Background
Preterm premature rupture of membranes complicates approximately 3% of pregnancies. Currently, in the absence of chorioamnionitis or placental abruption, expectant management, including antenatal steroids for lung maturation and prophylactic antibiotic treatment, is recommended. The benefits of individualized management have not been adequately explored.
Objective
This study aimed to compare the impact of 2 different management strategies of preterm premature rupture of membranes in 2 tertiary obstetrical centers on latency of >7 days, latency to birth, chorioamnionitis, funisitis, and short-term adverse maternal and neonatal outcomes.
Study Design
This was a multicenter retrospective study of women with singleton pregnancies with preterm premature rupture of membranes from 23 0/7 to 33 6/7 weeks of gestation between 2014 and 2018 and undelivered within 24 hours after hospital admission managed at Sunnybrook Health Sciences Center, Toronto, Canada (standard management group), and BCNatal (Hospital Clínic of Barcelona and Hospital Sant Joan de Déu Barcelona), Barcelona, Spain (individualized management group), following local protocols. The standard management group received similar management for all patients, which included a standard antibiotic regimen and routine maternal and fetal surveillance, whereas the individualized management group received personalized management on the basis of amniocentesis at hospital admission (if possible), to rule out microbial invasion of the amniotic cavity and targeted treatment. The exclusion criteria were cervical dilatation >2 cm, active labor, contraindications to expectant management (acute chorioamnionitis, placental abruption, or abnormal fetal tracing), and major fetal anomalies. The primary outcome was latency of >7 days, and the secondary outcomes included latency to birth, chorioamnionitis, and short-term adverse maternal and neonatal outcomes. Statistical comparisons between groups were conducted with propensity score weighting.
Results
A total of 513 pregnancies with preterm premature rupture of membranes were included in this study: 324 patients received standard management, and 189 patients received individualized management, wherein amniocentesis was performed in 112 cases (59.3%). After propensity score weighting, patients receiving individualized management had a higher latency of >7 days (76.0% vs 41.6%; P <.001) and latency to birth (18.1±14.7 vs 9.7±9.7 days; P <.001). Although a higher rate of clinical chorioamnionitis was suspected in the individualized management group than the standard group (34.5% vs 22.0%; P <.01), there was no difference between the groups in terms of histologic chorioamnionitis (67.2% vs 73.4%; P =.16), funisitis (57.6% vs 58.1%; P =.92), or composite infectious maternal outcomes (9.1% vs 7.9%; P =.64). Prolonged latency in the individualized management group was associated with a significant reduction of preterm birth at <32 weeks of gestation (72.1% vs 90.5%; P <.001), neonatal intensive care unit admission (75.6% vs 83.0%; P =.046), and neonatal respiratory support at 28 days of life (16.1% vs 26.1%; P <.01) compared with that in the standard management group. Moreover, prolonged latency was not associated with neonatal severe morbidity at discharge (survival without severe morbidity, 80.4% vs 73.5%; P =.09).
Conclusion
Individualized management of preterm premature rupture of membranes may prolong pregnancy and reduce preterm birth at <32 weeks of gestation, the need for neonatal support, and neonatal intensive care unit admissions, without an increase in histologic chorioamnionitis, funisitis, neonatal infection–related morbidity, and short-term adverse maternal and neonatal outcomes.
Why was this study conducted?
The optimal care for persons with preterm premature rupture of the membranes (PPROM) is still under debate. This study investigated the impact of 2 different management strategies on latency of >7 days, latency to birth, chorioamnionitis, and other maternal and neonatal outcomes.
Key findings
Individualized management, including amniotic fluid culture and targeted treatment, resulted in prolongation of pregnancy and a reduction in preterm birth at <32 weeks of gestation, neonatal intensive care unit admission, and the need for respiratory support, without increasing histologic chorioamnionitis, funisitis, neonatal infection–related morbidity, or other short-term adverse maternal or neonatal outcomes.
What does this add to what is known?
An individualized approach to managing PPROM may be explored as an alternative to standardized care.
Introduction
Preterm premature rupture of the membranes (PPROM) complicates approximately 3% of pregnancies and contributes to one-third of spontaneous preterm deliveries. Multiple factors contribute to PPROM, individually or in combination; however, an exaggerated inflammatory response or infection, which is largely subclinical, is often the primary process involved. Microbial invasion of the amniotic cavity (MIAC) is identified by amniocentesis in up to one-half of all PPROM cases, particularly at earlier gestational ages (GAs), and in 70% of patients with PPROM going into labor. PPROM is associated with high neonatal mortality and short- and long-term severe neonatal morbidities, such as periventricular leukomalacia, bronchopulmonary dysplasia, necrotizing enterocolitis (NEC), and retinopathy of prematurity (ROP). GA at the time of PPROM and GA at birth represent the primary predictors of short- and long-term neonatal outcomes, superseding the risk factor for intra-amniotic infection. , , The PPROM event is followed by a period of latency, or duration of rupture before the onset of labor, ranging from hours to several weeks. The median latency from PPROM to delivery is negatively correlated with GA at PPROM and oligohydramnios. ,
Because prematurity confers most of the fetal and neonatal risks in cases of PPROM, expectant management in the absence of contraindications, such as infection, placental abruption, and cord accidents, remains the most crucial risk reduction strategy and represents the current standard of care in patients with PPROM, according to most international guidelines. , This approach to PPROM management has not dramatically changed in the last 30 years. Although the latency period provides an opportunity to administer risk-reducing interventions before birth, which include antenatal corticosteroids and broad-spectrum antimicrobial therapy, a preferred antibiotic regimen is still not clear ( Supplementary Table 1 ). , An alternative approach to management includes individualizing antibiotic therapy based on the most common microorganisms isolated in the amniotic fluid to maximize the benefit of treatment, prevent antibiotic resistance, and potentially decrease the risk of superimposed infection associated with prolonged therapy and changes in vaginal microbiota. , Amniocentesis has the potential to determine and rule out MIAC , and select the subgroup of women with PPROM in which interventions to prolong pregnancy can be safely considered, particularly at an earlier GA. A randomized trial that included 47 participants reported a reduction of fetal distress in labor and a shorter neonatal hospital stay in the group of PPROM managed with amniocentesis than those managed without the knowledge of MIAC. However, despite the high success rate and low risk of amniocentesis when performed in cases of PPROM in the third trimester of pregnancy, current guidelines consider its benefit scarce and therefore do not routinely recommend its use. , ,
Here, we compared the impact of 2 different management strategies of PPROM (standard vs individualized management bundles) practiced in 2 tertiary referral centers on latency of >7 days, latency to birth, risk of chorioamnionitis and funisitis, and other short-term adverse maternal and neonatal outcomes, to suggest an alternative approach to the care of women presenting with PPROM.
Material and Methods
Study design and participants
This retrospective 2-center study was approved by the institutional review boards at both participating institutions (Sunnybrook, Toronto, Canada [approval number 3322; approval date February 27, 2019], and Hospital Clínic, Barcelona, Spain [approval number HCB/2018/0910; approval date October 23, 2018]). All consecutive women with singleton pregnancies admitted to the High-Risk Unit at Sunnybrook Health Sciences Centre, Toronto, Canada (standard management group), and BCNatal (Hospital Clinic of Barcelona and Hospital Sant Joan de Déu Barcelona), Barcelona, Spain (individualized management group), from 2014 to 2018, with the diagnosis of PPROM occurring from 23 0/7 to 33 6/7 weeks of gestation were included. This study period was chosen because there was no major change in the maternal and neonatal management protocols in either institution during this period.
PPROM was diagnosed as the presence of visual pooling of amniotic fluid during sterile speculum examination, a positive nitrazine test and/or ferning, or a positive result on the insulin-like growth factor–binding protein 1 or placental alpha macroglobulin-1 protein detection test. Pregnant women presenting at hospital admission with cervical dilatation >2 cm, with active labor or delivering within 24 hours from hospital admission, with a contraindication to expectant management (clinical chorioamnionitis, placental abruption, or abnormal fetal tracing), or with a fetus carrying major fetal genetic or structural abnormalities or deciding for termination of pregnancy were excluded.
Clinical chorioamnionitis was diagnosed on the basis of the Gibbs criteria by the presence of maternal fever (temperature >37.8°C) accompanied by one or more of the following criteria: maternal tachycardia (heart rate >100 bpm), uterine tenderness; foul-smelling amniotic fluid, fetal tachycardia (heart rate >160 bpm), and maternal leukocytosis (leukocyte count >15,000 cells/mm 3 ). Placental abruption was suspected when persistent vaginal bleeding occurred in the absence of placenta previa and in association with abdominal pain and/or abnormal fetal tracing. Abnormal fetal tracing was defined on the basis of persistent abnormal baseline, absence of variability, and presence of repetitive decelerations.
Exposure and outcomes
The primary exposure was the PPROM management protocol, which differed between the 2 groups as detailed in Figure 1 . In summary, the main differences between protocols were that pregnant women in the individualized management group received (1) amniocentesis for amniotic fluid culture to exclude MIAC, performed whenever possible to guide the decision regarding expectant management and choice of antibiotic treatment; (2) broader spectrum empirical treatment at presentation, followed by an individualized antibiotic treatment based on amniocentesis results when available; (3) a repeated course of tocolysis in the event of resumption of uterine activity when MIAC was excluded; and (4) a rescue course of corticosteroids when indicated ( Figure 1 ).
The primary outcome was latency of >7 days from PPROM, as it has been the outcome traditionally used to evaluate interventions in studies on preterm birth (PTB; eg, the effect of corticosteroids and tocolysis), as prolongation of pregnancy is associated with improved pregnancy outcomes. The secondary outcomes were latency to birth (calculated as the time measured in days from the diagnosis of PPROM-to-birth, excluding cases that delivered after 35 weeks [n=4]); PTB at <32 weeks of gestation; clinical and histologic chorioamnionitis and funisitis (including umbilical cord vasculitis) , ; maternal infectious composite morbidity (defined as the presence of one of more of the following: sepsis, wound infection, endometritis, postpartum hemorrhage, and intensive care unit [ICU] admission); and neonatal outcomes, such as admission to a level III neonatal ICU (NICU), neonatal survival (live-born neonates surviving until discharge), early lung disease (defined as need for positive pressure ventilation [PPV] at 28 days after birth), early (≤3 days) vs late (>3 days) sepsis, neonatal survival without severe morbidity (defined as neonates that survived until discharge without white matter disease [WMD], including the diagnosis at 34 weeks of corrected GA of grade III or IV intraventricular hemorrhage, periventricular leukomalacia, porencephalic cyst, or late ventriculomegaly), NEC (including medical stage I Bell, stage II Bell, and/or requiring surgery), and stage III ROP and/or late lung disease (defined as PPV and/or oxygen requirement at 28 weeks of life and at 36 weeks of corrected GA).
Data collection
Demographic information, risk factors for PPROM, and PTB were recorded for each patient. GA was calculated on the basis of the crown-to-rump length at first-trimester ultrasound. Latency to birth was measured in days from the diagnosis of PPROM-to-birth. In addition, pregnancy characteristics at admission, GA at PPROM, oligohydramnios (defined as a maximum vertical pocket of <2 cm in the individualized management group or as the largest pocket of amniotic fluid <2 cm in depth and <1 cm wide in the standard management group [Chamberlain classification ]), leukocytosis (defined as white blood cell [WBC] count >15,000 cells/mm 3 ), cervical length on ultrasound, amniotic fluid culture results (glucose concentrations, Gram stain, and cultures for genital mycoplasma [Mycoplasma IST 2; bioMérieux, Marcy-l’Étoile, France; for Ureaplasma spp. or Mycoplasma hominis ]), and aerobic (chocolate agar) and anaerobic (Schaedler agar for anaerobes and thioglycollate broth bacteria) results were collected. Microbial profiling using 16S ribosomal RNA, a common method for studying bacterial phylogeny and taxonomy, was used in a small number of cases. Inpatient management (corticosteroids, tocolysis, magnesium sulfate for neuroprotection, and antibiotic treatment), birth outcomes (GA at birth, onset of labor, mode of birth, and indication for birth), and neonatal data (birthweight, sex, and 5-minute Apgar score), and outcomes (neonatal death, early and late lung diseases, NEC, early- and late-onset sepses, and WMD) were reported.
Statistical analysis
Descriptive statistics, including means, standard deviation (SD), and percentages, were examined for all variables of interest.
To reduce confounding and balance some differences between the 2 groups in this observational study, inverse probability of treatment weighting (IPTW) using the propensity score was used. The stabilized IPTW was calculated using logistic regression on the management groups, by controlling for the admission criteria (GA at PPROM, oligohydramnios, and leukocytosis) and maternal risk factors for PPROM. The weights were later applied in group comparisons, multivariable logistic regression, and survival analysis on latency using the Cox model.
Weighted comparisons were made for all primary and secondary outcomes. Kaplan-Meier plot was used to examine the survival function of latency, and a log-rank test was used to compare the latency PPROM-to-birth interval between groups. Furthermore, multivariable analysis was performed to examine group differences in prolongation and latency, adjusting for GA at the time of PPROM, oligohydramnios at admission, and leukocytosis.
We used the observed means and SDs of latency as parameters to evaluate if sample size was able to detect differences in the primary outcome. The results showed that a sample of 158 (79 in each group) would achieve 80% power to detect a 7-day overall difference in latency between the 2 institutions, with the SDs of 10.1 and 14.9 for the 2 institutions.
Statistical analyses were performed using SAS (version 9.4; SAS Institute Inc, Cary, NC). All P values were 2-sided, and a P value of <.05 was considered statistically significant.
Results
Characteristics of the study population
A total of 513 patients with PPROM met the inclusion criteria, of whom 324 received standard management and 189 received individualized management. Table 1 illustrates the demographic variables, risk factors, and clinical characteristics of patients at hospital admission. After propensity score weighting, none of the variables were significantly different between groups ( Table 2 ), which indicated that the data were appropriately balanced for analysis. In both institutions, maternal comorbidities (ie, diabetes mellitus, chronic hypertension, or kidney diseases) were rare (<2%) and similar between groups. Therefore, they were not used in propensity score analysis.
| Characteristics | Before propensity score weighting | After propensity score weighting | ||||
|---|---|---|---|---|---|---|
| Standard management group (n=324) | Individualized management group (n=189) | P value | Standard management group (n=316) | Individualized management group (n=189) | P value | |
| Maternal age (y) | 32±5 | 33±6 | .11 | 33±5 | 33±7 | .93 |
| Gestational age at PPROM (wk) | 27.5±2.5 | 29.0±2.9 | <.001 | 28.0±2.6 | 27.8±3.1 | .41 |
| Nulliparity | 158 (48.8) | 80 (42.3) | .16 | 147 (46.5) | 86 (45.4) | .81 |
| Previous preterm birth | 42 (13.0) | 24 (12.7) | .93 | 40 (12.6) | 23(12.0) | .82 |
| Cerclage | 39 (12.0) | 16 (8.5) | .21 | 35(11.0) | 25 (13.3) | .43 |
| Oligohydramnios | 150 (46.9) | 53 (29.1) | <.001 | 128 (40.6) | 80 (42.4) | .71 |
| Leukocytosis | 68 (21.0) | 24 (12.7) | <.02 | 55 (17.4) | 30 (15.9) | .66 |
All patients received at least 1 dose of corticosteroids, and almost all patients at both sites received 1 complete course of antenatal corticosteroids (322 of 324 [99.4%] in the standard management group vs 183 of 189 [96.8%] in the individualized management group; P =.17). In addition, in the individualized management group, rescue doses were administered in 80 of 189 patients (42.3%). Amniocentesis was performed in almost two-thirds of the cases (112/189 [59.3%]) in the individualized management group with 34 of 112 cases (30.4%) resulting in a positive culture. Ureaplasma spp. was the most common isolated microorganism (22/34 [65%]) of the positive cultures ( Supplemental Table 2 ). In more than a third of the cases (77/189 [40.7%]), amniocentesis was either not feasible because of anhydramnios or in a very small number of cases because it was declined ( Figure 2 ). Cervical length was measured routinely in the individualized management group and only in 52 of 324 cases (16%) in the standard management group. Antibiotic regimens were different in the 2 institutions as shown in Figure 1 , and overall, the duration of the treatment was significantly longer in the standard management group than in the individualized management group (median [interquartile range (IQR)]: 7 [4–10] vs 5 [4–5] days; P <.001). In the individualized management group, the duration of antimicrobial treatment differed according to feasibility or the results of the amniocentesis ( Supplemental Table 3 ). Particularly, the duration of treatment was significantly longer in the subgroup with MIAC than in the subgroup with negative culture (median [IQR]: 6.5 [4–10] vs 5.0 [4–5] days; P <.001). Moreover, in the individualized management group, in 5 of 34 cases (14.7%) of MIAC, the antibiotic treatment was modified according to the result of amniocentesis ( Supplemental Table 2 ). Tocolysis was used in a significantly higher proportion of women in the individualized management group than in the standard management group, either within the first 48 hours after hospital admission (8/324 [2.5%] vs 74/189 [39.2%]; P <.001) or as an additional course of tocolysis in cases of resumption of uterine activity, once infection, placental abruption, and nonreassuring fetal or maternal status were excluded (0/324 [0%] vs 28/189 [14.8%]; P <.001). Both groups used nifedipine or indomethacin as tocolytic agents. In addition, the individualized group used atosiban (an oxytocin receptor antagonist, currently not licensed for use in North America). Magnesium sulfate was used for fetal neuroprotection at both institutions if birth occurred at <32 weeks of gestation, with a significantly increased use rate in the standard management group compared with the individualized management group (207/324 [63.9%] vs 71/189 [37.6%]; P <.001).
Duration of pregnancy after preterm premature rupture of membranes
A significantly higher proportion of patients were still pregnant 7 days after PPROM in the individualized management group than in the standard management group (76.0% vs 41.6%; P <.001) both after propensity score–weighted analysis ( Table 2 ) and multivariable logistic regression, controlling for GA, oligohydramnios, and leukocytosis ( Table 3 ). Similarly, latency to birth was significantly longer in the individualized management group than in the standard management group (mean±SD: 18.1±14.7 vs 9.7±9.7 days; P <.001), and consequently, patients in the individualized management group gave birth at a later GA compared with those in the standard management group ( P <.001) with a significant reduction of approximately 20% of spontaneous PTB before 32 weeks of gestation ( P <0.01) ( Table 2 ). A prolonged latency of >7 days and the latency to birth were noted, even within subgroups that received individualized management on the basis of different treatments received ( Figure 2 ).
| Variable | Standard management group (n=316) | Individualized management group (n=189) | P value |
|---|---|---|---|
| Latency of >7 d | 41.6 | 76.0 | <.001 |
| Latency to delivery (d) | 9.7±9.7 | 18.1±14.7 | <.001 |
| Gestational age at delivery (wk) | 29.4±2.7 | 30.4±2.9 | <.001 |
| Preterm birth at 32 wk | 90.5 | 72.1 | <.001 |
| Neonatal birthweight (g) | 1355±467 | 1521±551 | <.001 |
| Clinical chorioamnionitis | 22.0 | 34.5 | <.01 |
| Histologic chorioamnionitis | 73.4 | 67.2 | .16 |
| Funisitis | 58.1 | 57.6 | .92 |
| Maternal infectious composite morbidity (sepsis, wound infection, endometritis, postpartum hemorrhage, ICU admission) | 7.9 | 9.1 | .64 |
| NICU admission | 83.0 | 75.6 | .046 |
| Positive pressure ventilation or O 2 at 28 d | 26.1 | 16.1 | <.01 |
| Survival (survived until discharge or alive at birth) | 97.1 | 95.2 | .27 |
| Survival without severe morbidity (survived at discharge without any of grade II or IV IVH, WMD, stage II or III NEC, stage III ROP, or BPD [alive at discharge]) | 73.5 | 80.4 | .09 |
| Early-onset sepsis | 2.7 | 2.7 | .98 |
| Late-onset sepsis | 6.0 | 4.4 | .44 |
| Variable Parameter | Latency of >7 d | |
|---|---|---|
| Odds ratio (95% confidence interval) | P value | |
| Individualized management (vs standard management) | 4.64 (3.07–7.02) | <.001 |
| Gestational age at PPROM (wk) | 0.85 (0.79–0.92) | <.001 |
| Oligohydramnios at hospital admission | 0.71 (0.48–1.06) | .09 |
| Leukocytosis at hospital admission | 0.59 (0.35–0.98) | <.043 |
Figure 3 shows the survival curve of latency from PPROM-to-birth, stratified by institution. The log-rank test indicated significant differences in survival rates between the standard management group and the individualized management group ( P <.001). These differences persisted in the Cox model, after controlling for GA at PPROM, oligohydramnios, and leukocytosis (adjusted hazard ratio, 0.51 [95% confidence interval (CI), 0.42–0.63]; P <.001). The difference in latency was even more significant in the subgroup of patients where PPROM occurred before 28 weeks of gestation (median latency: 19 days [IQR, 11–36] in the individualized management group vs 6 days [IQR, 4–13] in the standard management group; P <.001).
Maternal and neonatal outcomes
The proportion induced labor was similar in the 2 institutions (87/324 [26.9%] vs 50/189 [26.5%]; P =.99); however, cesarean deliveries were significantly higher in the standard management group (152/324 [46.6%] vs 48/189 [25.4%]; P <.001).
Clinical chorioamnionitis was suspected more frequently in the individualized management group (34.5% vs 22.0%; P <.01); however, there was no difference between the groups in terms of histologic chorioamnionitis (67.2% vs 73.4%; P =.16) and funisitis (57.6% vs 58.1%; P =.92). Analogously, composite infectious maternal morbidity was low and similar in the 2 groups (9.1% vs 7.9%; P =.64). In contrast, neonates in the individualized management group had higher birthweights ( P <.001) with fewer level III NICU admissions than neonates in the standard management group ( P <.05) ( Table 2 ). Furthermore, neonates in the standard management group were more likely to receive respiratory support and PPV or oxygen at 28 days of life than neonates in the individualized management group ( P <.01). A prolonged latency in the individualized management group did not result in increased neonatal sepsis or severe morbidity at discharge ( Table 2 ).
Outcomes of the subgroup of patients who did not undergo amniocentesis
Table 4 shows the pregnancy and neonatal outcomes of the subgroup of patients in the individualized management group who did not undergo amniocentesis, compared with standard management group, with propensity score weighting Clinical characteristics of these subgroups are presented in Supplemental Table 4 .
| Subgroup of patients with PPROM who did not undergo amniocentesis | Standard management group (n=315) | Individualized management group (n=85) | P value |
|---|---|---|---|
| Latency of >7 d | 41.7 | 77.0 | <.001 |
| Latency (d) | 9.7±9.7 | 16.8±14.1 | <.001 |
| PTB at 32 wk | 90.5 | 79.2 | <.010 |
| Gestational age at delivery(wk) | 29.5±2.7 | 30.2±3.0 | <.04 |
| Neonatal birthweight (g) | 1356±468 | 1472±587 | .09 |
| Clinical chorioamnionitis | 21.7 | 31.2 | .07 |
| Maternal infectious composite morbidity (sepsis, wound infection, endometritis, postpartum hemorrhage, ICU admission) | 7.9 | 15.1 | <.047 |
| Histologic chorioamnionitis | 73.7 | 60.3 | .02 |
| Funisitis | 58.3 | 49.8 | .18 |
| NICU admission | 82.9 | 75.9 | .14 |
| Positive pressure ventilation or O 2 at 28 d | 25.8 | 13.5 | <.02 |
| Survival (survived until discharge or alive at birth) | 97.1 | 92.9 | .07 |
| Survival without severe morbidity (survived at discharge without any of grade II or IV IVH, WMD, stage II or III NEC, stage III ROP, or BPD [alive at discharge]) | 73.7 | 84.2 | .05 |
| Early-onset sepsis | 2.7 | 5.7 | .18 |
| Late-onset sepsis | 6.0 | 6.2 | .95 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
