosteroids have pro-vided an improved perspective on the risks and benefitsof treatments and have changed practice. (12)

Figure 2. Early (<8 days) postnatal corticosteroids for preventing chronic lung dis-ease in preterm infants. Effect on cerebral palsy. Reprinted with permission from Hallidayet al. (8)




References1. Cook DJ, Mulrow CD, Haynes RB. Systematic reviews: syn-thesis of best evidence for clinical decision. Ann Intern Med. 1997;126:376–3802. Sinclair JC, Bracken M, Horbar J. Introduction to neonatalsystematic reviews. Pediatrics. 1997;100:892–8953. Oxman AD, Cook DJ, Guyatt GH. Users’ guides to the medicalliterature: VI. How to use an overview. JAMA. 1994;272:1367–13714. Higgins JPT, Green S, eds. Cochrane Handbook for SystematicReviews of Interventions Version 5.0.2. Oxford, United Kingdom:The Cochrane Collaboration; 20095. Bero L, Rennie D. The Cochrane Collaboration: preparing,

maintaining, and disseminating systematic reviews of the effects ofhealth care. JAMA. 1995;274:1935–19386. Laupacis A, Sackett DL, Roberts RS. An assessment of clinicallyuseful measures of the consequences of treatment. N Engl J Med.1988;318:1728–17337. Bailar JC III. The promise and problems of meta-analysis.N Engl J Med. 1997;337:559–5618. Halliday HL, Ehrenkranz RA, Doyle LW. Early (� 8 days)postnatal corticosteroids for preventing chronic lung disease inpreterm infants. Cochrane Database Syst Rev. 2010;1:CD0011469. Schmidt B, Asztalos EV, Roberts RS, Robertson CM, SauveRS, Whitfield MF; Trial of Indomethacin Prophylaxis in Pre-terms (TIPP) Investigators. Impact of bronchopulmonary dyspla-sia, brain injury, and severe retinopathy on the outcome of ex-tremely low-birth-weight infants at 18 months: results from the trialof indomethacin prophylaxis in preterms. JAMA. 2003;289:1124–112910. Shinwell ES, Karplus M, Reich D, et al. Early postnatal dexa-methasone treatment and increased incidence of cerebral palsy.Arch Dis Child Fetal Neonatal Ed. 2000;83:F177–F18111. Committee on Fetus and Newborn. Postnatal corticosteroidsto treat or prevent chronic lung disease in preterm infants. Pediat-rics. 2002;109:330–33812. Ohlsson A. Randomized controlled trials and systematic re-views: a foundation for evidence-based perinatal medicine. ActaPaediatr. 1996;85:647–655

NeoReviews Quiz

1. You are reviewing the results of a randomized trial to determine the effect of a drug (treatment group) inpreventing the occurrence of a specific disease (outcome) as compared with that of a placebo (controlgroup). The results are tabulated below:

Study Group Disease Present Disease Absent Risk of Disease

Treatment (n�100) 41 59 0.41

Control (n�100) 52 48 0.52

Of the following, the number needed to treat (number of infants needed to be treated to prevent theoccurrence of the disease in a single patient) in this trial is closest to:

A. 7.B. 9.C. 11.D. 13.E. 15.

2. A systematic review is designed to identify, appraise, and synthesize research-based evidence from allrelevant studies. Invariably, studies brought together in the systematic review differ in some aspects,leading to heterogeneity in the clinical, methodological, and statistical domains. Of the following,methodological heterogeneity is most likely to represent variability in the:

A. Data analysis.B. Primary outcome.C. Study design.D. Study participants.E. Treatment interventions.

American Board of Pediatrics Neonatal-PerinatalMedicine Content Specifications• Understand the purpose of a systematic

review.• Understand the advantages of adding a

meta-analysis to a systematic review.• Interpret the results of a meta-analysis.• Identify the limitations of a systematic review.• Identify the limitations of a meta-analysis.





Roger F. Soll Meta-analysis in Neonatal Perinatal Medicine













Juan I. Remon, Aarti Raghavan and Akhil Maheshwari Polycythemia in the Newborn







Polycythemia in the NewbornJuan I. Remon, MD,*

Aarti Raghavan, MD,*

Akhil Maheshwari, MD*

Author Disclosure

Drs Remon, Raghavan,

and Maheshwari have

disclosed no financial



commentary does not


of an unapproved/


commercial

product/device.

AbstractNeonatal polycythemia, defined as a venous hematocrit �65% (0.65), is a commonproblem in newborns. Infants born postterm or small for gestational age, infants ofdiabetic mothers, recipient twins in twin-to-twin transfusion syndrome, and thosewho have chromosomal abnormalities are at higher risk. Although the cause ofpolycythemia is often multifactorial, most cases can be classified as having active(increased fetal erythropoiesis) or passive (erythrocyte transfusion) polycythemia. Byincreasing blood viscosity, polycythemia can impair microcirculatory flow in endorgans and can present with neurologic, cardiopulmonary, gastrointestinal, andmetabolic symptoms. In this article, we review the pathophysiology, clinical presen-tation, diagnosis, and management of polycythemia in the newborn.


1. List the causes of polycythemia and hyperviscosity in the neonate.2. Review the signs, symptoms, and diagnostic criteria for polycythemia and

hyperviscosity.3. Discuss the treatment of polycythemia and hyperviscosity in the neonate.

IntroductionPolycythemia (or more accurately, erythrocythemia), an abnormal elevation of the circu-lating red blood cell (RBC) mass, is seen frequently in newborns. Although neonatalpolycythemia usually represents a normal fetal adaptation to hypoxemia rather than a truehematopoietic defect, the abnormal increase in hematocrit increases the risk of hyper-viscosity, microcirculatory hypoperfusion, and multisystem organ dysfunction. In thisarticle, we review the definition, pathophysiology, clinical presentation, and managementof polycythemia in the newborn.

DefinitionIn healthy term infants, the hematocrit and hemoglobin concentrations in venous bloodobtained at birth are 50.2�6.9% (0.5�0.07) (mean � standard deviation) and15.9�1.86 g/dL (159�18.6 g/L), respectively. (1) Polycythemia is defined in newbornsas a venous hematocrit greater than 65% (0.65) or a hemoglobin value greater than22 g/dL (220 g/L). (2)(3)(4) Based on this definition, the incidence of polycythemia inhealthy newborns has been reported to be 0.4% to 5%. (5)(6)(7) Increased incidence isseen in certain high-risk groups such as postterm neonates, small-for-gestational ageinfants, infants of diabetic mothers, identical twins who share the same placenta anddevelop twin-to-twin transfusion, and infants who have chromosomal abnormalities.(7)(8)(9)

PathophysiologyAlthough the cause of polycythemia is often multifactorial, most patients can be classifiedas having active (increased fetal erythropoiesis) or passive (erythrocyte transfusion) poly-cythemia (Table 1). (5)(7)

*Department of Pediatrics, University of Illinois at Chicago, Chicago, IL.

Article hematology



Increased fetal erythropoiesis is frequently seen inconditions associated with hypoxia:

● Placental insufficiency due to preeclampsia, primaryrenovascular disease, chronic or recurrent placentalabruption, maternal cyanotic congenital heart disease,postdate pregnancy, maternal smoking, and maternalheavy alcohol intake. (10)(11)(12) The severity ofhematopoietic dysfunction in these conditions appearsto be proportional to the degree of placental insuffi-ciency and fetal growth restriction. (8) Whereas mildplacental dysfunction and consequent tissue hypoxiaare associated with increased erythropoietin concen-trations and polycythemia, more severe placental vas-culopathy may cause erythropoietin resistance andanemia. (13)(14)

● Endocrine abnormalities associated with increasedfetal oxygen consumption, such as congenital thyro-toxicosis or maternal diabetes with poor glycemiccontrol. (13)(14) Thyrotoxicosis is presumed to in-crease erythropoiesis through a direct effect on marrowprogenitor cells, increased erythropoietin expression,and the indirect effects of intrauterine growth restric-tion. (15)(16) In diabetic mothers who have poorglycemic control, maternal hyperglycemia is proposedto increase fetal erythropoiesis through fetal hyper-insulinemia, tissue hypoxia, and increased erythro-poietin concentrations. (17)(18) Although plasma lep-tin concentrations are frequently elevated in infantsof diabetic mothers, the actual concentrations do not

correlate with the severity of polycythemia, and currentdata do not support a causative role of leptin in thisprocess. (18)

● Genetic disorders, such as trisomy 13, trisomy 18,trisomy 21, and Beckwith-Wiedemann syndrome. Theincidence of polycythemia in infants who have Downsyndrome is 15% to 33%. (19)(20)(21) Although thecause of polycythemia in Down syndrome is notknown, high cord blood erythropoietin concentra-tions in affected infants has led to the speculation thatintrauterine hypoxemia may play a role. (9) Amongneonates who have trisomy 13 and trisomy 18, theprevalence of polycythemia has been estimated as 8%and 17%, respectively. (22)

Erythrocyte transfusion polycythemia can result fromplacental-fetal transfusion. Delayed cord clamping allowsfor delivery of an increased blood volume to the infant.When cord clamping is delayed more than 3 minutesafter birth, blood volume may increase by as much as30%. (23) Hutton and Hassan (24) analyzed data fromseven randomized studies and showed that neonates inthe late-clamping group had a higher incidence of asymp-tomatic polycythemia with a benign course (relative risk(RR), 3.82; 95% confidence interval (CI), 1.11 to 13.21).However, a more recent meta-analysis showed that delayedcord clamping did not cause a clinically significant changein hematocrit at 1 and 4 hours of age. (25)

Placental-fetal transfusion is likely influenced by grav-ity and the relative position of the delivered infant inrelation to the maternal introitus before cord clamping;raising or lowering the baby by 15 to 20 cm or more withthe cord intact appears to influence placental transfusion.(26) No randomized studies have examined this issue todate. Placental transfusion is augmented in infants whohave perinatal asphyxia, which causes an active shift ofthe blood volume from the placenta to the fetus. (27)Oxytocin administration to the mother can also increasethe volume of placental transfusion to the newborn. (28)

Twin-to-twin transfusion syndrome due to a vascularcommunication occurs in approximately 10% of mono-zygotic twin pregnancies. (29)

Polycythemia and HyperviscosityPolycythemia remains an area of interest due to its po-tential effects on the viscosity of blood and its flowproperties in the microcirculation. (5)(7)(8) Viscosity isa measure of the resistance of a fluid that is being de-formed by either shear stress or tensile stress. (30) Moresimply, viscosity refers to the “thickness” of a fluid andis a measure of the fluid’s internal resistance to flow.

Table 1. Conditions AssociatedWith Neonatal PolycythemiaIncreased Fetal Erythropoiesis

● Placental insufficiency due to preeclampsia, maternalchronic hypertension, chronic or recurrent placentalabruption, maternal cyanotic congenital heartdisease, postdate pregnancy, maternal smoking andheavy alcohol intake

● Endocrine abnormalities such as congenitalthyrotoxicosis or maternal diabetes with poorglycemic control

● Genetic disorders such as trisomy 13, trisomy 18,trisomy 21, and Beckwith-Wiedemann syndrome

Erythrocyte Transfusion

● Placental-fetal transfusion with delayed cordclamping, relative positioning of the delivered infantin relation to the maternal introitus before cordclamping, perinatal asphyxia, oxytocin administration

● Twin-to-twin transfusion syndrome

hematology polycythemia



Hyperviscosity is arbitrarily defined as a viscosity mea-surement of greater than 14.6 centipoise detected in aviscometer at a shear rate of 11.5/sec. (5)(7)(30) Viscos-ity rises linearly with increasing hematocrit until thehematocrit reaches 60% (0.6) but increases exponentiallywhen hematocrit equals or exceeds 70% (0.7).(5)(31)(32) Although the terms polycythemia and hy-perviscosity are often used interchangeably, they are notequivalent and show only modest concordance in clinicalcohorts. (5)(30) Furthermore, blood viscosity can alsorise with an increase in plasma proteins, platelets, leuko-cytes, and endothelial factors. (30)(32)

Hyperviscosity occurs in polycythemia due to thepresence of an abnormally large number of circulatingerythrocytes; the plasma viscosity in the newborn is al-most always normal. (5)(7) According to Poiseuille’slaw, flow velocities in the circulation are determined bythe resistance to flow, which varies with the viscosity ofthe blood and inversely with the fourth power of theradius of the blood vessel. This relationship can be ex-pressed by the equation R�8hL/�r4, where R repre-sents the resistance to blood flow, h is the viscosity, L isthe length of the vessel, and r is the radius of the vessel.(33)(34) Because resistance is affected by viscosity as wellas the caliber of the blood vessel, the effects of poly-cythemia on blood flow patterns are usually most pro-nounced in the microcirculation. (34) In these smallvessels, non-newtonian mechanisms such as rouleauxformation and increased erythrocyte-endothelial interac-tion are also active and further contribute to the alteredflow. (32)

Polycythemia and hyperviscosity are associated withdecreased blood flow to the brain, heart, lung, intestines,and carcass. (5)(7)(35)(36) Although renal blood flow isnot affected, renal plasma flow and glomerular filtrationrate are often diminished. (35) Hyperviscosity can alsoreduce pulmonary blood flow that, in turn, can causesystemic hypoxia. (36) In contrast to the effects of poly-cythemia on the kidney and lungs, reduced cerebralblood flow in polycythemia likely represents a vascularresponse to the increased arterial oxygen content (relatedto increased hemoglobin concentrations) rather thanhyperviscosity. (7)(37) Changes in blood flow may alsoalter the delivery of substrates (such as glucose) to organsthat are dependent on plasma flow. (7)(38)(39)

Clinical FindingsThe symptom complex associated with polycythemia isfrequently described by the term “hyperviscosity syn-drome,” although it is important to remember that only47% of infants who have polycythemia exhibit hypervis-

cosity, and only 24% of infants who have hyperviscosityhave a diagnosis of polycythemia. (6)(7)(40) Whereasmost patients who have polycythemia remain asymptom-atic, characteristic clinical features may be recognized asearly as 1 to 2 hours after birth as the hematocrit peakswith normal postnatal fluid shifts. (3) In some infantswho have high borderline hematocrits, symptoms may bedelayed until the second to third postnatal day, whenexcessive depletion of the extracellular fluid may lead tohemoconcentration and hyperviscosity. Infants who haveno symptoms by 48 to 72 hours of age are likely toremain asymptomatic. (3)(41)

Clinical features associated with neonatal polycythe-mia are generally nonspecific and include ruddy com-plexion, irritability, jitteriness, tremors, feeding difficul-ties, lethargy, apnea, cyanosis, respiratory distress, andseizures. (7) Neurologic symptoms occur in approxi-mately 60% of affected patients. (5)(7)(42) The cause ofthese symptoms is uncertain, but reduced cerebral bloodflow and altered tissue metabolism likely play importantroles. Neurologic signs may also be related to metabolicchanges such as hypoglycemia and hypocalcemia. Hypo-glycemia is the most common metabolic derangementand is observed in 12% to 40% of infants who havepolycythemia. Hypocalcemia is found in 1% to 11% ofneonates who have polycythemia, possibly related toelevated concentrations of calcitonin gene-related pep-tide (CGRP) in affected infants. (43) The pathophysiol-ogy of increased CGRP concentrations is not clear;CGRP may play a role in the normal postnatal circulatoryadaptation to extrauterine life, and this process is pre-sumed to be accentuated in infants who have polycythe-mia. (44)

Polycythemia and hyperviscosity have been implicatedas pathogenic factors in necrotizing enterocolitis (NEC),particularly in term or near-term neonates. (45)(46)(47)(48)(49) Historically, polycythemia has been identifiedin up to half of all term infants who have NEC.(46)(47)(48) Although altered splanchnic perfusion iswidely considered to cause gut mucosal injury in affectedinfants, recent data indicate that attempts to reduce thehematocrit with partial exchange transfusions (PET) alsocould contribute to the risk of NEC. (36)(42)(49)

Renal manifestations of polycythemia include de-creased glomerular filtration rates, oliguria, hematuria,proteinuria, and renal vein thrombosis. (5)(7) Thrombo-ses can also be seen in other sites. Thrombocytopenia canbe seen in up to one third of all cases, presumably due toplatelet consumption in the microvasculature. (50)(51)In neonates who have polycythemia due to increasederythropoiesis, thrombocytopenia may also be related to




“progenitor steal,” the diversion of hematopoietic pro-genitors toward erythropoiesis at the expense of otherlineages. (8)(18) Overt disseminated intravascular coag-ulation is rare. (51)

DiagnosisThe diagnosis of polycythemia requires detection of avenous hematocrit of at least 65% (0.65). (2)(3)(4)This cut-off finds its origins in studies on blood viscositythat show an exponential increase in viscosity above thisvalue. (3)(7) However, clinical evidence for this thresh-old remains scant; studies have shown only modestconcordance between a hematocrit greater than 65%(0.65) and actual demonstration of hyperviscosity. (7)Although blood viscosity could be a useful guide fordeciding appropriate management strategies in affectedpatients, hematocrits continue to be widely used surro-gate markers of hyperviscosity due to limited availabilityof tools for direct measurement of blood viscosity.

Detection of high hematocrits (�55% [0.55]) in cordblood could help predict the risk of polycythemia at2 hours postnatal age, but such “screening” has notgained acceptance because of the lack of data showingthat treatment of asymptomatic newborns who haveelevated hematocrits alters outcomes. (3) The possibilityof polycythemia should be considered in any infant ex-hibiting signs of hyperviscosity. Detection of a highhematocrit in the first few hours after birth should triggera follow-up measurement in a few hours to identify anyfurther rise with normal postnatal fluid shifts. (3)

Close attention must be paid to the technique ofsample collection while interpreting the test results. Cap-illary blood samples often show hematocrits that are 5%to 15% higher than venous samples, and, therefore, highhematocrit measurements in samples obtained by heel-sticks should be confirmed in a free-flowing venous sam-ple. (3)(52) The technique of sample analyses shouldalso be taken into account during serial evaluation ofresults because microcentrifuge hematocrit may beslightly higher (due to trapped plasma of about 2%) thanthat calculated from RBC volume and RBC count deter-mined by hematology analyzer. (1)

Neonates who have polycythemia should be evaluatedfor underlying causes such as intrauterine growth restric-tion, maternal diabetes, or birth asphyxia. Because clini-cal manifestations of hyperviscosity can overlap withother conditions, alternative causes for the symptomsshould always be carefully excluded. Infants also shouldbe monitored for systemic complications of polycythe-mia such as thromboses, NEC, hypoglycemia, hypocal-cemia, hyperbilirubinemia, and thrombocytopenia.

TreatmentThe management of polycythemia is controversial be-cause of the lack of evidence showing that aggressivetreatment improves long-term outcomes. Asymptomaticinfants whose central hematocrits are between 60%(0.60) and 70% (0.70) can be monitored closely andaggressively hydrated with adequate enteral intake oradministration of intravenous fluids. The hematocritshould be reassessed in 12 to 24 hours, and plasmaglucose and bilirubin and cardiorespiratory status shouldbe monitored. If the hematocrit decreases or remainsstable and the patient remains asymptomatic, monitoringcan be continued for a further 24 to 48 hours. In asymp-tomatic infants whose hematocrits are greater than 70%(0.70), the treatment options are controversial. Al-though traditional treatment has been PET, studies showa lack of difference in outcomes with continued hydra-tion and expectant management versus aggressive man-agement with PET. (36)(42) In the absence of a consen-sus, the decision to perform PET is usually taken on acase-by-case basis, with a careful analysis of risks andpotential benefits.

To perform PET, a precalculated volume of blood(calculated to reduce the central hematocrit to 50%[0.50] to 55% [0.55]) is replaced by an equivalent vol-ume of normal saline, 5% albumin, commercially avail-able solutions of human plasma protein fraction, or freshfrozen plasma. Colloid solutions do not offer any thera-peutic advantages over normal saline and, at least in somestudies, have been associated with a higher incidence ofNEC. (53) Because of its lower cost, ready availability,and absence of risk of transfusion-associated infection,normal saline is generally accepted as the replacementfluid of choice for PET in infants who have polycythemia.(54)(55)

The total blood volume to be exchanged is deter-mined as follows:

[Total blood volume � (patient’s hematocrit� desired hematocrit)]/(patient’s hematocrit)

where total blood volume�the patient’s weight in kilo-grams multiplied by a presumed blood volume of90 mL/kg. PET can be performed through a singleumbilical venous catheter using a pull-push technique(withdrawal of blood alternated with replacement offluid through a single catheter) or by withdrawing bloodfrom an umbilical or peripheral arterial catheter andadministering replacement fluid simultaneously throughan umbilical or peripheral venous catheter. Regardless ofthe sites used, small aliquots of 5 mL/kg or less should




be used for removal or delivery, with each step carriedout over 2 to 3 minutes.

Several randomized studies have evaluated the efficacyof PET in patients who have polycythemia (Table 2).Malan and de V Heese (n�49), (56) Goldberg andcolleagues (n�20), (57) Black and coworkers (n�94),(58) and Ratrisawadi and associates (n�42) (59) ran-domly assigned infants to receive either PET usingplasma or supportive care. Bada and colleagues (n�28)(60) compared PET using a commercially availablehuman plasma protein solution with supportive care.Kumar and Ramji (61) randomized 45 infants to periph-eral PET using normal saline or to routine medicalmanagement. None of these six studies documented abeneficial effect of PET on neurodevelopmental out-come. Dempsey and Barrington (42) systematically re-viewed five of these studies to investigate whetherPET was associated with improved short- and long-termoutcomes in neonates who had polycythemia. They doc-umented no improvement in long-term neurologic out-come (mental developmental index, incidence of devel-opmental delay, and incidence of neurologic diagnoses)after PET in symptomatic or asymptomatic infants.There was also no evidence of improvement in earlyneurobehavioral assessment scores (Brazelton neonatalbehavioral assessment scale). PET could have been asso-ciated with an earlier improvement in symptoms, but thedata were insufficient to calculate the size of the effect.

Ozek and coworkers (36) reviewed six randomizedtrials to determine the effect of PET on primary out-comes of mortality and neurodevelopmental outcomes at2 years and at school age. Secondary outcomes includedseizures, cerebral infarction, NEC (Bell stage 2 or greater),hypoglycemia, hyperbilirubinemia, and thrombocytope-nia. Only one study reported data on mortality, and nosignificant increase was noted with PET (RR, 5.23; 95%CI, 0.66, 41.26). Four studies reported neurodevelop-mental outcomes at 18 months or older, and no signifi-cant delay was reported in the PET group (typical RR,1.45; 95% CI, 0.83 to 2.54 including all studies andtypical RR, 1.35; 95% CI, 0.68 to 2.69 when onlyrandomized, controlled trials were included). How-ever, these results were based on data limited by poorfollow-up and did not account for patients who were lostto follow-up. The authors performed a worst case/bestcase scenario post hoc analysis, which showed a signifi-cant skewing toward or away from the association ofPET with poor neurodevelopmental outcomes and wasconsidered to reflect the wide distribution of data pointsrather than actual outcomes. The authors concluded that

there is no significant benefit of PET in asymptomaticpatients or those who have mild symptoms.

Patients who have polycythemia and show signs orsymptoms that may be related to hyperviscosity are fre-quently treated with PET, even though the practice isnot supported by high-quality evidence. The argumentsin favor of PET are based on the pathophysiology ofhyperviscosity syndrome because most of the symptomsare presumed to be related to altered microcirculatoryperfusion and tissue hypoxia. (5)(7) By replacing some ofthe circulating RBC mass with a crystalloid solution,PET is believed to reduce blood viscosity and improveend-organ perfusion. However, no randomized clinicalstudies demonstrate a clear benefit of PET in the treat-ment of symptomatic polycythemia. The groups led byBada, (60) Ratrisawadi, (59) and Kumar (61) random-ized only asymptomatic polycythemic infants, whilethose led by Black, (58) Goldberg, (57) and Malan (56)made no distinction between symptomatic and asymp-tomatic newborns. In nonrandomized clinical reports,PET has been shown to lower pulmonary vascular resis-tance, improve cerebral blood flow velocity, (63)(64)and possibly normalize cerebral hemodynamics and im-prove the clinical status of infants who have polycythe-mia. (65) However, the long-term benefits of PET re-main unclear.

Concerns remain about potential adverse events fol-lowing PET. PET was associated with an increased riskof NEC in the systematic reviews performed by bothDempsey (42) (RR, 8.68; 95% CI, 1.06 to 71.1) andOzek (36) (two studies: typical RR, 11.18; 95% CI, 1.49,83.64 and typical risk difference, 0.14; 95% CI, 0.05,0.22). Malan and de V Heese (56) reported that 1 oftheir 24 patients in the exchanged group developed NECwithin the first 24 hours after PET compared with noneof the control patients. Black and associates (58) re-corded NEC in 8 of their 43 infants in the PET groupcompared with none of the 50 control infants. PET alsodid not alter the frequency of hypoglycemia (two studies)or thrombocytopenia (one study). (36) Given the risks ofPET for polycythemia in the newborn and the lack ofevidence indicating clear benefit, we are generally reluc-tant to use PET in asymptomatic infants. We do offerPET for treatment of infants who have symptoms thatcould be ascribed to hyperviscosity, but this decision istaken cautiously, with a careful review of all risk factors.Routine use of PET in neonatal polycythemia is notsupported by current evidence, and further study isneeded to identify patients who would be more likely tobenefit from aggressive correction of polycythemia.




Tab

le2.

Clin

ical

Tria

lsof

Part

ialE

xcha

nge

Tran

sfus

ion

inN

eona

talP

olyc

ythe

mia

Stud

yRa

ndom

izat

ion/

Sam

ple

Size

Hem

atoc

rit(H

ct)

atEn

rollm

ent

Mod

eof

Exch

ange

/Re

plac

emen

tflu

idO

utco

me

Mea

sure

men

ts

Studiesincludedinthemeta-analysisbyOzeketal(36)

Studiesincludedinthemeta-analysisbyDempseyandBarrington(42)M

alan

and

deV

Hee

se(5

6)

Not

clea

rlyst

ated

;n�

49H

ct>

65%

(0.6

5)UV

CFF

PBr

azel

ton

neon

atal

scal

ePr

echt

lneu

rolo

gic

asse

ssm

ent

at10

days

Neu

rode

velo

pmen

tal

asse

ssm

ent

at8

mon

ths

Gold

berg

etal

(57)

Not

clea

rlyst

ated

;n�

20H

ct>

64%

(0.6

4)an

dhy

perv

isco

sity

UVC

FFP

Braz

elto

nne

onat

alsc

ale

at8

hour

s,24

hour

s,72

hour

s,an

d2

wee

ksBS

IDan

dne

urol

ogic

asse

ssm

ent

at8

mon

ths

Blac

ket

al(5

8)Ra

ndom

ized

;n�

93An

tecu

bita

lven

ous

Hct

>65

%(0

.65)

and

hype

rvis

cosi

ty

UVC

FFP

Neo

nata

lsig

nsBS

IDan

dne

urol

ogic

asse

ssm

ent

at1

and

2ye

ars

Cogn

itive

test

ing

at7

year

sBa

daet

al(6

0)N

otcl

early

stat

ed;

n�28

Radi

alar

tery

Hct

>65

%(0

.65)

Rout

eno

tst

ated

;Pl

asm

apr

otei

nso

lutio

n

Cere

bral

arte

ryDo

pple

rm

easu

rem

ent

BSID

orco

gniti

vete

stin

gat

30m

onth

sRa

tris

awad

iet

al(5

9)Ra

ndom

ized

;n�

105

Cent

ralv

enou

sH

ct>

65%

(0.6

5)Ro

ute

not

stat

edGa

selD

evel

opm

ent

at1.

5to

2ye

ars

Hak

anso

net

al(6

2)*

Not

clea

rlyst

ated

;n�

24H

yper

visc

osity

(und

efine

d)N

otcl

early

stat

ed;

FFP

BSID

at8

mon

ths

Kum

aran

dRa

mji

(61)

Rand

omiz

ed;

n�22

Perip

hera

lven

ous

Hct

>70

%(0

.70)

Perip

hera

l;N

orm

alsa

line

Neu

rolo

gic

defic

itsat

3,6,

9,12

,18

mon

ths

FFP�

fres

hfr

ozen

plas

ma,

BSI

D�

Bay

ley

Scal

eof

Infa

ntD

evel

opm

ent,

UV

C�

umbi

lical

veno

uslin

e,*

�ab

stra

cton

ly




American Board of Pediatrics Neonatal-PerinatalMedicine Content Specifications• Know the causes of neonatal polycythemia.• Know the clinical manifestations,

management, and outcomes of neonatalpolycythemia.

References1. Brugnara C, Platt OS. The neonatal erythrocyte and its disor-ders. In: Orkin SH, Nathan DG, Ginsburg D, Look AT, eds.Nathan and Oski’s Hematology of Infancy and Childhood. Vol. 1.7th ed. Philadelphia, PA: Elsevier; 2009:36–662. Gross GP, Hathaway WE, McGaughey HR. Hyperviscosity inthe neonate. J Pediatr. 1973;82:1004–10123. Ramamurthy RS, Berlanga M. Postnatal alteration in hematocritand viscosity in normal and polycythemic infants. J Pediatr. 1987;110:929–9344. Oh W, Lind J. Venous and capillary hematocrit in newborninfants and placental transfusion. Acta Paediatr Scand. 1966;55:38–485. Sarkar S, Rosenkrantz TS. Neonatal polycythemia and hyper-viscosity. Semin Fetal Neonatal Med. 2008;13:248–2556. Jeevasankar M, Agarwal R, Chawla D, Paul VK, Deorari AK.Polycythemia in the newborn. Indian J Pediatr. 2008;75:68–727. Rosenkrantz TS. Polycythemia and hyperviscosity in the new-born. Semin Thromb Hemost. 2003;29:515–5278. Black LV, Maheshwari A. Disorders of the fetomaternal unit:hematologic manifestations in the fetus and neonate. Semin Peri-natol. 2009;33:12–199. Henry E, Walker D, Wiedmeier SE, Christensen RD. Hemato-logical abnormalities during the first week of life among neonateswith Down syndrome: data from a multihospital healthcare system.Am J Med Genet A. 2007;143:42–5010. Mannino F. Neonatal complications of postterm gestation.J Reprod Med. 1988;33:271–27611. Awonusonu FO, Pauly TH, Hutchison AA. Maternal smokingand partial exchange transfusion for neonatal polycythemia. Am JPerinatol. 2002;19:349–35412. Halliday HL, Reid MM, McClure G. Results of heavy drinkingin pregnancy. Br J Obstet Gynaecol. 1982;89:892–89513. Snijders RJ, Abbas A, Melby O, Ireland RM, Nicolaides KH.Fetal plasma erythropoietin concentration in severe growth retar-dation. Am J Obstet Gynecol. 1993;168:615–61914. Weiner CP, Williamson RA. Evaluation of severe growth retar-dation using cordocentesis–hematologic and metabolic alterationsby etiology. Obstet Gynecol. 1989;73:225–22915. Bussmann YL, Tillman ML, Pagliara AS. Neonatal thyrotoxi-cosis associated with the hyperviscosity syndrome. J Pediatr. 1977;90:266–26816. Zimmerman D. Fetal and neonatal hyperthyroidism. Thyroid.1999;9:727–73317. Mimouni F, Miodovnik M, Siddiqi TA, Butler JB, Holroyde J,Tsang RC. Neonatal polycythemia in infants of insulin-dependentdiabetic mothers. Obstet Gynecol. 1986;68:370–37218. Nelson SM, Freeman DJ, Sattar N. Lindsay RS. Erythrocytosisin offspring of mothers with type 1 diabetes–are factors other thaninsulin critical determinants? Diabet Med. 2009;26:887–892

19. Weinberger MM, Oleinick A. Congenital marrow dysfunctionin Down’s syndrome. J Pediatr. 1970;77:273–27920. Miller M, Cosgriff JM. Hematological abnormalities in new-born infants with Down syndrome. Am J Med Genet. 1983;16:173–17721. Widness JA, Pueschel SM, Pezzullo JC, Clemons GK. Elevatederythropoietin levels in cord blood of newborns with Down’ssyndrome. Biol Neonate. 1994;66:50–5522. Wiedmeier SE, Henry E, Christensen RD. Hematological ab-normalities during the first week of life among neonates withtrisomy 18 and trisomy 13: data from a multi-hospital healthcaresystem. Am J Med Genet A. 2008;146:312–32023. Rabe H, Mercer JS. Early cord clamping protects at-risk neo-nates from polycythemia. Biol Neonate. 2004;86:10824. Hutton EK, Hassan ES. Late vs early clamping of the umbilicalcord in full-term neonates: systematic review and meta-analysis ofcontrolled trials. JAMA. 2007;297:1241–125225. Rabe H, Reynolds G, Diaz-Rossello J. A systematic review andmeta-analysis of a brief delay in clamping the umbilical cord ofpreterm infants. Neonatology. 2008;93:138–14426. Airey RJ, Farrar D, Duley L. Alternative positions for the babyat birth before clamping the umbilical cord. Cochrane DatabaseSyst Rev. 2010;10:CD00755527. Linderkamp O, Versmold HT, Messow-Zahn K, Muller-Holve W, Reigel KP, Betke K. The effect of intra-partum andintra-uterine asphyxia on placental transfusion in premature andfull-term infants. Eur J Pediatr 1978;127:91–9928. Rabe H, Wacker A, Hulskamp G, et al. A randomised con-trolled trial of delayed cord clamping in very low birth weightpreterm infants. Eur J Pediatr. 2000;159: 775–77729. Chalouhi GE, Stirnemann JJ, Salomon LJ, Essaoui M, QuibelT, Ville Y. Specific complications of monochorionic twin pregnan-cies: twin-twin transfusion syndrome and twin reversed arterialperfusion sequence. Semin Fetal Neonatal Med. 2010;15:349–35630. Somer T, Meiselman HJ. Disorders of blood viscosity. AnnMed. 1993;25:31–3931. Linderkamp O. Blood rheology in the newborn infant. Bail-lieres Clin Haematol. 1987;1:801–82532. Pearson TC. Hemorheology in the erythrocytoses. Mt SinaiJ Med. 2001;68:182–19133. Pozrikidis C. Numerical simulation of blood and interstitialflow through a solid tumor. J Math Biol. 2010;60:75–9434. Sirs JA. The flow of human blood through capillary tubes.J Physiol. 1991;442:569–58335. Herson VC, Raye JR, Rowe JC, Philipps AF. Acute renalfailure associated with polycythemia in a neonate. J Pediatr. 1982;100:137–13936. Ozek E, Soll R, Schimmel MS. Partial exchange transfusionto prevent neurodevelopmental disability in infants with polycy-themia. Cochrane Database Syst Rev. 2010;1:CD00508937. Rosenkrantz TS, Stonestreet BS, Hansen NB, Nowicki P,Oh W. Cerebral blood flow in the newborn lamb with polycythemiaand hyperviscosity. J Pediatr. 1984;104:276–28038. Rosenkrantz TS, Phillipps AF, Knox I, et al. Regulation ofcerebral glucose metabolism in normal and polycythemic newbornlambs. J Cereb Blood Flow Metab. 1992;12:856–86539. Rosenkrantz TS, Philipps AF, Skrzypczak PS, Raye JR. Cere-bral metabolism in the newborn lamb with polycythemia. PediatrRes. 1988;23:329–33340. Drew JH, Guaran RL, Cichello M, Hobbs JB. Neonatal wholeblood hyperviscosity: the important factor influencing later neuro-




logic function is the viscosity and not the polycythemia. ClinHemorheol Microcirc. 1997;17:67–7241. Shohat M, Merlob P, Reisner SH. Neonatal polycythemia: I.Early diagnosis and incidence relating to time of sampling. Pediat-rics. 1984;73:7–1042. Dempsey EM, Barrington K. Short and long term outcomesfollowing partial exchange transfusion in the polycythaemic new-born: a systematic review. Arch Dis Child Fetal Neonatal Ed. 2006;91:F2–F643. Saggese G, Bertelloni S, Baroncelli GI, Cipolloni C. Elevatedcalcitonin gene-related peptide in polycythemic newborn infants.Acta Paediatr. 1992;81:966–96844. Dong YL, Chauhan M, Green KE, et al. Circulating calcitoningene-related peptide and its placental origins in normotensive andpreeclamptic pregnancies. Am J Obstet Gynecol. 2006;195:1657–166745. Lambert DK, Christensen RD, Henry E, et al. Necrotizingenterocolitis in term neonates: data from a multihospital health-caresystem. J Perinatol. 2007;27:437–44346. Martinez-Tallo E, Claure N, Bancalari E. Necrotizing entero-colitis in full-term or near-term infants: risk factors. Biol Neonate.1997;71:292–29847. Wiswell TE, Robertson CF, Jones TA, Tuttle DJ. Necrotizingenterocolitis in full-term infants. A case-control study. Am J DisChild. 1988;142:532–53548. Wilson R, del Portillo M, Schmidt E, Feldman RA, Kanto WPJr. Risk factors for necrotizing enterocolitis in infants weighingmore than 2,000 grams at birth: a case-control study. Pediatrics.1983;71:19–2249. Maayan-Metzger A, Itzchak A, Mazkereth R, Kuint J. Necro-tizing enterocolitis in full-term infants: case-control study andreview of the literature. J Perinatol. 2004;24:494–49950. Acunas B, Celtik C, Vatansever U, Karasalihoglu S. Thrombo-cytopenia: an important indicator for the application of partialexchange transfusion in polycythemic newborn infants? Pediatr Int.2000;42:343–34751. Katz J, Rodriguez E, Mandani G, Branson HE. Normal coag-ulation findings, thrombocytopenia, and peripheral hemoconcen-tration in neonatal polycythemia. J Pediatr. 1982;101:99–10252. Rivera LM, Rudolph N. Postnatal persistence of capillary-venous differences in hematocrit and hemoglobin values in low-birth-weight and term infants. Pediatrics. 1982;70:956–957

53. Dempsey EM, Barrington K. Crystalloid or colloid for partialexchange transfusion in neonatal polycythemia: a systematic reviewand meta-analysis. Acta Paediatr. 2005;94:1650–165554. de Waal KA, Baerts W, Offringa M. Systematic review of theoptimal fluid for dilutional exchange transfusion in neonatal poly-cythaemia. Arch Dis Child Fetal Neonatal Ed. 2006;91:F7–F1055. Wong W, Fok TF, Lee CH, et al. Randomised controlled trial:comparison of colloid or crystalloid for partial exchange transfusionfor treatment of neonatal polycythaemia. Arch Dis Child FetalNeonatal Ed. 1997;77:F115–F11856. Malan AF, de V Heese H. The management of polycythaemiain the newborn infant. Early Hum Dev. 1980;4:393–40357. Goldberg K, Wirth FH, Hathaway WE, et al. Neonatal hyper-viscosity. II. Effect of partial plasma exchange transfusion. Pediat-rics. 1982;69:419–42558. Black VD, Lubchenco LO, Koops BL, Poland RL, Powell DP.Neonatal hyperviscosity: randomized study of effect of partialplasma exchange transfusion on long-term outcome. Pediatrics.1985;75:1048–105359. Ratrisawadi V, Plubrukarn R, Trakulchang K, Puapondh Y.Developmental outcome of infants with neonatal polycythemia.J Med Assoc Thai. 1994;77:76–8060. Bada HS, Korones SB, Kolni HW, et al. Partial plasma ex-change transfusion improves cerebral hemodynamics in symptom-atic neonatal polycythemia. Am J Med Sci. 1986;291:157–16361. Kumar A, Ramji S. Effect of partial exchange transfusion inasymptomatic polycythemic LBW babies. Indian Pediatr. 2004;41:366–37262. Hakanson DO. Neonatal hyperviscosity syndrome: long-termbenefit of partial plasma exchange transfusion. [Abstract]. PediatrRes. 1981;15:449A63. Murphy DJ Jr, Reller MD, Meyer RA, Kaplan S. Effects ofneonatal polycythemia and partial exchange transfusion on cardiacfunction: an echocardiographic study. Pediatrics. 1985;76:909–91364. Maertzdorf WJ, Tangelder GJ, Slaaf DW, Blanco CE. Effectsof partial plasma exchange transfusion on cerebral blood flowvelocity in polycythaemic preterm, term and small for date newborninfants. Eur J Pediatr. 1989;148:774–77865. Bada HS, Korones SB, Pourcyrous M, et al. Asymptomaticsyndrome of polycythemic hyperviscosity: effect of partial plasmaexchange transfusion. J Pediatr. 1992;120:579–585




NeoReviews Quiz

6. Polycythemia represents an abnormal elevation of the circulating red blood cell mass, which increases therisk of hyperviscosity, microcirculatory hypoperfusion, and multisystem organ dysfunction. Of the following,the hematocrit threshold that best defines polycythemia is:

A. 55% (0.50).B. 60% (0.60).C. 65% (0.65).D. 70% (0.70).E. 75% (0.75).

7. Polycythemia in neonates is often associated with metabolic complications. Of the following, the mostcommon metabolic abnormality associated with polycythemia in neonates is:

A. Hyperchloremia.B. Hyperkalemia.C. Hypernatremia.D. Hypocalcemia.E. Hypoglycemia.

8. The treatment of polycythemia includes partial exchange transfusion in which a precalculated volume ofblood is replaced by an equivalent volume of fluid. Of the following, the most accepted choice of fluid forpartial exchange transfusion is:

A. Albumin solution.B. Fresh frozen plasma.C. Lactated Ringer solution.D. Normal saline.E. Plasma protein fraction.





Juan I. Remon, Aarti Raghavan and Akhil Maheshwari Polycythemia in the Newborn













Amélia Miyashiro Nunes dos Santos and Cleide Enoir Petean Trindade Red Blood Cell Transfusions in the Neonate







Red Blood Cell Transfusions in the NeonateAmelia Miyashiro Nunes

dos Santos, MD, PhD,*

Cleide Enoir Petean

Trindade, MD, PhD†

Author Disclosure

Drs Nunes dos Santos

and Trindade have

disclosed no financial



commentary does not


of an unapproved/


commercial

product/device.

AbstractDespite recent trends to decrease allogeneic red blood cell (RBC) transfusion thresh-olds, such transfusions remain an important supportive and life-saving intervention forneonatal intensive care patients. In neonates, apart from concerns about transfusion-associated infections, many controversial questions regarding transfusion practicesremain unanswered. Moreover, neonates present specific clinical and immunologiccharacteristics that require selected blood component products. This article addressesmany of these issues from a medical perspective, with emphasis on the best bloodbanking techniques to provide RBC products for neonatal transfusions.


1. Describe RBC products available for use in neonates.2. Choose the best RBC products for use in preterm infants and in neonates who have

Rhesus hemolytic diseases or ABO incompatibility hemolytic diseases.

IntroductionTransfusion therapy in the neonate poses specific challenges, requiring consideration ofissues not found in older patients. The transfer of maternal antibodies across the placenta,the immature immune system, and the small blood volume of the fetus and neonate maketransfusion therapy a special issue in this group of patients.

Transfusion compatibility testing must take into account not only the mother’s bloodtype but also blood group antibodies against the fetus or neonate’s blood type. Theimmature immune system of fetuses and neonates makes them more susceptible totransfusion-associated cytomegalovirus infections and transfusion-associated graft versushost disease (TA-GVHD). Moreover, fetuses and neonates, especially those who areextremely low birthweight (�1,000 g) may be at risk of hemodynamic instability ifvolumes are not appropriately adjusted, hypothermia if blood is cold, and hyperkalemia iflarge volumes of blood product that have high concentrations of potassium are infusedrapidly. Also, particular attention must be paid when transfusing blood group O toneonates who have group A, B, or AB blood. The quantity of anti-A or anti-B antibodiespresent in the type O RBC product can produce hemolysis in the neonate. All of theseissues should be considered when choosing the RBC product for transfusion in neonates.

RBC ProductsMost RBC components available today are derived from the collection of 450 to 500 mLof whole blood into sterile plastic bags containing citrate-phosphate-dextrose (CPD)anticoagulant. Because RBCs, platelets, and plasma have different specific gravities, theycan be separated from each other by centrifugation. RBCs are separated from platelet-richplasma by performing centrifugation of the whole blood. RBCs are then collected into asterile satellite bag containing an anticoagulant solution, generally a nutrient additivesolution. A variety of additive solutions are used, containing a mix of glucose, adenine, andin some cases, mannitol. These solutions prolong the shelf life of the RBC product from21 days for packed RBCs in CPD to 35 days for RBCs in citrate-phosphate-dextrose-

*Associate Professor, Department of Pediatrics, Neonatal Division of Medicine, Federal University of Sao Paulo, Sao Paulo, SP,Brazil.†Emeritus Professor of Pediatrics, Department of Pediatrics, Division of Neonatology, Botucatu School of Medicine, Sao PauloState University-UNESP, Botucatu, Sao Paulo, Brazil.

Article hematology



adenine (CPDA-1) and to 42 days for RBCs in additivesolutions. (1)(2) Transfusion of 10 mL/kg of RBCs isexpected to raise the patient’s hematocrit by 9% (0.9) to10% (0.10) for CPDA-1 packed RBCs and by 7% (0.7) to8% (0.8) for RBCs preserved in additive solutions. (2)(3)

Currently, an increasing proportion of blood compo-nents are being collected by an automated apheresisinstrument, which draws blood into an external circuit,separates the components by centrifugation or filtration,collects the desired component, and returns the remain-ing blood components to the donor. RBCs collected byapheresis contain additive solution. (3)

RBC components for use in neonates are modifiedbecause of particular immunologic characteristics of thisgroup of patients. Leukoreduced RBCs and irradiatedRBCs are used in most countries for newborns, andwashed RBCs are used for certain special conditions. (4)

Leukocyte-reduced RBC productsMost RBC products are modified soon after collection bypassing them through special filters that adsorb whiteblood cells (WBCs), decreasing them from 109 per unitto 106 per unit. Currently, third-generation leukocytereduction filters consistently provide a 99.9% reductionof WBCs, with a final product of 1 to 5�106 WBCs perunit. (3)

Mukagatare and associates (5) reported that the im-plementation of universal leukocyte reduction signifi-cantly decreased the rate of all transfusion reactions from0.49% to 0.31% (P�0.001), the rates of febrile non-hemolytic transfusion reactions from 0.35% to 0.24%(P�0.002), and the rate of allergic reactions from 0.05%to 0.01% (P�0.001). However, the effect of leukocyte-reduced preparations on patient outcomes is still notevidence-based, (6) except for some benefits in patientsundergoing cardiac surgery. (7)(8) In preterm infants,leukocyte-depleted blood components have been used inmany countries, (4) based on a possible cause-effectrelationship between the implementation of universalWBC reduction and a lower incidence of bronchopul-monary dysplasia (odds ratio [OR], 0.42; 95% con-fidence interval [CI], 0.25 to 0.70), retinopathy ofprematurity (OR, 0.56; 95% CI, 0.33 to 0.93) andnecrotizing enterocolitis (OR, 0.39; 95% CI, 0.17 to0.93). (9) However, in neonatal settings, the primaryobjective of leukocyte reduction is to reduce transfusion-associated cytomegalovirus (CMV) infection. Preterminfants, seronegative neonates whose birthweights areless than 1,250 g and require blood component therapy,

or fetuses requiring intrauterine transfusions are at in-creased risk for posttransfusion CMV-related morbidityand mortality. (2)

CMV-seronegative blood is considered the gold stan-dard for preventing transfusion-associated CMV infec-tions, but such products often are difficult to obtain inareas where the prevalence of positive CMV antibodies ishigh. Because CMV is harbored within WBCs, leukocytereduction should decrease the risk of this infection.Whether leukocyte reduction is as efficacious as CMV-seronegative blood is controversial. A meta-analysis ofthe available controlled studies indicates that CMV-seronegative blood components are more efficaciousthan WBC-reduced blood components in preventingtransfusion-acquired CMV infection. (10) However, be-cause use of WBC-reduced components may have otheradvantages in neonates, they should be used, regardlessof CMV serologic status.

Irradiated RBC ProductsIrradiation inactivates donor lymphocytes present in theblood component, preventing them from proliferatingand thereby reducing the risk of TA-GVHD. TA-GVHDcan occur when viable donor lymphocytes are transfusedinto patients who have cellular immunodeficiency orwhen donor and recipient share similar human leukocyteantigens (HLAs), such as with blood donation fromfamily members. Neonates, especially preterm infants,are at high risk for TA-GVHD. Some neonatal centersirradiate all cellular blood products for infants youngerthan 4 months of age; others irradiate only blood prod-ucts given to preterm infants whose birthweight is 1.2 kgor less. (4) TA-GVHD has been reported after intrauter-ine transfusion in preterm infants and in term infants whoreceived exchange transfusion. (11)(12)

Blood components are usually subjected to a mini-mum of 25 Gy gamma irradiation. (13) Irradiation in-creases potassium leakage from RBCs during storage,and irradiated RBCs have a shortened shelf life. (14) TheUnited States Food and Drug Administration recom-mends a 28-day expiration for irradiated RBCs becausein vivo recovery of irradiated RBCs is decreased com-pared with nonirradiated RBCs at 42 days of storage.(15) Therefore, it is preferable to use irradiated RBCsclose to the administration rather than prolonged refrig-erator storage products, especially for neonates, who maynot be able to tolerate a large potassium load. Irradiationof blood products has not been shown to prevent post-transfusion CMV infection. (14)(16)

hematology red blood cell transfusions



Washed RBC ProductsSaline washing can be used to remove plasma or additivesolution from blood components or to reduce the po-tassium content in the supernatant of stored RBCs.This process involves the addition of saline to cellularblood components, followed by centrifugation, removalof supernatant, and resuspension of the cells in salinesolution or plasma. Washed RBC products are used mostfrequently in neonatal settings, for intrauterine trans-fusions or exchange transfusions, or when transfusingmore than 20 mL/kg of RBCs that have been stored formore than 14 days. (1)(3) Washed products must beused within 4 hours of processing if stored at roomtemperature or within 24 hours if stored in the refriger-ator because the washing process itself creates an opensystem. (17)

Selection of the RBC ProductsIntrauterine Transfusion

The primary indication for prenatal RBC transfusions ismanagement of hemolytic disease of the fetus and new-born. Maternal antibodies directed against the fetalRBCs are produced when fetal RBCs are positive for acertain antigen, usually Rhesus D (RhD), and pass intothe blood circulation of a mother who is negative for thatparticular antigen. Passage of maternal immunoglobulinG antibodies across the placenta into the fetal circulationcan cause mild-to-severe hemolytic anemia and fetalhydrops.

During the past few decades, prenatal care strategiesfor patients who have rhesus hemolytic diseases haveevolved significantly, including the introduction of RhDprophylaxis, the development of Doppler ultrasono-graphic techniques to detect fetal anemia, and treatmentwith intrauterine blood transfusions. Such developmentshave led to a dramatic decrease in perinatal mortality.(18)

The first intrauterine transfusion was performed byLiley in 1963 by the intraperitoneal route. (19) DonorRBCs administered by that route are absorbed into thecirculation via the diaphragmatic lymphatics. This methodresults in slower restoration of fetal hemoglobin thanobserved with blood administered directly into the fetalcirculation and, as a result, is no longer the route ofchoice for fetal therapy, except in cases requiring trans-fusions before 18 to 20 weeks of gestation, when intra-vascular access is not possible. (20)

Depending on disease severity, either transfusions orexchange transfusions can be performed (in utero) viathe umbilical vein. The RBCs chosen for intrauterinetransfusions for Rhesus hemolytic disease are typically

group O, RhD-negative, CMV-reduced risk (CMV-negative antibody or leukocyte-reduced), and gamma-irradiated to avoid TA-GVHD. The selected RBC unitsfor intrauterine transfusions must be compatible withboth the fetus and the mother. If mother and fetus havethe same ABO group, then group-identical units may bechosen. Fresh RBC product (usually within 72 hoursafter collection) can be used to maintain 2,3-diphos-phoglycerate concentrations and avoid high potassiumconcentrations. Group O RBCs should be saline-washedto eliminate plasma that contains anti-A and anti-B anti-bodies, if the fetal blood type is A, B, AB, or unknown.The RBC product can be resuspended in either normalsaline or AB plasma at the desired hematocrit. The RBCsshould be negative for blood group antigens to which themother has made antibodies and crossmatch-compatiblewith the mother’s plasma or serum because any bloodgroup antibodies in the fetal circulation will be derivedfrom the mother. (21) RBCs used for exchange trans-fusions or intrauterine transfusions should be negativefor hemoglobin S screening to optimize oxygen-carryingcapacity. Washed and irradiated maternal blood can beused when there is no suitable blood donor. (22) Cellsare packed to a hematocrit of 75% (0.75) to 85%(0.85) to prevent volume overload. The volume ofRBCs to be transfused to achieve a hematocrit incrementof 10% (0.10) can be roughly calculated by multiplyingfetal weight (g) estimated on ultrasonography by a factorof 0.02. (23) Usually, a final target hematocrit of 40%(0.40) to 50% (0.50) is used. (18) The rate of RBCinfusion should be about 5 to 10 mL/min. (24)

Exchange Transfusion After BirthCurrently, with the use of both intensive phototherapymanagement and immune globulin intravenous, ex-change transfusion rarely is indicated in Rhesus or ABOhemolytic diseases. (25) Exchange transfusions, whenindicated, (25) are performed with double-volume trans-fusion (160 mL/kg) using irradiated, leukocyte-reducederythrocytes (cross-matched against the mother andcompatible with the infant), usually via an umbilical veincatheter.

For Rhesus hemolytic diseases, ABO-compatibleRhD-negative or group O RhD-negative RBCs resus-pended in normal saline or neonate-compatible plasmaare used for small-volume or exchange transfusion.

Type O Rh-compatible RBCs must be used for ex-change transfusions or RBC transfusions in neonateswho have ABO hemolytic diseases because of the pres-ence of maternal antibodies (anti-A or anti-B) against theneonate’s RBCs in the neonatal circulation until 4 to




6 months of age. Therefore, type O Rh-compatible RBCssuspended in neonatal-compatible plasma are indicatedfor exchange transfusion. For example, in a group A Rh-positive neonate who has ABO-incompatible hemolyticdisease, group O Rh-positive (or -negative) RBCs sus-pended in plasma type A must be used for exchangetransfusion. If type A RBCs are used, the anti-A anti-body transferred from the maternal to fetal circulationwill cause hemolysis of the transfused erythrocytes and asubstantial increase in the bilirubin concentration, in-creasing the risk of kernicterus.

Although exchange transfusion for hemolytic diseasesof the newborn is used rarely, it may be beneficial in somecases of inborn errors of metabolism, especially ornithinetranscarbamylase deficiency. This disease is the mostcommon urea cycle defect in which hyperammonemiaand different nitrogenous precursors of urea accumulate,leading to coma and a high mortality rate.

Transfusion in the NeonateBefore receiving any RBC transfusions, neonates mustbe typed (ABO and Rh group) and screened for pas-sively transferred RBC antibody of maternal origin. If theantibody screen is negative, no further antibody screensare required during the hospitalization until the child is4 months old, and ABO/Rh-compatible RBCs may beprovided. If the antibody screen is positive, RBCs thatare compatible with the maternally derived antibodymust be provided until the maternal antibody is nolonger demonstrable. (1)(2)

In many countries, leukocyte-reduced, gamma-irradiated, and ABO-Rh-compatible RBC products areused for transfusions in neonates, especially those whosebirthweights are less than 1,200 g. When group O RBCsare used for transfusing type A or B neonates, centrifu-gation before transfusion to remove the excess of plasmathat contains anti-A and anti-B antibodies can reduce therisk of hemolysis in the neonate.

Directed donations from family members, who oftenhave never donated blood before, are not recommended.This blood product presents a higher risk of infectiousdisease and TA-GVHD than those provided by standardblood donors.

Mothers are not the ideal donors for their infantsbecause maternal plasma frequently contains a variety ofantibodies against RBCs, leukocytes, platelets, and HLAantigens expressed by neonatal cells. When maternalRBCs are used for transfusion in the neonate, washingwith saline solution can remove the excess of plasma withthese antibodies. (1) Moreover, the product must begamma-irradiated to prevent TA-GVHD.

Improvements in prenatal and postnatal care have ledto a decline in the number of exchange transfusions. As aconsequence, the most frequently transfused neonatesare preterm infants, especially those of very low birth-weight (�1,500 g). In this group of patients, phlebot-omy blood losses and a slower rate of erythropoiesis dueto low concentrations of erythropoietin result in progres-sive anemia during the first few months after birth. (26)However, the optimal hemoglobin threshold for RBCtransfusions in preterm infants is unknown. (27) Restric-tive guidelines for RBC transfusions have been devel-oped in an effort to limit transfusion-associated risks.(28)(29)(30)(31)(32) Some studies showed the efficacyof restrictive guidelines to reduce transfusion rates inpreterm infants without increasing the morbidity andmortality rate. (30)(32) However, scientifically basedrisk/benefit evidence for short-term clinical outcomeand long-term neurodevelopmental outcome of usingrestrictive or liberal guidelines is still lacking. (27)(28)(29)(30)(33) Accordingly, varying transfusion practiceshave been described in children and neonates betweeninstitutions and countries. (4)(34)(35) The guidelinesfor RBC transfusions in preterm neonates were publishedin NeoReviews by Ohls. (36)

To limit donor exposure, aliquots can be preparedfrom a single dedicated blood donor for small-volumetransfusions. Pediatric bags or sterile connecting devicescan be used to make small aliquots from an adult bloodunit. Transfusions of small-volume RBCs stored in addi-tive anticoagulant-preservative solutions (for 42 days) orCPDA-1 (for 35 days) decrease blood donor exposuresin newborns. (37)(38) Strauss and colleagues (37) no-ticed a reduction from 6.5�3.7 to 1.9�0.6 (P�0.005)in the number of donors to which infants were exposedfor preterm infants transfused with RBCs stored for up to42 days compared with neonates that received erythro-cytes preserved for up to 7 days. In this study, no majorclinical changes or significant differences in infants’ pHvalues and plasma sodium, potassium, lactate, and glu-cose values were observed. Usually, the transfusion of amaximum of 20 mL/kg of RBCs, slowly infused over4 hours, is not followed by acidosis or hyperkalemiaunless the neonate already has impaired renal functionand a high potassium concentration. (1)(38) Fernandesda Cunha and associates, (38) using a sterile connectingdevice for transfusion of small aliquots of leukocyte-reduced and gamma-irradiated RBCs preserved inCPDA-1 for up to 28 days after collection, noted a 70%reduction of donor exposure in very low-birthweight(�1,500 g) infants.

The use of restrictive guidelines and “dedicated” RBC




units for small-volume (10 to 20 mL/kg) RBC transfu-sions in very low-birthweight infants has progressivelydecreased donor exposure and, therefore, the known andunknown risks of allogeneic transfusions. However, afteran increased storage time, RBCs undergo numerouschanges that may affect their functions. The primaryhypothesis is that the longer blood is stored, the lesseffective is the oxygen delivery to tissues. Moreover,transfusion of older stored blood to critically ill infantsmay trigger immunosuppressive or proinflammatoryeffects, resulting in higher organ dysfunction and mor-bidity or mortality rates. (13) A multicenter, double-blind, randomized, controlled trial is in place to deter-mine if RBCs stored for 7 days or less (fresh RBCs)compared with current standard transfusion practice de-creases major nosocomial infection and organ dysfunc-tion in neonates admitted to the neonatal intensive careunit. (39)

Controversial Transfusion Practicesin Neonates

Delayed Umbilical Cord ClampingThe beneficial effects of delaying clamping of the umbil-ical cord have been questioned. Early clamping of thecord (5 to 10 seconds after birth), compared with laterclamping, can decrease the neonatal blood volume by20 to 40 mL/kg. (40)

Strauss and colleagues (41) randomized 105 neonatesfrom 30 to 36 weeks of gestation to early cord clamping(maximum of 15 seconds) and delayed cord clamping(1 minute). Circulating RBC volume/mass (P�0.04)and weekly hematocrit values (P�0.005) were higherafter delayed clamping, but the number of RBC transfu-sions was similar between groups (P�0.70). Require-ments for mechanical ventilation were similar. More(P�0.03) neonates needed phototherapy after delayedclamping, but initial bilirubin values and extent of pho-totherapy did not differ.

Rabe and associates, (42) in a systematic review ofdelayed cord clamping based on 10 studies involving 454preterm infants, found that the major benefits of theintervention were higher circulating blood volume dur-ing the first 24 hours after birth, less need for bloodtransfusions (P�0.004), and lower incidence of intra-ventricular hemorrhage (P�0.002).

Hutton and Hassan (43) published a meta-analysis of15 controlled trials involving term infants that comparedcord clamping delayed for at least 2 minutes (1,001newborns) and early clamping immediately after birth(911 newborns) in terms of benefits over ages 2 to 6

months. In this study, late cord clamping improved thehematocrit and iron status and reduced the relative riskof anemia: 0.53 (95% CI, 0.40 to 0.70). Neonates whohad late cord clamping were at increased relative risk ofasymptomatic polycythemia: 3.91 (95% CI, 1.00 to15.36).

In a randomized study of 72 infants from 24 to31 weeks of gestation, Mercer and coworkers (44) showedthat infants in the delayed cord clamping group (30 to45 seconds) had significantly less intraventricular hem-orrhage and late-onset sepsis compared with the earlycord clamping (�10 sec) group, with an advantage formale infants for both outcomes. A follow-up evaluationat 7 months of corrected age showed no differences inthe Bayley Scales of Infant Development scores betweendelayed and early groups. However, in a subset analysis ofmale neonates, regression model analysis adjusted forgestational age, intraventricular hemorrhage, broncho-pulmonary dysplasia, sepsis, and male sex showed thatinfants in whom cord clamping was delayed had highermotor scores. (45)

Delayed cord clamping in preterm infants who arewell at birth seems to be safe and to confer some ben-efit. However, larger and longer term follow-up in-vestigations are required to confirm the benefits beforethe practice can be recommended for routine use. (46)

Autologous Cord Blood TransfusionAttempts have been made to collect, store, and transfuseblood from the placenta/umbilical cord, especially inlow-resource countries. However, several and significantproblems must be solved before this practice can berecommended routinely, including collection of insuffi-cient volumes of placental blood for meaningful transfu-sions, clots forming during both collection and storage,and bacterial contamination. Volumes collected from theplacenta vary greatly among individual infants, averagingabout 20 mL/kg of whole blood, which is sufficient foronly one to two transfusions. (47) A review article (47)reported an analysis of 154 papers searched on thePubMed and MEDLINE database related to the use ofcord blood for transfusion purposes. The percentage ofcontaminated units of cord blood collections was lessthan 5%, but Ademokun and associates (48) reported a12% incidence of bacterial contamination. These prob-lems and the high costs have led authors to questionwhether, on balance, autologous/placental RBC trans-fusions offer clinically significant benefits. (49)




References1. Strauss RG. Blood banking issues pertaining to neonatal redblood cell transfusions. Transfus Sci. 1999;2:7–192. Fasano R, Luban NL. Blood component therapy. Pediatr ClinNorth Am. 2008;55:421–453. Galal SA. Therapeutic techniques. Selection of blood compo-nents for neonatal transfusion. NeoReviews. 2005;6:e351–e3554. New HV, Stanworth SJ, Engelfriet CP, et al. Neonatal transfu-sions. Vox Sang. 2009;96:62–855. Mukagatare I, Monfort M, de Marchin J, Gerard C. The effect ofleukocyte-reduction on the transfusion reactions to red blood cellsconcentrates [French]. Transfus Clin Biol. 2010;17:14–196. Vamvakas EC, Blajchman MA. Transfusion-related immuno-modulation (TRIM): an update. Blood Rev. 2007;21:327–3487. Vamvakas EC, Blajchman MA. Transfusion-related mortality:the ongoing risks of allogeneic blood transfusion and the availablestrategies for their prevention. Blood. 2009;113:3406–34178. van de Watering LM, Hermans J, Houbiers JG, et al. Beneficialeffects of leukocyte depletion of transfused blood on postoperativecomplications in patients undergoing cardiac surgery: a randomizedclinical trial. Circulation. 1998;97:562–5689. Fergusson D, Hebert PC, Lee SK, et al. Clinical outcomesfollowing institution of universal leukoreduction of blood transfu-sions for premature infants. JAMA. 2003;289:1950–195610. Vamvakas EC. Is white blood cell reduction equivalent toantibody screening in preventing transmission of cytomegalovirusby transfusion? A review of the literature and meta-analysis. Trans-fus Med Rev. 2005;19:181–19911. Flidel O, Barak Y, Lifschitz-Mercer B, Frumkin A, MoglilnerBM. Graft versus host disease in extremely low birth weight neo-nate. Pediatrics. 1992;89:689–69012. Ruhl H, Bein G, Sachs UJ. Transfusion-associated graft-versus-host disease. Transf Med Rev. 2009;23:62–7113. Moroff G, Leitman SF, Luban NL. Principles of blood irradi-ation, dose validation, and quality control. Transfusion. 1997;37:1084–109214. Zubair AC. Clinical impact of blood storage lesions. Am JHematol. 2010;85:117–12215. Davey RJ, McCoy NC, Yu M, Sullivan JA, Spiegel DM,Leitman SF. The effect of prestorage irradiation on posttransfusionred cell survival. Transfusion. 1992;32:525–52816. Wu Y, Zou S, Cable R, et al. Direct assessment of cytomega-lovirus transfusion-transmitted risks after universal leukoreduction.Transfusion. 2010;50:776–78617. Cambell-Lee SA. Packed RBCs and related products. In:Hillyer CD SL, Ness PM, Anderson KC, eds. Blood Banking andTransfusion Medicine Basic Principles and Practice. Philadelphia,PA: Churchill Livingstone; 2003:145–152

18. Moise KJ Jr. Management of rhesus alloimmunization in preg-nancy. Obstet Gynecol. 2008;112:164–17619. Liley AW. Intrauterine transfusion of foetus in haemolyticdisease. Br Med J. 1963;2:1107–110920. Fox C, Martin W, Somerset DA, Thompson PJ, Kilby MD.Early intraperitoneal transfusion and adjuvant maternal immuno-globulin therapy in the treatment of severe red cell alloimmuniza-tion prior to fetal intravascular transfusion. Fetal Diagn Ther. 2008;23:159–16321. Wu Y, Stack G. Blood product replacement in the perinatalperiod. Semin Perinatol. 2007;31:262–27122. Cassandra DJ, Strauss RG. Transfusion of Neonates and Pedi-atric Patients Technical Manual. 14th ed. Bethesda, MD: AmericanAssociation of Blood Banks; 200223. Giannina G, Moise KJ Jr, Dorman K. A simple method toestimate volume for fetal intravascular transfusions. Fetal DiagnTher. 1998;13:94–9724. British Committee for Standards in Haematology. Transfusionguidelines for neonates and older children. Br J Haematol. 2004;124:433–45325. American Academy of Pediatrics Subcommittee on Hyper-bilirubinemia. Management of hyperbilirubinemia in the new-born infant 35 or more weeks of gestation. Pediatrics. 2004;114:297–31626. Luban NL. Management of anemia in the newborn. EarlyHum Dev. 2008;84:493–49827. Strauss RG. How I transfuse red blood cells and platelets toinfants with the anemia and thrombocytopenia of prematurity.Transfusion. 2008;48:209–21728. Bell EF, Strauss RG, Widness JA, et al. Randomized trial ofliberal versus restrictive guidelines for red blood cell transfusion inpreterm infants. Pediatrics. 2005;115:1685–169129. Miyashiro AM, Santos N, Guinsburg R, et al. Strict red bloodcell transfusion guideline reduces the need for transfusions in very-low-birthweight infants in the first 4 weeks of life: a multicentretrial. Vox Sang. 2005;88:107–11330. Kirpalani H, Whyte RK, Andersen C, et al. The PrematureInfants in Need of Transfusion (PINT) study: a randomized, con-trolled trial of a restrictive (low) versus liberal (high) transfusionthreshold for extremely low birth weight infants. J Pediatr. 2006;149:301–30731. Venancio JP, Santos AM, Guinsburg R, Peres Cde A, ShinzatoAR, Lora MI. Strict guideline reduces the need for RBC transfu-sions in premature infants. J Trop Pediatr. 2007;53:78–8232. Mimica AF, dos Santos AM, da Cunha DH, et al. A very strictguideline reduces the number of erythrocyte transfusions in pre-term infants. Vox Sang. 2008;95:106–11133. Bell EF. When to transfuse preterm babies. Arch Dis ChildFetal Neonatal Ed. 2008;93:F469–F47334. Slonim AD, Joseph JG, Turenne WM, Sharangpani A, LubanNL. Blood transfusions in children: a multi-institutional analysis ofpractices and complications. Transfusion. 2008;48:73–8035. dos Santos AM, Guinsburg R, Procianoy RS, et al. Variabilityin red blood cell transfusion practices among Brazilian neonatalintensive care units. Transfusion. 2010;50:150–15936. Ohls RK. Transfusions in the preterm infants. NeoReviews.2007;8:e377–e38637. Strauss RG, Burmeister LF, Johnson K, Cress G, Cordle D.Feasibility and safety of AS-3 red blood cells for neonatal transfu-sions. J Pediatr. 2000;136:215–21938. Fernandes da Cunha DH, Nunes Dos Santos AM, Kopelman

American Board of Pediatrics Neonatal-PerinatalMedicine Content Specification• Know the clinical indications of potential

complications, and their prevention, ofusing blood products in neonataltransfusions, including transfusionreactions and graft-versus-host disease.




BI, et al. Transfusions of CPDA-1 red blood cells stored for up to28 days decrease donor exposures in very low-birth-weight prema-ture infants. Transfus Med. 2005;15:467–47339. Fergusson D, Hutton B, Hogan DL, et al. The Age of RedBlood Cells in Premature Infants (ARIPI) randomized controlledtrial: study design. Transfus Med Rev. 2009;23:55–6140. Mercer JS. Current best evidence: a review of the literature onumbilical cord clamping. J Midwif Womens Health. 2001;46:402–41441. Strauss RG, Mock DM, Johnson KJ, et al. A randomizedclinical trial comparing immediate versus delayed clamping of theumbilical cord in preterm infants: short-term clinical and laboratoryendpoints. Transfusion. 2008;48:658–66542. Rabe H, Reynolds G, Diaz-Rossello J. A systematic review andmeta-analysis of a brief delay in clamping the umbilical cord ofpreterm infants. Neonatology. 2008;93:138–14443. Hutton EK, Hassan ES. Late vs early clamping of the umbilicalcord in full-term neonates: systematic review and meta-analysis ofcontrolled trials. JAMA. 2007;297:1241–1252

44. Mercer JS, Vohr BR, McGrath MM, Padbury JF, Wallach M,Oh W. Delayed cord clamping in very preterm infants reduces theincidence of intraventricular hemorrhage and late-onset sepsis: arandomized, controlled trial. Pediatrics. 2006;117:1235–124245. Mercer JS, Vohr BR, Erickson-Owens DA, Padbury JF, Oh W.Seven-month developmental outcomes of very low birth weightinfants enrolled in a randomized controlled trial of delayed versusimmediate cord clamping. J Perinatol. 2010;30:11–1646. Nicholl RM, Richards S, Ray S, Gowda R. Question 1. Isdelayed clamping of the umbilical cord in moderately pretermbabies beneficial? Arch Dis Child. 2010;95:235–23747. Khodabux CM, Brand A. The use of cord blood for transfusionpurposes: current status. Vox Sang. 2009;97:281–29348. Ademokun JA, Chapman C, Dunn J, et al. Umbilical cordblood collection and separation for haematopoietic progenitor cellbanking. Bone Marrow Transplant. 1997;19:1023–102849. Strauss RG, Widness JA. Is there a role for autologous/placental red blood cell transfusions in the anemia of prematurity?Transfus Med Rev. 2010;24:125–129

NeoReviews Quiz

3. Most red blood cell products are modified soon after collection by passing them through special filters thatadsorb white blood cells. Of the following, the primary objective of leukocyte reduction for transfusion inpreterm infants is to decrease the incidence of:

A. Bronchopulmonary dysplasia.B. Cytomegalovirus infection.C. Necrotizing enterocolitis.D. Patent ductus arteriosus.E. Retinopathy of prematurity.

4. Gamma irradiation of blood products inactivates donor lymphocytes, which reduces the risk of transfusion-associated graft-versus-host disease. Of the following, gamma irradiation of blood products is most likelyto result in neonatal:

A. Hypercalcemia.B. Hyperchloremia.C. Hyperglycemia.D. Hyperkalemia.E. Hypernatremia.

5. Delayed clamping of the umbilical cord (>1 minute after birth), as compared with early clamping of theumbilical cord (<15 seconds after birth), increases neonatal blood volume by 20 to 40 mL/kg bodyweight.Of the following, based on a systematic review of controlled studies of delayed versus early clamping of theumbilical cord, the major benefit of delayed clamping in preterm infants is reduced need for:

A. Antibiotic treatment.B. Blood transfusions.C. Intensive phototherapy.D. Mechanical ventilation.E. Vasopressor treatment.





Amélia Miyashiro Nunes dos Santos and Cleide Enoir Petean Trindade Red Blood Cell Transfusions in the Neonate












DOI: 10.1542/neo.12-1-e1 2011;12;e1 NeoReviews

Alistair G.S. Philip Something Old, Something New







CommentarySomething Old, Something New

Author Disclosure

Dr Philip has disclosed no financial

relationships relevant to this article.

This commentary does not contain a



product/device.

In 2010, we commemorated our 10thanniversary of NeoReviews by collect-ing previously published Historical Per-spectives into a book titled “Milestonesin Neonatal-Perinatal Medicine.”

The past decade was one of sub-stantial growth for NeoReviews. At thistime, I would like to acknowledge theimportant contribution made by one ofthe founding editors, Dr Bill Hay. We allowe him a debt of gratitude. He was re-sponsible for setting the tone of thejournal, maintaining that it should havea strong academic flavor and not be-come a chat room. Until the presenttime, he has continued as coeditor, buthe now relinquishes this position. Join-ing the editorial staff as Associate Ed-itor is Dr Joe Neu, who is currently onthe Editorial Board of NeoReviews. Heis well known to many of our readersand has been a frequent contributorof articles for the journal. Among manyother activities, he was recently theChair of the Organization for Neonatol-ogy Training Program Directors. I feel

sure that he will be a very positiveinfluence on the journal.

As we move into a new year, it isperhaps worth reflecting on the currentstatus of NeoReviews. During the pastyear, in addition to the main articles,we have consistently provided learningopportunities in the form of Index ofSuspicion in the Nursery (IOSITN), Stripof the Month, and Visual Diagnosis. Wealso completed an initial series of arti-cles on neonatal informatics and planto publish more on this topic in thecoming months. In almost every month,we have published a perspective piece(Historical, Educational, or Interna-tional). We have also published a num-ber of pieces under the heading “CoreConcepts” and another of these (“TheBiology of Hemoglobin”) is published inthis issue.

Among the 20 major areas listedfor the American Board of PediatricsNeonatal-Perinatal Medicine (NPM) Con-tent Specifications is “Core Knowledgein Scholarly Activities.” Because thisarea accounts for a sizeable percentageof the questions (about 7%) on theNPM sub-board examination, we willpay specific attention to it in the com-ing months. Previous issues have cov-ered some of the topics included under“Scholarly Activities,” but future con-tributions will have a more definitivefocus. This issue contains the first ofthese contributions, addressing “Meta-analysis in Neonatal Perinatal Medicine.”

With this first issue of 2011, weintroduce a new section that we believewill be of interest to all who practiceclinical neonatology. It is, unfortu-nately, a fact of life that most of us willbe accused of malpractice at somepoint in our careers. Although manysuch accusations are unfounded, thereare other situations in which errors ofjudgment may be committed for a va-riety of reasons. One of the best ways toavoid being sued is to communicateeffectively with parents, but it is alsoworth being reminded of situations thatresulted in a bad outcome that couldnot be defended, so we can try to avoiddoing the same thing ourselves.

We have titled this new section“Legal Briefs” because the messages arederived from actual legal cases andbecause the information concerningthese cases will be presented in a brief(or concise) manner. In contrast toIOSITN, where the focus is on the pres-ence of findings that should lead to adiagnosis, Legal Briefs will emphasizethe absence of appropriate investiga-tions or interpretation – at least fromthe lawyer’s perspective! Being re-minded of our collective deficienciesshould help to avoid repetition of suchmistakes in the future and improvepatient care and outcomes.

Alistair G.S. Philip, MDEditor-in-Chief, NeoReviews

commentary



DOI: 10.1542/neo.12-1-e1 2011;12;e1 NeoReviews

Alistair G.S. Philip Something Old, Something New













Kamran Shahid, Kamakshya P. Patra, Amber Dawson and Eric Thomas Index of Suspicion in the Nursery • Case: An Increased Number of Wet Diapers







The reader is encouraged to writepossible diagnoses for each casebefore turning to the discussion.We invite readers to contributecase presentations and discussions.Please inquire first by contactingDr. Philip at [email protected].

Author Disclosure

Drs Shahid, Patra, Dawson, and

Thomas have disclosed no financial

relationships relevant to this case.

This commentary does not contain a



product/device.

Case: An Increased Number of Wet DiapersCase PresentationA 10-day-old term neonate has 8 to10 diapers per day. She sucks vigor-ously during feeding but vomits im-mediately. She was born to a 17-year-old primigravida woman by cesareansection and had no perinatal compli-cations. There are no findings ofnote on the family history. The in-fant exhibits temperature instability,poor feeding, and abnormal facies.On physical examination, the babyhas a blank stare and her temperatureis 36.7°C, heart rate is 178 beats/min,respiratory rate is 60 beats/min, bloodpressure is 60/36 mm Hg, and oxy-gen saturation is 97% on room air.Capillary refill is 4 seconds. She has asunken anterior fontanelle and drymucous membranes. Chest examina-tion reveals no abnormalities, cardiacevaluation documents normal heart

sounds and no murmur, and abdomi-nal examination findings are unre-markable. In addition to the blankstare, the infant fails to close her eyesappropriately.

Ophthalmologic examination re-veals severe bilateral optic nerve hypo-plasia. The metabolic profile revealsa serum sodium of 176 mEq/L(176 mmol/L), serum potassium of4.2 mEq/L (4.2 mmol/L), blood ureanitrogen of 64 mg/dL (22.8 mmol/L), creatinine of 1.5 mg/dL (132.6�mol/L), glucose of 104 mg/dL(5.8 mmol/L), and calcium of10 mg/dL (2.5 mmol/L). Completeblood count and C-reactive proteinvalues are normal. Blood and urinecultures yield no growth. Ultrasonog-raphy of kidneys shows normal find-ings. Further investigations reveal thediagnosis.

index of suspicion in the nursery



Case DiscussionBased on the polyuria and increasedserum osmolarity, the resident physi-cian makes a tentative diagnosis ofdiabetes insipidus (DI). Further labo-ratory investigations reveal a serum os-molality of 317 mOsm/kg and urineosmolality of 813 mOsm/kg. Thyroidfunction test results and serum cortisolvalues are normal. Because of bilateraloptic nerve hypoplasia and mid-facialanomalies, central DI is suspected andmagnetic resonance imaging of thebrain is obtained (Figure), whichshows absence of the septum pelluci-dum and hypoplasia of the optic appa-ratus.

An intravenous (IV) normal salinebolus of 10 mL/kg is administered,followed by maintenance IV fluids. AFoley catheter is placed to monitor theurine output, and IV fluids are ad-justed according to electrolyte results.

After receiving a dose of desmo-pressin (synthetic form of vasopres-sin) subcutaneously, the infant’surine output decreases to 2.5 mL/kgper hour, serum osmolality decreases,and urine osmolality and urine so-dium increase. Repeat serum sodiummeasures 146 mEq/L (146 mmol/L). Coagulation studies yield nor-

mal results. Intranasal desmopressin(DDAVP) therapy is initiated, as isbreastfeeding.

Differential DiagnosisThe differential diagnosis of polyuriaincludes DI, diabetes mellitus, chronicrenal failure, urinary tract infection,and solute (mannitol, glucose, saline)diuresis. Hyperglycemia, ketonuria,and glucosuria differentiate diabetesmellitus from DI. In chronic renal fail-ure, in contrast to DI, azotemia doesnot reverse with hydration.

The Condition/PathophysiologyPolydipsia, polyuria, hypernatremiaand dehydration are the hallmarksof DI. DI results from deficiency ofantidiuretic hormone (ADH) or va-sopressin and can be either central ornephrogenic. The hypothalamic os-moreceptors, the supraoptic or para-ventricular nuclei, and superior por-tion of the supraopticohypophysealtract are the sites for ADH secretion.Any condition affecting these sites(infection, intracranial hemorrhage,neurosurgery, primary or secondarytumors or infiltrative diseases, andidiopathic DI) results in central DIdue to decreased ADH secretion.

Nephrogenic DI results from re-nal tubular resistance to vasopressin.It arises from a vasopressin receptorAQP2 water channel defect, whichis transmitted by X-linked or auto-somal recessive inheritance. Drugs(amphotericin B, lithium), electro-lyte abnormalities, sickle cell diseaseand trait, Fanconi syndrome, and re-nal tubular acidosis are some of thecauses of nephrogenic DI.

DiagnosisDI should be considered in any de-hydrated neonate who has polyuria,increased serum osmolality, hyper-natremia, and a urinary concentrationdefect. Affected infants often presentwith dehydration, fever, poor feeding,failure to thrive, irritability, and sei-zures. They are “hungry feeders” butvomit immediately.

Septo-optic dysplasia (de Morsiersyndrome) is a rare congenital malfor-mation involving partial or completeabsence of the septum pellucidum andoptic nerve hypoplasia. The malfor-mation may manifest with seizures,blindness, nystagmus, hypotonia, andoccasionally panhypopituitarism. Otherneuroradiologic findings may includeschizencephaly and holoprosencephaly.

TreatmentIntranasal DDAVP is the treatmentof choice for central DI. Aqueousvasopressin or DDAVP also can beadministered IV or subcutaneously.Chlorpropamide, clofibrate, and thi-azide diuretics may be used in con-junction with DDAVP but are rarelyadministered in neonates. Breast-feeding is preferred because of lowsolute content. For acute episodes,oral or parenteral hydration is neededto reverse the hyperosmolality anddehydration. Cerebral sinus throm-bosis is a complication of hypernatre-mic dehydration in DI, and aggressiverehydration and neurologic monitor-ing are necessary in acute episodes.Thiazide diuretic, amiloride, and indo-

Figure. Brain MRI showing absence of the septum pellucidum (blue arrow) andhypoplasia of the optic apparatus (yellow arrow).




methacin have been used in nephro-genic DI. Genetic counseling is impor-tant because many of the cases arehereditary. The long-term prognosisof DI depends on the cause.

Lessons for the Clinician● DI should be considered in any

dehydrated neonate who exhibitspolyuria, hypernatremia, and uri-nary concentration defect.

● Central DI results from deficiency ofvasopressin, and intranasal DDAVP isthe treatment of choice.

● Septo-optic dysplasia is character-ized by hypoplasia of the optic nerveand agenesis of the corpus callosum.

(Kamran Shahid, MD, Kamak-shya P. Patra, MD, Amber Dawson,MD, Eric Thomas, MD, LouisianaState University, Shreveport, LA)

Suggested Reading

Alon U, Chan JCM. Hydrochlorothiazide-amiloride in the treatment of nephro-genic diabetes insipidus. Am J Nephrol.1985;5:9–13

DeVeber G, Andrew M, Adams C, et al.Cerebral sino-venous thrombosis in chil-dren. N Engl J Med. 2001;345:417–423

Leung AKC, Robson WLM, Halperin ML.Polyuria in childhood. Clin Pediatr.1991;11:634–640

Sener RN. Septo-optic dysplasia associatedwith cerebral cortical dysplasia. J Neuro-radiol. 1996;23:245–247

Wang LC, Cohen ME, Duffner PK. Eti-ologies of central diabetes insipidus inchildren: Pediatr Neurol 1994;11:273–277

American Board of PediatricsNeonatal-Perinatal MedicineContent Specifications• Know the specific

hormonal factorsthat influence waterbalance in newborninfants.

• Know the effects of arginine vasopressin(antidiuretic hormone) on sodium andwater balance.





Kamran Shahid, Kamakshya P. Patra, Amber Dawson and Eric Thomas Index of Suspicion in the Nursery • Case: An Increased Number of Wet Diapers













Charlotte Clock and Leonardo Pereira Strip of the Month: January 2011







Strip of the Month: January 2011Charlotte Clock, MD,*

Leonardo Pereira, MD,

MCR*

Author Disclosure

Dr Clock has disclosed

no financial


to this article. Dr

Pereira has disclosed

that he is the primary

investigator for a

research trial for

ProteoGenix, Inc,

Costa Mesa, Calif.

This commentary does

not contain a

discussion of an

unapproved/


commercial

product/device.

Electronic Fetal Monitoring Case Review SeriesElectronic fetal monitoring (EFM) is a popular technology used to establish fetal well-being. Despite its widespread use, terminology used to describe patterns seen on themonitor has not been consistent until recently. In 1997, the National Institute of ChildHealth and Human Development (NICHD) Research Planning Workshop publishedguidelines for interpretation of fetal tracings. This publication was the culmination of2 years of work by a panel of experts in the field of fetal monitoring and was endorsed in2005 by both the American College of Obstetricians and Gynecologists (ACOG) and theAssociation of Women’s Health, Obstetric and Neonatal Nurses (AWHONN). In 2008,ACOG, NICHD, and the Society for Maternal-Fetal Medicine reviewed and updated thedefinitions for fetal heart rate patterns, interpretation, and research recommendations.Following is a summary of the terminology definitions and assumptions found in the 2008NICHD workshop report. Normal values for arterial umbilical cord gas values andindications of acidosis are defined in Table 1.

Assumptions from the NICHD Workshop

● Definitions are developed for visual interpretation, assuming that both the fetal heart rate(FHR) and uterine activity recordings are of adequate quality

● Definitions apply to tracings generated by internal or external monitoring devices● Periodic patterns are differentiated based on waveform, abrupt or gradual (eg, late

decelerations have a gradual onset and variable decelerations have an abrupt onset)● Long- and short-term variability are evaluated visually as a unit● Gestational age of the fetus is considered when evaluating patterns● Components of fetal heart rate FHR do not occur alone and generally evolve over time

DefinitionsBaseline Fetal Heart Rate

● Approximate mean FHR rounded to increments of 5 beats/min in a 10-minute segmentof tracing, excluding accelerations and decelerations, periods of marked variability, andsegments of baseline that differ by �25 beats/min

● In the 10-minute segment, the minimum baseline duration must be at least 2 minutes(not necessarily contiguous) or the baseline for that segment is indeterminate

● Bradycardia is a baseline of �110 beats/min; tachycardia is a baseline of �160 beats/min

● Sinusoidal baseline has a smooth sine wave-like undulating pattern, with waves havingregular frequency and amplitude

Baseline Variability

● Fluctuations in the baseline FHR of two cycles per minute or greater, fluctuations areirregular in amplitude and frequency, fluctuations are visually quantitated as the ampli-tude of the peak to trough in beats per minute

● Classification of variability:Absent: Amplitude range is undetectableMinimal: Amplitude range is greater than undetectable to 5 beats/minModerate: Amplitude range is 6 to 25 beats/minMarked: Amplitude range is �25 beats/min

*Assistant Professor, Division of Maternal-Fetal Medicine, Oregon Health & Sciences University, Portland, Ore.

strip of the month



Accelerations

● Abrupt increase in FHR above the most recently deter-mined baseline

● Onset to peak of acceleration is �30 seconds, acme is�15 beats/min above the most recently determinedbaseline and lasts �15 seconds but �2 minutes

● Before 32 weeks’ gestation, accelerations are definedby an acme �10 beats/min above the most recentlydetermined baseline for �10 seconds

● Prolonged acceleration lasts �2 minutes but �10 min-utes

Late Decelerations

● Gradual decrease in FHR (onset to nadir �30 seconds)below the most recently determined baseline, withnadir occurring after the peak of uterine contractions

● Considered a periodic pattern because it occurs withuterine contractions

Early Decelerations

● Gradual decrease in FHR (onset to nadir �30 seconds)below the most recently determined baseline, withnadir occurring coincident with uterine contraction

● Also considered a periodic pattern

Variable Decelerations

● Abrupt decrease in FHR (onset to nadir �30 seconds)● Decrease is �15 beats/min below the most recently

determined baseline lasting �15 seconds but �2 min-utes

● May be episodic (occurs without a contraction) orperiodic

Prolonged Decelerations

● Decrease in the FHR �15 beats/min below the mostrecently determined baseline lasting �2 minutes but�10 minutes from onset to return to baseline

Decelerations are tentatively called recurrent if they oc-cur with �50% of uterine contractions in a 20-minuteperiod.

Decelerations occurring with �50% of uterine contrac-tions in a 20-minute segment are intermittent.

Sinusoidal Fetal Heart Rate Pattern

● Visually apparent, smooth sine wavelike undulatingpattern in the baseline with a cycle frequency of 3 to5/minute that persists for �20 minutes.

Uterine Contractions

● Quantified as the number of contractions in a 10-minute window, averaged over 30 minutes.

Normal: �5 contractions in 10 minutesTachysystole: �5 contractions in 10 minutes

InterpretationA three-tier Fetal Heart Rate Interpretation system hasbeen recommended as follows:

● Category I FHR tracings: Normal, strongly predictiveof normal fetal acid-base status and require routinecare. These tracings include all of the following:

�Baseline rate: 110 to 160 beats/min�Baseline FHR variability: Moderate�Late or variable decelerations: Absent�Early decelerations: Present or absent�Accelerations: Present or absent

● Category II FHR tracings: Indeterminate, require evalu-ation and continued surveillance and reevaluation. Exam-ples of these tracings include any of the following:

�Bradycardia not accompanied by absent variability�Tachycardia�Minimal or marked baseline variability�Absent variability without recurrent decelerations�Absence of induced accelerations after fetal stimula-

tion

Table 1. Arterial Umbilical Cord Gas ValuespH PCO2 (mm Hg) PO2 (mm Hg) Base Excess

Normal* >7.20 <60 >20 <�10(7.15 to 7.38) (35 to 70) (�2.0 to �9.0)

Respiratory Acidosis <7.20 >60 Variable <�10Metabolic Acidosis <7.20 <60 Variable >�10Mixed Acidosis <7.20 >60 Variable >�10

* Normal ranges from Obstet Gynecol Clin North Am. 1999;26:695

strip of the month



�Recurrent variable decelerations with minimalor moderate variability

�Prolonged decelerations�Recurrent late decelerations with moderate vari-

ability�Variable decelerations with other characteristics,

such as slow return to baseline● Category III FHR tracings: Abnormal, predictive of

abnormal fetal acid-base status and require promptintervention. These tracings include:

�Absent variability with any of the following:y Recurrent late decelerations

y Recurrent variable decelerationsy Bradycardia

�Sinusoidal pattern

Data from Macones GA, Hankins GDV, Spong CY, HauthJ, Moore T. The 2008 National Institute of Child Healthand Human Development workshop report on electronicfetal monitoring. Obstet Gynecol. 2008;112:661–666.

We encourage readers to examine each strip in thecase presentation and make a personal interpretation ofthe findings before advancing to the expert interpreta-tion provided.

strip of the month



Case PresentationHistory

A 21-year-old primigravida at 32–6/7 weeks’ gestationpresents to labor and delivery with preterm labor.Cervical examination reveals 3 cm dilation. Her preg-nancy has been complicated by maternal repaired trans-position of the great vessels with ventricular inver-sion and first-trimester bleeding. The fetus had normal

echocardiographic findings at 20 weeks’ gestation.The patient has no evidence of infection, so she re-ceives nifedipine tocolysis, betamethasone for fetalmaturity, and penicillin for group B Streptococcusprophylaxis. The delivery plan is made for an assistedsecond stage of labor due to maternal cardiac disease.A fetal heart tracing is obtained upon admission(Fig. 1).

Figure 1. EFM Strip #1.

strip of the month



Findings on EFT Strip #1 are:

● Variability: Moderate● Baseline Rate: 140 beats/min● Episodic Pattern: Accelerations● Periodic Pattern: None noted● Uterine Contractions: Occurring every 2 minutes, last-

ing 45 seconds● Interpretation: Category I tracing● Differential Diagnosis: Normal tracing, with evidence

of fetal well-being indicated by moderate variabilityand accelerations

● Action: No intervention required for fetal status. Closemonitoring will continue for the preterm labor andblood pressure changes that can occur with the use ofnifedipine. Nifedipine is associated with a decrease in

maternal blood pressure that can be associated withuteroplacental insufficiency. (1) In our institution, thedose is held if the maternal blood pressure is less than90/60 mm Hg to avoid this complication.

Progression of LaborTocolysis is discontinued after 48 hours. The womanremains stable for another 24 hours, when she beginshaving increased contractions and variable decelerations.On physical examination, she is 6 cm dilated, completelyeffaced, and �2 station. She is afebrile, and other exam-ination findings are within normal parameters. A fetalheart tracing is obtained as her contractions begin toincrease (Fig. 2).



strip of the month



Findings on EFM Strip #2 are:

● Variability: Moderate● Baseline Rate: 140 beats/min● Episodic Pattern: None● Periodic Pattern: Recurrent variable decelerations● Uterine Contractions: Occurring every 3 minutes,

lasting 45 seconds● Interpretation: Category II tracing● Differential Diagnosis: Variable decelerations are usually

associated with intermittent cord occlusion. They arenormally well tolerated, but if they continue to be repet-itive or more severe, acidemia can develop and can occurmore quickly in preterm or growth-restricted infants.

● Action: A plan is made to proceed with augmentationof labor due to her advanced cervical dilation. TheFHR tracing requires close surveillance and re-evaluation.

The decision is made to proceed with augmentation oflabor because she has completed her corticosteroidcourse, has advanced cervical dilation, and is havingvariable decelerations. Labor progresses rapidly after rup-ture of membranes and oxytocin augmentation, and hercervix is 9 cm dilated. An intrauterine pressure catheter(IUPC) is placed because of recurrent variable decelera-tions, but blood returns through the catheter, and it isremoved. The fetal tracing is shown in Figure 3.



strip of the month




● Variability: Minimal to moderate● Baseline Rate: 130 beats/min● Episodic Pattern: None● Periodic Pattern: Recurrent variable decelerations● Uterine Contractions: Occurring every 3 minutes,

lasting 45 seconds● Interpretation: Category II● Differential Diagnosis: Unchanged● Action: Overall, the tracing is reassuring due to mod-

erate variability, but the variables are continuing,

which prompts consideration of IUPC placement andamnioinfusion. Amnioinfusion has been associatedwith a decrease in variable decelerations, cesarean sec-tion for fetal distress, and neonatal acidemia. (2) Thereturn of blood in the catheter in this case most likely isdue to placement behind the placenta. It is importantto recognize this complication, not start an amnioinfu-sion, and remove the catheter.

One hour later, the patient has a prolonged deceleration,and the decision is made to move to the operating roomfor further evaluation and possible operative delivery.The fetal tracing is shown in Figure 4.



strip of the month




● Variability: Moderate● Baseline Rate: 120 beats/min● Episodic Pattern: Prolonged bradycardia● Periodic Pattern: Recurrent variable decelerations● Uterine Contractions: Occurring every 2 minutes● Interpretation: Category II, based on recurrent vari-

able decelerations with periods of moderate variability,followed by a prolonged bradycardia

● Differential Diagnosis: Prolonged cord occlusion,uteroplacental underperfusion, or uteroplacental dys-function resulting in fetal hypoxia

● Action: Cervical examination shows no evidence ofcord prolapse. Resuscitative measures should be un-dertaken, including administration of intravenousfluids and oxygen, position changes, and possiblydiscontinuation of oxytocin. The patient is transferredto the operative room for further evaluation and pos-sible cesarean delivery if the bradycardia does not re-solve.

In the operating room, the bradycardia resolves, butvariable decelerations continue. Thirty minutes later, thepatient is examined, and her cervix is still 9 cm dilated.A final fetal heart tracing is obtained (Fig. 5).



strip of the month




● Variability: Moderate● Baseline Rate: 130 beats/min● Episodic Pattern: None● Periodic Pattern: Recurrent variable decelerations with

slow return to baseline● Uterine Contractions: Occurring every 2 minutes● Interpretation: Category II● Differential Diagnosis: The slow return to baseline is

concerning for the development of fetal acidosis● Action: The decision is made to proceed with cesarean

delivery due to the repetitive variable decelerationswith late return because vaginal delivery is not immi-nent, with no cervical change over the past 90 minutes.

Fifteen minutes later, a primary cesarean delivery is per-formed. There is a single nuchal cord.

OutcomeA viable male infant is delivered by cesarean section. Heweighs 4 lbs 4 oz and has Apgar scores of 5 at 1 minuteand 7 at 5 minutes. The arterial blood gas findings areconsistent with respiratory acidosis (Table 2). The babyreceives continuous positive airway pressure. He is then

taken to the neonatal intensive care unit, where he has anormal course for a preterm infant.

This infant’s FHR tracing could be explained by thepresence of a nuchal cord. Single nuchal cords are seen in upto 30% of normal pregnancies. They are rarely associatedwith significant neonatal morbidity or mortality. Results aremixed on their association with abnormal FHR patterns,low birthweight, and umbilical artery pH of 7.10 or less.The poor sensitivity of ultrasonography in diagnosing nu-chal cord and poor predictive value of adverse outcomes donot justify routine screening for its presence. (3)(4)

References1. Nassar AH, Aoun J, Usta IM. Calcium channel blockers for themanagement of preterm birth: a review. Am J Perinatol. 2010 Jul16. Epub ahead of print2. Regi A, Alexander N, Jose R, Lionel J, Varghese L, Peedicavil A.Amnioinfusion for relief of recurrent severe and moderate variabledecelerations in labor. J Reprod Med. 2009;54:295–3023. Larson JD, Rayburn WF, Crosby S, Thurnau GR. Multiplenuchal cord entanglements and intrapartum complications. Am JObstet Gynecol. 1995;173:1228–12314. Peregrine E, O’Brien P, Jauniaux E. Ultrasound detection ofnuchal cord prior to labor induction and the risk of cesarean section.Ultrasound Obstet Gynecol. 2005; 25:160–164

Table 2. Arterial Umbilical Cord Gas ValuespH PCO2 (mm Hg) PO2 (mm Hg) Base Excess

Normal* >7.20 <60 >20 <�10(7.15 to 7.38) (35 to 70) (�2.0 to �9.0)

Respiratory Acidosis <7.20 >60 Variable <�10Metabolic Acidosis <7.20 <60 Variable >�10Mixed Acidosis <7.20 >60 Variable >�10Patient 7.10 81 17 �8.7

* Normal ranges from Obstet Gynecol Clin North Am. 1999;26:695


strip of the month




Charlotte Clock and Leonardo Pereira Strip of the Month: January 2011












A newborn male born at 35 weeks’ gestation displays respiratory distress, flaccid abdominal musculature, and undescended testes. fig. 1

Prenatal History: 29-‐year-‐old G3 P1011à2 woman who has a healthy 8-‐year-‐old son Rubella-‐nonimmune, group B Streptococcus-‐positive Social history negative for medications, alcohol, tobacco Family history negative for renal or genitourinary anomalies Antenatal diagnosis of bladder outlet obstruction with posterior urethral valves, left cystic kidney, possible right talipes equinovarus, and oligohydramnios Amniocentesis revealed normal 46XY karyotype Placement of a fetal vesicoamniotic shunt at 23 weeks’ gestation Repeat ultrasonography performed at 31 weeks’ gestation revealed a decompressed bladder, left kidney with mild dilation and multiple small cysts, irregular abdominal musculature, and normal amniotic fluid volume Repeat ultrasonography at 33 weeks’ gestation showed decreased amniotic fluid volume, prompting hospitalization of the mother for intravenous fluid hydration and corticosteroid administration.

Birth History and Presentation: Due to concerns for recurrent, self-‐resolved fetal heart rate decelerations, the infant was delivered via cesarean section at 35 weeks’ gestation Apgar scores were 6 and 8 at 1 and 5 minutes, respectively Infant developed increased work of breathing in the delivery room and was intubated Birthweight: 2,590 g (50th percentile) Length: 47.5 cm (50th to 90th percentile) Occipitofrontal circumference: 27 cm (<3rd percentile) Vital Signs: Temperature: 36.8°C Heart rate: 152 beats/min Respiratory rate: 40 breaths/min Blood pressure: 68/37 mm Hg Physical Examination: Intubated, pink, and well perfused Relatively small chest, with subcostal retractions, fair aeration bilaterally Regular heart rate and rhythm without murmur, normal pulses Bulging abdomen, poor abdominal musculature that is flaccid to palpation and nontender, palpable loops of bowel Palpable left flank mass Vesicoamniotic tube to the right of umbilicus draining urine Normal penis with bilateral undescended testes Right talipes equinovarus Radiologic Findings: Bilateral cystic dysplastic kidneys with echogenic cortex and multiple small cortical cysts, bilateral dilation of the renal pelvis and ureters, decompressed urinary bladder

Differential Diagnosis: Respiratory Distress, Flaccid Abdominal Musculature, and Cryptorchidism: Megacystis megaureter Megacystis-‐microcolon-‐intestinal hypoperistalsis syndrome Neurogenic bladder Posterior urethral valves Prune belly syndrome Severe primary vesicoureteral reflux Ureteropelvic junction obstruction Urethral obstruction Actual Diagnosis: Prune belly syndrome

Prune belly syndrome (PBS), also known as Eagle-‐Barrett Syndrome, is a rare condition characterized by the triad of absent or incomplete abdominal musculature, undescended testes, and urinary tract abnormalities. The incidence is estimated at approximately 3.8 cases per 100,000 live births, with a marked male predominance. (1) Recent epidemiologic data obtained from the Kids Inpatient Database found 50% of PBS patients were white, 31% black, and 10% Hispanic. (1) Affected patients are often born preterm (43% of infants), and prematurity is associated with higher inpatient mortality. (1) Although the exact inheritance pattern is unknown, PBS is most likely transmitted in a sex-‐influenced, autosomal recessive pattern,,with some familial cases reported. (2) Of note, affected females are seen more commonly in the familial (28%) than the nonfamilial (5%) forms of PBS. (2) PBS has been observed in association with trisomy 13, 18, and 21. (3-‐5) In addition, one affected infant had a large deletion of the long arm of chromosome 6. (6) The two primary theories regarding the pathogenesis of PBS are: aberrant mesodermal development (7,8) and early urethral obstruction (9,10). The mesodermal defect theory proposes that the constellation of anomalies is due to a defect in the intermediate and lateral plate mesoderm, which affects the development of the mesonephric and paramesonephric ducts as well as the abdominal musculature and urinary tract (8). The urethral obstruction theory proposes that distal obstruction in early gestation results in distension of the bladder and ureters, urinary ascites, and degeneration of the abdominal muscles. (11,12) Patients who have PBS can present with a variety of renal and urologic complications, including cystic renal disease, renal dysplasia, renal insufficiency,

hypospadias, micropenis, urachal anomalies, and ureterocele. (1) Anomalies of the urinary tract can lead to impaired voiding, vesicoureteral reflux, repeat urinary tract infections, pyelonephritis, and renal scarring during childhood. (13) Among the extrarenal manifestations are pulmonary hypoplasia due to oligohydramnios and skeletal deformities (45%) such as talipes equinovarus, hip dysplasia, kyphoscoliosis, torticollis, and pectus excavatum. (13) In addition to pulmonary hypoplasia, many patients who have PBS develop chronic respiratory dysfunction due to associated skeletal anomalies and abdominal weakness; they are prone to recurrent respiratory infections and are at increased risk of respiratory complications after exposure to general anesthesia. (14) Cardiac abnormalities are seen in 10% of patients and include ventricular septal defects, tetralogy of Fallot, patent ductus arteriosus, and atrial septal defects. (13) Gastrointestinal complications are seen in up to 30% of patients and include malrotation, intestinal atresias (often colonic) or stenoses, volvulus, and chronic constipation. (12,15,16) Although central nervous system anomalies are rare (5%), (1) many patients exhibit developmental delay and growth retardation as children. (17) PBS is often diagnosed by prenatal ultrasonography as early as 13 weeks’ gestation or upon physical examination at birth. (18,19) The severity of PBS is classified as grade I: oligohydramnios, lung hypoplasia, and Potter facies; grade 2: moderate-‐to-‐severe involvement of the fetal urinary system without lung hypoplasia and without Potter facies; and grade 3: mild renal impairment. (19) The prognosis varies primarily according to the severity of renal impairment. Approximately 30% of patients who survive the neonatal period develop chronic renal insufficiency necessitating dialysis or transplantation during childhood or adolescence. (13,20,21) The overall mortality is estimated at 36% to 60% (1,22), with most deaths occurring during the initial hospitalization or neonatal period. Promising treatments include in utero placement of a vesicoamniotic shunt, which has been successful in restoring amniotic fluid volume and reducing neonatal mortality, although long-‐term outcomes in terms of chronic renal insufficiency are not yet available. (23) Postnatal treatment includes orchiopexy, abdominoplasty, renal transplant, and relief of bladder outlet obstruction (vesicostomy early and later Mitrofanoff). (24) In conclusion, PBS is a rare disease primarily affecting males that is characterized by a triad of congenital anomalies (deficient abdominal musculature, cryptorchidism, and urinary tract abnormalities) and related complications. Early diagnosis and treatment can lead to improved perinatal outcomes, although long-‐term data regarding chronic renal disease are not yet available. References: 1. Routh JC, Huang L, Retik AB, Nelson CP. Contemporary epidemiology and characterization of newborn males with prune belly syndrome. Urology. 2010;76:44-‐48

2. Ramasamy R, Haviland M, Woodard JR, Barone JG. Patterns of inheritance in familial prune belly syndrome. Urology. 2005;65:1227 3. Beckmann H, Rehder H, Rauskolb R. Prune belly sequence associated with trisomy 13. Am J Med Genet. 1984;19:603-‐604 4. Nivelon-‐Chevallier A, Feldman JP, Justrabo E, Turc-‐Carel C. Trisomy 18 and prune belly syndrome [in French]. J Genet Hum. 1985;33:469-‐474 5. Baird PA, Sadovnick AD. Prune belly anomaly in Down syndrome. Am J Med Genet. 1987;26:747-‐748 6. Fryns JP, Vandenberghe K, Van den Berghe H. Prune-‐belly anomaly and large interstitial deletion of the long arm of chromosome 6. Ann Genet. 1991;34:127 7. Barnhouse DH. Prune belly syndrome. Br J Urol. 1972;44:356-‐360 8. Stephens FD, Gupta D. Pathogenesis of the prune belly syndrome. J Urol. 1994;152:2328-‐2331 9. Hoagland MH, Hutchins GM. Obstructive lesions of the lower urinary tract in the prune belly syndrome. Arch Pathol Lab Med. 1987;111:154-‐156. 10. Moerman P, Fryns JP, Goddeeris P, Lauweryns JM. Pathogenesis of the prune-‐belly syndrome: a functional urethral obstruction caused by prostatic hypoplasia. Pediatrics. 1984;73:470-‐475 11. Wheatley JM, Stephens FD, Hutson JM. Prune-‐belly syndrome: ongoing controversies regarding pathogenesis and management. Semin Pediatr Surg. 1996;5:95-‐106. 12. Herman TE, Siegel MJ. Prune belly syndrome. J Perinatol. 2009;29:69-‐71 13. Bogart MM, Arnold HE, Greer KE. Prune-‐belly syndrome in two children and review of the literature. Pediatr Dermatol. 2006;23:342-‐345 14. Crompton CH, MacLusky IB, Geary DF. Respiratory function in the prune-‐belly syndrome. Arch Dis Child. 1993;68:505-‐506 15. Wright JR Jr, Barth RF, Neff JC, Poe ET, Sucheston ME, Stempel LE. Gastrointestinal malformations associated with prune belly syndrome: three cases and a review of the literature. Pediatr Pathol. 1986;5:421-‐448 16. Smythe AR 2nd. Ultrasonic detection of fetal ascites and bladder dilation with resulting prune belly. J Pediatr. 1981;98:978-‐980 17. Crankson S, Ahmed S. The prune belly syndrome. Aust N Z J Surg. 1992;62:916-‐921 18. Woods AG, Brandon DH. Prune belly syndrome. A focused physical assessment. Adv Neonatal Care. 2007;7:132-‐143 19. Papantoniou N, Papoutsis D, Daskalakis G, et al. Prenatal diagnosis of prune-‐belly syndrome at 13 weeks of gestation: case report and review of literature. J Matern Fetal Neonatal Med. 2010;23:1263-‐1267 20. Fischbach M. Ask the expert. Is peritoneal dialysis (CAPD or APD) appropriate for small children with prune belly syndrome and terminal renal failure? Pediatr Nephrol. 2001;16:936-‐937 21. Gonzalez R, Reinberg Y, Burke B, Wells T, Vernier RL. Early bladder outlet obstruction in fetal lambs induces renal dysplasia and the prune-‐belly syndrome. J Pediatr Surg. 1990;25:342-‐345 22. Druschel CM. A descriptive study of prune belly in New York State, 1983 to 1989. Arch Pediatr Adolesc Med. 1995;149:70-‐76

23. Biard JM, Johnson MP, Carr MC, et al. Long-‐term outcomes in children treated by prenatal vesicoamniotic shunting for lower urinary tract obstruction. Obstet Gynecol. 2005;106:503-‐508. 24. Strand WR. Initial management of complex pediatric disorders: prunebelly syndrome, posterior urethral valves. Urol Clin North Am. 2004;31:399-‐415

Documents

NEOREVIEWS 1.11