The preterm birth syndrome: issues to consider [607652]

The preterm birth syndrome: issues to consider
in creating a classification system
Robert L. Goldenberg, MD; Michael G. Gravett, MD; Jay Iams, MD; Aris T. Papageorghiou, MBChB, MRCOG;
Sarah A. Waller, MD; Michael Kramer, MD; Jennifer Culhane, PhD, MPH; Fernando Barros, PhD;Augustin Conde-Agudelo, MD, MPH; Zulfiqar A. Bhutta, MBBS, FRCP, FRCPCH, FCPS, PhD;Hannah E. Knight, MSc; Jose Villar, MD, MSc, MPH, FRCOG
In the first article of this series, the po-
tential benefits of a classification sys-
tem for preterm birth were articulatedand a brief history of attempts to classifypreterm birth was presented.
1In this ar-
ticle, our goal is to raise many of the is-sues that need to be addressed and thedecisions that need to be made to create apreterm birth classification system. As in
the other articles in this series, the au-thors were brought together as a directresult of the Global Alliance to PreventPrematurity and Stillbirth (GAPPS)meeting with instructions to determinethe need for such a classification system,to define the issues related to creating apreterm birth classification system, and
to present a prototype classification sys-tem for general consideration.In addition to discussing the issues
that need to be resolved before a clas-
sification system can be created, wealso intend to cause readers to considerconceptual issues that may have hin-dered progress toward better under-standing preterm birth. These includethe assumption that the clinical pre-sentation for delivery defines distinctcauses and acceptance of the arbitrarygestational age boundaries that defineprematurity. In writing this commen-tary, we began with many diverse opin-ions regarding the development of aclassification system for preterm birth.We found that, by isolating each issueand posing a specific question regard-ing the issue, we could better under-stand the principles on which to base aclassification system and, eventually,came to a consensus on each of the is-sues. We have tried to identify and em-phasize clearly superior options amongthe possible choices, although notingother potential options and the ratio-nale for our choices.
What is the reason for creating
this classification systemfor preterm birth?
There are many reasons to classify preterm
births and to consider various systems ofclassification. In this article, we focus onthe decisions involved in creating a classi-
fication system for use in both popula-tion surveillance and research, so thatwhen specific types of preterm births arediscussed, studied, or compared acrosspopulations or over time, categorieshave consistent definitions that arewidely understood and accepted.A comprehensive classification system for preterm birth requires expanded gestationalboundaries that recognize the early origins of preterm parturition and emphasize fetalmaturity over fetal age. Exclusion of stillbirths, pregnancy terminations, and multifetalgestations prevents comprehensive consideration of the potential causes and presenta-tions of preterm birth. Any step in parturition (cervical softening and ripening, decidual-membrane activation, and/or myometrial contractions) may initiate preterm parturition,and should be recorded for every preterm birth, as should the condition of the mother,fetus, newborn, and placenta, before a phenotype is assigned.
Key words: classification, phenotype, preterm birth
From the Department of Obstetrics and Gynecology (Dr Goldenberg), Drexel University,
Philadelphia, PA; Department of Obstetrics and Gynecology (Drs Gravett and Waller), Universityof Washington, Seattle, and Global Alliance to Prevent Prematurity and Stillbirth, SeattleChildren’s, Seattle, WA; Division of Maternal Fetal Medicine, Department of Obstetrics andGynecology (Dr Iams), The Ohio State University Medical Center, Columbus OH; NuffieldDepartment of Obstetrics and Gynaecology (Drs Papageorghiou, Knight, and Villar), and OxfordMaternal and Perinatal Health Institute, Green Templeton College, University of Oxford, Oxford,UK; Departments of Pediatrics and of Epidemiology, Biostatistics, and Occupational Health (DrKramer), McGill University Faculty of Medicine, Montreal, Canada; Department of Pediatrics(Dr Culhane), University of Pennsylvania, Philadelphia, PA; Post-Graduate Course in Health andBehavior (Dr Barros), Catholic University of Pelotas, RS, Brazil; Perinatology Research Branch,Eunice Kennedy Shriver National Institute of Child Health and Human Development/National
Institutes of Health/Department of Health and Human Services (Dr Conde-Agudelo), Bethesda,MD and Detroit, MI; and the Division of Women and Child Health (Dr Bhutta), The Aga KhanUniversity, Karachi, Pakistan.
Received May 29, 2011; revised Aug. 27, 2011; accepted Oct.19, 2011.
This project was supported by the Bill and Melinda Gates Foundation and the Global Alliance to
Prevent Prematurity and Stillbirth, an initiative of Seattle Children’s, and by INTERGROWTH-21stGrant ID 49038 from the Bill and Melinda Gates Foundation to the University of Oxford, for whichwe are very grateful.
The authors report no conflict of interest.Reprints not available from the authors.0002-9378/$36.00  © 2012 Mosby, Inc. All rights reserved.  doi: 10.1016/j.ajog.2011.10.865
See related editorial, page 99www.AJOG.org Clinical Opinion
FEBRUARY 2012 American Journal of Obstetrics &Gynecology 113

What should the gestational age
boundaries in a classificationof preterm births be?
The lower and upper gestational age
boundaries for defining preterm birthare variably defined. Although most geo-graphic areas base their preterm birthrates on live births (usually excludingstillbirths), the boundaries at both endsare arbitrary. For example, if a lower ges-tational age boundary for defining a pre-term birth is used at all, the cutoffs rangefrom 20
0/7to 22 or even 28 weeks. How-
ever, as demonstrated in the first paperin this series, the risk factors, causes, andrecurrence risks for spontaneous birthsat 16-19 weeks do not differ substantiallyfrom those births occurring at 20-24weeks.
2-11Thus, if the objective is to ex-
plore the full range of preterm birth,there is no reason to exclude births at16-19 weeks from the classification sys-tem. Regardless of the lower cutoff cho-sen, for comparison purposes across sitesor over time, some clearly defined, scien-tifically sound, lower gestational age cut-off that defines preterm birth should beused.
Similarly, there is now abundant evi-
dence that many infants born at 37 or 38weeks of gestation experience increasedneonatal mortality and even lifetimemorbidity related to immaturity of oneor more organs as compared with infantsborn at /H1135039 weeks.
12-15The historical
choice of 37 weeks as the upper gesta-tional age cutoff for defining a pretermbirth was arbitrary and may no longerserve a useful purpose, because it doesnot coincide with functional maturity.For this reason, we believe that defining a
preterm birth as any occurring before 39weeks would be more appropriate. Forresearch and reporting purposes, ex-tending both the lower and upperboundaries of preterm birth should beconsidered.
Regardless of the final gestational age
cutoffs for defining preterm birth, therewas universal agreement (the authorsagreed) that gestational age data shouldbe collected and recorded in narrow cat-egories (eg, no more than 1 week) to al-low flexibility in later categorization. Di-viding the preterm births into severalgestational age groups may also be use-
ful.
16-18The exact thresholds matter less
than the common use of a universal sys-tem of gestational age groupings, so thatdifferences in the gestational age distri-bution of preterm birth can be under-stood. Finally, because menstrual dating of
gestational age is often inaccurate, we be-lieve gestational age estimation should,whenever possible, be corroborated by anearly, high quality ultrasound and the bestobstetric estimate be used for all gesta-tional age determinations in the classifica-tion system.
19
What information will be
collected in this pretermbirth classification system?
Because the system we envision will be
used for research and population sur-veillance, we propose to classify the pre-term birth at some time after delivery,with as much information available as pos-sible. The clinical record should be the pri-mary source of information. This recordshould include antepartum and intrapar-tum data, a record of all prior pregnancies,medical history, a patient and physician in-terview when preterm birth has beenscheduled and where the reason for deliv-ery is not completely clear, and finally, agross and microscopic placental evalua-tion and, for stillbirths, an autopsy or pa-thology report. Without examining eachof these sources, an important potentialcause or phenotype might be missed.
Should a classification system be
based on phenotype or cause?
Because the cause of a specific case of
preterm birth is rarely known with anydegree of certainty, the authors agreedthat the optimal classification systemshould primarily be based on the clinicalphenotype, defined in this study as oneor more characteristics of the mother, fe-tus, placenta, and the presentation fordelivery. We also agree that more than 1phenotype may be present in a singlecase of preterm delivery and that eachphenotype present should be recordedso that the choice of a single category isnot forced. Finally, the actual method ofdelivery (spontaneous or instrumentalvaginal or cesarean birth) should not bepart of the phenotypic classification sys-tem. Data about method of deliveryshould be noted and collected separately.
One question raised in relation to
those preterm births considered as“spontaneous” in many prior studies iswhether the findings at initial clinicalpresentation (eg, contractions, preterm
premature rupture of membranes [P-PROM], bleeding, or advanced dilation),or the likely pathway leading to the finalpresentation (eg, short cervix or polyhy-dramnios), should be primary. In this mat-ter, we were influenced by evidence sug-gesting that a common “phenotype” ofspontaneous preterm birth is primarilycharacterized by progressive cervical ef-facement, after which P-PROM, persistentmild contractions, prolapsed membranes,or bleeding could be the acute reason forseeking care.
20Based on this consider –
ation, we believe that the most useful clas-sification system will not only capture in-formation about the clinical presentationon admission (contractions, P-PROM,bleeding, advanced cervical dilation with-out P-PROM or significant contractions,or none of these for a provider initiated de-livery), but will also be based on conditionsand observations during pregnancy, in-cluding significant maternal infection,short cervical length, increased or de-creased amniotic fluid volume, as well asrelevant clinical, laboratory, and placentalfindings.
Should risk factors be part of
the classification system?
The next issue is whether risk factors
should be part of the classification sys-tem. We believe that distal determinantsthat have no clear causal pathway to pre-term birth, such as low socioeconomicstatus, ethnicity, smoking, or illicit druguse, should be collected in a systematicway, but should not be part of the classi-fication system. Some classification sys-tems include potential causes, like stress,unspecified immune, or allergic path-ways, with no clear means of defininghow a specific case gets so classified.
21At
this point, unless a condition can beclearly defined and there is a reasonablyclear pathway from that condition to thepreterm birth, we believe it should beconsidered a potential risk factor butshould not constitute a phenotype in aClinical Opinion Obstetrics www.AJOG.org
114 American Journal of Obstetrics &Gynecology FEBRUARY 2012

classification system. One such example
is whether the method of conception (as-sisted reproductive technologies [ART]vs spontaneous) should be considered arisk factor for preterm birth or merit aseparate phenotypic category? Becausethere is no clear etiologic pathway link-ing ART to increased risk of pretermbirth among singleton or even multifetalgestations, we believe that the method ofconception should be considered as arisk factor for preterm birth but shouldnot constitute a phenotype.
Should pregnancy terminations
and stillbirths be included?
The issue of whether to include preg-
nancy terminations (live born or still-born), occurring at or above the lowergestational age limit in the classification,is controversial. Various stillbirth classi-fication systems handle these cases dif-ferently, with many systems excludingthem.
22An important question in this
study is whether the reason for the termi-nation makes a difference. Terminationsoccur electively but also for diverse rea-sons, such as severe growth retardation,absence of amniotic fluid, P-PROM, ad-vanced cervical dilatation, or a majoranomaly detected at a previable gestationalage above the lower threshold for definingpreterm birth. Some of these fetuses are ei-ther live born or die before delivery.Should these deliveries be included in theclassification system at all, and if so, shouldsome be considered spontaneous, indi-cated, or elective terminations? Our pref-erence is to include all births above thelower gestational age threshold for pre-term birth, whether it was a termination ornot, and within the system, to classify ter-minations of pregnancy by the phenotypesused for all other preterm births. A systemthat includes some terminations but notothers would likely be confusing for all.
Stillbirth is also a difficult issue. In many
data sets used to study preterm birth, still-births are not combined with the livebirths. Preterm births are reported only forlive born infants. This is an important con-sideration because, in developed countries,as many as 50% of stillbirths occur before28 weeks and 80% or more of stillbirths arepreterm.
23,24Also, the pathologic pro –
cesses leading to preterm stillbirths are of-ten similar to those for live born pretermbirths (eg, chorioamnionitis or abrup-tion).
25-27In fact, many intrapartum still –
births occur during preterm labor after adecision that a live fetus in distress is tooimmature to salvage by cesarean delivery.Further confusion is added when a still-birth that occurs in the antepartum periodpresents in preterm labor or with P-PROM.With these considerations in mind, the au-thors agreed that the classification systemforpreterm birth should include all pre-
term stillbirths.
How do we deal with multiple births?
Should multiple births be combinedwith singletons in the same classificationsystem, or should they be consideredseparately? And if separately, shouldtwins and higher-order multiples beconsidered together? If multiples areconsidered separately, should they beclassified using the same system used forsingletons? If a single system were usedfor classifying both singleton and multi-ple preterm births, multiplicity could bepart of a preterm birth phenotype. Thus,there are many questions related to theinclusion or exclusion of multiples inthis system. Perhaps the most importantinfluence on the group was the senti-ment that all preterm births should beincluded in this preterm birth classifica-
tion system. Therefore, our recommen-dation is to create a single classificationsystem, with multiples included in thesystem as 1 potential phenotype for pre-term birth. The number of fetusesshould, of course, be noted. In addition,there are issues related to multiples thatdo not apply to singletons that could beconsidered subcategories within themultiples phenotype, including vanish-ing twin, twin/twin transfusion, fetal de-mise of 1 of multiples, and the type ofplacentation. These characteristics couldbe considered fetal and placental condi-tions in association with the multiplesphenotype.
What should the definition of
indicated and spontaneousbirths be and how do we drawa distinction between them?
The most common classifications divide
all live born preterm births into sponta-neous vs indicated deliveries.28-33How-
ever, review of papers using these catego-ries reveals that these terms are neitherwell defined nor consistently used. Anindicated preterm birth is often definedas one that occurred because continua-tion of the pregnancy risked the health ofthe mother and/or fetus, but the degreeof risk is variably defined, affected by lo-cal circumstances, and may arise from acomplication of pregnancy that had a“spontaneous” onset (eg, infection afterruptured membranes). Thus, these termsneed further exposition for any classifi-cation system to be acceptable to mostusers. We believe that coming to a clearconsensus on this issue is one of the mostimportant requirements to create awidely accepted classification system.
How do we classify P-PROM,
spontaneous dilation, and bleeding?
Classification of preterm (/H11021 37 weeks) pre-
mature (before the onset of labor) rupture
of the fetal membranes (P-PROM) is a par-
ticularly difficult issue.31,34Most women
with confirmed P-PROM enter spontane-ous preterm labor within several hours ordays, depending on the gestational age andcause of rupture, but some remain unde-livered for many days without infection orother complications. In women who donot labor spontaneously, labor might beinduced or a cesarean delivery performedfor many reasons, most commonly be-cause of clinical or laboratory evidenceor fear of infection. Should births in thelatter category be classified as “indi-cated” (because the mothers were not inspontaneous labor), or instead be classi-fied as “spontaneous” (because the pro-cess that led to the preterm delivery—theP-PROM—was spontaneous)? To un-derstand preterm birth, it seems clearthat the phenotypic classification systemshould include information about thepresentation at delivery, and this wouldinclude P-PROM, regardless of whetherit was followed by spontaneous labor oran induction. Forcing it into a spontane-ous or indicated category will likely re-flect a physician management decisionand, thus would not help to define a pre-term birth phenotype.
A spontaneously dilating cervix with-
out contractions may lead to deliverywww.AJOG.org Obstetrics Clinical Opinion
FEBRUARY 2012 American Journal of Obstetrics &Gynecology 115

with few or no contractions, usually at
early gestational ages. At later— but stillpreterm— gestational ages, the findingof advanced cervical dilatation may befollowed by a cesarean delivery becauseof fear of spontaneous membrane rup-ture, followed by head entrapment incases of breech presentation or a pro-lapsed cord. In both instances, parturi-tion is present without any indication ofactive labor. Should these cases be classi-fied as spontaneous, because the dilationoccurred spontaneously, or as indicated,because active labor was not present? Aswith P-PROM, for the purposes of phe-notypic classification, the important in-formation is that the patient presentedwith a dilated cervix, not that she beforced into a specific spontaneous or in-dicated category. These discussions sug-
gest that categorical assignment of all pre-term births into one of the traditionalcategories as spontaneous or indicatedcontributes to confusion rather than clar-ity in the creation of a useful classificationsystem.
Similar issues arise when bleeding is
the initial or dominant manifestation ofparturition. Bleeding may be associatedwith a placental abruption, placenta pre-via, or no obvious pathology. Each con-dition may have different bleeding pat-terns in timing and volume over thecourse of pregnancy. As with P-PROM,induction of labor, or cesarean birth forbleeding because of an ongoing abrup-tion or a placenta previa might be classi-fied as a spontaneous or indicated pre-term birth. Because there was no laborand delivery was accomplished after aprelabor cesarean delivery or inductionof labor, the preterm birth could be con-sidered medically indicated. Conversely,should it be considered spontaneous, be-cause the precipitating event followed aspontaneously occurring maternal con-dition? The discussion surrounding thisissue again led the authors to concludethat attempts to assign preterm births
related to bleeding into spontaneousand indicated groups would be artifac-
tual. Women who present for deliverywith bleeding, either with an abruptionor a previa or without a clearly definedcause, can be characterized phenotypi-cally as having one of those clinical con-ditions, as well as having either signs ofspontaneous initiation of parturition(contractions, cervical effacement, or P-PROM) or a nonspontaneous initiationof parturition (induction or prelabor ce-sarean birth). This discussion empha-sized the need for the classification sys-tem to have several potential phenotypiccomponents, including the maternalcondition, the fetal condition, and thepresentation at delivery.
How should we define and classify
indicated preterm births?
For this classification system, mainte-
nance of the existing terminology relatedto what are customarily called indicatedpreterm births, proved confusing.
35-37
Thus, we chose to define a category of(indicated) preterm birth as one inwhich parturition was initiated by the
caregivers. This designation would applyto a preterm birth in which there was noevidence that any part of the parturi-tional process had begun (ie, little cervi-cal shortening or effacement and nofluid leakage, persistent contractions orbleeding, and specifically, little likeli-hood that birth would have occurredwithin the next several days, unless initi-ated by the obstetric care giver).
However, even if this definition is ac-
cepted, other questions remain. For ex-ample, should medically indicated pre-term births be defined as those followinga cesarean delivery or induction of laboronly for urgent maternal or fetal indica-tions (eg, clearly defined maternal or fe-tal distress as evidenced by severe pre-eclampsia or a dangerously abnormalfetal heart rate pattern)? If so, how do weclassify physician-initiated deliverieswith “softer” indications, such as mildpreeclampsia or mild fetal growth re-striction, in which there is clearly somediscretion in timing of the delivery? Arethese preterm births as “indicated” asthose in the prior group, or should wecall these deliveries “discretionary”?What are the threshold events that de-mark the boundary between indicatedand discretionary, and how are they af-fected by the gestational age and avail-ability of neonatal care?
38
Classification of scheduled births be-
fore 39 weeks that lack any obvious med-ical indication, regardless of the mode ofdelivery, is also an issue of recent con-cern.
39Should these still be considered
“indicated,” because the physician choseto deliver and there was no spontaneousmaternal process leading to labor or de-livery? Should these be called “iatro-genic” or designated as being performedfor “social reasons”? In any case, weagreed that, for this classification system,provider initiated deliveries be subdi-vided into 3 or 4 groups, with headingssuch as urgent, discretionary, iatrogenic,and/or social.
There were a number of remaining
questions. For example, how do we clas-sify preterm deliveries where the motherentered the hospital before term withcontractions or slight cervical change butwithout active labor? If her labor was“augmented” by amniotomy or oxytocinor a cesarean birth performed, is this tobe categorized as a spontaneous pretermbirth or an indicated, discretionary, oriatrogenic preterm birth? For the classi-fication system, we must be able to dis-tinguish between (1) essentially socialor convenience inductions of labor inwomen with minimal signs of active la-bor and (2) appropriate augmentation ofspontaneous dysfunctional labor. Thus,an important issue is whether the classi-fication system should attempt to deter-mine the reason and perhaps appropri-ateness for the physician’s decision toinitiate delivery? We agree that both thetype of indication, such as urgent, discre-tionary, and iatrogenic or social, and themedical or social conditions leading tothe decision to initiate a preterm deliveryshould be captured in this classificationsystem.
Other important issues
At times, signs of spontaneous parturi-tion will occur in pregnancies compli-cated by preeclampsia, maternal illness,fetal growth restriction, and fetal dis-tress, although these conditions mightnot be part of another obvious pheno-type that led to the preterm birth.
40If
preeclampsia is present in a pretermbirth that follows spontaneous onset oflabor, should this birth be still be classi-fied as a spontaneous preterm birth? Weagree that these births should still be clas-Clinical Opinion Obstetrics www.AJOG.org
116 American Journal of Obstetrics &Gynecology FEBRUARY 2012

sified as having signs of spontaneous
parturition, but the type of informationdiscussed previously should be collected,as it will allow an examination of the linkbetween various maternal and fetal con-ditions and different preterm deliverypresentations.
Another important issue is how to best
integrate placental pathologic and otherlaboratory information into a classifica-tion system based on phenotype. For theplacenta, if histologic chorioamnionitis,signs of abruption, or of placental dys-function (as might be indicated by largeareas of infarction or necrosis) are pres-ent, how should these findings affectthe classification? For laboratory tests,would an elevated white count or a pos-itive blood or amniotic fluid culture beincluded in an infection-related pheno-type? These questions must also be con-sidered before a classification system canbe developed. Discussion surroundingthis issue led us to recommend that theplacental findings be included as part ofthe phenotype of preterm birth with 4potential components: infection, hem-orrhage, infarction, or no pathology.Important laboratory findings, such asevidence of infection/inflammation onamniocentesis, would become part of thephenotype within the category on im-portant maternal pregnancy-relatedconditions.
In this study, we use polyhydramnios
as an example of how the classificationsystem might work. Women with appar-ently similar degrees of polyhydramniosmay present with spontaneous contrac-tions, develop P-PROM, or dilate theircervix.
41Still others will be induced or
undergo cesarean delivery performed forfear of prolapsed cord or a ruptureduterus. In discussing whether all suchcases should be considered spontaneousbecause the process started with a mater-nal or fetal condition or whether thespontaneous classification should be re-served for only those cases that presentedwith contractions or P-PROM, it be-came clear that the classification sys-
tem should capture the presence of poly-hydramnios, the presence or absence of afetal anomaly, whether there was evi-dence of spontaneous parturition, and ifnot, whether the physician initiated de-livery was urgent, discretionary, or iatro-genic. A classification system with thesecharacteristics would allow analysis of allcases of polyhydramnios, for example, asa single group, regardless of whetherthe woman presented for delivery withP-PROM, labor, a spontaneously dilatedcervix, or for one of many reasons wasinduced, had a termination, or had a ce-sarean birth before the appearance ofsigns of spontaneous parturition.
This discussion also led us to add a
fourth component to the classification sys-tem, one dealing with the fetal condition.Thus, the presence of a fetal demise, fetaldistress, fetal growth restriction, a congen-ital anomaly, multifetal pregnancy, andpoly- or oligohydramnios may influencewhen a delivery occurs and should be in-cluded in the fetal component of the clas-sification system.
Definitions
For this classification system to achieve itsgoals, virtually all of the maternal and fetalconditions, presentations at delivery, andplacental findings that may comprise aphenotype must be rigorously defined. Forexample, how much hydramnios must oc-cur and when must it occur for polyhy-dramnios to be considered a component ofthe phenotype of a preterm birth?
Moving toward a
classification system
From the foregoing discussion, the issues
and components of a preterm birth phe-notypic classification system are comingmore clearly into focus. After much discus-sion, we agree that a preterm phenotypecould be defined as having the following 4components: (1) the presence of impor-tant maternal pregnancy related condi-tions; (2) important fetal conditions; (3)clinical presentation for delivery, includ-ing evidence of spontaneous parturition;and (4) placental findings. Risk factors forpreterm birth, such as smoking, could becollected but would not be part of the phe-notype. We recognize that the dividing linebetween significant maternal conditionsand maternal risk factors is not always clearand that various characteristics might beput in one or the other category with somedegree of arbitrariness. The use of thewords “spontaneous” and “indicated” tocategorize the presentation for deliveryhave not been used in a consistent fashionand would be better replaced by less con-fusing, more descriptive terms. Neverthe-less, by carefully defining the presentationfor delivery, the concept of 2 broad catego-ries of preterm births—those followingspontaneous signs of parturition and
those cases where there was none—shouldbe retained, with cases where the birth pro-cess is of maternal/fetal origin, including
shortening cervix, P-PROM, contrac-
tions, and bleeding classified as spontane-ous. All other births in which deliverywould likely not have happened withinseveral days without the intervention of a
caregiver, should be placed in the sec-
ond, provider-initiated group. Severalbroad categories of maternal conditionsshould be noted as part of the phenotypes,usually based on information available be-fore presentation for delivery, includingclinical categories, such as shortened cervixand polyhydramnios.
Conclusions
Preterm birth is a syndrome defined bytime and clearly is not a distinct clinicalphenotype. Births at gestational ages lessthan 20 weeks and many of those at 37and 38 weeks share with births at 20-36weeks several etiologic and prognosticfeatures that suggest these boundariesare artificial and therefore, should be re-considered.
2-15,42,43Because the cause of
many preterm births is unknown, wealso believe that, at least for the near fu-ture, preterm birth classification systemswill need to focus on phenotype ratherthan suspected cause. These phenotypes,whenever possible, should be based, atleast in part, on maternal and/or fetal an-tecedent events, such as a shortening cer-vix or fetal death, with the understand-ing that presentation at delivery,including P-PROM, bleeding, contrac-tions, or cervical dilation, may all besymptoms of the underlying process andmay not be primary in determining orlabeling the pathway leading to the pre-term delivery. Finally, when the care-giver initiates a preterm delivery, a dis-tinction should be made between casesin whom such interventions are clearlyindicated, those in whom the timing ofintervention is discretionary, and thosewww.AJOG.org Obstetrics Clinical Opinion
FEBRUARY 2012 American Journal of Obstetrics &Gynecology 117

without a clinical indication. Thus, cre-
ating a classification system for pretermbirth involves making many choices,some of which are clearly controversial.The issues described in this article aresome that should be considered in creat-ing a classification system for pretermbirth phenotypes.
f
REFERENCES
1.Kramer MS, Papageorghiou A, Culhane JF,
et al. Challenges in defining and classifying the
preterm birth syndrome. Am J Obstet Gynecol2012;206:108-12.2.Goldenberg RL, Mayberry SK, Copper RL,
DuBard MB, Hauth JC. Pregnancy outcomesfollowing a second trimester loss. Obstet Gyne-col 1993;81:444-6.3.Rasmussen S, Irgens LM, Skjaerven R,
Melve KK. Prior adverse pregnancy outcomeand the risk of stillbirth. Obstet Gynecol2009;114:1259-70.4.Iams JD, Berghella V. Care for women with
prior preterm birth. Am J Obstet Gynecol2010;203:89-100.5.Nuovo GJ, Cooper LD, Bartholomew D. His-
tologic, infectious, and molecular correlates ofidiopathic spontaneous abortion and perinatalmortality. Diagn Mol Pathol 2005;14:152-8.6.Srinivas S, Ma Y, Sammel MD, et al. Placental
inflammation and viral infection are implicated insecond trimester pregnancy loss. Am J ObstetGynecol 2006;195:797-802.7.Ancel PY, Saurel-Cubizolles MJ, Di Renzo
GC, Papiernik E, Breart G. Risk factors for14-21 week abortions: a case-control study inEurope. The Europop Group. Hum Reprod2000;15:2426-32.8.Leitich H, Bodner-Adler B, Brunbauer M,
Kaider A, Egarter C, Husslein P. Bacterial vagi-nosis as a risk factor for preterm delivery: ameta-analysis. Am J Obstet Gynecol 2003;189:139-47.9.Llahi-Camp JM, Rai R, Ison C, Regan L, Tay-
lor-Robinson D. Association of bacterial vagino-sis with a history of second trimester miscar-riage. Hum Reprod 1996;11:1575-8.
10.Nelson DB, Bellamy S, Nachamkin I, Ness
RB, Macones GA, Allen-Taylor L. First trimesterbacterial vaginosis, individual microorganismlevels, and risk of second trimester pregnancyloss among urban women. Fertil Steril 2007;88:1396-403.11.Edlow AG, Srinivas SK, Elovitz MA. Second
trimester loss and subsequent pregnancy out-comes: what is the real risk? Am J Obstet Gy-necol 2007;197:581.e1-6.12.Tita ATN, Landon MB, Spong CY, et al.
Timing of elective cesarean delivery at term andneonatal outcomes. N Engl J Med 2009;360:
111-20.13.Zhang X, Kramer MS. Variations in mortality
and morbidity by gestational age among infantsborn at term. J Pediatr 2009;154:358-62.14.McIntire DD, Leveno KJ. Neonatal mortality
and morbidity rates in late preterm births com-pared with births at term. Obstet Gynecol2008;111:35-41.15.Bastek JA, Sammel MD, Paré E, Srinivas
SK, Posencheg MA, Elovitz MA. Adverse neo-natal outcomes: examining the risks betweenpreterm, late preterm, and term infants. Am JObstet Gynecol 2008;199:367.e1-8.16.Chen A, Feresu SA, Barsoom MJ. Hetero-
geneity of preterm birth subtypes in relation toneonatal death. Obstet Gynecol 2009;114:516-22.17.Moutquin JM. Classification and heteroge-
neity of preterm birth. Br J Obstet Gynaecol2003;110(suppl 20):30-3.18.Ancel PY, Saurel-Cubizolles MJ, Di Renzo
GC, Papiernik E, Breart G. Very and moderatepreterm births: are the risk factors different?Br J Obstet Gynaecol 1999;106:1162-70.19.Behrman RE, Butler AS, eds. Preterm birth:
causes, consequences, and prevention. Com-mittee on understanding premature birth andassuring healthy outcomes. Institute of Medi-cine of the National Academies. Washington,DC: The National Academies Press; 2006:79.20.Celik E, To M, Gajewska K, Smith GC, Ni-
colaides KH. Cervical length and obstetric his-tory predict spontaneous preterm birth: devel-opment and validation of a model to provideindividualized risk assessment. Ultrasound Ob-stet Gynecol 2008;31:549-54.21.Klebanoff MA, Shiono PH. Top down, bot-
tom up and inside out: reflections on pretermbirth. Paediatr Perinat Epidemiol 1995;9:125-9.22.Flenady V, Froen JF, Pinar H, et al. An eval-
uation of classification systems for stillbirth.BMC Pregnancy Childbirth 2009;9:24.23.Goldenberg RL, Koski JF, Boyd BW, Cutter
GR, Nelson KG. Fetal deaths in Alabama 1974to 1983: a birth weight-specific analysis. ObstetGynecol 1987;70:831-5.24.Flenady V, Middleton P, Smith G, et al. Still-
birth: the way forward in high income countries.Lancet 2011;377:1703-17.25.Korteweg FJ, Gordijn SJ, Timmer A, Holm
JP, Ravise JM, Erwich JJ. A placental cause ofintra-uterine fetal death depends on the perina-tal mortality classification system used. Pla-centa 2008;29:71-80.26.Khong TY. The placenta in stillbirth. Curr
Diag Pathol 2006;12:161-72.27.Pinar H, Carpenter M. Placenta and umbil-
ical cord abnormalities seen with stillbirth. ClinObstet Gynecol 2010;53:656-72.28.Savitz DA, Blackmore CA, Thorp JM. Epi-
demiologic characteristics of preterm delivery:etiologic heterogeneity. Am J Obstet Gynecol
1991;164:467-71.29.Ananth CV, Joseph KS, Oyelese Y, Demis-
sie K, Vintzileos AM. Trends in preterm birth andperinatal mortality among singletons: UnitedStates, 1989 through 2000. Obstet Gynecol2005;105:1084-91.30.Barros F, Velez Mdel P. Temporal trends of
preterm birth subtypes and neonatal outcomes.Obstet Gynecol 2006;107:1035-41.31.Pickett KE, Abrams B, Selvin S. Defining
preterm delivery–the epidemiology of clinicalpresentation. Paediatr Perinat Epidemiol 2000;14:305-8.32.Savitz DA, Dole N, Herring AH, et al. Should
spontaneous and medically indicated pretermbirths be separated for studying aetiology?Paediatr Perinat Epidemiol 2005;19:97-105.33.McElrath TF, Hecht JL, Dammon O, et al.
Pregnancy disorders that lead to delivery beforethe 28th week of gestation: an epidemiologicapproach to classification. Am J Epidemiol2008;168:980-9.34.Furman B, Shoham-Vardi I, Bashiri A, Erez
O, Mazor M. Clinical significance and outcomeof preterm prelabor rupture of the membranes:population-based study. Eur J Obstet GynecolRepro Biol 2000;92:209-16.35.Ananth CV, Vintzileos AM. Medically in-
duced preterm birth: recognizing the impor-tance of the problem. Clin Perinatol 2008;35:53-67.36.Ananth CV, Joseph KS, Kinzler WL. The in-
fluence of obstetric interventions on trends intwin stillbirths, United States1989-1999. J Ma-tern Fetal Neonatal Med 2004;15:380-7.37.Klebanoff MA. Conceptualizing categories
of preterm birth. Prenatal Neonatal Med 1998;3:13-5.38.Kierse MJNC. Elective induction, selective
deduction and cesarean section. Birth 2010;37:252-6.39.Clark SL, Miller DD, Belfort MA, Dildy GA,
Frye DK, Meyers JA. Neonatal and maternaloutcomes associated with elective term deliv-ery. Am J Obstet Gynecol 2009;200:156.e1-4.40.Ananth CV, Savitz D, Luther E, Bowes W.
Preeclampsia and preterm birth subtypes inNova Scotia, 1986 to 1992. Am J Perinatol1997;14:17-23.
41.Many A, Hill LM, Lazebnik N, Martin JG. The
association between polyhydramnios and pre-term delivery. Obstet Gynecol 1995;86:389-91.42.Kirby RS, Wingate M. Late preterm birth
and neonatal outcome: is 37 weeks’ gestation athreshold level or a road marker on the highwayof perinatal risk? Birth 2010;37:169-71.43.Fleischman AR, Oinuma M, Clark SL. Re-
thinking the definition of “term pregnancy.” Ob-stet Gynecol 2010;116:136-9.
Clinical Opinion Obstetrics www.AJOG.org
118 American Journal of Obstetrics &Gynecology FEBRUARY 2012