Erythromycin therapy in preterm premature rupture of the membranes: A prospective, randomized trial of 220 patients Brian M. Mercer, MD,. Michael L. Moretti, MD,. Rebecca R. Prevost, PharmD," and Baha M. Sibai, MD"
Memphis, Tennessee OBJECTIVES: The use of prophylactic antibiotics in the management of preterm premature rupture of the membranes has not been adequately studied. The purpose of this study was to evaluate the efficacy of oral erythromycin therapy in the prolongation of latency and reduction of infectious morbidity after preterm premature rupture of membranes. STUDY DESIGN: In this randomized, prospective, double-blind, placebo-controlled study, 220 women at 20 to 35 weeks' gestation were evaluated. Subjects received oral erythromycin 333 mg (n = 106) or indistinguishable placebo (n = 114) every 8 hours from randomization to delivery. RESULTS: Prolongation of latency was identified with erythromycin therapy (p = 0.02), particularly for those destined to have chorioamnionitis (p = 0.003) and those with oligohydramnios (p = 0.01). No decrease in the incidence of maternal or neonatal infectious morbidity was seen. CONCLUSIONS: Oral erythromycin delays, but does not prevent, the onset of clinical infection when administered to women with preterm premature rupture of membranes. This regimen does not decrease neonatal morbidity and mortality. (AM J OBSTET GVNECOL 1992;166:794-802.)
Key words: Preterm premature rupture of fetal membranes, erythromycin, antibiotics, infection
Preterm premature rupture of the membranes is a major cause of maternal, fetal, and neonatal morbidity.l.' Neonatal morbidity is mainly related to preterm birth as a result of the short latency from membrane rupture to delivery. Demographic studies have also demonstrated a correlation between preterm premature rupture of the membranes and maternal and neonatal infection, and neonatal sequelae have been demonstrated to increase with maternal infection before delivery.2.5 There have been several studies (controlled and uncontrolled) describing the use of antibiotics in the management of patients with preterm premature rupture of the membranes. In these studies demethylchlortetracycline,6 vaginal nitrofurazone,7 penicillin, kanamycin,S and ampicillin 9 • 10 have been evaluated for their efficacy in reducing maternal and neonatal morbidity. Unfortunately, these studies have included heterogeneous groups of patients at various gestational ages with various risk factors. In addition, some of the agents studied are not currently considered appropriate for use during pregnancy.6. B Moreover, these studies have been nonblinded 9 • 10 or retrospective II or have demons trated no clinical efficacy.7 From the Departments of Obstetrics and Gynecology" and Pharmacy,b University of Tennessee, Memphis. Received for publication April 25 , 1991 .. revised September 13, 1991 .. accepted September 30, 1991. Reprint requests: Brian M. Mercer, MD, Department of Obstetrics and Gynecology, University of Tennessee, Memphis, 853 jefferson Ave., Room £102, Memphis, TN 38103.
611134061
794
In a recent prospective, double-blind study the use of parenteral mezlocillin followed by oral ampicillin until delivery was associated with prolonged latency, increased birth weight, and reduced incidences of chorioamnionitis, endometritis, and suspected neonatal sepsis. 12 In addition, a prospective study conducted at this institution by Amon et al. 9 found that the combination of parenteral and oral ampicillin in patients with premature rupture of the membranes was associated with an increased latency period and reduced neonatal infectious morbidity. At the time of submission of this article there are no published randomized studies of the use of erythromycin in patients with preterm premature rupture of the membranes. The purpose of this prospective, double-blind, placebo-controlled study was to assess whether the use of oral antibiotic therapy in women with preterm premature rupture of the membranes would increase the latency period and reduce the incidence of maternal and neonatal infectious morbidity.
Material and methods This clinical trial was performed at the University of Tennessee, Memphis, a tertiary care perinatal referral center, from March I, 1989, to Aug. 1, 1990. The study was approved by the university'S institutional review board. Each woman who was admitted with a history of preterm premature ru pture of the membranes had a sterile speculum examination for confirmation of membrane rupture and evaluation of cervical dilatation. Mem-
Volume 166 Number 3
Erythromycin therapy in preterm premature rupture of membranes
brane rupture was confirmed by the visualization of amniotic fluid passing from the cervical os or by the presence of a pool of fluid in the posterior fornix that gave positive results with Nitrazine or ferning. Cervical cultures for Neisseria gonorrhoeae and Chlamydia trachomatis and vaginal cultures for group B streptococci were obtained at the time of initial examination. Digital vaginal examinations were not performed in the absence of advanced cervical dilatation . Gestational age was determined by menstrual history, early clinical examination, or ultrasonographic assessment before 24 weeks' gestation. Fetal heart rate and uterine contraction monitoring was performed for evidence of fetal distress and labor. Blood was drawn for complete blood cell count and C-reactive protein. Ultrasonographic examination was performed for estimation of fetal weight, amniotic fluid assessment, and amniocentesis. Amniotic fluid from amniocentesis was assessed for pulmonary maturity (fetal surfactant index ~47, lecithin/sphingomyelin ratio ~2 : I considered mature), Gram stain, culture, and sensitivity. Vaginal pool fluid was sent for pulmonary maturity studies (fetal surfactant index) if amniocentesis was not feasible or was unsuccessful. Catheterized urine was obtained for microscopy, culture, and sensitivity. Women were excluded from study in the presence ofthe following: membrane rupture of >72 hours' duration, cervical dilatation > 4 cm, progressive labor, vaginal bleeding, temperature ~99.0° F, active infection requiring antibiotic therapy, antibiotic therapy within 1 week before admission, active hepatic disease, erythromycin allergy, cervical cerclage, or a medical condition necessitating delivery. Women with fetuses having intrauterine growth retardation «10th percentile), congenital anomalies, or evidence of fetal distress were also excluded from randomization. Women requiring parenteral tocolysis for preterm labor on admission were excluded from randomization unless tocolysis was successful. Consenting women between 20 weeks' and 34 weeks' 6 days' gestation were thell randomized. The randomization scheme was generated by computerized random number tables and administered by the inpatient pharmacy at this institution. To reduce the chance of differing gestational ages within the two study groups, randomization was stratified at 30 weeks' gestational age. Theenteric-coated erythromycin base and placebo were supplied by Boots Pharmaceuticals (Lincolnshire, III .) and packaged by the inpatient pharmacy in indistinguishable plastic wraps. At the time of randomization the patients were assigned by the pharmacist to receive either erythromycin base, :\33 mg, or an indistinguishable placebo tablet every 8 hours until delivery. All investigators, patient care givers, and patients were blinded to the randomization scheme until after completion of the study and data analysis. Subsequent to a minimum of 12 hours of observation,
795
patients not showing evidence of infection, fetal distress, or infection were transferred to the antepartum ward. They were restricted to bed rest with bathroom privileges and were counseled to remain in the hospital until delivery. Digital pelvic examinations were prohibited until such time as a decision was made to allow delivery. Women with immature pulmonary maturity studies (fetal surfactant index < 47 or lecithin/sphingomyelin ratio 100.4 F and continued until the patient remained afebrile for 48 hours. Endometritis was diagnosed in the presence of a persistent temperature of 0 ~ 100.4 F with abnormal uterine tenderness, in the absence of another identifiable source of infection. Neonates were cared for by neonatal subspecialists blinded to the maternal treatment arm. Preterm infants who were delivered >48 hours after membrane rupture and any infant of a mother with intrapartum fever received parenteral ampicillin and gentamicin, pending results of blood cultures and urine Wellcogen testing for group B streptococci; daily complete blood cell count and C-reactive protein testing were also done. Antibiotics were then discontinued if the infant re0
796 Mercer et al.
March 1992 Am J Obstet Gynecol
Table I. Maternal demographic findings on admission
I
Placebo
Erythromycin
I
(n = 114) . (n = 106)
75.5 Race (% black) 74.6 24.1 ± 5.6 23.7 ± 5.7 Age (yr, mean ± SO) 156 ± 37 159 ± 36 Weight (Ib, mean ± SO) 3 ± 1.8 Gravidity (median ± SO) 2 ± 1.4 Gestation at premature rupture 29.8 ± 3.6 30.3 ± 3.4 of membra nest (wk, mean ± SO) 13.9 ± 17.7 13.2 ± 18.7 Latency from premature rupture of membranes to admission (hr, mean ± SO)
Table II. Maternal morbidity after conservative therapy of preterm premature rupture of membranes at 20 to 34 weeks' gestation Placebo
Chorioamnionitis Suspected chorioamnionitis Endometritis Cesarean delivery Abruptio placentae
(n = 112)*
Erythromycin (n = 105)*
(%)
(%)
19.6 14.3
17.1 17.1
6.3 16.1 9.8
6.7 24.8 14.2
*Three patients lost to follow-up (placebo, n = 2; erythromycin, n = 1).
mained clinically stable, without hypothermia, elevated C-reactive protein level, or increased or decreased white blood cell count. Neonatal sepsis was diagnosed when the infant appeared clinically infected, and either cultures (blood, urine, or cerebrospinal fluid) were positive or radiographic findings were diagnostic (pneumonia, necrotizing enterocolitis). Suspected neonatal sepsis was diagnosed when clinical findings of sepsis could not be confirmed on microbiologic or radiographic assessment or in the presence of an abnormal white blood cell count or C-reactive protein level in an otherwise healthy-appearing neonate. Infants diagnosed as having suspected or confirmed sepsis received a full course of parenteral antibiotics. Cranial ultrasonographic examinations were routinely performed on all infants weighing < 1750 gm and on any infant demonstrating suspicious neurologic findings. Statistical analysis was performed by a statistician blinded to the study code with the SAS system for elementary statistical analysis (SAS Institute Inc. , Cary, N.C.). All patients were maintained within their randomization group for analysis, regardless of compliance and results of cultures performed on admission. Comparative data analysis was performed with Fisher's
exact test and the Student t test where appropriate. Life-table analysis with the Wilcoxon statistic was performed for temporal data. A p value of
Memphis, Tennessee OBJECTIVES: The use of prophylactic antibiotics in the management of preterm premature rupture of the membranes has not been adequately studied. The purpose of this study was to evaluate the efficacy of oral erythromycin therapy in the prolongation of latency and reduction of infectious morbidity after preterm premature rupture of membranes. STUDY DESIGN: In this randomized, prospective, double-blind, placebo-controlled study, 220 women at 20 to 35 weeks' gestation were evaluated. Subjects received oral erythromycin 333 mg (n = 106) or indistinguishable placebo (n = 114) every 8 hours from randomization to delivery. RESULTS: Prolongation of latency was identified with erythromycin therapy (p = 0.02), particularly for those destined to have chorioamnionitis (p = 0.003) and those with oligohydramnios (p = 0.01). No decrease in the incidence of maternal or neonatal infectious morbidity was seen. CONCLUSIONS: Oral erythromycin delays, but does not prevent, the onset of clinical infection when administered to women with preterm premature rupture of membranes. This regimen does not decrease neonatal morbidity and mortality. (AM J OBSTET GVNECOL 1992;166:794-802.)
Key words: Preterm premature rupture of fetal membranes, erythromycin, antibiotics, infection
Preterm premature rupture of the membranes is a major cause of maternal, fetal, and neonatal morbidity.l.' Neonatal morbidity is mainly related to preterm birth as a result of the short latency from membrane rupture to delivery. Demographic studies have also demonstrated a correlation between preterm premature rupture of the membranes and maternal and neonatal infection, and neonatal sequelae have been demonstrated to increase with maternal infection before delivery.2.5 There have been several studies (controlled and uncontrolled) describing the use of antibiotics in the management of patients with preterm premature rupture of the membranes. In these studies demethylchlortetracycline,6 vaginal nitrofurazone,7 penicillin, kanamycin,S and ampicillin 9 • 10 have been evaluated for their efficacy in reducing maternal and neonatal morbidity. Unfortunately, these studies have included heterogeneous groups of patients at various gestational ages with various risk factors. In addition, some of the agents studied are not currently considered appropriate for use during pregnancy.6. B Moreover, these studies have been nonblinded 9 • 10 or retrospective II or have demons trated no clinical efficacy.7 From the Departments of Obstetrics and Gynecology" and Pharmacy,b University of Tennessee, Memphis. Received for publication April 25 , 1991 .. revised September 13, 1991 .. accepted September 30, 1991. Reprint requests: Brian M. Mercer, MD, Department of Obstetrics and Gynecology, University of Tennessee, Memphis, 853 jefferson Ave., Room £102, Memphis, TN 38103.
611134061
794
In a recent prospective, double-blind study the use of parenteral mezlocillin followed by oral ampicillin until delivery was associated with prolonged latency, increased birth weight, and reduced incidences of chorioamnionitis, endometritis, and suspected neonatal sepsis. 12 In addition, a prospective study conducted at this institution by Amon et al. 9 found that the combination of parenteral and oral ampicillin in patients with premature rupture of the membranes was associated with an increased latency period and reduced neonatal infectious morbidity. At the time of submission of this article there are no published randomized studies of the use of erythromycin in patients with preterm premature rupture of the membranes. The purpose of this prospective, double-blind, placebo-controlled study was to assess whether the use of oral antibiotic therapy in women with preterm premature rupture of the membranes would increase the latency period and reduce the incidence of maternal and neonatal infectious morbidity.
Material and methods This clinical trial was performed at the University of Tennessee, Memphis, a tertiary care perinatal referral center, from March I, 1989, to Aug. 1, 1990. The study was approved by the university'S institutional review board. Each woman who was admitted with a history of preterm premature ru pture of the membranes had a sterile speculum examination for confirmation of membrane rupture and evaluation of cervical dilatation. Mem-
Volume 166 Number 3
Erythromycin therapy in preterm premature rupture of membranes
brane rupture was confirmed by the visualization of amniotic fluid passing from the cervical os or by the presence of a pool of fluid in the posterior fornix that gave positive results with Nitrazine or ferning. Cervical cultures for Neisseria gonorrhoeae and Chlamydia trachomatis and vaginal cultures for group B streptococci were obtained at the time of initial examination. Digital vaginal examinations were not performed in the absence of advanced cervical dilatation . Gestational age was determined by menstrual history, early clinical examination, or ultrasonographic assessment before 24 weeks' gestation. Fetal heart rate and uterine contraction monitoring was performed for evidence of fetal distress and labor. Blood was drawn for complete blood cell count and C-reactive protein. Ultrasonographic examination was performed for estimation of fetal weight, amniotic fluid assessment, and amniocentesis. Amniotic fluid from amniocentesis was assessed for pulmonary maturity (fetal surfactant index ~47, lecithin/sphingomyelin ratio ~2 : I considered mature), Gram stain, culture, and sensitivity. Vaginal pool fluid was sent for pulmonary maturity studies (fetal surfactant index) if amniocentesis was not feasible or was unsuccessful. Catheterized urine was obtained for microscopy, culture, and sensitivity. Women were excluded from study in the presence ofthe following: membrane rupture of >72 hours' duration, cervical dilatation > 4 cm, progressive labor, vaginal bleeding, temperature ~99.0° F, active infection requiring antibiotic therapy, antibiotic therapy within 1 week before admission, active hepatic disease, erythromycin allergy, cervical cerclage, or a medical condition necessitating delivery. Women with fetuses having intrauterine growth retardation «10th percentile), congenital anomalies, or evidence of fetal distress were also excluded from randomization. Women requiring parenteral tocolysis for preterm labor on admission were excluded from randomization unless tocolysis was successful. Consenting women between 20 weeks' and 34 weeks' 6 days' gestation were thell randomized. The randomization scheme was generated by computerized random number tables and administered by the inpatient pharmacy at this institution. To reduce the chance of differing gestational ages within the two study groups, randomization was stratified at 30 weeks' gestational age. Theenteric-coated erythromycin base and placebo were supplied by Boots Pharmaceuticals (Lincolnshire, III .) and packaged by the inpatient pharmacy in indistinguishable plastic wraps. At the time of randomization the patients were assigned by the pharmacist to receive either erythromycin base, :\33 mg, or an indistinguishable placebo tablet every 8 hours until delivery. All investigators, patient care givers, and patients were blinded to the randomization scheme until after completion of the study and data analysis. Subsequent to a minimum of 12 hours of observation,
795
patients not showing evidence of infection, fetal distress, or infection were transferred to the antepartum ward. They were restricted to bed rest with bathroom privileges and were counseled to remain in the hospital until delivery. Digital pelvic examinations were prohibited until such time as a decision was made to allow delivery. Women with immature pulmonary maturity studies (fetal surfactant index < 47 or lecithin/sphingomyelin ratio 100.4 F and continued until the patient remained afebrile for 48 hours. Endometritis was diagnosed in the presence of a persistent temperature of 0 ~ 100.4 F with abnormal uterine tenderness, in the absence of another identifiable source of infection. Neonates were cared for by neonatal subspecialists blinded to the maternal treatment arm. Preterm infants who were delivered >48 hours after membrane rupture and any infant of a mother with intrapartum fever received parenteral ampicillin and gentamicin, pending results of blood cultures and urine Wellcogen testing for group B streptococci; daily complete blood cell count and C-reactive protein testing were also done. Antibiotics were then discontinued if the infant re0
796 Mercer et al.
March 1992 Am J Obstet Gynecol
Table I. Maternal demographic findings on admission
I
Placebo
Erythromycin
I
(n = 114) . (n = 106)
75.5 Race (% black) 74.6 24.1 ± 5.6 23.7 ± 5.7 Age (yr, mean ± SO) 156 ± 37 159 ± 36 Weight (Ib, mean ± SO) 3 ± 1.8 Gravidity (median ± SO) 2 ± 1.4 Gestation at premature rupture 29.8 ± 3.6 30.3 ± 3.4 of membra nest (wk, mean ± SO) 13.9 ± 17.7 13.2 ± 18.7 Latency from premature rupture of membranes to admission (hr, mean ± SO)
Table II. Maternal morbidity after conservative therapy of preterm premature rupture of membranes at 20 to 34 weeks' gestation Placebo
Chorioamnionitis Suspected chorioamnionitis Endometritis Cesarean delivery Abruptio placentae
(n = 112)*
Erythromycin (n = 105)*
(%)
(%)
19.6 14.3
17.1 17.1
6.3 16.1 9.8
6.7 24.8 14.2
*Three patients lost to follow-up (placebo, n = 2; erythromycin, n = 1).
mained clinically stable, without hypothermia, elevated C-reactive protein level, or increased or decreased white blood cell count. Neonatal sepsis was diagnosed when the infant appeared clinically infected, and either cultures (blood, urine, or cerebrospinal fluid) were positive or radiographic findings were diagnostic (pneumonia, necrotizing enterocolitis). Suspected neonatal sepsis was diagnosed when clinical findings of sepsis could not be confirmed on microbiologic or radiographic assessment or in the presence of an abnormal white blood cell count or C-reactive protein level in an otherwise healthy-appearing neonate. Infants diagnosed as having suspected or confirmed sepsis received a full course of parenteral antibiotics. Cranial ultrasonographic examinations were routinely performed on all infants weighing < 1750 gm and on any infant demonstrating suspicious neurologic findings. Statistical analysis was performed by a statistician blinded to the study code with the SAS system for elementary statistical analysis (SAS Institute Inc. , Cary, N.C.). All patients were maintained within their randomization group for analysis, regardless of compliance and results of cultures performed on admission. Comparative data analysis was performed with Fisher's
exact test and the Student t test where appropriate. Life-table analysis with the Wilcoxon statistic was performed for temporal data. A p value of
