Preterm premature rupture of the membranes near the limit of fetal viability is an uncommon complication of pregnancy, affecting approximately 4 in 1000 gravidas. However, maternal, fetal, and neonatal complications resulting from this condition are significant and include chorioamnionitis, pulmonary hypoplasia, restriction deformities, fetal loss, and complications of extreme prematurity among surviving infants. In this article, we review the literature regarding pregnancy outcomes after preterm premature rupture of the membranes near the limit of viability, and the data on traditional and nontraditional interventions to improve outcomes. An approach to patients who present with preterm premature rupture of the membranes near the limit of viability will be proposed.

Key words: periviable; preterm premature rupture of membranes; previable

Article Outline

Etiology

Clinical course

Perinatal morbidity and mortality

Pulmonary hypoplasia

Ultrasound to predict fetal outcomes

Contractures/skeletal deformities

Management

Antibiotics/corticosteroids/tocolytics

Nontraditional therapies: cervical sealants and amniofusion

Recommendations

Summary

References

Few clinical situations in obstetrics are as challenging as preterm premature rupture of the membranes (PROM) near the limit of fetal viability. Although the term “midtrimester PROM” carried a similar connotation to periviable PROM in the 1970s and 1980s,^{[1], [2], [3], [4] and [5]} currently the majority of infants born at 24-26 weeks' gestation will survive with aggressive medical care. Since Taylor and Garite¹ reported on perinatal outcomes after PROM <26 weeks in 1984, practice patterns regarding antenatal corticosteroid administration and antibiotics have significantly changed. Although the incidence of PROM <24 weeks' gestational age is low (0.37%), the complications of this condition for both the mother and fetus are significant.^{[6], [7], [8], [9], [10], [11], [12] and [13]} In the preparation of this article, we chose to focus on publications since 1994 with PROM between 14-24 weeks' gestational age (defined herein as “periviable PROM”). In this article, we review the origin and clinical course of periviable PROM. In addition, we summarize the available data regarding maternal and fetal outcomes and treatments to improve outcomes resulting from this morbid condition. In general, when evaluating any of the outcomes of these trials, it should be noted that all of these studies were retrospective and included only those patients amenable to expectant treatment on admission. In addition, many of these publications excluded patients with advanced labor, fibroids, previa, bleeding, or acute infection, and those who ultimately terminated the pregnancy based on prognosis alone.^{[6], [7], [8], [9], [10], [11], [12] and} 13 S. Falk, L. Campbell and A. Lee-Parriz et al., Expectant management in spontaneous preterm premature rupture of membranes between 14 and 24 weeks' gestation, J Perinatol 24 (2004), pp. 611–616. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (5)^[13] This potentially exaggerates the overall latency, and underestimates the maternal and infant morbidities subsequent to periviable PROM as a whole.

For Editors' Commentary, see Table of Contents

Etiology

The etiology of periviable PROM is multifactorial. It is hypothesized that a weakness in the chorioamniotic membrane occurs as a result of either membrane stretch or degradation of the extracellular matrix.¹⁴ Multiple retrospective investigations have identified risk factors for early preterm PROM (Table 1). Specific risk factors include a history of cervical insufficiency, antepartum bleeding, multiple gestations, previous PROM or preterm labor, tobacco use, cervical cerclage, and amniocentesis. Although each of these factors is associated with periviable PROM, their predictive value is low. The rate of periviable PROM after a genetic amniocentesis has historically been reported as 1-2%.^{[15] and [16]} However, recent data from large trials evaluating complications after genetic amniocentesis suggested the risk to be much lower, with the overall risk of pregnancy loss <24 weeks to be 1 in 1600.¹⁷ An investigation by Kilpatrick et al¹⁸ specifically compared risk factors between pregnancies complicated by PROM at 14-24 weeks and matched term deliveries. Significant risk factors for periviable PROM included a history of preterm delivery (odds ratio [OR], 15.2; 95% confidence interval [CI], 16.1–37.8), cervical incompetence in the incident pregnancy (OR, 12.7; 95% CI, 3.5–46.4), cerclage (OR, 10.3; 95% CI, 2.8–38.8), history of PROM (OR, 7.3; 95% CI, 2.5–22.3), history of cervical incompetence in a previous pregnancy (OR, 3.8; 95% CI, 1.2–11.8), and tobacco use (OR, 2.0; 95% CI, 1.1–3.9). Neither a history of chlamydia (OR, 0.7; 95% CI, 0.4–1.3) or gonorrhea (OR, 1.2; 95% CI, 0.5–2.8), nor a current infection with chlamydia (OR, 1.0; 95% CI, 0.5–2.1), gonorrhea (OR, 0.6; 95% CI, 0.1–5.3), or bacterial vaginosis (OR, 1.6; 95% CI, 0.7–3.6) were associated with periviable PROM.

Clinical course

The clinical characteristics of women with PROM near the limit of viability are presented in Table 2. The majority of women in these studies received antibiotic treatment (generally ampicillin ± erythromycin), with >50% receiving antenatal corticosteroids once the limit of viability was achieved. Among the included studies, the gestational age at membrane rupture ranged from 14–24 weeks (mean, 19.8–22.7 weeks) with a mean gestational age at delivery of 23.2–27.1 weeks. Pregnancy outcomes after conservative management of PROM near the limit of viability are presented in Table 3. Median latency to delivery varied between 6-13.0 days (range, 0–161 days), with mean latency of 14.1-39.4 days. Limited data are available about the impact of gestational age at rupture on the length of latency with PROM near the limit of viability. Based on data from 1 investigation, latency does not appear to vary by gestational age at rupture with median latencies of 8 days (1-161) with PROM <20 weeks, 4.5 days (2-106) with PROM between 20-21 weeks, and 12.0 days (1-112) with PROM between 23-23 weeks.¹³ These data appear to be similar to the historical investigations of midtrimester PROM that reported mean latency ranging from 12-21.5 days,^{[1], [3], [4], [5], [19], [20] and [21]} but reflect the skewed nature of latency in this circumstance, which results from the rare occurrence of very prolonged latency after early preterm PROM. Older investigations noted that at least 50% of patients will deliver within 1 week of periviable PROM^{[1], [3], [5] and [19]} with close to 75% delivered within 2 weeks. One recent investigation reported that 38% of women delivered within 1 week and 69% delivered within 5 weeks of periviable PROM.⁹ These data are likely more clinically useful for patient counseling.

Table 4 presents maternal outcomes after periviable PROM. Because maternal outcomes with PROM <24 weeks are likely similar to those of midtrimester PROM, we present data from studies back through 1984, and provide summary estimates for these studies and for the subset of more recent publications with PROM ≤24 weeks. We found no statistically significant differences between the earlier and later cohorts except for a lower frequency of chorioamnionitis (31.8% vs 41.2%; P < .01) and a higher frequency of abruption (9.3% vs 4.3%; P < .01) after 1994. Overall, the most common maternal morbidity reported after periviable PROM is chorioamnionitis, with approximately 37% of women developing this complication. In addition, 11% will also develop endometritis. Of note, maternal sepsis and maternal death appear to be rare, with sepsis complicating 1 of every 100 pregnancies with periviable PROM, and only 1 investigation reporting a single maternal death (1/619 pregnancies) caused by sepsis.¹⁹

Perinatal morbidity and mortality

Fetal death after PROM at or near the limit of viability is common at 31.6% (Table 5). Although survival has been reported with membrane rupture as early as 11 weeks' gestational age,²² a significant difference in survival can be found with PROM <22 weeks compared with PROM occurring >22 weeks (Table 6). It is important to reiterate that currently available data likely overstate the survival through exclusion of those not amenable to continued conservative treatment and those electing pregnancy terminations for persistent fluid leakage, oligohydramnios, or abnormal ultrasound findings.

A summary of neonatal morbidities after conservative management of periviable PROM are presented in Table 7. Respiratory complications are frequent with 66% developing respiratory distress syndrome. Additional morbidities such as bronchopulmonary dysplasia and sepsis are also frequently present. Long-term outcomes including intact survival (survival without minor or major impairments) have been reported; however, most investigations have limited long-term follow-up and it is not possible to differentiate whether the reported outcomes are optimistic (underreporting of deaths) or pessimistic (disproportionate follow-up of those with morbidity).

To estimate the potential risks and benefits of latency for periviable PROM, Everest et al⁶ evaluated outcomes among neonatal survivors with prolonged (≥2 weeks in duration) periviable PROM in 2008. Of 98 women with PROM occurring <24 weeks, 40 (41%) delivered a liveborn infant after a latency of at least 14 days. Ten additional women delivered a stillbirth after prolonged latency with PROM <24 weeks. Among the 40 liveborn infants, the reported mortality was 30% (n = 12). Overall, 10 newborns were diagnosed with pulmonary hypoplasia, but only 1 of the neonatal deaths had an autopsy performed.

Few data are available regarding the prognosis of PROM after midtrimester genetic amniocentesis. Many of the retrospective trials either do not comment on this, or these patients are excluded from the analysis. Morales and Talley²¹ reported that 8 of 138 total women with PROM <25 weeks had an amniocentesis preceding rupture of membranes. Of these 8 women, 7 eventually resealed their membranes and had a term delivery. Gold et al¹⁶ reported on 603 cases of genetic amniocentesis, with 7 (1.2%) having subsequent rupture of membranes. In all 7 cases there was cessation of leakage of fluid with reaccumulation of a normal amount of amniotic fluid. A term delivery occurred in 6 of the 7, with 1 pregnancy ending with an intrauterine fetal death at 25 weeks.

Pulmonary hypoplasia

Pulmonary hypoplasia results from a number of conditions that are associated with fetal lung compression and oligohydramnios. Pulmonary hypoplasia is a serious complication of periviable PROM and mortality from this condition ranges between 50-100%.^{[6], [12], [23] and [24]} On pathological assessment, there is a failure of proliferation of alveoli with a corresponding decrease in lung volume. Normal development of the lung occurs in 5 stages: embryonic, pseudoglandular, canalicular, saccular, and alveolar. Periviable PROM occurs in the late pseudoglandular phase (weeks 8-16) and the early canalicular phase (weeks 16-28) of development. In the pseudoglandular phase, the segmental bronchi divide and develop into the intrasegmental bronchial tree. In the canalicular phase there is development of terminal bronchioles with flattening of epithelium and development of type II pneumocytes. Lethal pulmonary hypoplasia rarely develops subsequent to oligohydramnios >26 weeks' gestation.^{[25] and [26]}

Many of the investigations included in this review did not routinely perform autopsies to determine the cause of neonatal death, and clinical definitions of pulmonary hypoplasia have been inconsistent. Hence, the exact incidence of this disease is likely underreported. Current data suggest that approximately 9-20% of newborns delivered after periviable PROM will develop pulmonary hypoplasia. Factors contributing to the development of pulmonary hypoplasia include gestational age at PROM, duration of latency, and residual amniotic fluid volume.^{[24] and [27]} Blott and Greenough²⁸ described a cohort of 30 pregnancies complicated with PROM between 15-28 weeks with latency of at least 14 days. Eight (27%) infants had pulmonary hypoplasia and 14 (46%) had limb deformities. Newborns were grouped according to the presence or absence of pulmonary hypoplasia and compressive limb abnormalities. Compared with infants without these complications, those with pulmonary hypoplasia had an earlier mean gestational age of PROM (17 vs 24 weeks) and longer mean latency (10 vs 5 weeks). A similar trend was found for limb deformities with earlier mean gestational age of PROM (18 vs 24 weeks) and longer mean latency (13 vs 5 weeks) noted for the infants with limb deformities. The relationship between pulmonary hypoplasia and latency is confounded by the fact that time is required for pulmonary hypoplasia to develop, and those delivered soon after PROM cannot have this condition.

Ultrasound to predict fetal outcomes

Several authors have evaluated the role of amniotic fluid volume on perinatal outcomes. Hadi et al⁷ evaluated the presence or absence of oligohydramnios (no visible 2-cm pocket of fluid). A total of 178 singleton pregnancies complicated by PROM between 20-25 weeks were included. In all, 107 (60.1%) were deemed to have at least 1 visible fluid pocket and 71 (39.9%) did not. The frequency of chorioamnionitis was higher (33.8 vs 21.5%; P = .07) and perinatal survival was lower (9.9% vs 85.0%; P < .01) in the group with severe oligohydramnios. For those who delivered between 26-34 weeks (n = 104), the frequency of chorioamnionitis was higher for those with severe oligohydramnios (69.2% vs 24.1%; P < .001), as was the incidence of perinatal death (69.2% vs 2.1%; P < .001). These results are similar to those reported by Xiao et al,¹¹ who reported improved fetal survival among patients with an amniotic fluid index ≥5 cm after periviable PROM at 14-24 weeks' gestation (80.0% vs 30.8%; P < .01). In addition, both Muris et al⁹ and Grisaru-Granovsky et al⁸ reported a higher mean amniotic fluid index among survivors after PROM near the limit of viability (P < .05 for both).

The patient with persistent oligohydramnios appears to be at particular risk for pulmonary hypoplasia. Rotschild et al²³ found less frequent pulmonary hypoplasia among newborns delivered with normal amniotic fluid volumes after PROM (6.7%) compared with those with oligohydramnios (20.9%). Overall, persistent oligohydramnios has a sensitivity of 52-100%, specificity of 41-82%, positive predictive value of 22-64%, and a negative predictive value of 89-100% for pulmonary hypoplasia.^{[23], [24], [27] and [29]}

Considerable effort has been directed toward assessment of fetal lung growth by ultrasound to predict pulmonary hypoplasia. In 1985, Callen et al³⁰ demonstrated a progressive lag of thoracic growth in a fetus that developed pulmonary hypoplasia. Subsequently, various fetal measurements have been assessed such as thoracic circumference, lung length, and thoracic area. Some authors have attempted to correct for overall fetal size by including another fetal parameter such as femur length, heart area, or abdominal circumference. Table 8 presents results from studies that evaluated the predictive value of ultrasound for lethal pulmonary hypoplasia.^{[29], [31], [32], [33], [34] and [35]} All of these studies were performed in patients at risk for pulmonary hypoplasia, including patients with PROM. Although sonographic measurements appear to be useful in the prediction of pulmonary hypoplasia, no one test is clearly preferable. Preliminary study of the role of estimated lung volumes by 3-dimensional ultrasound has been encouraging but not definitive.³⁵

Contractures/skeletal deformities

The pattern of fetal restriction deformities occurring subsequent to periviable PROM are similar to those seen with Potter syndrome, and include abnormal facies (low-set ears, hypertelorism, receding chin, flattened nose, wrinkled skin), abnormal limb positions,³⁶ and flattened hands and feet. The incidence of skeletal deformities after periviable PROM ranges from 1.5-38%. Rotschild et al²³ reported on 88 newborns with prolonged PROM (≥7 days) at <29 weeks. The overall incidence of deformities was 20.4%, with increasing latency and severe oligohydramnios identified as significant risk factors. Fortunately, many of these are easily treated with physiotherapy and do not require surgery.³⁷

Management

Antibiotics/corticosteroids/tocolytics

For women who present with PROM at ≥24 weeks' gestational age, broad-spectrum antibiotics and antenatal corticosteroids have been shown to increase latency and improve neonatal outcomes.³⁸ However, even with broad-spectrum antibiotics, 55% of patients with PROM >24 weeks will deliver within 48 hours, and 86% will deliver within 21 days. For many patients with periviable PROM, this may not be enough time to achieve fetal viability. To date, no randomized trials have specifically addressed the role of broad-spectrum antibiotic therapy in periviable PROM. Although the risks and benefits of antibiotic therapy for women who desire conservative management of periviable PROM are unproven, their use has increased over time from 0% in 1984 to 97% in 2008. Compared with patients not treated with antibiotics on admission, Morales and Talley²¹ reported a lower frequency of chorioamnionitis with antenatal ampicillin (12.9 vs 30.3%; P = .06). Both Xiao et al¹¹ and Grisaru-Granovsky et al⁸ reported on neonatal survival with antenatal antibiotic treatment after periviable PROM. In the study by Grisaru-Granovsky et al,⁸ 22 (88%) of 25 women with periviable PROM received ampicillin and erythromycin on admission and were continued on antibiotics until delivery. The exact antibiotic regimen was not specified by Xiao et al,¹¹ however, 25 (89.3%) of 28 patients received some form of antibiotics before delivery. Overall, 47 (88%) of 53 patients were given therapy with improved survival noted for those who received antibiotics (51.1% vs 0%; P = .02). Although these results are encouraging, the current evidence is insufficient to guide either the use or timing of antibiotics in the setting of PROM near the limit of viability. It is unlikely that antibiotic therapy initiated long after membrane rupture will significantly enhance pregnancy outcomes.

A similar trend is observed in the use of antenatal corticosteroids. Although no specific recommendations exist for the use of antenatal corticosteroids in the setting of periviable PROM, many investigations report the administration of antenatal corticosteroids after viability has been reached. This can skew the association of steroids with survival because the achievement of viability alone would be associated with improved outcomes. Although a paucity of data is available on this topic, at least 4 publications have reported on infant survival according to steroid exposure^{[3], [8], [11] and [21]} and are presented in Table 9. From these results, antenatal corticosteroid exposure appears to be associated with significant improvements. The role of rescue steroids remote from an initial course of antenatal steroids has not been determined.

The use of tocolytics with PROM is controversial,³⁹ and its value in periviable PROM is not known. At best, tocolysis can offer the benefit of a few days of pregnancy prolongation and is unlikely to have any clinical impact when given before the limit of fetal viability. It has been postulated that tocolytic therapy might confer a brief prolongation of pregnancy and thereby allow more time for antibiotic therapy, however, this particular benefit is speculative.

Nontraditional therapies: cervical sealants and amniofusion

Several investigators have evaluated the use of cervical plugs or membrane sealants in the management of periviable PROM.^{[40], [41], [42], [43], [44] and [45]} Although the exact therapy used in each of these investigations are different (gelatin sponge, platelets, cryoprecipitate, fibrin) the goal of each was to repair a defect in the amniotic membrane and allow for reaccumulation of amniotic fluid. In general, these investigations are limited in size and used other therapies including cerclage, tocolytics, and/or antibiotics. In addition, 2 of these investigations enrolled patients with iatrogenic PROM.^{[42] and [43]} None offered a control group for comparison of outcomes. The reported survival for those treated ranged from 30-54%. The largest trial of cervical sealants, performed by Sciscione et al,⁴⁵ evaluated an intracervical fibrin sealant in 12 women with spontaneous PROM <24 weeks' gestation. The mean gestational age at delivery was 27 weeks, and the mean latency was 48 days. Survival was 54%. Adverse events were not reported. Although the findings of these studies may be encouraging, the risks associated with these interventions remain unknown. Absent comparative data that suggest benefit without significant maternal risks, these treatments should not be incorporated into routine medical practice.

Transabdominal amniofusion for periviable PROM was evaluated in a trial by Vergani et al.⁴⁶ A total of 18 women with PROM ≤25 weeks with oligohydramnios were treated with serial amniofusion and were compared with a historical cohort of 16 untreated patients. The authors reported no adverse events in the amniofusion-treated group, and found a lower incidence of pulmonary hypoplasia compared with control subjects (46% vs 86%; P = .03). Although the findings of this investigation are interesting, the frequency of pulmonary hypoplasia in the control group was unusually high and the biologic plausibility of amniofusion without correction of the membrane rupture is questionable. At present, the benefits of amniofusion for periviable PROM are unproven and the risks remain undetermined. Additional research is required before this therapy can be considered for routine clinical use.

Recommendations

Based on the available literature, we offer an algorithm for the management of PROM near the limit of viability (Figure). The patient with periviable PROM and no indication for immediate delivery should be extensively counseled regarding the risks and benefits of conservative management. This should include a realistic estimate of the potential neonatal outcomes and maternal risks. For patients who determine the risks of conservative management outweigh the potential benefits, delivery can be achieved with induction or by dilatation and evacuation. The optimal approach for delivery should be individualized based on patient characteristics, patient preference, available facilities, and physician experience.

For the patient who desires conservative management, no data delineate an optimal approach. Initial care may include broad-spectrum antibiotics with strict bed rest and pelvic rest in conjunction with close surveillance for infection or abruption as an inpatient or outpatient. Women who are discharged should return immediately for fever, abdominal pain, a suspicious vaginal discharge, or bleeding. Hospitalization should be considered once viability is achieved to allow for early intervention for infection, abruption, labor, and abnormal fetal surveillance. Antenatal corticosteroids should be administered at this time.

After an initial ultrasound assessment, repeated evaluation can be performed every 2 weeks to asses amniotic fluid status (and the possibility of secondary membrane sealing) and fetal lung growth. If pulmonary hypoplasia becomes suspected before viability the patient may choose to reconsider the decision of expectant management.

Summary

Preterm PROM before fetal viability is a rare complication of pregnancy, but carries significant maternal morbidity, neonatal morbidity, and neonatal mortality. For the gravida who develops this obstetric complication, extensive counseling should be performed regarding the risks and benefits of conservative management. For a patient who desires conservative management of PROM before the limit of viability, serial assessments should be performed for signs of infection or labor. In addition, interval ultrasound examinations should be performed to asses for pulmonary hypoplasia. If these conditions arise, then conservative management may no longer be desirable for the patient. Limited data are available to guide the use of antenatal therapies such as antibiotics and corticosteroids to improve outcomes. However, the use of antenatal antibiotics to improve latency followed by antenatal corticosteroids once viability has been achieved is a reasonable approach. Future research should address what, if any, benefit cervical sealants confer for periviable PROM, and the role of tocolytic therapy and antibiotic therapy to improve latency and neonatal outcomes.

Reprints not available from the authors.