At term, the onset of labor occurs within 24 hours after membrane rupture in 80% to 90% of patients. Among patients with PROM prior to term, latent periods occur longer; latent periods of more than 24 hours occur in 60% to 80%, of more than 72 hours in about 15% to 25%, and of 7 days or more in 20% to 40%. There is an inverse relationship between gestational age and the proportion of patients with latent periods longer than 3 days. For pregnancies between 25 and 32 weeks, 33% had latent periods longer than 3 days, whereas for pregnancies of 33 to 34 and 35 to 36 weeks, the corresponding values were 16% and 4.5%, respectively.

The value of tocolytics in PPROM remains controversial. None of eight prospective trials showed a decrease in neonatal morbidity, whether tocolysis was used prophylactically (for all patients in admission regardless of uterine activity) or therapeutically (for patients who developed uterine contractions). Only three of the eight showed a prolongation of the latency period; an additional study showed a prolongation for gestation less than 28 weeks. In three of these studies steroids were used; in another three studies antibiotics were used.

A comparison of short-term versus long-term tocolysis in preterm PROM at 26 to 35 weeks showed an adverse effect of “long-term tocolysis.” Patients with preterm PROM at 26 to 35 weeks were randomized to receive either an intravenous _-mimetic drug for less than 48 hours versus until delivery. All patients received corticosteroids, and group B streptococci (GBS) and gonococci were treated. There was no significant difference in the latent period or in neonatal infection, but there was a significant increase in both chorioamnionitis and endometritis with long-term tocolysis. Accordingly, use of tocolytics in patients with preterm PROM remains controversial, but the bulk of the evidence shows no benefit. However, if tocolytics are used, such as during transfer to a tertiary care center or obtaining benefits of corticosteroids, the course of tocolytics should be limited to less than 48 hours.

The risks of PROM have generally been viewed as those of infection versus those of prematurity. The most common clinically evident complication among pregnancies with PROM before 37 weeks is respiratory distress syndrome (RDS), which, in general, is found in 10% to 40% of neonates. Bona fide neonatal sepsis is documented in less than 10%, and amnionitis (based on clinical criteria) occurs in approximately 3% to 30%. Subclinical infection, based on positive amniotic fluid culture or histologic inflammation of the cord or membranes, is seen much more often, in up to 80% at very early gestational ages with PPROM. Endometritis develops in up to 30%. Abruption after PROM is reported in 4.0% to 6.0% of cases, severalfold higher than the rate of 0.5% to 1.0% in the general population. Pulmonary hypoplasia is a serious fetal complication occurring in preterm PROM. Pulmonary hypoplasia is more common when there is very early preterm PROM, especially when this occurs in the presence of prolonged PROM and with severe oligohydramnios. There is nearly a 100% probability of lethal pulmonary hypoplasia when PROM occurs before 23 weeks and when there is severe oligohydramnios. With later gestational age at the onset of preterm PROM, the likelihood of pulmonary hypoplasia decreases. Notably with preterm PROM more than 24 to 26 weeks, even with oligohydramnios, pulmonary hypoplasia is rare. When preterm PROM occurs less than 25 weeks with severe oligohydramnios lasting more than 14 days, the likelihood of lethal pulmonary hypoplasia is estimated to be 80%. At the other extreme, when preterm PROM occurs at more than 25 weeks and when there is either no severe oligohydramnios or severe oligohydramnios for less than 5 days, then the predicted probability of lethal pulmonary hypoplasia is only 2%. These data provide important information for counseling patients with mid-trimester PROM.

The initial evaluation is likely to reveal amniotic fluid egressing from the vagina. The differential diagnosis of rupture of membranes includes loss of mucus plug, vaginal discharge associated with infection, and urinary loss. If the patient is not going to be delivered immediately, then a digital examination should be deferred as examination may introduce bacteria into the uterus and shorten the latent phase. A sterile speculum exam may demonstrate pooling of fluid in the posterior vaginal vault. Direct observation of fluid leaking from the cervical os confirms ruptured membranes. The normal pH of the vagina is between 4.0 and 4.7 in pregnancy, whereas the pH of the amniotic fluid is 7.1 to 7.3. Nitrazine paper changes to a dark blue from yellow with a pH above 6.5. Nitrazine paper to diagnose amniotic fluid in the vagina has an overall accuracy of approximately 93%, but false-positive results can result from blood, semen, alkaline urine, bacterial vaginosis, and trichomoniasis.

The diagnosis of PROM may also be confirmed by observing arborization or “ferning” of dried amniotic fluid on a slide. This method has an overall accuracy of diagnosis of PROM of approximately 96%. False positives occur with contamination by semen or cervical mucus. False negatives can result from a dry swab, contamination with blood at a 1:1 dilution, or not allowing sufficient time for the fluid to dry on the slide. Amniotic fluid arborization is unaffected by meconium at any concentration and is unaffected by pH alteration.

Ultrasound examination has been used widely, since oligohydramnios suggests PROM, but there have been no evaluations of its sensitivity and specificity. A multicenter clinical trial compared fetal fibronectin detection with standard tests for detection of rupture of membranes at term. Fetal fibronectin showed an excellent sensitivity (98.2%) but a low specificity, leading to speculation that fetal fibronectin in cervicovaginal secretions may be a marker for impending labor, even without frank rupture of the membranes. An immunoassay for placenta alpha microglobulin 1 (abundant in amniotic fluid, but barely detectable in normal cervicovaginal secretions) has been approved for detecting PROM. Initial reports have found a high diagnostic accuracy.

When the diagnosis of ruptured membranes is unclear by these tests, a transabdominal dye injection is sometimes performed. Indigo carmine blue (1 mL diluted in 9 mL of sterile normal saline solution) is injected into the amniotic fluid, and a sponge is placed into the vagina and inspected 30 minutes later for the dye. Methylene blue should not be used because of reported methemoglobinemia in the fetus. This test is invasive, and the accuracy of diagnosis is not established.

Determination of the fetal age and maturity status is useful in developing a treatment plan. The patient’s history and early milestones of the pregnancy should be used. Ultrasound examination of the fetus can be limited because of the decreased fluid surrounding the fetus, and some measurements, especially of the abdomen and head, may be altered with oligohydramnios after PROM. As part of the overall decision process as to delivery, assessment of fetal lung status may be incorporated, for example, if gestational age is 32 to 34 weeks or if there is uncertainty regarding gestational age because of possible growth restriction. Amniotic fluid may be collected by amniocentesis or by collection from the vaginal pool. Vaginal pool collection is less accurate for lecithin/sphingomyelin (L/S) ratio determination. Whereas phosphatidylglycerol (PG) production by vaginal bacteria has been described, there has been an excellent correlation between PG detection in amniotic fluid obtained vaginally and transabdominally.

In addition to documentation of ruptured membranes, the sterile speculum exam can evaluate the degree of cervical dilation and can exclude the possibility of a fetal extremity or umbilical cord prolapsing through the cervix. Endovaginal ultrasound may be used safely in patients with preterm PROM as it does not increase the risk of infection.

When the diagnosis of PROM is made, a rectovaginal culture should be taken for GBS unless the GBS status is already known. Appropriate antibiotics (usually i.v. penicillin G) for prevention of GBS infection should be given pending culture results (see Chapter 1).

All patients with preterm PROM should be evaluated for possible evidence of chorioamnionitis. Physical exam includes maternal or fetal tachycardia, uterine tenderness, and detection of a purulent, foul-smelling discharge. Temperature elevation is often a late sign of chorioamnionitis, especially in preterm PROM. In Table 16.1, we show the positive and negative predictive values of several tests for intrauterine
infection in PPROM. Most of these tests have modest performance characteristics, thus limiting their clinical utility. Amniocentesis may be performed to evaluate for an intrauterine infection, for example, if there are suggestive clinical signs of infection. Because of the high likelihood of subclinical infection and the association of intrauterine infection with cerebral palsy, there is a growing enthusiasm for early detection of subclinical infection. Accordingly, amniocentesis may become more widely used. Analyses of amniotic fluid for possible infection include Gram stain, glucose concentration, and culture. Gram stain does not identify colonization with genital mycoplasmas. A low amniotic fluid glucose predicts a positive amniotic fluid culture. When the glucose is greater than 20 mg per dL, the likelihood of a positive culture is less than 5%; when glucose is less than 5 mg per dL, the likelihood of a positive culture approaches 90%. Although not widely available, an elevated interleukin 6 (IL-6) in amniotic fluid may be the most sensitive predictor of intrauterine infection. A biophysical profile of 6 or less has been shown in several studies to correlate with intrauterine infection. Most newborns who are delivered after clinical chorioamnionitis do not show clinical infection, possibly because of common use of empiric antibiotic therapy.

Both invasive and noninvasive tests have been assessed. As shown in Table 16.1, none of these tests is ideal, particularly because of their modest positive predictive values.