OBJECTIVE: To determine whether prophylactic antibiotic administration in women presenting with premature rupture of membranes (PROM) at term can alter the rates of maternal and neonatal infections.
METHODS: In a randomized, controlled, nonblinded trial, women at low risk with singleton term pregnancies presenting with PROM were assigned to either antibiotic administration or no treatment. Main outcomes evaluated were rates of maternal infection (chorioamnionitis and endometritis) and neonatal infection.
RESULTS: A total of 161 patients were evaluated, 78 in the antibiotic group and 83 in the control group. Maternal infection was significantly lower when antibiotics were administered (2.6% compared with 13.2%; relative risk 0.89, 95% confidence interval 0.81–0.98; P=.013). All cases of maternal infection occurred in women with more than 12 hours of PROM. Newborns of mothers receiving antibiotics had fewer infections (3.8%) compared with those in the control group (6.0%), but the difference was not statistically significant (P=.375).
CONCLUSION: Prophylactic use of antibiotics in PROM at term significantly reduced the risk of maternal infection in our population.
CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov, www.clinicaltrials.gov, NCT01633294.
LEVEL OF EVIDENCE: I
Department of Obstetrics, Gynecology, and Reproductive Medicine and the Department of Pediatrics, Hospital de Santa Maria, Lisboa, Portugal.
Corresponding author: Filipa Faria Vaz Passos, MD, Rua Dom António Ribeiro, N° 5–6° Dto, 1495–049 Algés; e-mail: firstname.lastname@example.org.
The authors thank Professor Carlos Calhaz-Jorge for his help analyzing and interpreting the data of our study and for revising this manuscript critically for important intellectual content, and Teresa Fonseca and Maria Clara Barata for their help with the statistical analyses.
Financial Disclosure The authors did not report any potential conflicts of interest.
The optimal management of women with term premature rupture of membranes (PROM) remains a controversial issue in obstetrics, although it complicates approximately 8% of term pregnancies.1 Numerous studies have demonstrated that prolongation of latency more than 24 hours is associated with an increased incidence of chorioamnionitis and neonatal sepsis.2,3 The risk of maternal and neonatal infectious complications is reduced in PROM when antibiotic prophylaxis is used. However, there is no consensus regarding its use in term rupture of membranes.1 Therefore, many authors recommend induction of labor when active labor does not begin spontaneously soon afterward.4,5 Others prefer to wait for spontaneous labor, as long as there is no evidence of fetal or maternal infection, in the hope of lowering the risk of a cesarean delivery.6,7 The TermPROM study showed that expectant management was associated with a significantly increased incidence of clinical chorioamnionitis, postpartum fever, and longer maternal hospital stay compared with labor induction, but there was no difference in neonatal infection rate and cesarean delivery rate.8 Therefore, induction of labor could be an efficient strategy to reduce infectious morbidity associated with term PROM. Alternatively, we could routinely use antibiotics for women at the time of term PROM. The use of antibiotics in PROM at term has been addressed in a Cochrane review in 2002, which included only two well-designed randomized trials by Cararach et al9 and by Ovalle et al.10 The authors' conclusion is that routine antibiotics for term PROM reduce maternal infectious morbidity but have no neonatal benefits.11 They also suggest that the modest benefit for women seen within these trials might be greater in circumstances in which the duration of membrane rupture was more prolonged (related to either a policy of expectant management or a delayed induction).
Therefore, the aim of this study is to determine whether antibiotics administered routinely in women presenting with PROM later than 37 0/7 weeks of gestation can alter the rates of maternal and neonatal infection.
MATERIALS AND METHODS
This randomized, controlled, nonblinded trial was conducted at the Department of Obstetrics, Gynecology, and Reproductive Medicine, Lisbon University Hospital, from October 2008 to January 2012. The Ethics Committee of the institution approved the study protocol. Women were eligible for entry into the study if they had a term (37 0/7 weeks of gestation or more) singleton pregnancy, a vertex presentation, ruptured membranes for less than 12 hours, and a negative group B streptococcus culture performed between 35 and 37 weeks of gestation. Criteria precluding enrollment were active labor, absence of group B streptococcus culture, or indication for group B streptococcus antibiotic prophylaxis (such as maternal group B streptococcus colonization between 35 and 37 weeks of gestation, group B streptococcus bacteriuria, previous newborn with group B streptococcus sepsis) or contraindication to expectant management (such as fetal distress, meconium staining of the amniotic fluid, or chorioamnionitis) or to vaginal delivery. The diagnosis of PROM was established if gross amniotic fluid leakage was observed after a vaginal sterile speculum was inserted, or in the presence of anamnesis suggestive of fluid leakage and ultrasonographic diagnosis of oligohydramnios previously not documented. Digital examination was not performed unless the mother felt regular painful contractions, ie, at least two in 10 minutes. Patients who met the criteria for inclusion in the study after providing written informed consent were randomly assigned to the antibiotic group or to the control group, according to a computer-generated randomization list. The treatment arm assignments were contained in sequentially numbered opaque envelopes. Patients randomized to receive antibiotic treatment were administered ampicillin 1 g every 6 hours and gentamicin 240 mg every day intravenously. This combination of antibiotics is the standard regimen for treating chorioamnionitis at our center. In the control group, if clinical findings of chorioamnionitis were detected, the same combination of antibiotics was immediately initiated. Maternal blood analyses, including white blood cell count and C-reactive protein levels, were performed twice per day until delivery. Maternal body temperature and heart rate were monitored four times per day. Continuous electronic fetal monitoring was initiated after admission to the labor unit. The decision for initiating labor induction was made at the discretion of the physicians responsible for the patient. In the absence of regular uterine contractions, oxytocin or prostaglandins were used for induction of labor, depending on Bishop score. If Bishop score was 5 or more, oxytocin was initiated and the infusion rate was titrated to contractions, according to local hospital practice; if Bishop score was less than 5, misoprostol (50 micrograms) was inserted into the posterior vaginal fornix and repeated every 6 hours if labor had not started (maximum three doses), or it was followed by oxytocin 6 or more hours later if labor had started. The mode of delivery (ie, vaginal compared with cesarean delivery) was decided considering the safest option for mother and fetus at the discretion of the physicians responsible. In both groups, women with cesarean delivery received a prophylactic dose of antibiotics before skin incision (cefoxitin 2 g intravenously).
For all women entering the study, the following data were recorded: demographic and reproductive characteristics; time from PROM to admission at labor unit; time from PROM to initiation of spontaneous labor (latency time); time from PROM to induction of labor; total time from rupture of membranes to delivery; mode of delivery; and cesarean delivery rate for fetal distress. Collected data related to newborns included: Apgar score at 1 and 5 minutes; birth weight; early-onset neonatal infections; and death. Gestational age was determined by ultrasound examination data before 20 weeks of gestation. Chorioamnionitis was diagnosed on a clinical basis that included maternal fever (tympanic temperature 38.0°C or higher) and two or more of the following findings: uterine tenderness; foul-smelling amniotic fluid; maternal heart rate more than 100 beats per minute; baseline fetal heart rate more than 160 beats per minute for at least 10 minutes; and blood white cell count more than 15,000/mm3 or C-reactive protein more than 2 mg/dL with no other potential source of infection. Postpartum endometritis was diagnosed when any of the following was present: temperature 38.0°C or higher (detected 6 hours apart, after 24 hours postdelivery in the absence of other maternal causes); foul-smelling uterine secretion; and uterine tenderness together with the previously described laboratory signs of infection.
According to our neonatal intensive care unit protocols, all the newborns with perinatal risk factors for early-onset infection or suspicious clinical signs of infection have a septic screening that includes white blood cell count and C-reactive protein. The risk factors are: rupture of membranes more than 18 hours; maternal fever (auricular temperature 38°C or higher during labor); chorioamnionitis as defined; positive maternal septic screening with C-reactive protein more than 2 mg/dL, white blood cell count more than 15,000/mm3, and neutrophilia; untreated maternal urinary tract infection; or family history of previous offspring with group B streptococcus infection. The suspicious clinical signs considered are fever, hypothermia, bradycardia, tachycardia, respiratory distress, apnea, hypoxemia, paleness, or hypotonia. Early neonatal infection was diagnosed if the newborn had laboratory parameters suggestive of bacterial infection during the first 72 hours of life (C-reactive protein more than 1 mg/dL, absolute neutropenia or neutrophilia according to Manroe,12 or thrombocytopenia less than 80,000/mm3), in which case a blood culture was taken and empirical antimicrobial therapy was started. Because neonatal blood cultures have a low sensitivity and because culture-proven sepsis has become rare,13,14 in our study a positive result was not required for the diagnosis of early-onset infection.
Two independent obstetric investigators and one pediatric investigator individually revised all processes to confirm the diagnosis of chorioamnionitis, endometritis, and neonatal infection according to the outlined diagnostic criteria. The pediatric investigator was blind to the treatment arm assignment when reviewing neonatal data. Primary outcomes were maternal infection rate (chorioamnionitis or puerperal endometritis) and neonatal infection rate (early onset sepsis, meningitis, and pneumonia) in women with and without antibiotic prophylaxis.
According to Spybrook et al,15 the sample size was calculated to provide a statistical power of 80% to detect a difference of 50% or more in the rate of maternal and neonatal infections between groups. We expected baseline rates of maternal and neonatal infections of 5.8% and 4.3% in the control group and 2.9% and 2.1% in the antibiotic group, respectively.9,16 Using Optimal Design Software considering our single-level trial, assuming a significance level of 0.05 and a balanced sample, the minimum number of patients would be approximately 128 in total. The results were evaluated according to the intention-to-treat analysis, excluding only those women for whom outcome data were not available.
All data management and analysis were performed using IBM SPSS Statistics 19.0. All tests of significance were two-tailed and used the α level of 0.05. Comparisons between study groups were conducted with Student t test or Mann-Whitney U test, depending on the violation of the normality assumption (Kolmogorov-Smirnov). For categorical variables we used χ2 tests or Fisher exact tests, as appropriate.
During the 3-year study period, 161 patients were enrolled in the trial, 78 in the antibiotic group and 83 in the control group. Five women were excluded because study records were incomplete and clinical records were missing. Six other women in the control group had protocol violations: intrapartum administration of ampicillin because of subclinical chorioamnionitis not meeting the study definition (n=4) or PROM lasting more than 12 hours (n=2). Nevertheless, according to the intention-to-treat analysis, these six women remained in the control study sample (Fig. 1), with four of them achieving criteria of chorioamnionitis after administration of ampicillin. Antibiotic prophylaxis was initiated at admission to the labor unit, a mean time of 145 minutes (minimum 30 minutes to maximum 690 minutes) after PROM.
Maternal demographic and labor characteristics are shown in Table 1. There were no significant differences between the groups. The 161 women had a mean (±standard deviation) length of total time from rupture of membranes to delivery of 17±8 hours in total and in each group. Approximately 26.7% of women delivered less than 12 hours after PROM and 16.8% delivered more than 24 hours after PROM (Table 2).
Spontaneous delivery occurred in 34.6% (27 of 78) and in 42.2% (35 of 83) of women in the antibiotic and control groups (Table 1). Three women in the antibiotic group had no information about labor induction. Sixty-one percent of women in antibiotic group (48 of 78) and 58% (48 of 83) in control group had labor induction, which was initiated at a mean time of 9.2±5.5 hours after rupture of membranes.
There were no significant differences in the rate of cesarean deliveries or in the rate of cesarean delivery for fetal distress. We also compared the length of hospitalization for mothers and neonates. Both were released from the hospital together a mean of 2.7 days after PROM.
The incidence of maternal infection was significantly lower in women who received antibiotics (2.6 compared with 13.2%; relative risk 0.89, 95% confidence interval [CI] 0.81–0.98; P=.013); the same was observed regarding the incidence of chorioamnionitis (2.6% compared with 10.8%; relative risk 0.92, 95% CI 0.84–0.99; P=.037). There were only two cases of endometritis, both in the control group; one occurred after retained placenta, which was treated with manual extraction, and another occurred after cesarean delivery. The incidence of endometritis was not statistically different between the groups (0% compared with 2.4%; P=.500). The number needed to treat to prevent one case of maternal infection was nine (95% CI 5–46). There was no case of maternal infection in the subgroup of women who delivered in the first 12 hours after rupture of membranes, 38.5% (5 of 13) occurred in the 12- to 18-hour interval and 38.5% (5 of 13) occurred in the 18- to 24-hour interval.
In women with labor induction (n=96), maternal infection was significantly more frequent in the control group than in the antibiotic group (8 of 48 [16.7%] women compared with 1 of 48 [2.1%] women; relative risk 0.85, 95% CI 0.74–0.97; P=.03), but there was no difference in neonatal infection rate (Fig. 1). Concerning neonatal data, mean birth weight was similar in antibiotic and control groups (3,261±406 g compared with 3,281±503 g; P=.78). Apgar scores were similar between the groups. Newborns of mothers receiving antibiotics had less neonatal infection (3 of 78; 3.8%) compared with those in the control group (5 of 83; 6%), but the difference was not statistically significant (P=.375). All those considered to have early neonatal infection had negative blood cultures. There were no cases of meningitis or pneumonia. Also, no neonatal mortality occurred in either group.
Our results suggested that antibiotic prophylaxis reduces the rate of maternal infection in women with PROM at term; however, antibiotic prophylaxis had no significant effect on neonatal infection rate. In 1998, Cararach et al,9 using the same study protocol but including women at 36 weeks of gestation or more (n=733), found no difference in maternal infection rate but found a significant difference in neonatal infectious complications. In the same year, Ovalle et al,10 with a population of 105 women and with a similar study protocol, although using cefuroxime-clindamycin in the antibiotic arm, found a significant reduction in the rate of maternal infection-related morbidity but no difference in neonatal outcomes. These contradictory results and the absence of more empirical evidence regarding prophylactic antibiotics since the trials of Cararach and Ovalle highlight the importance of studying this issue.
Our study failed to demonstrate a reduction of 50% in the rate of neonatal infection, although fewer cases of infection occurred in the offspring of women receiving antibiotics. With improved obstetric management, the incidence of early-onset infection in term newborns has been decreasing but it still is an important cause of mortality or serious morbidity.17,18 Over the course of the past decade, the widespread use of intrapartum chemoprophylaxis has resulted in a remarkable decrease in group B streptococcus early-onset infection to a rate of 0.3 of 1,00019 and overall early-onset infection incidence to 0.76 cases per 1,000 live births.20 In our study, the offspring of women in the control group had a rate of neonatal infection higher (6%) than the previously reported ranges. This can be related to the criteria used to diagnose early-onset infection. In epidemiologic studies, positive blood cultures were needed for diagnosis, whereas others using clinical or biochemical markers report rates of early-onset sepsis as high as 4.3%,16 which is closer to the ones we found.
The low rate of puerperal endometritis in both groups can be partly explained by the use of prophylactic antibiotics before skin incision in women undergoing cesarean delivery, supposedly at higher risk for this complication. Even so, a lower, although not significant, rate of endometritis was verified in the antibiotic group (0 compared with 2). The case of puerperal endometritis occurring after manual extraction of placenta could introduce potential bias, but the difference of maternal infections between groups remained significant after excluding this case from statistical analysis.
Our 11% rate of maternal infection in the control group is a middle value between the 5.8% in the study by Cararach and the 16% in the study by Ovalle et al, which used similar criteria, although it is worth adding that our criteria for chorioamnionitis diagnosis were more restrictive. Furthermore, only 27% of the 161 patients delivered in the first 12 hours after PROM. This was much lower than the 50% previously reported.21 Although the majority of women delivered, as expected, in the first 24 hours after PROM, we observed 17% of patients with total time from rupture of membranes to delivery lasting more than 24 hours. These findings might explain the benefit of antibiotic prophylaxis in reducing maternal infection in our population. The study by Cararach9 did not find a statistically significant difference of maternal infection, but only 5% and 6% of women in the antibiotic group and control group, respectively, delivered after 24 hours of PROM. It should be noted that there was no case of maternal infection in the subgroup of women who delivered in the first 12 hours after rupture of membranes. Also, in induced labor, maternal infection was more frequent in the subgroup of women without antibiotic prophylaxis.
One of the shortcomings of our study is that no histologic examinations of the placentas were performed. In fact, there are no universally accepted clinical criteria for diagnosing chorioamnionitis. The term is a histologic one; the clinical diagnosis is usually based on a combination of fever and other clinical signs or symptoms. We tried to define chorioamnionitis as the mandatory presence of fever and at least two more clinical suggestive criteria to minimize false diagnosis. Another limitation that should be noted is that our study was nonblinded and not placebo-controlled, resembling the previous studies.9,10 The fact that the treatment regimens were not blinded creates the possibility of a bias of ascertainment. However, we strictly defined clinical criteria to diagnose chorioamnionitis and two independent investigators revised all processes and uniformly applied these same criteria to confirm the results. Moreover, they also observed that all cases of chorioamnionitis had not only the mandatory presence of fever but also positive biochemical infection parameters, and all are objective criteria of infection. Concerning the absence of a placebo treatment, because the diagnosis of infection was out of voluntary control, interference is not likely. Regarding the neonatal data, the pediatric investigator was blind to the arm assignment when reviewing the processes, which means that there was no interference in the diagnosis of neonatal infection. Strengths of our study include a randomized controlled trial with an intention-to-treat analysis.
In this single-institution study, the incidences of chorioamnionitis and neonatal infection in women with term PROM not submitted to antibiotic prophylaxis were 11% and 6%, respectively. Antibiotic prophylaxis clearly reduced maternal infectious morbidity in women with PROM at term. Regarding neonatal infection and puerperal endometritis, larger studies are needed to corroborate the apparent benefit of antibiotic prophylaxis. The generalization of a prophylactic treatment should take into account the number needed to treat to prevent one case of infection. Additionally, because of increasing problems with bacterial resistance and the exceptional but potentially life-threatening risks of maternal anaphylaxis, it is important to ensure judicious use of antibiotics. A number needed to treat of nine surely justifies this risk. These types of recommendations must be based on the characteristics of the population, notably on the incidence of infection. We had a high incidence of neonatal infection (6%) in the control group and no apparent adverse effects from the use of antibiotics for women or their newborns. Our groups were homogeneous and no demographic or labor characteristic could be inferred as a determinant of infection, other than the use of antibiotics, which enables targeting therapy to a specific susceptible group of women. Therefore, our results suggest the importance of antibiotic prophylaxis in women with term PROM, irrespective of the policy of expectant or induction management.
1. Premature rupture of membranes. Practice Bulletin No. 80. American College of Obstetricians and Gynecologists. Obstet Gynecol 2007;109:1007–19.
2. Gunn GC, Mishell DR Jr, Morton DG. Premature rupture of fetal membranes. A review. Am J Obstet Gynecol 1970;106:469–83.
3. Lanier LR Jr, Scarbrough RW Jr, Fillingim DW, Baker RE Jr. Incidence of maternal and fetal complications associated with rupture of membranes before onset of labor. Am J Obstet Gynecol 1965;93:398–404.
4. Wagner MV, Chin VP, Peters CJ, Drexler B, Newman LA. A comparison of early and delayed induction of labor with spontaneous rupture of the membranes at term. Obstet Gynecol 1989;74:93–7.
5. Rydhström H, Ingemarsson I. No benefit from conservative management in nulliparous women with premature rupture of the membranes (PROM) at term. A randomized study. Acta Obstet Gynecol Scand 1991;70:543–7.
6. Conway DI, Prendiville WJ, Morris A, Speller DC, Stirrat GM. Management of spontaneous rupture of the membranes in the absence of labor in primigravid women at term. Am J Obstet Gynecol 1984;150:947–51.
7. Grant JM, Serle E, Mahmood T, Sarmandal P, Conway DI. Management of prelabour rupture of the membranes in term primigravidae: report of a randomized prospective trial. Br J Obstet Gynaecol 1992;99:557–62.
8. Hannah ME, Ohlsson A, Farine D, Hewson SA, Hodnett ED, Myhr TL, et al.. Induction of labor compared with expectant management for prelabor rupture of the membranes at term. TERMPROM Study Group. N Engl J Med 1996;334:1005–10.
9. Cararach V, Botet F, Sentis J, Almirall R, Pérez-Picañol E. Administration of antibiotics to patients with rupture of membranes at term: a prospective, randomized, multicentric study. Collaborative Group on PROM. Acta Obstet Gynecol Scand 1998;77:298–302.
10. Ovalle A, Gomez R, Martinez MA, Giglio MS, Bianchi R, Diaz J, et al.. Antibiotic treatment of patients with term premature rupture of membranes: a randomized clinical trial. Prenat Neonatal Med 1998;3:599–606.
11. Flenady V, King J Antibiotics for prelabor rupture of membranes at term. The Cochrane Database of Systematic Reviews 2002, Issue 3. Art. No.: CD001807. DOI: 10.1002/14651858.CD001807
12. Manroe BL, Weinberg AG, Rosenfeld CR, Browne R. The neonatal blood count in health and disease. I. Reference values for neutrophilic cells. J Pediatr 1979;95:89–98.
13. Ottolini MC, Lundgren K, Mirkinson LJ, Cason S, Ottolini MG. Utility of complete blood count and blood culture screening to diagnose neonatal sepsis in the asymptomatic at risk newborn. Pediatr Infect Dis J 2003;22:430–4.
14. Connell TG, Rele M, Cowley D, Buttery JP, Curtis N. How reliable is a negative blood culture result? Volume of blood submitted for culture in routine practice in a children's hospital. Pediatrics 2007;119:891–6.
15. Spybrook J, Bloom H, Congdon R, Hill C, Martinez A, Raudenbush S. Optimal design for longitudinal and multilevel research: Documentation for the optimal design software version 3.0 (2011). Available at: http://pikachu.harvard.edu/od/
. Retrieved on February 1, 2012.
16. Popowski T, Goffinet F, Maillard F, Schmitz T, Leroy S, Kayem G. Maternal markers for detecting early onset neonatal infection and chorioamnionitis in cases of premature rupture of membranes at or after 34 weeks of gestation: a two-center prospective study. BMC Pregnancy Childbirth 2011;11:26.
17. Ungerer R, Lincetto O, McGuire W, Saloojee H, Gulmezoglu AM. Prophylactic versus selective antibiotics for term newborn infants of mothers with risk factors for neonatal infection. The Cochrane Database of Systematic Reviews 2004, Issue 4. Art. No.: CD003957. DOI: 10.1002/14651858.CD003957.pub2
18. Polin RA, Committee on Fetus and Newborn. Management of neonates with suspected or proven early-onset bacterial sepsis. Pediatrics 2012;129:1006–15.
19. Perinatal group B streptococcal disease after universal screening recommendations–United States, 2003–2005. MMWR Morb Mortal Wkly Rep 2007;56:701–5.
20. Weston E, Pondo T, Lewis M, Martell-Cleary P, Morin C, Jewell B, et al.. The burden of invasive early-onset neonatal sepsis in the United States, 2005–2008. Pediatr Infect Dis J 2011;30:937–41.
© 2012 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
21. Myers V. Premature rupture of membranes at or near term. In: Berghella, editor. Obstetric Evidence Based Guidelines. 1st ed. Philadelphia (PA): Informa; 2007. p. 162–4.