Preterm premature rupture of membranes: diagnosis, evaluation and management strategies
Dr H. N. Simhan, 300 Halket Street, Pittsburgh, Pennsylvania 15213, USA.
Preterm premature rupture of the membranes (PPROM) is responsible for one-third of all preterm births and affects 120,000 pregnancies in the United States each year. Effective treatment relies on accurate diagnosis and is gestational age dependent. The diagnosis of PPROM is made by a combination of clinical suspicion, patient history and some simple tests. PPROM is associated with significant maternal and neonatal morbidity and mortality from infection, umbilical cord compression, placental abruption and preterm birth. Subclinical intrauterine infection has been implicated as a major aetiological factor in the pathogenesis and subsequent maternal and neonatal morbidity associated with PPROM. The frequency of positive cultures obtained by transabdominal amniocentesis at the time of presentation with PPROM in the absence of labour is 25–40%. The majority of amniotic fluid infection in the setting of PPROM does not produce the signs and symptoms traditionally used as diagnostic criteria for clinical chorioamnionitis. Any evidence of infection by amniocentesis should be considered carefully as an indication for delivery. Documentation of amniotic fluid infection in women who present with PPROM enables us to triage our therapeutic decision making rationally. In PPROM, the optimal interval for delivery occurs when the risks of immaturity are outweighed by the risks of pregnancy prolongation (infection, abruption and cord accident). Lung maturity assessment may be a useful guide when planning delivery in the 32- to 34-week interval. A gestational age approach to therapy is important and should be adjusted for each hospital's neonatal intensive care unit. Antenatal antibiotics and corticosteroid therapies have clear benefits and should be offered to all women without contraindications. During conservative management, women should be monitored closely for placental abruption, infection, labour and a non-reassuring fetal status. Women with PPROM after 32 weeks of gestation should be considered for delivery, and after 34 weeks the benefits of delivery clearly outweigh the risks.
Premature rupture of the fetal membranes (PROM) is defined as the rupture of the amniotic membranes with release of the amniotic fluid more than 1 hour prior to the onset of labour. PROM may be subdivided into term PROM (TPROM, i.e. PROM after 37 weeks of gestation) and preterm PROM (PPROM, i.e. PROM prior to 37 weeks of gestation). PPROM occurs in approximately 3% of pregnancies and is responsible for a third of all preterm births.1
DIAGNOSIS AND EVALUATION
The diagnosis of PPROM is made by clinical suspicion, patient history and simple testing. Patient history has an accuracy of 90% for the diagnosis of PPROM and should not be ignored.2 Numerous tests have been recommended for the evaluation of PPROM but two tests have withstood the test of time: nitrazine (Bristol Myers Squibb, Princeton, New Jersey) paper testing and ferning (also referred to as amniotic fluid crystallisation testing) of the vaginal pool. Freidman and McElin2 found that if a combination of patient history, nitrazine testing and ferning were used to evaluate a patient for PROM, the accuracy of at least two positive tests was 93.1%.
The ferning test should be performed on midvaginal or posterior fornix fluid. Contamination with cervical mucus can cause a false-positive result. The slide should be allowed to dry in air for a minimum of 10 minutes. The false-negative rate increases with less time or with flame drying. Ferning is unaffected by meconium at any concentration or changes in pH. Small amounts of blood were not found to affect ferning but blood and amniotic fluid mixed in equal amounts resulted in no ferning.3 Amniotic fluid specimens were found to be fern and nitrazine positive, up to two weeks after amniocentesis. The nitrazine test was found to be more susceptible to alteration by contamination. The presence of bacterial vaginosis (BV), cervicitis, semen, alkaline urine, blood, soap and antiseptic solutions was found to alter the nitrazine test, mainly resulting in a false positive.3,4
During diagnosis and evaluation of PPROM, a digital cervical examination should be avoided. Lewis et al.5 compared the digital cervical examination with the sterile speculum examination and found that latency was shortened significantly by cervical examination at any gestational age (2.1 [4.0] vs 11.3 [13.4] days, P < 0.0001). Alexander et al.6 evaluated data from the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units network (NICHD MFMU) randomised controlled trial on antibiotics after PPROM, and found that one or two cervical examinations in patients with PPROM, between 24 and 32 weeks, was associated with a shorter latency (3 vs 5 days, P < 0.009) but did not worsen either maternal or neonatal outcome.
INFECTION AND PPROM
There are several different risk factors that may be associated with PPROM. These include placental abruption, excessive collagen degradation (or decreased membrane collagen content), localised membrane defects, excessive membrane stretch (uterine over distension), precocious programmed amniotic cell death and choriodecidual infection.1,2,7,8 Subclinical intrauterine infection has been implicated as a major aetiological factor in the pathogenesis and consequential maternal and neonatal morbidity in PPROM.
The rate of positive culture obtained by transabdominal amniocentesis at the time of presentation with PPROM, in the absence of labour is approximately 25–40%.9–17 Amniotic fluid infection is closely associated with in utero fetal inflammation.18 Serious newborn and infant adverse events are intimately tied with this fetal inflammatory process.19–22 Reliance on clinical criteria alone in order to diagnose amniotic fluid infection, while common practice, may not be particularly useful. Clinical chorioamnionitis is present on admission in 1–2% of women who present with PPROM and subsequently develops in 3–8% of women.14,23 The majority of cases of amniotic fluid infection in the setting of PPROM do not produce the signs and symptoms traditionally used as diagnostic criteria for clinical chorioamnionitis.
The diagnostic and therapeutic goals of the management of the patient with PPROM focus on a reduction in neonatal morbidity. As a result, hospitalised bed rest/expectant management, a broad spectrum of antibiotics, antenatal corticosteroids and serial maternal/fetal surveillance are currently employed. These therapies are designed to prolong pregnancy and reduce neonatal morbidity.
The current understanding of the contribution of infection and inflammation to the morbidity associated with preterm birth strengthens the notion that intrauterine infection (clinical or subclinical) is an indication for delivery (and, by extension, a contraindication to expectant management or pregnancy prolongation). Documentation of amniotic fluid infection in the patient who presents with PPROM will influence the choice of management.
It is important to define subclinical amniotic fluid infection as a positive amniotic fluid culture. A culture result may take as long as 48 hours to return as positive or negative. In an effort to provide more rapid information with respect to amniotic fluid infection status, several short term tests have been evaluated. The tests that are currently available for clinical use are Gram stain, white blood cell count, leucocyte esterase and glucose concentration. Romero et al.24 initially evaluated the leucocyte esterase assay (LEA) and Gram stain. For the prediction of a positive amniotic fluid culture (prevalence of positive cultures = 33.9%), the LEA had a sensitivity of 19%, a specificity of 86.7%, a positive predictive value (PPV) of 42.3% and a negative predictive value (NPV) of 67.6%. The Gram stain had a sensitivity of 36.2%, specificity of 94.7%, PPV of 77.8% and NPV of 74.3%. When both tests were combined, a significant increase in sensitivity to 50% was observed. This was associated with a drop in specificity to 81.4%.24 In a subsequent study, Romero et al.17 evaluated amniotic fluid glucose, white blood cell count, interleukin-6 and Gram stain. The combination of the clinically available tests (glucose <10 mg/dL, white blood cell ≤30 cells/mm3 and Gram stain), in which if any one of the three tests is positive the ‘combination’ test is considered positive, provides a sensitivity of 76.2%, specificity of 60.3%, PPV of 61.0% and NPV of 80.4%.17 In addition to amniotic fluid culture, these three rapid tests and leucocyte esterase could also be obtained using amniotic fluid samples.
MANAGEMENT CONSIDERATIONS: TOCOLYSIS, ANTIBIOTICS AND CORTICOSTEROIDS
The perinatal complications of PPROM change with gestational age at rupture requiring a gestational age approach to treatment. There is little maternal benefit to conservative management, but there can be significant neonatal benefit, especially in the late second and early third trimester. The benefits of conservative management are mainly in prolonging pregnancy, which has the potential to decrease gestational age-related morbidity associated with preterm birth. This must be balanced with the risks of conservative management, which include cord prolapse, placental abruption, perinatal infection, emergent delivery for a non-reassuring fetal status and fetal death.
Tocolytics have been evaluated in the conservative management of PPROM and have been found to be of limited value. Two studies have evaluated the benefits of ritodrine. Christensen et al.25 performed a randomised double-blind study of women with PPROM, between 28 and 36 weeks, comparing ritodrine tocolysis to placebo and found latency was prolonged for 24 hours, with no obvious clinical benefit. Levy and Warsof26 performed a prospective randomised controlled trial comparing ritodrine to no tocolytic administration in women with PPROM, from 25 to 34 weeks, without any evidence of labour and found an increase in latency (10 vs 3.5 days, P= 0.033), with 47.6% of the treated women versus 14.2% of those untreated having a latent period of more than one week. They did not examine neonatal outcome and had no evidence that the increased latency had any clinical benefit. Weiner et al.27 evaluated the cost-effectiveness of tocolysis with ritodrine, terbutaline or magnesium sulphate in women with PPROM at less than 34 weeks and found some prolongation of pregnancy (105.2  vs 62.1  hours, P= 0.06) without any significant reduction in cost or improvement in perinatal outcome. A prospective randomised controlled trial comparing magnesium sulphate tocolysis to no tocolysis in women with PPROM (24–34 weeks) by How et al.28 found no significant improvement in perinatal outcomes. In this study both groups received corticosteroids and antibiotics.
In the setting of PPROM ≤32 weeks, in order to achieve 48 hours of corticosteroid administration, tocolytics may be used in the absence of clinically apparent or subclinical amniotic fluid infection, placental abruption, non-reassuring fetal status or other maternal/fetal contraindications to labour inhibition.
The literature currently does not support the use of maintenance or prophylactic labour inhibiting agents beyond the initial 48-hour steroid window. In the setting of recurrent labour in PPROM in the extremely preterm (≤28 weeks) gestation, amniocentesis should be performed for the detection of amniotic fluid infection. Labour inhibition may be considered in this group in the absence of subclinical amniotic fluid infection.
The benefits of intrapartum chemoprophylaxis for the prevention of early onset group B streptococcal infection of the neonate are well established.29,30 Intrapartum chemoprophylaxis should be initiated in any patient with unknown group B streptococcal status or a history of a positive culture during the present pregnancy (treatment is not initiated if there has been a negative anovaginal culture in the past two weeks). Therapeutic options include the following: intravenous penicillin 5 million unit bolus followed by 2.5 million units every 4 hours; intravenous ampicillin 2 g bolus followed by 1 g every 4 hours; intravenous erythromycin 500 mg every 6 hours; or intravenous clindamycin 900 mg every 8 hours (in the presence of penicillin allergy).29 The Centers for Disease Control (CDC) in the United States has recommended that cefazolin (Ancef), 2 g intravenous as a loading dose followed by 1 g every 8 hours, should be used in women with an uncertain penicillin allergy or with a minor allergic reaction, such as a rash.29
The latency period in PPROM (time interval from rupture of membranes to delivery) is alarmingly brief; Mercer et al.31 found that expectantly managed women had a median latency period of 2.9 days. In this trial, the addition of one week of broad spectrum antibiotic therapy increased median latency to 6.1 days. As a result of this generally brief latency period, the major risks to the fetus after PPROM are those related to immaturity. In the absence of clinically obvious intra-amniotic infection, fetal distress or placental abruption, prolongation of pregnancy in order to reduce the risk of immaturity has been the main goal of conservative management of PPROM. The neonatal benefit of prolongation of pregnancy lies in a reduction in the risk of sequelae of immaturity with advancing gestational age. Prolongation of pregnancy carries the risk of intra-amniotic infection leading to neonatal sepsis, altered neonatal haemodynamic transition and increased risk of adverse neurodevelopmental outcomes, including cerebral palsy. There is also a high risk of placental abruption and prolapse/compression of the umbilical cord leading to fetal distress or stillbirth. Antibiotics are valuable not only for their antimicrobial spectrum but also to modulate the maternal and fetal inflammatory response that leads to labour and neonatal morbidity. Adjunctive antibiotic therapy has been recommended in the conservative management of PPROM, with the aim of preventing or treating ascending intrauterine infection, thereby prolonging pregnancy and decreasing maternal and neonatal infectious morbidity. Numerous studies have been performed in order to evaluate the benefit of antibiotics and significant evidence has been generated, which shows that adjunctive antibiotics are beneficial in the conservative management of PPROM.31–39
Therapies have varied according to antibiotic choice, route of delivery and duration of treatment. Many follow recommendations of the NICHD MFMU study, which advocates intravenous therapy for 48 hours using ampicillin and erythromycin.31 This is followed by a limited course of oral therapy of both drugs for one week. This trial found that antibiotics improved neonatal health by reducing the risk of death, respiratory distress syndrome (RDS), early sepsis, severe intraventricular haemorrhage (IVH) and severe necrotising enterocolitis (NEC) (down from 53% to 44%, P < 0.05). There was also a decrease in amnionitis and an increase in the likelihood of prolongation of pregnancy for longer than one week.
The administration of antepartum glucocorticoids to women at risk of preterm birth has clearly been shown to reduce perinatal morbidity and mortality. The National Institutes of Health consensus conference of 1994 recommended corticosteroid administration in women with PPROM before 30–32 weeks, using a single course of either betamethasone (two doses of 12 mg intramuscularly, 24 hours apart) or dexamethasone (four doses of 6 mg intramuscularly, 12 hours apart). The consensus was that corticosteroids reduced the incidence of RDS, IVH and neonatal mortality.40 For some, uncertainty existed as to whether corticosteroids might increase the risk of neonatal infection, but two randomised controlled trials found no such association. Lewis et al.35 studied women with PPROM between 24 and 34 weeks of gestation who received antibiotics and corticosteroids and found a reduced risk of RDS (18.4 vs 43.6%, P= 0.03) without evidence of increased neonatal infection (3 vs 5%, P= not significant). A recent meta-analysis by Harding et al.41 found that corticosteroid administered to women with PPROM reduced the risks of NEC (relative risk [RR] = 0.21, 95% CI 0.05–0.82), IVH (RR = 0.47, 95% CI 0.31–0.7), RDS (RR = 0.56, 95% CI 0.46–0.7) without any significant increased risk in maternal infection (RR = 0.86, 95% CI 0.61–1.2) or neonatal infection (RR = 1.05, 95% CI 0.66–1.68). This confirms that a single course of antepartum glucocorticoids is highly beneficial in PPROM prior to 32 weeks of gestation. If the patient has documentation of immaturity (negative fetal maturity testing of the vaginal pool) or there is no fluid for testing, corticosteroid therapy could be considered in women with PPROM from 32 to 34 weeks. The benefits of repeated corticosteroids courses have not been determined and are not recommended.
TIMING OF DELIVERY
Controversy exists regarding the optimum gestational age at which expectant management is discontinued and delivery is expedited. The optimum interval for delivery occurs when the risks of immaturity are outweighed by the risks of pregnancy prolongation in the setting of PPROM (infection, abruption and cord accident). Lung maturity assessment may be useful to plan the timing of delivery in the 32–34 weeks of gestation range. An amniocentesis performed at 32 weeks is helpful with respect to the diagnosis of infection and confirmation of fetal lung maturity (FLM). Both a retrospective cohort study and a randomised controlled trial support the use of amniocentesis among women with PPROM to plan the mode and timing of delivery.14,16 In the randomised controlled trial by Cotton et al.,14 fetal distress, manifested as a non-reassuring fetal heart rate, was more frequent in the no amniocentesis group (P < 0.05). The number of days the infant remained in the hospital was significantly less in the amniocentesis group (median = 8.5 days, range: 2–88 days) than in the no amniocentesis group (median = 22 days, range: 2–110 days, P < 0.01). This difference in neonatal hospital days appeared to be mainly due to a slower resolution of the multiple problems of immaturity.14
There are no data to suggest that studies of lung maturity in the presence of PPROM are any different from those used in pregnancies with intact membranes. The choice of clinical tests to be used is provider preference dependent.
Those women with a reassuring fetal status (those that are not in advanced labour and without signs or symptoms of placental abruption, clinical amnionitis or subclinical amniotic fluid infection) can be offered conservative management. Corticosteroids and antenatal antibiotics should be used and the woman should be placed on modified bed rest.
Women with PPROM between 32 and 33 weeks of gestation should have amniotic fluid collected via transabdominal amniocentesis for lung maturity studies and infection-related studies, if possible. The Amniostat-FLM (Irvine Scientific, Santa Ana, California), is a simple slide agglutination test that evaluates the presence of phosphatidylglycerol (PG). It can be performed on vaginal pool fluid and is not affected by the presence of blood or meconium. The test has a PPV of 95–100% (test indicates the fetus has lung maturity) and an NPV of 35–51%.42
The TDX–FLM assay (Abbott Laboratories, Abbott Park, Illinois) is a very reliable and reproducible test for amniotic fluid FLM testing. It is a fluorescent polarisation assay that measures the segregation of a fluorescent dye between surfactant and albumin in amniotic fluid using an automated technique. The manufacturer recommends using a level of ≤70 as consistent with FLM and the test is reported to have a PPV of 96–100% (test is positive if FLM is present) and an NPV of 47–61%.43
If testing indicates lung maturity, delivery should be considered. If there is no fluid available for testing or lung-maturity testing is negative, the women can be offered expectant management with antenatal corticosteroids and antibiotics. Delivery should be considered after corticosteroid benefit has been obtained. If testing indicates amniotic fluid infection, delivery should be considered.
There appears to be no role of expectant management in any women with PPROM beyond 34 weeks of gestation. Several studies have evaluated expectant management with labour induction and found increased risks of chorioamnionitis, neonatal infection and increased length of hospital stay in women managed expectantly.44–46 Immediate induction and delivery should be encouraged in this patient population.
AMNIOCENTESIS: LIKELIHOOD OF SUCCESS AND RISK OF COMPLICATIONS
There are nine English language trials in the literature that report the likelihood of success for transabdominal amniocentesis in the presence of PPROM. The success rates range from 49% to 98% with an average of 72% (686/950).13 These studies were published between 1979 and 1996. There is a clear trend towards increased success of amniocentesis with the date of publication of the study, confirming that, as ultrasound technology has developed and training in invasive procedures has improved, the likelihood of success is likely to improve. Seven of these nine studies were published prior to 1990 so the likelihood of success today may be greater than 72%.
There is only one English language trial in the literature that specifically addresses the frequency of complications of amniocenteses performed in the presence of PPROM. Yeast et al.47 found zero cases of fetal injury after amniocentesis in 91 women, and that these women had no shorter latency interval than a matched group of women who did not have amniocentesis.
ANTENATAL FETAL SURVEILLANCE IN PPROM
Antenatal fetal surveillance is recommended during conservative management of PPROM. Regimens vary and are usually based on expert opinion. The two most common testing modalities used are the non-stress test and the biophysical profile. The goal of testing is to predict adverse fetal outcome from the two major sources of fetal compromise: umbilical cord compression (secondary to oligohydramnios or anhydramnios) and chorioamnionitis.48–51
PPROM affects over 120,000 pregnancies in the United States annually and is associated with significant maternal morbidity and neonatal morbidity and mortality. Healthcare costs are significantly affected by the prolonged maternal hospital stay, the need for frequent testing and the resulting neonatal costs as a result of prolonged neonatal intensive care for the newborn. Management requires an accurate diagnosis and determination of gestational age. A gestational age approach to therapy is important and should be adjusted for each hospital's neonatal intensive care unit outcomes. Antenatal antibiotics and corticosteroid therapies have clear benefits and should be offered to all women without contraindications. These women are at high risk of infection, and amniocentesis can be used to evaluate early markers of infection and provide a sample of amniotic fluid for culture. Any evidence of infection following amniocentesis should be considered carefully as an indication for delivery. During conservative management, women should be monitored closely for placental abruption, infection, labour and a non-reassuring fetal status. Women with PPROM after 32 weeks should be considered for delivery and after 34 weeks of gestation the benefits of elective delivery appear to outweigh the risks.