OBJECTIVE: To examine maternal and neonatal outcomes in expectant management of spontaneous preterm premature rupture of membranes (PPROM) before 24 weeks.
STUDY DESIGN: Patients presenting with spontaneous PPROM from 14 to 23 completed weeks' gestation between January 1, 1995 and December 31, 1999 were reviewed. A total of 108 pregnancies were evaluated; 57 patients elected expectant management.
RESULTS: Median latency from rupture of membranes (ROM) to delivery was 6 days; the overall survival rate was 26.3%. In ROM <20 weeks, a twin and a triplet pregnancy with loss of the presenting fetuses yielded the only survivors. In patients with ROM from 20 to 21 and 22 to 23 weeks, survival rates were 2/16 (12.5%) and 11/20 (55.0%), respectively. In all, 18/57 (31.6%) of patients developed chorioamnionitis. There was no maternal sepsis or death. There were three cases of pulmonary hypoplasia, all in patients with ROM <20 weeks.
CONCLUSIONS: Neonatal survival in spontaneous PPROM before 20 weeks is rare, irrespective of latency from ROM to delivery. When PPROM occurs from 20 to 24 weeks, survival improves with increasing gestational age at ROM and at delivery.
Midtrimester preterm premature rupture of membranes (PPROM) confronts both clinicians and patients with a difficult dilemma. The aim of expectant management is to extend latency and to maximize gestational age at delivery. However, prolonged rupture of membranes (ROM) is associated with an increased risk of maternal or fetal infectious morbidity, pulmonary hypoplasia, skeletal deformities,1 and Potter's facies.2
Advances in obstetric and neonatal care continue to improve outcomes for extremely premature infants. In the current medical context, the accepted limit of viability is 24 weeks of gestation. Previous studies of early PPROM, performed between 1979 and 2001,3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 included patients with PPROM up to 30 weeks of gestation and none evaluated multiple gestations, which in recent years have had a great impact on the preterm delivery rate. This study examines the maternal and neonatal outcomes for expectant management of spontaneous PPROM from 14 to 24 weeks. We specifically evaluate how maternal and neonatal outcomes vary according to gestational age at membrane rupture.
MATERIALS AND METHODS
The study is a retrospective chart review. The study was approved by the Institutional Review Board of Brigham and Women's Hospital. The study group included all patients that presented to Brigham and Women's Hospital with premature rupture of membranes at 14 weeks to 23 completed weeks' gestation between January 1, 1995 and December 31, 1999. Gestational age was established by menstrual history; if the estimated date of delivery (EDD) calculated from the last menstrual period differed by more than 8% from the EDD calculated by the first ultrasound in the pregnancy, the ultrasound date was used.19 For the purpose of analysis, we subdivided the groups into patients with ROM at 14 to 19 completed weeks of gestation, 20 to 21 completed weeks, and 22 to 23 completed weeks.
The diagnosis of rupture of membranes was made by two or more of the following: pooling of amniotic fluid on sterile speculum examination, positive nitrazine paper test, or positive fern test; or by provider documentation.
Subjects were excluded if ROM occurred within 30 days of amniocentesis, since women with amniocentesis-related rupture have been previously reported to be a discrete group with longer latency periods and better survival rates than the rest of the population.20 A woman was also excluded if she chose pregnancy termination (38/96; 39.6%), was lost-to-follow-up, or was not eligible for expectant management. Specifically, women were not included if at the time of presentation they had: a fetal demise; chorioamnionitis, defined by temperature of ≥100.4°F, uterine tenderness or irritability, purulent vaginal discharge, or fetal tachycardia; preterm labor or cervical incompetence, defined by cervical dilation of more than 1 cm with or without regular uterine contractions, respectively. Women who delivered in ≤24 hours from time of rupture of membranes were considered to have insufficient time to elect expectant management and were excluded.
Multiple gestations were included; each pregnancy was treated as a unit and the most favorable neonatal outcome was used in the survival analysis. The few discrepancies in outcome among siblings are reported. All multiple gestations were di- or tri-amnionic and di- or tri-chorionic.
Clinical management was determined by the patients' attending physician. These physicians included academic perinatologists and general obstetricians in private practices and health care maintenance organizations. Data were collected on the use of antibiotics, corticosteroids and tocolytic agents.
Maternal and neonatal outcomes were recorded from the medical records. The major obstetric outcomes were maternal morbidity and mortality, interval to delivery, and indication for and mode of delivery. Fetal complications were recorded, including pulmonary hypoplasia and skeletal abnormalities. Pulmonary hypoplasia was diagnosed by postmortem examination. Serious neonatal complications recorded included: chronic pulmonary disease, retinopathy, intraventricular hemorrhage, and necrotizing enterocolitis. Neonates were followed until time of discharge from the hospital.
Statistical analyses included χ2 analysis and Wilcoxon rank sum test. Analyses were performed using the SAS Statistical Software. A p-value of <0.05 was considered statistically significant.
Between January 1, 1995 and December 31, 1999, 95 patients presented with PPROM between 14 weeks and 23 completed weeks and met the inclusion criteria. In all, 57 patients elected expectant management; 38 patients chose to terminate the pregnancy and were excluded from the analysis.
The total number of subjects was 57. The mean maternal age was 31.8 years. A total of 20 patients (35.1%) were primigravidas. Among the multigravidas (n=37), three women (8.1%) had a history of previous PPROM; ROM during the previous pregnancy occurred at less than 24 weeks in one case (2.7%), and between 24 and 36 completed weeks in the other two cases (5.4%). In total, 21 pregnancies (36.8%) were conceived through assisted reproductive technology. The pregnancies included 42 (73.7%) singletons, eight (14.0%) twins, and seven (12.3%) triplets. Seven (12.3%) patients had a cerclage. The mean gestational age at diagnosis of ROM was 20.3 weeks. The median latency from gestational age at membrane rupture to delivery was 6 days (range 1–161 days).
Antenatal interventions were as follows: 12.3% of patients received antepartum antibiotics; 7.0% received tocolytics; 33.3% were given corticosteroids. Group B streptococcus cultures were positive in 19% of subjects; no significant effect was seen on neonatal survival (p=0.4). Indications for delivery are listed in Table 1. The most common reasons for delivery were: spontaneous labor, 29 patients (50.9%); and chorioamnionitis, 18 patients (31.6%). The mode of delivery was cesarean section for 11 (19.3%) of the subjects.
Our study population had few maternal complications. There were no cases of maternal sepsis or death. In all, 17.5% of patients had a postpartum fever and 10.5% had a retained placenta and required dilation and curettage.
Neonatal outcome, including rates of intrauterine fetal demise, stillbirth, death prior to hospital discharge, and survival are listed in Table 2. Fewer than half (47.4%) of the pregnancies resulted in a liveborn infant. Overall, the proportion of pregnancies in which at least one infant survived to discharge was 26.3%. Discrepancies in outcome for multiple pregnancies are noted in Table 2.
The proportion of pregnancies with a surviving infant for singleton and multiple gestations was 23.8 and 33.3%, respectively. The multiple gestations with a surviving infant included a triplet pregnancy with rupture at 14.0 weeks, delivery of the presenting fetus at 15.9 weeks, and delivery of two healthy infants at 37.0 weeks; and a twin pregnancy with rupture at 15.7 weeks, delivery of the presenting fetus at 18.4 weeks, and delivery of an infant at 29.9 weeks with no chronic complications. If these two pregnancies are excluded, the survival rate for multiple gestations is 25%, similar to that for singletons.
In the 21 subjects with PPROM at less than 20 weeks, median latency was 8 days; the range was 1 to 161 days. Six patients delivered after 24 weeks. Of these patients, one had an intrauterine fetal demise and three delivered infants with pulmonary hypoplasia who died shortly after birth. The only two pregnancies in this category, which resulted in surviving infants, were the unusual twin and triplet gestations described above.
In the group for whom ROM occurred at 20 to 21 completed weeks, median latency was 4.5 days; the range was 2 to 106 days. In total, two (12.5%) of the pregnancies resulted in a surviving infant. In the group for whom ROM occurred at 22 to 23 completed weeks, median latency was 12.0 days; the range was 1 to 118 days. In all, 11 (55.0%) of the pregnancies resulted in an infant surviving to discharge from the hospital. There was no significant difference in latency between the two groups (p=0.21).
The relationship between survival and gestational age at ROM and at delivery is illustrated in Table 3 and Figure 1. As noted above, only two pregnancies with ROM at less than 20 weeks resulted in surviving infants. Among women with ROM after 20 weeks, only one of the 22 women delivering at less than 24 weeks had a surviving infant; in infants delivered at greater than 24 weeks survival was 70%; and in infants delivered at greater than 28 weeks survival was 100%.
Neonatal complications included three cases of neonatal sepsis, all of which were in the 22-to-23-week group. Additionally, there were three cases of pulmonary hypoplasia, all of which occurred in the <20-week ROM group, and two infants with skeletal abnormalities, one in the <20-week group, and the other in the 22-to-23-week group.
A total of 21 infants from 57 pregnancies, including siblings in multiple gestations, survived to time of discharge from the hospital. Among the survivors, two (9.5%) were discharged to chronic care facilities and 10 (47.6%) others had chronic prematurity-associated complications.
Our data indicate that for women with spontaneous PPROM before 24 weeks gestation, a survival may be rare prior to 20 weeks of gestation. In pregnancies in which ROM occurred before 20 weeks, the prognosis was poor. In this group, later age at delivery did not confer a survival advantage. When ROM occurred between 20 and 24 weeks, the prognosis improved as gestational age at ROM increased. For these patients, later gestational age at ROM correlated with advanced age at delivery and improved neonatal survival rates.
The overall survival rate was 15/57 (26.3%). Two pregnancies in which membrane rupture occurred before 20 weeks resulted in surviving infants. Both of these were multiple gestations with early loss of the presenting fetus, which is an unusual event.21 Four other pregnancies continued until delivery between 24 and 34 weeks; among these there were one intrauterine fetal demise and three infants with pulmonary hypoplasia. This is an earlier upper limit of vulnerability for pulmonary hypoplasia than reported by Nimrod et al.,22 who found that risk of this complication was more common in infants with rupture prior to 26 weeks of gestation and a duration of rupture of greater than 5 weeks.
PPROM after 20 weeks showed a stepwise improvement in survival, with 2/16 (12.5%) between 20 and 22 weeks and 11/20 (55.0%) between 22 and 24 weeks. Increased gestational age at ROM was associated with advanced age at delivery. However, 60% of the surviving infants had serious complications, such as chronic pulmonary disease or retinopathy.
Recent advances in obstetric care have included the use of antepartum antibiotics, which may prolonged latency to delivery in pregnancies with PPROM.23 Concurrent developments in perinatal care have included administration of antenatal corticosteroids and use of surfactant in premature infants, which have contributed to decreasing the gestational age limit of viability and improving outcomes in premature infants. Since the mid-1980 s, there has been a concomitant trend toward consideration of conservative management in early PPROM patients. Previous studies are summarized in Table 4.
Our study had a small study population, as did previous reports. This is due to the infrequency of PPROM in the second or early third trimester, as well the even smaller number of subjects who met inclusion criteria.
The accepted limit of viability during our study period was 24 weeks, we therefore chose this as the upper limit of gestational age at ROM. In previous reports, including recent ones, the maximum gestational age at ROM examined ranged from 25 to 37 weeks. Reported survival rates ranged from 22 to 71%. Our survival rate of 26.3%, which was lower than other studies, corresponds to our use of this earlier upper limit at ROM. The effort to decrease the limit of viability continues. Recent data reported by McElrath et al.24 suggest that one-third of infants delivered at 23 weeks will survive, but that all suffer from significant complications of prematurity.
Assisted reproductive technology has increased the rate of multiple gestation. Twin gestations have been found to have an increased incidence of PPROM.25 We therefore considered it important to include multiple gestations, which have been excluded by most previous studies. In our population, the proportion of pregnancies with survival of an infant to discharge was similar for singleton and multiple pregnancies.
Antibiotics and tocolytics were not given consistently in our study. In a randomized study, Mercer has reported an antibiotic protocol for patients with ROM from 24 to 32 weeks which increased latency from time of ROM to delivery by 3.2 days.21 Fortunato et al. used tocolysis and antibiotics in patients in patients with PPROM from 15 to 27 weeks; the mean latency in that study was 22 days,26 which exceeded other studies. These results suggest that the use of an antibiotic protocol in PPROM before 24 weeks merits further study.
The introduction of widespread use of corticosteroids and surfactant in the early 1990 s has improved survival in premature infants. These developments may increase survival and, therefore, distinguish our study and concurrent studies from those performed earlier. Steroid use is noted in at least some of the study participants in all studies except Beydoun.27 Use of surfactant is documented only in the study by Farooqi.28 In our study population, steriods were given to all patients upon reaching 24 weeks' gestation.
Our study population was drawn from an academic tertiary care hospital with 9000 deliveries per year during the study period. We were able to identify all patients who were evaluated in the hospital for PPROM. We excluded patients who delivered within 24 hours of ROM, and therefore could not elect expectant management. Our population did not include patients who were referred directly for pregnancy termination or were evaluated solely in a clinician's office and were managed expectantly at home. The latter group would most likely include patients with fetal loss prior to 24 weeks. Therefore, if these pregnancies had been captured, our overall survival rate would likely be decreased.
The occurrence of spontaneous PPROM prior to viability presents patients and clinicians with a formidable decision about whether to choose expectant management or terminate the pregnancy. While there are maternal infectious risks incurred by increased latency to delivery, in our population we found these to be infrequent and without serious sequelae. The decision therefore turns on the fetal outcome. Our data indicate that 20 weeks of gestation at ROM is a turning point in neonatal survival. After 20 weeks, the chance of delivery at a later gestation age enhances survival and decreases long-term sequelae of prematurity. It is our hope that the information provided in this study will further aid clinicians to counsel patients who find themselves in this unfortunate situation.