Preterm premature rupture of membranes (PROM) complicates approximately 25% to 30% of all preterm births and is one of the leading causes of prematurity in the United States.1,2 Possible underlying mechanisms of preterm PROM include uterine overdistention, ischemia, hemorrhage, stress, and, perhaps most importantly, infection or inflammation or both. In a study by Goncalves et al,3 among women with preterm PROM, the rate of positive amniotic fluid cultures at admission was 32.4%; however, at the time of labor onset, as many as 75% of women will have microbial invasion,4 suggesting that this invasion occurs during the latency period.5 Moreover, intra-amniotic inflammation is present in approximately 40% of patients with preterm PROM and is a risk factor for impending preterm delivery and adverse neonatal outcomes regardless of culture results.6 Despite this, the majority of women presenting with preterm PROM do not have overt clinical infection diagnosed,3 and current paradigm recommendations include expectant management if no maternal or fetal contraindications exist.7,8
Neonatal risks after preterm PROM are primarily attributed to prematurity and include sepsis, respiratory distress, intraventricular hemorrhage, necrotizing enterocolitis, and neonatal death.7 However, it is unclear whether immediate delivery after presentation with preterm PROM remote from term reduces these risks,9–11 or whether the presence of preterm PROM increases neonatal mortality risk. Today, pregnant patients at risk for preterm delivery regardless of the underlying etiology are routinely counseled about adverse neonatal risks related to prematurity, yet often the presence or absence of preterm PROM is not incorporated into this discussion. The aim of our study was to compare neonatal mortality rates between premature newborns delivered in the setting of preterm PROM to those delivered without membrane rupture before delivery.
MATERIALS AND METHODS
A cross-sectional study of neonates delivered and recorded by the California Perinatal Quality Care Collaborative (www.cpqcc.org) was performed. The collaborative includes 127 member hospitals, representing more than 90% of all neonates cared for in California neonatal intensive care units. Member hospitals include private and public institutions, as well tertiary and community hospitals. Membership is offered to any hospital in California that provides neonatal intensive care. The collaborative is supported in part by the David and Lucile Packard Foundation and by the California Department of Public Health, Maternal, Child, and Adolescent Health Program.
The California Perinatal Quality Care Collaborative collects clinical data in a prospective fashion for neonates born at member hospitals by use of an expanded version of the Vermont Oxford Dataset. The collaborative conducts yearly data abstractor trainings at locations throughout California. Each record has a variety of range and logic checks at the time of data collection and before data closeout. Records with excessive missing data are audited. Maternal obstetric data are collected by each neonatal intensive care unit from maternal records and submitted concurrently with neonatal outcome data.
The diagnosis of ruptured membranes is made clinically by the treating obstetricians in the maternal record. The dataset does not account for individual clinical findings such as vaginal pooling of amniotic fluid, nitrazine testing, or fern appearance of amniotic fluid during microscopic visualization. For the present study, preterm PROM was defined by the database as rupture of membranes occurring at a gestational age less than 37 weeks. Preterm PROM was then subcategorized as occurring less than (“recent preterm PROM”) or more than (“prolonged preterm PROM”) 18 hours before delivery. The 18-hour latency period is currently recorded as a dichotomous variable by the California Perinatal Quality Care Collaborative. Maternal demographic and clinical data including age, race, hypertension, prenatal care, antenatal steroid administration, intrauterine infection, and mode of delivery were recorded. Intrauterine infection was diagnosed clinically by the treating obstetrician. Neonatal mortality rates were compared for neonates born without preterm PROM and those with recent and prolonged preterm PROM. Neonatal mortality is defined by the California Perinatal Quality Care Collaborative as death at any time before hospital discharge. Neonates with fetal anomalies were excluded.
We also compared early neonatal sepsis rates according to preterm PROM status. Early neonatal sepsis was defined as a positive blood culture or cerebrospinal fluid culture in the first 3 days of life. Sepsis rates were compared between neonates born after preterm PROM and those with either recent or prolonged preterm PROM. For this analysis, all surviving neonates and those nonsurviving neonates with proven sepsis were included.
Statistical analysis was performed using SAS 9.1. Categorical variables were analyzed using χ2 tests. In comparing neonatal mortality and sepsis rates between the group with no preterm PROM and the groups with recent preterm PROM and prolonged preterm PROM, χ2 tests were used for each combination of comparisons. Each group was subdivided by 2-week gestational age at delivery intervals. Multivariable logistic regression was used to account for possible confounding factors. Risk adjustment variables included maternal age, maternal race, prenatal care, maternal hypertensive disease, delivery mode, multiple gestations, intrauterine infection, antenatal steroid use, birth weight, gestational age, neonatal sex, and early neonatal sepsis. Odds ratios (ORs) for mortality between the groups were calculated. An α level of less than 0.05 was used as the cut-off for statistical significance. When using the χ2 test for comparing the three groups of no preterm PROM, recent preterm PROM, and prolonged preterm PROM, a Bonferroni correction was used with α level of less than 0.017 as the cut-off for statistical significance. The study was approved by the Stanford Institutional Review Board.
All 17,501 non-anomalous neonates born between 24 0/7 and 34 0/7 weeks of gestation from 127 California Perinatal Quality Care Collaborative participating neonatal intensive care units between 2005 and 2007 were included in the analysis. The majority (94.6%) of women received prenatal care, defined as any obstetrical care before the admission during which birth occurred, and 71% received antenatal corticosteroids before delivery (Table 1). Women in the preterm PROM group were more likely to be African American, have an intrauterine infection, receive antenatal corticosteroids, and deliver vaginally (Table 1).
Preterm PROM rates decreased with advancing gestational age, complicating 45% of all preterm deliveries at 24 to 26 weeks of gestation but only 27% at 32 to 34 weeks of gestation (Table 2). The overall mortality rate was higher in neonates born in the setting of preterm PROM (7.4% compared with 5.9%; P=.001); however, when analyzed by 2-week gestational age groups, there were no differences between those neonates born with and without membrane rupture before delivery (Table 2). Neonatal mortality rates also decreased with increasing gestational age, complicating less than 5% of all pregnancies beyond 28 weeks of gestation. Among neonates born between 24 and 26 weeks of gestation, the presence of prolonged preterm PROM was associated with decreased neonatal mortality when compared with either no preterm PROM or recent preterm PROM (18% compared with 29% compared with 31%; P<.001; Table 3).
There was no difference in overall mortality between neonates when comparing all neonates born with preterm PROM and those without preterm PROM before delivery when adjusting for possible confounding factors, including maternal age, race, prenatal care, maternal hypertension, intrauterine infection, birth weight, fetal sex, multiple gestations, mode of delivery, and antenatal corticosteroid administration (Table 4). However, there were differences in mortality for select gestational ages when considering the duration of preterm PROM. At 24 to 26 weeks of gestation, the prolonged preterm PROM group had improved survival, with ORs for neonatal mortality of 1.67 (confidence interval [CI] 1.18–2.29) and 1.79 (CI 1.25–2.56) when comparing no preterm PROM and recent preterm PROM with prolonged preterm PROM (Table 4). However, at 28 to 30 weeks of gestation, preterm PROM of less than 18 hours conferred a protective effect when compared with preterm PROM more than 18 hours (OR 0.44; CI 0.22–0.88).
Early neonatal sepsis rates decreased with advancing gestational ages (Table 5) and complicated less than 4% of all pregnancies beyond 28 weeks of gestation. Early sepsis was significantly higher in the prolonged preterm PROM group when compared with the no preterm PROM group at gestational ages 26 to 34 weeks. This difference in sepsis was not seen when comparing neonates born with recent and prolonged preterm PROM and no difference in neonatal sepsis was seen between the three groups at 24 to 26 weeks of gestation.
The objective of our study was to estimate the effect of preterm PROM on neonatal mortality. Including more than 17,000 neonatal outcomes in our cohort allowed us to stratify mortality based on 2-week gestational age intervals, an important factor when considering underlying subclinical infection and inflammation rates at various gestational ages. Previous studies analyzing more limited cohorts combined data across larger gestational age intervals10,12 and likely incorporated different underlying subclinical infections and inflammatory rates in their analysis. Moreover, applying a strict definition for neonatal sepsis allowed us to study the potential effect of this variable.
In this cohort, and contrary to what one might expect, neonatal mortality rates were not increased by the presence of ruptured membranes before delivery at any gestational age. Also, the effects of a latency period of more than 18 hours were not consistent across gestational ages, with recent preterm PROM having an increased mortality OR at 24 to 26 weeks but a decreased OR at 28 to 30 weeks when compared with prolonged preterm PROM (Table 3). We speculate that the reduced mortality rates associated with prolonged preterm PROM at the earliest gestational ages may be attributable to longer exposure to antenatal corticosteroids and antibiotics. Although antenatal corticosteroid administration rates were recorded by the California Perinatal Quality Care Collaborative, the number of doses administered, and exposure interval were not quantified. Moreover, prenatal maternal antibiotic data are not routinely recorded by the California Perinatal Quality Care Collaborative. It is also possible that preterm PROM itself is a confounder. Women with preterm PROM access the health care system at a higher rate than women with more occult signs of preterm labor. Unfortunately, the retrospective nature of the dataset and the data contained therein do not allow for a definitive assessment of this possibility. Finally, we were unable to assess the potential effect of group B streptococci status on neonatal mortality because culture results are not universally obtained by clinicians in this setting. Previous studies have shown that adverse neonatal outcomes in the presence of preterm PROM can differ based on both group B streptococci status and antenatal antibiotic administration.13,14
In our cohort, sepsis rates were altered by preterm PROM status. Overall, early neonatal sepsis rates decreased with increasing gestational ages. A latency period more than 18 hours was not associated with increased sepsis when compared with a latency period less than 18 hours. This finding is consistent with other studies9,15 and supports the fact that subclinical infection or inflammation or both often exist at the time patients present with preterm PROM. However, neonatal sepsis rates were higher in the prolonged preterm PROM group in every gestational age except 24 to 26 weeks when compared with the no preterm PROM group.
Our study is not without limitations. The cohort included both tertiary and community hospitals, and differences in practice were not accounted for in our analysis. However, including data from a wide range of settings allowed us to study a more generalizable patient population. Certain inherent biases of a large database such as the California Perinatal Quality Care Collaborative may have also limited our study. Although we controlled for maternal age, maternal race, birth weight, intrapartum infection, maternal hypertension, mode of delivery, corticosteroid administration, and prenatal care, and excluded neonates with anomalies, it is possible that other factors not routinely recorded by the California Perinatal Quality Care Collaborative may have played a role in neonatal mortality. For example, we do not have data on the indication for delivery in those pregnancies not complicated by preterm PROM and thus cannot comment on how this may have affected neonatal survival. Also, we defined preterm PROM for our query as any “rupture of membranes” before delivery. Currently, the California Perinatal Quality Care Collaborative does not record whether this rupture occurs before the onset of labor and thus there may be variation in the documentation of this outcome by the 127 individual institutions.
The 18-hour definition of latency was chosen for our study because this outcome is recorded as a dichotomous variable by the California Perinatal Quality Care Collaborative. The lack of difference in neonatal mortality between the recent and prolonged preterm PROM group at most gestational ages in our cohort is consistent with other studies. A recent Cochrane review including seven preterm PROM studies evaluating a total of 353 women in an “early birth group” and 339 women in an “expectant management group” failed to show a difference in perinatal mortality.9 Also, in a study by Manuck et al15 of 306 pregnancies complicated by preterm PROM between 22 and 34 weeks of gestation, neonatal outcomes did not worsen with longer latency after controlling for gestational age at delivery. In our cohort, it is unclear whether a longer latency period would have yielded similar results; however, our study's primary objective was not to assess the effect of latency on neonatal outcomes, and larger prospective studies are needed to adequately study this variable as a continuum.
In summary, our data suggest that neonatal mortality is not affected by the presence or absence of membrane rupture before delivery. We hope that clinicians incorporate these data as they counsel patients in this setting. Future prospective studies are needed to confirm this finding and to better assess the neonatal effects of latency across different gestational ages.
1. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75–84.
2. Mercer BM. Preterm premature rupture of the membranes. Obstet Gynecol 2003;101:178–93.
3. Gonçalves LF, Chaiworapongsa T, Romero R. Intrauterine infection and prematurity. Ment Retard Dev Disabil Res Rev 2002;8:3–13.
4. Romero R, Quintero R, Oyarzun E, Wu YK, Sabo V, Mazor M, et al. Intraamniotic infection and the onset of labor in preterm premature rupture of the membranes. Am J Obstet Gynecol 1988;159:661–6.
5. Romero R, Espinoza J, Kusanovic JP, Gotsch F, Hassan S, Erez O, et al. The preterm parturition syndrome. BJOG 2006;113(suppl 3):17–42.
6. Shim SS, Romero R, Hong JS, Park CW, Jun JK, Kim BI, et al. Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 2004;191:1339–45.
7. American College of Obstetricians and Gynecologists (ACOG). Practice Bulletin No. 80: premature rupture of membranes. Clinical management guidelines for obstetrician-gynecologists. Obstet Gynecol 2007;109:1007–19.
8. Ramsey PS, Nuthalapaty FS, Lu G, Ramin S, Nuthalapaty ES, Ramin KD. Contemporary management of preterm premature rupture of membranes (preterm premature rupture of membranes): a survey of maternal-fetal medicine providers. Am J Obstet Gynecol 2004;191:1497–502.
9. Buchanan SL, Crowther CA, Levett KM, Middleton P, Morris J. Planned early birth versus expectant management for women with preterm prelabour rupture of membranes prior to 37 weeks' gestation for improving pregnancy outcome. The Cochrane Database of Systematic Reviews 2010, Issue 3. Art. No.: CD004735. DOI: 10.1002/14651858.CD004735.pub3.
10. Pasquier JC, Picaud JC, Rabilloud M, Claris O, Ecochard R, Moret S, et al. Neonatal outcomes after elective delivery management of preterm premature rupture of the membranes before 34 weeks' gestation (DOMINOS study). Eur J Obstet Gynecol Reprod Biol 2009;143:18–23.
11. Pasquier JC, Bujold E, Rabilloud M, Picaud JC, Ecochard R, Claris O, et al. Effect of latency period after premature rupture of membranes on 2 years infant mortality (DOMINOS study). Eur J Obstet Gynecol Reprod Biol 2007;135:21–7.
12. Ramsey PS, Lieman JM, Brumfield CG, Carlo W. Chorioamnionitis increases neonatal morbidity in pregnancies complicated by preterm premature rupture of membranes. Am J Obstet Gynecol 2005;192:1162–6.
13. Mercer BM, Miodovnik M, Thurnau GR, Goldenberg RL, Das AF, Ramsey RD, et al. Antibiotic therapy for reduction of infant morbidity after preterm premature rupture of the membranes. A randomized controlled trial. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. JAMA 1997;278:989–95.
14. Hutzal CE, Boyle EM, Kenyon SL, Nash JV, Winsor S, Taylor DJ, et al. Use of antibiotics for the treatment of preterm parturition and prevention of neonatal morbidity: a metaanalysis. Am J Obstet Gynecol 2008;199:620 e1–8.
© 2010 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
15. Manuck TA, Maclean CC, Silver RM, Varner MW. Preterm premature rupture of membranes: does the duration of latency influence perinatal outcomes? Am J Obstet Gynecol 2009;201:414 e1–6.