|Home | About | Journals | Submit | Contact Us | Français|
Epidemiologists have grouped the multiple disorders that lead to preterm delivery before the 28th week of gestation in a variety of ways. The authors sought to identify characteristics that would help guide how to classify disorders that lead to such preterm delivery. They enrolled 1,006 women who delivered a liveborn singleton infant of less than 28 weeks' gestation at 14 centers in the United States between 2002 and 2004. Each delivery was classified by presentation: preterm labor (40%), prelabor premature rupture of membranes (23%), preeclampsia (18%), placental abruption (11%), cervical incompetence (5%), and fetal indication/intrauterine growth restriction (3%). Using factor analysis (eigenvalue=1.73) to compare characteristics identified by standardized interview, chart review, placental histology, and placental microbiology among the presentation groups, the authors found 2 broad patterns. One pattern, characterized by histologic chorioamnionitis and placental microbe recovery, was associated with preterm labor, prelabor premature rupture of membranes, placental abruption, and cervical insufficiency. The other, characterized by a paucity of organisms and inflammation but the presence of histologic features of dysfunctional placentation, was associated with preeclampsia and fetal indication/intrauterine growth restriction. Disorders leading to preterm delivery may be separated into two groups: those associated with intrauterine inflammation and those associated with aberrations of placentation.
The burdens on families and society associated with extreme premature delivery have prompted efforts to find ways to reduce the occurrence of this disorder (1–4). Some epidemiologists suggest that, if different clinical presentations share etiologies, then grouping them together maximizes the power of epidemiologic studies (5–7) and can recommend common therapeutic interventions. Others feel that, until the evidence is stronger, it is best to study individual entities (2, 8–10). Finally, some are not yet convinced there is an advantage either way (4).
Our large prospective study of pregnancies that ended with a livebirth before the 28th week of gestation allowed us to explore the heterogeneity/homogeneity of clinically defined subgroups of these pregnancies. In addition to the traditional information about maternal demographic and clinical characteristics, we gathered details about placental microbiology and histology. These additional data provided insight into the intrauterine conditions that existed prior to delivery, thereby helping us classify the antecedent conditions associated with extremely preterm delivery.
During the years 2002–2004, 1,249 women delivering before 28 weeks’ gestation at 1 of 14 participating institutions in 11 cities in 5 states consented to participate in the Extremely Low Gestational Age Newborns (ELGAN) Study, which was designed to identify characteristics and exposures that increase the risk of structural and functional neurologic disorders. Approximately 260 (17%) women were either missed or did not consent to participate. Because the epidemiology of preterm delivery in singleton pregnancies differs from that of multiple gestation pregnancies (11), we limited our study to the 1,006 (81%) singleton deliveries (Table 1). Each of the individual institutional review boards approved the enrollment and consent processes.
After delivery, a trained research nurse interviewed each mother in her native language using a structured questionnaire. The medical record provided information about events following admission.
The clinical circumstances that led to each maternal admission and ultimately to each preterm delivery were operationally defined by using data from the maternal interview and data abstracted from the medical record. Preterm labor was defined as progressive cervical dilation with regular contractions and intact membranes. The diagnosis of prelabor premature rupture of membranes (pPROM) was defined as the presence of vaginal pooling with either documented nitrazine-positive testing or ferning prior to regular uterine activity. Preeclampsia was defined as new-onset hypertension and proteinuria of sufficient severity to warrant delivery for either a maternal or fetal indication/intrauterine growth restriction (IUGR). For a diagnosis of cervical insufficiency, a woman had to present with cervical dilation of greater than 2 cm, in the absence of membrane rupture and detected or perceived uterine activity. Prolapse of the fetal membranes was not required. Placental abruption was defined as presentation with a significant amount of vaginal bleeding (either documented in the medical record or a postpartum hematocrit of <24%) and a clinical diagnosis of placental abruption in the absence of cervical change. Painful uterine contractions were not required. Among the presentations in the category of fetal indication/IUGR were nonreassuring fetal testing, oligohydramnious, Doppler abnormalities of umbilical cord blood flow, and severe intrauterine growth restriction based on antepartum ultrasound examination. For the purposes of this study, the initial complication that caused presentation to medical attention is the focus of analysis. Complications that occurred after presentation were not considered independent and as such were not assumed to have a separate etiology.
The gestational age estimates were based on a hierarchy of the quality of available information. Most desirable were estimates based on the dates of embryo retrieval or intrauterine insemination or an ultrasound examination before the 14th week of gestation (62%). When these were not available, reliance was placed sequentially on an ultrasound examination at 14 or more weeks (29%), menstrual dating without ultrasound confirmation (7%), and gestational age recorded in the log of the neonatal intensive care unit (1%). The birth weight z score is the number of standard deviations the infant's birth weight is above or below the median weight of infants at the same gestational age in a standard data set (12).
Delivered placentas were placed in a sterile examination basin and transported to a sampling room. Eighty-two percent of the samples were obtained within 1 hour of delivery.
Details about the microbiologic and histologic analyses are presented elsewhere (13, 14). Briefly, midway between the cord insertion and the edge of the placental disk, a piece of chorion and the underlying trophoblastic tissue was obtained under sterile conditions, placed in liquid nitrogen, and then stored at −80°C. In the central microbiology laboratory, frozen samples were allowed to thaw at room temperature, homogenized, diluted, and plated on selective and nonselective media. After incubation, the various colony types were enumerated, isolated, and identified by established criteria (15).
In keeping with the guidelines of the College of American Pathologists’ conference (16), representative sections were taken from all abnormal areas, as well as routine sections of the umbilical cord, a membrane roll, and full-thickness sections from the center and a paracentral zone of the placental disk. After exercises were completed to minimize observer variability, placentas were reviewed at each site by a study pathologist, who completed a structured data collection form without knowledge of clinical details.
The main generalized null hypothesis that we evaluated is that pregnancy disorders leading to preterm delivery do not differ in demographic, clinical, microbiologic, or histologic characteristics. We tested this hypothesis in our search for the antecedents of each of the 6 pregnancy disorders. All P values presented are based on chi-squared analyses for homogeneity.
We also explored the possibility that common features characterize groups of these pregnancy disorders leading to preterm delivery. In our search for groups of pregnancy disorders, we relied on nonhierarchical cluster approaches and chose factor analysis (with varimax rotation) because it is the most transparent, identifying and quantifying interrelations and the structural basis for what is common to each factor as well as what is common to each disorder.
For the correlation matrix that served as the basis in this paper, we selected all 15 variables from Tables 11–5 that distinguished one pregnancy disorder from the rest. Four of these are demographic variables: maternal identification as White, maternal age <21 years, maternal education less than college, and self and/or spouse support; 6 are clinical characteristics: any tobacco during pregnancy, body mass index of ≥30 kg/m2, primigravidity, conception assistance, self-report of vaginitis during pregnancy, and the prescription of an antibiotic prior to hospitalization; 3 are newborn characteristics: birth weight z score of <−2, gestational age of <25 weeks, and male gender; 2 are histologic characteristics of the placenta: grade 3+ chorionic plate inflammation and increased syncytial knots; and 3 are bacteriologic characteristics of the placenta: recovery of an organism, recovery of a vaginal organism, and more than one organism recovered.
Another factor analysis (not shown), based on only 9 variables that discriminated 2 or more pregnancy disorders from the rest, provided a first factor (eigenvalue=1.52) similar to that obtained with 15 variables and with similar discriminating information.
Overall, preterm labor accounted for the largest proportion (40%) and fetal indication/IUGR accounted for the smallest proportion (3%) of deliveries. In between were pPROM (23%), preeclampsia (18%), placental abruption (11%), and cervical insufficiency (5%). Among the deliveries in our sample, abruption tended to occur preferentially among Whites, women less than 21 years of age, and women who did not graduate from high school. Similarly, fetal indication/IUGR occurred most commonly among women who identified themselves as Black and among women who did not rely on a partner, family, or friends for financial support (Table 1).
Smoking, both prior to and during pregnancy, was least common among women delivering pregnancies complicated by preeclampsia. Exposure to passive smoke was relatively uncommon in our sample among women presenting with cervical insufficiency. Among those delivering at less than 28 weeks' gestation, a high prepregnancy body mass index (i.e., body mass index of ≥30 kg/m2) was most common among women with the complications of preeclampsia and fetal indication/IUGR. Preeclampsia was overrepresented among primigravidas and women with intended pregnancies, although placental abruption and cervical insufficiency were most common among women who sought conception assistance. The rate of vaginitis was higher among women who presented with cervical insufficiency. These women and women who presented with fetal indication/IUGR were most likely to have received an antibiotic at some point prior to admission (Table 2).
Preeclampsia and fetal indication/IUGR were observed at older gestational ages. Birth weight and, even more prominently, birth weight z scores were lower among the pregnancies complicated by preeclampsia and fetal indication/IUGR. Male babies were overrepresented among deliveries complicated by cervical insufficiency, and female babies were overrepresented among those delivered in the setting of preeclampsia. The mortality rate was highest among infants of pregnancies complicated by preeclampsia and fetal indication/IUGR, even after controlling for maternal race, gestational age, and a complete course of antenatal steroids (Tables 3 and and44).
A total of 943 (94%) of the placentas were available for histologic examination. Inflammation in the chorionic plate, the fetal membranes, and the umbilical cords was most frequent in placentas delivered after pPROM and cervical insufficiency, somewhat less frequent in placentas delivered after preterm labor and placental abruption, and least frequent in placentas delivered after preeclampsia and for fetal indication/IUGR.
Placentas of pregnancies complicated by preeclampsia and fetal indication/IUGR were most likely to have infarcts and an abundance of syncytial knots, both of which are considered indicators of poor placentation (Table 5).
Compared with the placentas of pregnancies complicated by preeclampsia or fetal indication/IUGR, those delivered of pregnancies complicated by preterm labor, pPROM, placental abruption, and cervical insufficiency were much more likely to harbor an organism, especially those common to the vagina (i.e., Prevotella bivia, Lactobacillus spp., Peptostreptococcus magnus, or Gardnerella vaginalis). More than 30% of the placentas of these pregnancies harbored multiple organisms. In contrast, only one-quarter of pregnancies delivered for preeclampsia or fetal indication/IUGR had any microorganism present, and the majority of these tended to be single-organism culture.
To assess the magnitude of possible contamination, we conducted separate analyses of placentas delivered by cesarean section (data not shown). The rate of vaginal organism recovery from placentas delivered following labor, pPROM, placental abruption, and cervical insufficiency, although reduced, remained many times higher than the rate among placentas from preeclampsia and fetal indication/IUGR pregnancies (Table 6).
We plotted the frequency of organism recovery along the x-axis (as an indicator of infection/inflammation) and the frequency of finding increased syncytial knots along the y-axis (as an indicator of impaired placentation). As expected from Tables 5 and and6,6, preeclampsia and fetal indication/IUGR cluster together far removed from the cluster of preterm labor, pPROM, placenta abruption, and cervical insufficiency. A plot of the percentage of placentas that harbored an organism on one axis and the percentage of placentas with grade 3+ chorionic plate inflammation on the other axis also clearly separated these two groups as did a plot of the percentage of placentas with grade 3+ chorionic plate inflammation on one axis and the percentage of placentas with increased syncytial knots on the other axis (not shown) (Figure 1).
The first factor, with an eigenvalue of 1.73, was positively associated with organism recovery from the placenta, placental inflammation, and early gestational age and negatively associated with growth restriction and the finding of syncytial knots (Table 7). The second factor, with an eigenvalue of 1.07, was negatively associated with younger maternal age, lower educational attainment, and primiparity but positively associated with self and/or partner support. The third factor, with an eigenvalue of 0.60, was negatively associated with not attending or completing college but positively associated with identification as White, self and/or partner support, primiparity, and fertility treatment.
The mean first factor scores of each pregnancy complication can be divided into 2 groups (Table 8). One, characterized by high negative scores, identifies preeclampsia and fetal indication/IUGR. The other, characterized by low to moderate positive scores, identifies preterm labor, pPROM, placental abruption, and cervical insufficiency. The mean second and third factor scores did not provide additional information that discriminated 2 or more pregnancy disorders from the rest.
Our search for the antecedents of pregnancy disorders that lead to preterm delivery fueled our search for features common to these disorders. Although the 6 pregnancy disorders that lead to delivery much before term have distinct clinical presentations, our findings suggest that they can be divided into 2 broad groups on the basis of common characteristics. One group, characterized by the presence of infection and inflammation and the absence of indicators of impaired placentation, includes preterm labor, pPROM, placental abruption, and cervical insufficiency. “Spontaneous” preterm delivery has been defined as the group of entities that are neither maternal nor fetal indication/IUGR, and it typically includes complications such as preterm labor and pPROM. We prefer, however, to group these causes of preterm delivery under the description “inflammatory,” because it more accurately summarizes the observations made in this analysis, and to include placental abruption and cervical insufficiency under this umbrella.
The other group, characterized by infarcts and increased syncytial knots in the placenta, the relative absence of inflammation, and a higher neonatal mortality rate, includes preeclampsia and the entity identified as fetal indication/IUGR. The presence of increased syncytial knots and infarcts has been viewed as evidence of maternal-placental insufficiency (17, 18). Low-birth-weight z scores, another indicator of placental function, are characteristic of infants delivered for fetal indication/IUGR and preeclampsia (19, 20). Intrauterine growth restriction has been attributed to a restricted arteriolar supply of the uterine-placental interface (21). Delivery for preeclampsia and fetal indication/IUGR has been labeled “indicated” or “nonspontaneous” delivery. We, however, prefer to identify this group with the more pathophysiologically accurate term of “placental dysfunction.”
This study is unique in culturing placental stroma rather than amniotic fluid or fetal membranes. Nevertheless, the results are similar, recovering a diverse group of organisms with no one organism or class of organisms predominating (22–24). We also observed that vaginal microorganisms occur preferentially in the placentas of women with preterm labor, pPROM, placental abruption, and cervical insufficiency regardless of whether delivery is vaginal or abdominal, suggesting that these organisms are not merely contaminants associated with the route of delivery but are biologically significant.
Our study is not unique in finding elevated frequencies of histologic inflammation in placentas from pregnancies complicated by preterm labor and pPROM (25, 26), nor is it unique in finding evidence that placental abruption is characterized by inflammatory processes (27–29). However, our observations allow us to go the next step and suggest that the patterns of histologic changes and microbes recovered from placentas after both cervical insufficiency and abruption are similar to those of placentas delivered after preterm labor or pPROM.
We observed that demographic variables discriminate among the 6 pregnancy complications. As others have also observed, we found that maternal age (30–33), educational attainment (34), and race (35, 36) differ by cause of preterm delivery. Unlike others, we did not find that marital status (37, 38) and limited financial resources (39, 40) discriminate among the causes of preterm delivery.
Characteristics that are more specific to the physiology of the pregnancy itself such as maternal smoking, body mass index, and conception assistance also discriminate among the 6 disorders that lead to preterm delivery. Consistent with prior work, we observed that smoking was associated with reduced risk of preeclampsia (41, 42) and an increased risk of pPROM and placental abruption (8, 43–45). Similarly, we confirmed a positive association between increased body mass index and both fetal indication/IUGR and preeclampsia (46, 47).
Our data support the hypothesis that the intrauterine environment influences the risk of mortality in the preterm neonate (48). Neonatal mortality was higher in pregnancy disorders associated with placental dysfunction than in those disorders associated with intrauterine inflammation.
In summary, we have found support for classifying the disorders leading to delivery well before term in 2 broad groups. What we call the “inflammatory” group and others call “spontaneous preterm delivery” includes preterm labor, pPROM, placental abruption, and cervical insufficiency and is characterized by evidence of infection and inflammation. The “placental dysfunction” group, otherwise classified as “indicated delivery,” includes preeclampsia and fetal indication/IUGR and is characterized by the relative absence of inflammation and the presence of infarcts, as well as increased syncytial knots in the placenta. Further work will be needed to suggest whether intervening at a point before the final common pathway of each group can potentially reduce the burdens that these disorders place on families and society.
Author affiliations: Division of Maternal-Fetal Medicine, Brigham and Women's Hospital, Boston, Massachusetts (T. F. McElrath); Department of Pathology, Beth Israel Deaconess Medical Center, Boston, Massachusetts (J. L. Hecht); Division of Newborn Medicine, Tufts Floating Hospital for Children, Boston, Massachusetts (O. Dammann); University of North Carolina, Chapel Hill, North Carolina (K. Boggess); Channing Laboratory, Brigham and Women's Hospital, Boston, Massachusetts (A. Onderdonk); Department of Obstetrics and Gynecology, Bay State Medical Center, Springfield, Massachusetts (G. Markenson); Department of Maternal-Fetal Medicine, Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina (M. Harper); Department of Obstetrics and Gynecology, University of Massachusetts Memorial Medical Center, Worcester, Massachusetts (E. Delpapa); and Neuroepidemiology Unit, Children's Hospital of Boston, Boston, Massachusetts (E. N. Allred, A. Leviton).
The ELGAN project was supported by a cooperative agreement with the National Institute of Neurologic Diseases and Stroke (U01 NS 400069-01). Dr. McElrath was supported by the Women's Reproductive Health Research Program, National Institutes of Child Health and Development (K12 HDO1255).
The ELGAN Study participating centers in alphabetical order (site principal investigator, pathologist, and perinatologist) were as follows: Baystate Medical Center, Springfield, Massachusetts (Bhavesh Shah, Solveg Pflueger, Glenn Markenson); Beth Israel Deaconess Medical Center, Boston, Massachusetts (Camilia R. Martin, Jonathon L. Hecht, Bruce Cohen); Brigham and Women's Hospital, Boston, Massachusetts (Linda J. Van Marter, Harvey Kliman, Thomas F. McElrath); Children's Hospital, Boston, Massachusetts (Alan Leviton); Massachusetts General Hospital, Boston, Massachusetts (Robert Insoft, Drucilla Roberts, Laura Riley); Tufts New England Medical Center, Boston, Massachusetts (Cynthia Cole/John Fiascone, Ina Bhan, Sabrina Craigo/Theresa Marino); University of Massachusetts Memorial Health Center, Worcester, Massachusetts (Francis Bednarek, Gamze Ayata, Ellen Delpapa); Yale-New Haven Hospital, New Haven, Connecticut (Richard Ehrenkranz, Miguel Reyes-Múgica/Eduardo Zambrano, Keith P. Williams); Forsyth Hospital, Baptist Medical Center, Winston-Salem, North Carolina (T. Michael O'Shea, Dennis W. Ross, Maggie Harper); University Health Systems of Eastern Carolina, Greenville, North Carolina (Stephen Engelke, John Christie, Hamid Hadi); North Carolina Children's Hospital, Chapel Hill, North Carolina (Carl Bose, Chad Livasy, Kim Boggess); DeVos Children's Hospital, Grand Rapids, Michigan (Mariel Portenga, Barbara Doss, Curtis Cook); Sparrow Hospital, Lansing, Michigan (Padmani Karna, Gabriel Chamyan, Steve Roth); University of Chicago Hospital, Chicago, Illinois (Michael D. Schreiber, Aliya Husain, Mahmoud Ismail); William Beaumont Hospital, Royal Oak, Michigan (Daniel Batton, Chung-Ho Chang, Robert Lorenz).
Conflict of interest: none declared.