全文

Abstract

Objective

Preterm prelabor rupture of membranes (PPROM) before 27 weeks’ gestation is associated with severe perinatal complications, but quantitative estimates are lacking. The aim of this study was to report and predict outcomes of pregnancies complicated by early PPROM and to study antepartum risk factors that might predict perinatal death in future patients.

Study design

We performed a retrospective cohort study of women with PPROM between 13⁺⁰ weeks and 27⁺⁰ weeks’ gestation between 1994 and 2009 in three perinatal centers.

Main outcome measures

Perinatal mortality, composite neonatal morbidity and premature delivery. A model to predict these outcomes was developed from antepartum variables.

Results

We identified 314 women with PPROM before 27 weeks, including 28 multiple pregnancies. Six pregnancies (2%) were terminated before 24 weeks’ gestation, and three were lost to follow up, leaving 305 pregnancies for analysis. Overall, there were 166 perinatal deaths (49%). The perinatal mortality rate decreased with increasing gestational age at PPROM (from 70% in the group PPROM 13–20 weeks to 27% in the group PPROM 24–27 weeks). Of the 170 surviving neonates, 70 suffered from serious morbidity (41%). Early gestational age at PPROM, long interval between PPROM and birth and positive vaginal culture (any bacteria) were associated with perinatal mortality.

Conclusion

Perinatal mortality in PPROM before 27 weeks occurred in half of the cases and among those who survive approximately 40% suffered serious morbidity. Antenatal parameters can be helpful to predict perinatal mortality.

Keywords: Preterm birth, Preterm rupture of membranes, Midtrimester PPROM, Perinatal mortality, Neonatal morbidity.

1. Introduction

Preterm prelabor rupture of membranes (PPROM) before 27 weeks’ gestation occurs in approximately 0.5% of all pregnancies [1]. It is associated with severe complications, such as premature birth, pulmonary hypoplasia and fetal death [2]. Iatrogenic PPROM may be triggered by invasive procedures such as amniocentesis, chorionic villus sampling, cervical surgery or mechanical trauma. Midtrimester PPROM after amniocentesis occurs in up to 1% of all procedures [3]. The pathogenesis of spontaneous PPROM is not well understood: possible risk factors are history of preterm labor or PPROM, cervical insufficiency, smoking, multiple gestation and antepartum bleeding [4], [5], and [6].

Complications of second trimester PPROM have been studied previously, but the results are inconsistent [7], [8], [9], [10], [11], [12], [13], [14], and [15]. Overall, the prognosis for perinatal survival and morbidity after early PPROM seems poor, with survival rates between 14% and 70% (Table 1). Prediction of neonatal outcome in pregnancies with early PPROM is virtually impossible, due to heterogeneity of previous studies, small sample sizes and increased health care compared to the era in which previous studies were reported (1970s to 1980s).



It is, however, crucial to counsel the woman accurately on the risks in a case of early PPROM, in order to decide whether or not to terminate pregnancy. The primary aim of the current study was to report our findings on perinatal outcome in pregnancies complicated by early PPROM and to provide tools for counseling. Furthermore, we investigated if the risk of perinatal death could be predicted from antepartum variables.

2. Materials and methods

We performed a retrospective cohort study in the obstetric departments of the Máxima Medical Centre in Veldhoven (MMC) in the period 1994–2005, the Academic Medical Centre in Amsterdam (AMC) in the period 1996–2009, and Maastricht University Medical Centre (MUMC) in the period 1997–2001. The difference in time periods between the three centers can be explained by logistic reasons.

2.1. Patients

Women with PPROM before 27 weeks’ gestation were eligible for this study. Exclusion criteria were contractions at presentation (subsequently resulting in PPROM), cervical insufficiency necessitating cervical cerclage, or pregnancies with known lethal fetal anomalies. Women were identified from local electronic databases in which the moment of rupture of the membranes (ROM) and moment of birth were registered.

Women were managed expectantly as inpatients. Routinely daily temperature measurements and cardiotocography were performed and blood samples were taken at least once weekly. Corticosteroids were administered for fetal lung maturation from 25 weeks onwards, and preterm labor was arrested with tocolytics. Prophylactic antibiotics, erythromycin 250 mg 4 times daily for 10 days, were administered in two centres (MUMC and MMC). In one centre (AMC) antibiotics were given only when clinical signs of infection were present (maternal temperature > 37.8 °C (100 °F) or fetal tachycardia). This policy differed between the centres because the national guideline does not provide a clear recommendation on the use of antibiotics in cases of PPROM [1].
2.2. Outcome

We recorded the following outcome measurements: perinatal mortality, premature delivery, and neonatal morbidity, i.e. pulmonary hypoplasia, respiratory distress syndrome (RDS), intraventricular hemorrhage (IVH), necrotizing enterocolitis (NEC), chronic lung disease (CLD) and sepsis.

Pulmonary hypoplasia was defined as neonatal death within 24 h after birth, due to respiratory failure not attributable to other causes and impossibility of postpartum ventilation, if possible confirmed by histological examination of the lungs. RDS was defined as clinical signs of respiratory distress in prematurely born infants, with impaired oxygenation and characteristic radiologic signs (air bronchograms, ground glass appearance) [16] and [17]. CLD, (formerly named bronchopulmonary dysplasia (BPD), was defined as infants with oxygen dependency at either 28 days of life or 36 weeks’ gestation [18].

IVH was defined as hemorrhage in the germinal matrix, ventricles, or cerebral parenchyma, observed by ultrasound examination or MRI. Ultrasound examination is routinely performed in all neonates born prior to 32 weeks’ gestation, or in neonates with neurologic symptoms [19].

A diagnosis of NEC was made on the presence of the characteristic clinical features of abdominal distention, with or without rectal bleeding, and abdominal radiographic findings associated with pneumatosis intestinalis (this last finding is an abnormal gas pattern with dilated loops consistent with ileus) [20].

Neonatal sepsis was classified as suspected or proven (caused by any pathogen) and defined as a neonatal infection with cardiorespiratory instability or a positive blood culture caused by any pathogen. Clinical infection includes symptoms like positive findings on clinical examination, imaging, or laboratory tests. Laboratory signs of infection were increased C-reactive protein (CRP), leukocytosis or leukocytopenia [21].

For pregnancy outcome and neonatal outcome measurements, a subdivision was made for different categories: PPROM between 13 and 20 weeks, 20 to 24 weeks and 24 to 27 weeks.

 

We calculated the rates of each outcome measure, expressed as percentage, mean with standard deviation or median with interquartile ranges. Kaplan–Meier curves were constructed, indicating time to delivery in relation to gestational age at PPROM.

2.4. Prediction model development

2.4.1. Potential predictors

For the estimation of the individual risk of perinatal death, we assumed the following variables as potential predictors: maternal age, gestational age at PPROM, interval between PPROM and birth, anhydramnios, positive vaginal culture (any bacteria) and positive vaginal culture for GBS (group B streptococcus).

2.4.2. Model building

To account for missing values, we used multiple imputation techniques. We introduced all potential predictors in a multivariable logistic regression model and used backward stepwise elimination to reduce the amount of predictors per dataset, using a liberal p-value of 0.20.

2.4.3. Internal validation

We adjusted the model using bootstrapping techniques to reduce the probability of overfitting (i.e. the model performs particularly well on the data that we used to develop the model, but is often very disappointing in future patients) [22].

2.4.4. Model performance

To quantify the performance of the final model, we assessed the discriminative ability, the calibration, and the overall performance. The discriminative ability is the models’ ability to distinguish cases from non-cases. It ranges from 0.5 (no discrimination) to 1.0 (perfect discrimination). A Hosmer and Lemeshow (H-L) goodness-of-fit statistic was computed. A high H-L statistic will yield a low p-value and provides evidence of lack of fit. The overall performance, or the accuracy of the model, was quantified by computing the Brier score [22].

Analyses were done using Microsoft Excel, SPSS version 18.0 for Windows (SPSS inc, IL, Chicago, USA) and R version 2.12.2.

return to article outline

3. Results

PPROM before 27 weeks was identified in 314 women (158 cases from AMC, 36 from MUMC and 120 from MMC), of whom three were excluded as their outcome was unknown. Six (1.9%) women requested termination of pregnancy before 24 weeks’ gestation (all singleton pregnancies with PPROM between 15+0 and 20+5 weeks). Of the remaining 305 women, 25 were twin pregnancies and three triplet pregnancies, leaving 336 fetuses eligible for analysis (Table 2).

 

In the total group, 70 (21%) neonates were born within 48 h after PPROM, whereas 56 (17%) neonates were still unborn 50 days after PPROM. The median interval between PPROM and delivery was 10 days (IQR 3 days; 33 days). Of all neonates, 238 (71%) were born before 27 weeks and only 8 (2.4%) after 37 weeks.

In the early gestational age group (PPROM 13–20 weeks), significantly more women were still pregnant 50 days after PPROM, compared to the subcategory PPROM 24–27 weeks (35% versus 1.4%; RR 0.31 (95% CI 0.23–0.40); p < 0.0001), with an absolute risk reduction (ARR) of 66%.

The Kaplan Meier curve expressing time to delivery (Fig. 1) shows that the earlier the GA at PPROM, the more likely the chance of continuation of pregnancy for several days or weeks.


The overall perinatal mortality rate was 166/336 (49%), of which 93 (28%) were stillbirths. Table 3 shows the mortality rates in the different subgroups. The relation between gestational age at PPROM and perinatal mortality is shown in a Kaplan Meier curve (Fig. 2).

Pulmonary hypoplasia was diagnosed in 11 of the neonates who died within 24 h post-delivery (3.3%).

Of the 170 surviving neonates (alive at seven days after birth), 70 (41%) suffered serious morbidity (RDS grade 3 or 4, IVH, NEC, CLD or (suspicion of) sepsis). Some neonates had multiple morbidities.

One hundred neonates (59% of the surviving neonates) survived without any of these endpoints, which is 30% of the total group of 336 neonates. Table 4 summarizes the outcome measurement per subgroup. Survival rates without major morbidity ranged from 20% to 37% amongst the subgroups.

3.1. Prediction model

From the preselected candidate predictor variables, GA at PPROM, interval between PPROM and birth, and positive vaginal culture (any bacteria) were selected in the multivariable logistic regression analysis (Table 5). Ten different bacteria were identified in positive vaginal cultures (e.g. E. coli, Enterobacter cloacae, Gardnerella vaginalis, Klebsiella and Proteus mirabilis). The model performance was good. The area under the original ROC curve (Fig. 3) was 91.0% (95% CI: 87.9–94.1), after correction for optimisms it was 88.8% (95% CI: 85.7–91.9), which indicates a good expected discriminative ability. The calibration of the model was good: the calibration plot indicates that predicted probabilities equal observed frequencies. The H-L goodness-of-fit test (p = 0.69), confirms this.

4. Comments

We reported a perinatal mortality rate of 49% in this retrospective cohort study. Of the survivors, 41% suffered serious morbidity, but 59% of surviving neonates had no serious morbidity. In other words, after PPROM prior to 27 weeks’ gestational age, the overall chance of survival without severe morbidity is 30%. Our perinatal mortality rate is comparable with previous studies [7], [8], [9], [10], [11], [12], [13], [14], and [15], but perinatal mortality rates varying from 25% [12] to 86% [8] have been reported. An important remark should be made. The perinatal mortality rate decreases with increasing gestational age at PPROM. The perinatal mortality rate is 71% in the subgroup PPROM between 13 and 20 weeks, 59% in the subgroup PPROM between 20 and 24 weeks and only 27% in the group PPROM between 24 and 27 weeks. The morbidity rate, however, does not improve with increasing gestational age at PPROM.

Approximately 20% of women delivered within 48 h after ROM. Latency seems related to gestational age at ROM, with early GA at ROM being related to a longer latency.

A previous study from Farooqi et al. reports a mean latency period varying from 12 to 72 days [10]. We found a latency period of 25 days. The effect of latency on perinatal outcome is not entirely clear and the effects of a prolonged latency period do not seem to be consistent across gestational ages [23]. In our study, latency did not lead to an improvement in perinatal survival. This may be explained by the fact that although the latency period in the group with the earliest PPROM appeared to be the longest, many of these fetuses could not benefit from this latency as 70% were born before 27 weeks and 59% even before 24 weeks. The earlier the GA at PPROM, the higher the perinatal mortality rate (71% in subgroup PPROM 13–20 weeks versus 27% in subgroup 24–27 weeks).

A recently published study from France by Azria et al. [24], on PPROM between 15 and 25 weeks, focused on pregnancies which are terminated after early PPROM, instead of only pregnancies that have been continued. The authors hypothesized that perinatal outcomes are better in settings where a low risk group is selected and TOP is frequently performed. The incidence of TOP is much higher in the French study (50%), compared to the incidence of 2% in our study. In the Netherlands we seem to be more conservative in many pregnancy-related problems. The result of this French study is that the neonatal major morbidity and mortality were not lower in the center with higher rates of TOP, which is the opposite of what the authors expected. In our data, however, perinatal survival and major neonatal morbidity were much better in the conservatively managed group than in the French study. Nevertheless, the authors of the French study do find the perinatal risks after PPROM very high.

In our large retrospective study on women with PPROM before 27 weeks, we were able to study over 300 women. Since this was a retrospective study, it has its limitations. Because of the low incidence of early PPROM we decided to collect data over the period 1994–2009, which in itself may have influenced the outcome due to the improvement in neonatal care. We were unable, however, to demonstrate an improvement in perinatal mortality or morbidity over these years.

Referral bias probably explains the rather unfavorable outcome of iatrogenic PPROM in 45% of pregnancies in our study. This in contrast to a study by Borgida et al., who reported a perinatal survival rate of 91% after iatrogenic PPROM and 9% after spontaneous PPROM [25]. We expect that this difference can, at least partly, be explained by the difference between the definition of PPROM between Borgida's study and our study. Borgida et al. considered both women with persistent leak of fluid, and women with a normal amount of amniotic fluid on ultrasonic examination, as having ruptured membranes. In our study, women with transient loss of amniotic fluid after an invasive procedure were probably not referred to these high care centers and could thus not be included. The mean latency period between iatrogenic PPROM and delivery was 65 days in our study and 124 days in Borgida's study.

Due to the retrospective character of our study, some data were missing. Still, we were able to collect data on perinatal death in over 99% of women. Pulmonary hypoplasia was a difficult item to report. Since the majority of parents declined autopsy postpartum, we decided to define this item as respiratory failure not attributable to other causes and impossibility of postpartum ventilation.

We were able to construct a prediction model based on four antepartum parameters; early GA at PPROM, short interval between PPROM and delivery, positive vaginal culture (any bacteria) and no use of antibiotics during admission. The AUC and the H-L goodness-to-fit statistics suggest that the model seems reliable, but this prediction model is based on retrospective data collection and has to be externally validated.

The results of this study can be helpful in future counseling of women with early PPROM and help parents in their decision on whether or not to terminate pregnancy after early PPROM. However, the factor ‘interval between PPROM and delivery’ cannot possibly be predicted when PPROM occurs.

Based on the results from the prediction model, we advise giving prophylactic antibiotics to all women with early PPROM and treating any bacteria in the vaginal culture, since both might contribute to better perinatal outcome. In our study, the use of prophylactic antibiotics in case of PPROM differed between the three perinatal care centers. Previous literature has shown that the use of antibiotics (erythromycin or a combination of ampicillin and erythromycin followed by amoxicillin and erythromycin) in women with PPROM might lead to a reduction in neonatal morbidity [26] and [27].

In conclusion, in cases of PPROM before 27 weeks’ gestation, the risk of perinatal death in the total group is 49%. Looking at the outcome per age category, there seems to be a logical improvement in perinatal survival with increasing gestational age. There seems to be a high risk of serious morbidity in the neonate and only 30% survive without major complications. Antepartum variables seem to be useful in the prediction of the individualized risk of neonatal mortality and morbidity, which in itself is important for objective counseling of women with early PPROM.

Acknowledgements

None.

References

[1] Nederlandse Vereniging van Obstetrie en Gynaecologie [Rupture of the membranes before onset of labour]. [Online link: http://nvog-documenten.nl/index.php?pagina=/richtlijn/pagina.php&fSelectTG_62=75&fSelectedSub=62&fSelectedParent=75, accessed March 9th 2011.

[2] J.M. Ernest. Neonatal consequences of preterm PPROM. Clin Obstet Gynecol. 1998;41:827-831 Crossref.

[3] A. Tabor, J. Philip, M. Madsen, J. Bang, E.B. Obel, B. Nørgaard-Pedersen. Randomised controlled trial of genetic amniocentesis in 4606 low-risk women. Lancet. 1986;1(8493):1287-1293 Crossref.

[4] J.H. Harger, A.W. Hsing, R.E. Tuomala, et al. Risk factors for preterm premature rupture of fetal membranes: a multicenter case-control study. Am J Obstet Gynecol. 1990;163:130-137 Crossref.

[5] J.D. Iams, R.L. Goldenberg, P.J. Meis, et al. The length of the cervix and the risk of spontaneous premature delivery. National Institute of Child Health and Human Development Maternal Fetal Medicine Unit Network. N Engl J Med. 1996;334:567-572

[6] American College of Obstetricians and Gynecologists Committee on Practice Bulletins-Obstetrics. Society for Maternal-Fetal Medicine, ACOG Joint Editorial Committee. ACOG Practice Bulletin #56: Multiple gestation: complicated twin, triplet, and high-order multifetal pregnancy. Obstet Gynecol. 2004;104:869

[7] T.A. Manuck, A.G. Eller, M.S. Esplin, G.J. Stoddard, M.W. Varner, R.M. Silver. Outcomes of expectantly managed preterm premature rupture of membranes occurring before 24 weeks of gestation. Obstet Gynecol. 2009;114:29-37

[8] G. Tews, O. Shebl, T. Ebner, M. Sommergruber, K. Jesacher. Premature rupture of membranes with oligo- or anhydramnios before 24 weeks of gestation and the chances of fetal survival. Wien Klin Wochenschr. 2004;116:692-694 Crossref.

[9] H. Dewan, J.M. Morris. A systematic review of pregnancy outcome following preterm premature rupture of membranes at a previable gestational age. Aust N Z J Obstet Gynaecol. 2001;41:389-394 Crossref.

[10] A. Farooqi, P.A. Holmgren, S. Engberg, F. Serenius. Survival and 2-year outcome with expectant management of second-trimester rupture of membranes. Obstet Gynecol. 1998;92:895-901 Crossref.

[11] F. Botet, V. Cararach, J. Sentis. Premature rupture of membranes in early pregnancy. Neonatal prognosis. J Perinat Med. 1994;22:45-52 Crossref.

[12] J.M. Klein. Neonatal morbidity and mortality secondary to premature rupture of membranes. Obstet Gynecol Clin North Am. 1992;19:265-280

[13] G. Body, F. Forveille, J. Lemseffer, et al. Spontaneous rupture of the membranes before 28 weeks of amenorrhea. Obstetrical and perinatal outcome. Apropos of 28 cases. J Gynecol Obstet Biol Reprod (Paris). 1991;20:93-100 [Article in French]

[14] C.A. Major, J.L. Kitzmiller. Perinatal survival with expectant management of midtrimester rupture of membranes. Am J Obstet Gynecol. 1990;163:838-844 Crossref.

[15] S.N. Beydoun, S.Y. Yasin. Premature rupture of the membranes before 28 weeks: conservative management. Am J Obstet Gynecol. 1986;155:471-479

[16] A.J. Rudolph, C.A. Smith. Idiopathic respiratory distress syndrome of the newborn. J Pediatr. 1960;57:905-921 Crossref.

[17] J.M. Rennie, N.R.C. Roberton. Textbook of Neonatology. 3rd ed. (Churchill Livingstone, Edinburgh, 1999) 481

[18] H.L. Halliday, R.A. Ehrenkranz, L.W. Doyle. Early (<8 days) postnatal corticosteroids for preventing chronic lung disease in preterm infant. Cochrane Database Syst Rev. 2010;1:CD001146

[19] J.J. Volpe. Intracranial hemorrhage: Germinal matrix-intraventricular hemorrhage. Neurology of the Newborn 4th ed. (WB Saunders, Philadelphia, 2001) 428

[20] M.J. Bell, J.L. Ternberg, R.D. Feigin, et al. Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging. Ann Surg. 1978;187:1 Crossref.

[21] B. Goldstein, B. Giroir, A. Randolph. International Consensus Conference on Pediatric Sepsis. International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med. 2005;6:2-8 Crossref.

[22] E.W. Steyerberg, A.J. Vickers, N.R. Cook, et al. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21:128-138 Crossref.

[23] Y.J. Blumenfeld, H.C. Lee, J.B. Gould, E.S. Langen, A. Jafari, Y.Y. El-Sayed. The effect of preterm premature rupture of membranes on neonatal mortality rates. Obstet Gynecol. 2010;116:1381-1386 Crossref.

[24] E. Azria, O. Anselem, T. Schmitz, V. Tsatsaris, M.V. Senat, F. Geoffinet. Comparison of perinatal outcome after pre-viable preterm prelabour rupture of the membranes in two centres with different rates of termination of pregnancy. BJOG. 2012;119:449-457 Crossref.

[25] A.F. Borgida, A.A. Mills, D.M. Feldman, J.F. Rodis, J.F. Egan. Outcome of pregnancies complicated by ruptured membranes after genetic amniocentesis. Am J Obstet Gynecol. 2000;183:937-939 Crossref.

[26] B.M. Mercer, M. Miodovnik, G.R. Thurnau, et al. Antibiotic therapy for reduction of infant morbidity after preterm premature rupture of the membranes. A randomized controlled trial. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. JAMA. 1997;278:989-995 Crossref.

[27] S. Kenyon, M. Boulvain, J. Neilson. Antibiotics for preterm premature rupture of membranes. Cochrane Database Syst Rev. 2001;(4):CD001058

return to article outline

Footnotes

a Department of Obstetrics and Gynecology, Maastricht University Medical Centre, GROW – School for Oncology and Developmental Biology, The Netherlands Department of Obstetrics and Gynecology, Maastricht University Medical Centre, GROW – School for Oncology and Developmental Biology, The Netherlands

b Department of Obstetrics and Gynecology, Martini Hospital, Groningen, The Netherlands Department of Obstetrics and Gynecology, Martini Hospital, Groningen, The Netherlands

c Department of Epidemiology, Maastricht University Medical Centre, The Netherlands Department of Epidemiology, Maastricht University Medical Centre, The Netherlands

d Department of Obstetrics and Gynecology, VieCuri Medical Centre, Venlo, The Netherlands Department of Obstetrics and Gynecology, VieCuri Medical Centre, Venlo, The Netherlands

e Department of Obstetrics and Gynecology, Academic Medical Centre, Amsterdam, The Netherlands Department of Obstetrics and Gynecology, Academic Medical Centre, Amsterdam, The Netherlands

f Department of Obstetrics and Gynecology, Máxima Medical Centre, Veldhoven, The Netherlands Department of Obstetrics and Gynecology, Máxima Medical Centre, Veldhoven, The Netherlands

g Department of Neonatology, Máxima Medical Centre, Veldhoven, The Netherlands Department of Neonatology, Máxima Medical Centre, Veldhoven, The Netherlands

  Corresponding author at: Maastricht University Medical Centre, GROW – School for Oncology and Developmental Biology, Department of Obstetrics and Gynecology, Postbus (P.O. Box) 5800, 6202 AZ Maastricht, The Netherlands. Tel.: +31 433874800; fax: +31 433875730. 

Article information

PII: S0301-2115(13)00272-8

DOI: 10.1016/j.ejogrb.2013.06.012

© 2013 Published by Elsevier B.V.