Materials and Methods

This observational cohort secondary analysis of the randomized controlled trial of magnesium sulfate for the prevention of cerebral palsy, conducted through the Eunice Kennedy Shriver National Institute of Child Health and Development’s Maternal-Fetal Medicine Units Network, was approved by the Institutional Review Board at Columbia University Medical Center (New York, NY).

The parent trial enrolled women in 20 centers across the United States from 1997 to 2004 and sought to evaluate whether antenatal magnesium sulfate administration decreased the rate of cerebral palsy or death. Of a total of 2241 women at high risk for preterm delivery between 24 weeks 0 days and 31 weeks 6 days randomized to magnesium sulfate or placebo, 86.7% of women enrolled had PPROM. Women were eligible if they were at high risk for spontaneous preterm delivery because of the rupture of membranes or because of advanced preterm labor with dilation of 4–8 cm and intact membranes; women were excluded if delivery was anticipated within 2 hours.

Antibiotic administration and criteria for delivery were not standardized as part of the parent trial; however, in accordance with the current recommended guidelines, antibiotics were routinely administered and expectant management was pursued until 34 weeks unless a medical indication for delivery arose. Similarly, antenatal corticosteroid administration for fetal lung maturity occurred routinely as part of the current recommended guidelines. Multiple courses of antenatal corticosteroids was defined as > 2 courses.

The current study included nonanomalous, live singleton pregnancies with PPROM. The diagnosis of PPROM was based on at least 2 of the following 3 criteria being met: (1) pooling of amniotic fluid in the vaginal vault, (2) a positive nitrazine test, and (3) the ferning of vaginal fluid. PPROM could also be diagnosed if there was indigo carmine pooling in the vagina after amnioinfusion or if visible leakage of amniotic fluid from the cervix was noted. Patients with missing outcome data were excluded.

The exposure of interest was prolonged latency, which was defined as delivery ≥ 4 weeks after diagnosis of PPROM. For analyses, latency was categorized by number of weeks: 0 (0–6 days), 1 (7–13 days), 2 (14–20 days), 3 (21–27 days), and ≥ 4 weeks (≥ 28 days).

The primary outcome evaluated was neonatal sepsis, defined in the parent trial as proven sepsis in which blood, cerebrospinal fluid, and/or urine cultures were positive and infection was suspected on physical examination, or, in the absence of positive cultures, there was evidence of hemodynamic collapse and X-ray findings supporting infection in a neonate with a clinical presentation consistent with sepsis. The diagnosis of neonatal sepsis could occur at any point during the initial hospitalization and was not limited to a specific period of time after birth.

Patient demographic variables and other characteristics were compared using the χ ² test for categorical variables and a Student t test for continuous variables. A Wilcoxon rank sum test was used to compare median differences. We then fit a logistic regression model to control for the following possible confounders: gestational age (GA) at rupture of membranes, maternal age, parity, race, exposure to magnesium sulfate, body mass index (BMI), prenatal care, drug use, tobacco use, and administration of multiple courses of antenatal corticosteroids. Patients were considered to have used recreational drugs during the pregnancy if they tested positive on urine toxicology studies or reported using heroin, cocaine, marijuana, or other illicit drugs during their pregnancy. Significance was set at a value of P < .05. All analyses were performed using SAS 9.4 (SAS Institute, Cary, NC).

Our sample size was fixed by the number of patients enrolled in the initial randomized controlled trial and by our exclusions. A power analysis based on the fixed sample size revealed a minimum detectable effect size for PPROM latency ≥ 4 weeks of a < 0.6 or > 1.5 risk ratio for the primary outcome based on a power of 80% and a type I error of 5%.

Results

Of the 2444 infants included in the parent study, 2088 had PPROM; of these, 326 were excluded for multiple gestation, 100 for missing data, 58 for major congenital anomalies, and 8 for stillbirth, leaving 1596 patients included in the analysis. A total of 1390 patients had a < 4 week interval; among these, 762 patients had a latency interval of 0 weeks, 354 had PPROM for 1 week, 184 for 2 weeks, 90 for 3 weeks, and 206 patients had an interval of ≥ 4 weeks.

All patients included in this analysis were enrolled in the study within 48 hours of PPROM. Of the patients included in the analysis, 87 had PPROM at < 24 weeks, 611 had PPROM at 24–28 weeks, 386 had PPROM at 28–30 weeks, and 441 women had PPROM at 31–32 weeks.

Patients in the < 4 week latency group had an earlier mean GA at rupture of membranes than the ≥ 4 weeks latency group ( Table 1 ); additionally, these women were more likely to be white and receive multiple courses of steroids. Women in the < 4 week latency group were more likely to be African American and use drugs. At delivery, patients with latency ≥ 4 weeks delivered at higher mean gestational ages and higher birthweights, whereas the rates of chorioamnionitis and cesarean delivery were similar ( Table 2 ).

In the unadjusted analysis for the primary outcome, patients in the PPROM ≥ 4 weeks group were significantly less likely to develop neonatal sepsis (6.8% vs 17.2%, risk ratio, 0.40, 95% confidence interval [CI], 0.24–0.66). Figure 1 plots the proportion of neonatal sepsis by weeks of latency among all infants that developed sepsis and demonstrates that neonatal sepsis is more likely to occur among the women with shortest latency.

Proportion of neonatal sepsis by latency period

Prolonged PPROM latency was associated with a decreased risk of neonatal sepsis. Figure plots the proportion of neonatal sepsis by weeks of latency among all infants that developed sepsis. Among all infants that developed sepsis, 133 cases (52.6%) occurred in the shortest PPROM latency group, 66 cases (26.1%) occurred in the 1 week latency group, 25 cases (9.9%) occurred in the 2 week latency group, 15 cases (5.9%) in the 3 week latency group, 5 (2%) in the 4 week latency group, 3 (1.2%) in the 5 and 6 week latency groups, 2 cases (0.8%) in the 7 week latency group, and 1 case (0.4%) in the ≥ 8 week latency group. A χ ² test was used to compare all groups with the shortest latency group ( P < .01) for all comparisons.

PPROM , preterm premature rupture of membranes.

Drassinower et al. Prolonged PPROM and the risk of neonatal sepsis. Am J Obstet Gynecol 2016 .

The median latency time in patients with neonatal sepsis was 6.5 days (interquartile range [IQR], 3.4–12.1) compared with 7.9 days (IQR, 3.4–17.2) in patients without neonatal sepsis ( P < .01). Furthermore, the risk for sepsis decreased with each increased week of latency.

Figure 2 plots the rate of neonatal sepsis by week of latency and demonstrates that this rate is stable in the first 3 weeks and then decreases in the PPROM ≥ 4 weeks group ( P < .01). Logistic regression was performed to control for possible confounders. The model included covariates known or suspected to be associated with the primary outcome and covariates found to significantly differ in the univariate analysis ( Table 3 ). In the adjusted model, PPROM ≥ 4 weeks retained significance and was associated with a decreased risk of neonatal sepsis (adjusted odds ratio [aOR], 0.21, 95% CI, 0.10–0.41) ( Table 4 ). The logistic regression model also demonstrated that later GA at rupture of membranes was associated with decreased risk neonatal sepsis, whereas Hispanic race was associated with increased risk.

Rate of neonatal sepsis by latency period

The rate of neonatal sepsis was stable weeks 0–3 of PPROM latency and decreased in the latency ≥ 4 weeks group. This figure plots the rate of neonatal sepsis by week of latency, which was 17.5% (133 cases) in the shortest PPROM latency group, 18.6% (66 cases) in the 1 week latency group, 13.6% (25 cases) in the 2 week latency group, 16.7% (15 cases) in the 3 week latency group, and 6.8% (14 cases) in the ≥ 4 week latency group. A χ ² test was used to compare all groups with the shortest latency group; the P value was not significant for PPROM latency for weeks 1–3 compared with 0 weeks but was P < .01 when the ≥ 4 week latency group was compared with the shortest latency group.

PPROM , preterm premature rupture of membranes.

Drassinower et al. Prolonged PPROM and the risk of neonatal sepsis. Am J Obstet Gynecol 2016 .

Table 3

Univariate analysis of demographic factors associated with neonatal sepsis

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Midurethral slings: evidence-based medicine vs the medicolegal system
2. The relationship between variations in cesarean delivery and regional health care use in the United States
3. Cancer of the uterus and treatment of stress incontinence: a pilot study
4. Reply

Stay updated, free articles. Join our Telegram channel

Join