|
|
| |||||||||||||||||||||||||||||||||||||||||||||
Background Bacterial vaginosis has been associated with preterm birth. In clinical trials, the treatment of bacterial vaginosis in pregnant women who previously had a preterm delivery reduced the risk of recurrence.
Methods To determine whether treating women in a general obstetrical population who have asymptomatic bacterial vaginosis (as diagnosed on the basis of vaginal Gram's staining and pH) prevents preterm delivery, we randomly assigned 1953 women who were 16 to less than 24 weeks pregnant to receive two 2-g doses of metronidazole or placebo. The diagnostic studies were repeated and a second treatment was administered to all the women at 24 to less than 30 weeks' gestation. The primary outcome was the rate of delivery before 37 weeks' gestation.
Results Bacterial vaginosis resolved in 657 of 845 women who had follow-up Gram's staining in the metronidazole group (77.8 percent) and 321 of 859 women in the placebo group (37.4 percent). Data on the time and characteristics of delivery were available for 953 women in the metronidazole group and 966 in the placebo group. Preterm delivery occurred in 116 women in the metronidazole group (12.2 percent) and 121 women in the placebo group (12.5 percent) (relative risk, 1.0; 95 percent confidence interval, 0.8 to 1.2). Treatment did not prevent preterm deliveries that resulted from spontaneous labor (5.1 percent in the metronidazole group vs. 5.7 percent in the placebo group) or spontaneous rupture of the membranes (4.2 percent vs. 3.7 percent), nor did it prevent delivery before 32 weeks (2.3 percent vs. 2.7 percent). Treatment with metronidazole did not reduce the occurrence of preterm labor, intraamniotic or postpartum infections, neonatal sepsis, or admission of the infant to the neonatal intensive care unit.
Conclusions The treatment of asymptomatic bacterial vaginosis in pregnant women does not reduce the occurrence of preterm delivery or other adverse perinatal outcomes.
Bacterial vaginosis affects approximately 800,000 pregnant women per year in the United States, and women with bacterial vaginosis are more likely than women without bacterial vaginosis to have a preterm delivery or a low-birth-weight infant.6,7,8,9,10 If the treatment of bacterial vaginosis were to reduce this risk, as many as 80,000 preterm births, leading to 4000 perinatal deaths and 4000 infants with neurologic abnormalities, might be prevented in the United States each year.9 Among women with bacterial vaginosis who had a prior preterm delivery, the use of metronidazole alone11,12 or metronidazole plus erythromycin13 reduced the risk of recurrent preterm delivery. However, the treatment of bacterial vaginosis with vaginal clindamycin cream in pregnant women at lower risk for preterm delivery did not reduce the incidence of preterm delivery.14,15 To determine whether screening for bacterial vaginosis and systemic treatment of the condition would reduce the risk of preterm delivery, we conducted a trial of metronidazole therapy in pregnant women with asymptomatic bacterial vaginosis.
Methods
Subjects and Screening
We screened women who had completed between 8 weeks 0 days of gestation and 22 weeks 6 days of gestation for bacterial vaginosis and Trichomonas vaginalis infection. Women were ineligible for screening if they reported any of the following: increased vaginal discharge with itching, burning, or odor; an allergy to metronidazole; current abuse of ethanol; antibiotic therapy within the previous 14 days; an intention to receive antenatal care or to deliver the infant at a location where the follow-up visit could not be completed or from which information on delivery could not be obtained; planned antibiotic therapy before delivery (excluding intrapartum antibiotic prophylaxis); current or planned cervical cerclage; preterm labor before screening; current or planned tocolytic-drug therapy; fetal death or known life-threatening fetal anomaly; multifetal gestation; or medical illnesses (such as hypertension, preexisting diabetes mellitus, or asthma) that required long-term or intermittent drug therapy.
One Dacron swab, taken from the junction of the upper third and lower two thirds of the lateral vaginal wall, was rolled on a glass slide and then touched to a pH stick (ColorpHast pH stick, Curtin Matheson, Grand Prairie, Tex.). The slides from women whose vaginal pH was higher than 4.4 were shipped to the laboratory of one of the authors, where they underwent Gram's staining, with the results interpreted according to the criteria of Nugent et al.16 The scoring system is detailed in Table 1. In accordance with our previous work,8 we defined bacterial vaginosis as a Gram's-staining score of 7 or higher in conjunction with a vaginal pH higher than 4.4. Slides from women with a vaginal pH of 4.4 or lower were discarded, because, according to our definition, these women did not have bacterial vaginosis. An additional swab was inoculated into Diamond's medium for the isolation of T. vaginalis.
|
Randomization and Follow-Up Visits
The women underwent ultrasonography, if they had not already done so, to confirm the gestational age of the fetus, as estimated from the last menstrual period. At randomization, vaginal samples were obtained for measurement of pH, for Gram's staining, and for T. vaginalis cultures; the results were reported to the biostatistical coordinating center but not to the clinical site.
After these specimens were obtained, the women were randomly assigned in a double-blind manner to receive eight capsules, each of which contained either 250 mg of metronidazole or a lactose placebo. The capsules were prepared by placing either a generic metronidazole tablet or a placebo tablet in a capsule and filling the remainder with lactose, so that they were identical in appearance. The capsules were ingested in the presence of study personnel. The women were given an additional eight capsules that contained the same substance as previously assigned, to be taken 48 hours later. In a meta-analysis, a similar two-dose regimen was found to have an effectiveness similar to that of the standard seven-day regimen of metronidazole,17 and in a pilot study we found that this regimen reduced the Gram's-staining score to less than 7 in 100 percent of 33 women and to 4 or less in 89 percent of the women. The urn method of randomization,18 with stratification according to clinical center, was used to create the computer-generated randomization sequence.
One follow-up visit was scheduled between 24 weeks 0 days of gestation and 29 weeks 6 days of gestation, at least 14 days after the initial visit. The types of specimens that were collected at the base-line visit were collected again at the follow-up visit. The personnel in the clinic were again unaware of the results of the assays. All women were treated again with the same two-dose regimen received initially, regardless of the results of follow-up Gram's staining. Study personnel questioned the women about whether they had taken the second dose, which was to be taken 48 hours after the initial dose; about any side effects of the first two doses; and about whether they had received clinically indicated antibiotics after randomization.
Assessment of Outcome
The gestational age of the fetus at the time of randomization was determined from the last menstrual period, provided that the estimate based on the last menstrual period and the estimate based on the ultrasound results agreed within 7 days, if ultrasonography was performed at less than 20 weeks' gestation, or within 14 days, if it was performed at or after 20 weeks' gestation. When there was disagreement between the two estimates, gestational age at randomization was determined from the results of the first ultrasonographic study performed during pregnancy, and gestational age at delivery was determined from the length of time between randomization and delivery. Preterm birth was defined as delivery at less than 37 completed weeks (259 days) of gestation.
In addition to the base-line and follow-up visits, the women received the usual prenatal care at their institutions. After delivery, study personnel reviewed all prenatal, delivery, and postpartum records and abstracted the date of delivery, birth weight of the infant, and details of any antibiotic therapy received after randomization through the postpartum period and the dates of and indications for the therapy. Also noted were visits and admissions to the hospital, preterm labor, the use of tocolytic drugs, preterm premature rupture of the membranes (at least one hour before the onset of labor and before 37 weeks' gestation), clinical intraamniotic infection, postpartum endometritis, and neonatal sepsis.
Statistical Analysis
We compared continuous variables using the Wilcoxon rank-sum test and compared categorical variables using chi-square or Fisher's exact tests. Prolongation of pregnancy was assessed by life-table methods, with women entering the life table at the gestational age at randomization and continuing until they gave birth, were lost to follow-up, or reached 37 weeks' gestation, whichever came first. Event-free survival curves were estimated with use of the Kaplan–Meier method, with adjustment to account for differing gestational ages at entry.19 The statistical significance of the difference between the survival curves was assessed with use of the proportional-hazards-model score function test. Before the study started, the group sequential method of Lan and DeMets with the modified O'Brien–Fleming spending function was chosen for the adjustment of the significance level in interim analyses.20 Two interim analyses were performed, with data corresponding to 14 percent and 49 percent of the total planned sample. Therefore, in the final analysis of preterm delivery, two-tailed P values of 0.049 or less, rather than 0.05 or less, were considered significant. For other comparisons, a P value of 0.05 or less was considered significant. An independent data and safety monitoring committee reviewed the interim results.
Results
A total of 29,625 women were considered for a screening examination from May 30, 1995, through January 5, 1998 (Figure 1). Of these, 21,965 women completed the screening examination, 6540 had bacterial vaginosis without trichomoniasis, and 1953 were randomly assigned to receive placebo or metronidazole. The characteristics of the women in the two groups were similar (Table 2). Data on the week of gestation at delivery were missing for 34 women (1.7 percent), 13 in the metronidazole group and 21 in the placebo group (P=0.19).
|
|
A full course of treatment consisted of 32 capsules divided into four doses; women who did not complete the follow-up visit were assumed to have taken no doses after the first. Because the women were not contacted after the follow-up visit, information was not collected regarding compliance with respect to the final (fourth) 2-g dose. For the first three doses, the mean number of capsules taken by the women for whom this information was available was 21.4 in the metronidazole group and 21.7 in the placebo group (P=0.12). All 24 capsules in the first three doses were taken by 78.8 percent of the women in the metronidazole group and 81.8 percent of the women in the placebo group; no women in the metronidazole group and only one in the placebo group did not take any capsules.
A total of 1757 of the 1953 women (90.0 percent) returned for the follow-up visit and provided information on side effects. The reasons for failure to return were loss of contact (114 women), a decision by the woman not to continue in the study (38 women), delivery before the scheduled visit (27 women), and miscellaneous reasons (17 women); there was no significant difference between the groups in the proportion of women who did not have a follow-up visit. Side effects were significantly more common in the metronidazole group (21.6 percent) than in the placebo group (9.1 percent). This finding was attributable primarily to a higher rate of gastrointestinal symptoms (19.7 percent vs. 7.5 percent), particularly vomiting (9.7 percent vs. 2.8 percent), in the metronidazole group. A total of 12.0 percent of the women assigned to metronidazole and 4.9 percent of those assigned to placebo were treated for vaginal yeast infections with topical antifungal drugs (P<0.001).
Occurrence of Preterm Delivery
Outcome data were available for 1919 of the 1953 women (98.3 percent) (Table 3). The frequency of delivery before 37 weeks' gestation did not differ significantly between the metronidazole group and the placebo group (relative risk in the metronidazole group, 1.0; 95 percent confidence interval, 0.8 to 1.2). Similarly, there were no significant differences between the groups in terms of the rate of delivery before 35 or 32 weeks' gestation. The two groups did not differ significantly with regard to low birth weight (<2500 g), very low birth weight (<1500 g), or preterm delivery attributable to spontaneous labor or spontaneous rupture of the membranes. The placebo group and the metronidazole group were compared in a survival analysis (Figure 2). Additional analyses did not identify any subgroup of women in whom metronidazole significantly reduced the occurrence of preterm delivery (Table 4).
|
|
|
Among the women who had follow-up Gram's staining after the first course of treatment, bacterial vaginosis was still present in 188 of 845 women in the metronidazole group (22.2 percent) and 538 of 859 women in the placebo group (62.6 percent). Among the 1687 women in both groups who had follow-up Gram's staining and for whom information on delivery was available, preterm birth occurred in 77 of 718 women who had bacterial vaginosis at follow-up (10.7 percent) and 103 of 969 women whose bacterial vaginosis remitted (10.6 percent) (P=0.95), regardless of treatment.
Other Pregnancy-Related and Neonatal Complications
Treatment with metronidazole did not reduce the occurrence of admission to the hospital for preterm labor or preterm premature rupture of membranes, receipt of tocolytic drugs, vaginal infections that required treatment, clinical intraamniotic infection, or postpartum endometritis (data not shown). The groups did not differ significantly with regard to the passage of meconium, fetal death or neonatal death during the stay in the nursery, admission to the neonatal intensive care unit, or the presence of neonatal sepsis (data not shown).
Discussion
In this clinical trial, the treatment of asymptomatic bacterial vaginosis with metronidazole did not reduce the risk of preterm delivery in women at low risk for preterm delivery or women with a history of preterm delivery. Our results agree with those of McDonald et al.,12 who also reported no reduction in the risk of preterm delivery among pregnant women with bacterial vaginosis who were treated with metronidazole. However, our results with regard to women with a prior preterm delivery disagree with those of several other studies, all of which found a lower risk of recurrent preterm delivery among women with bacterial vaginosis who were treated with metronidazole11,12 or metronidazole and erythromycin13 than among those who received placebo. However, our study differs from the others in several ways. Two studies were of women with previous preterm delivery,11,13 whereas we studied a general obstetrical population; in the third study, bacterial vaginosis was diagnosed on the basis of a positive culture of Gardnerella vaginalis rather than Gram's staining.12
Several criticisms might be made of our study. The therapy consisted of a short course of metronidazole — two 2-g doses taken 48 hours apart at randomization and two more doses at 24 to less than 30 weeks' gestation. We chose this regimen to improve compliance; at least half the therapy could be given in the presence of study personnel. In contrast to our regimen of four 2-g doses, the regimens of metronidazole used in other studies were administered over four days12 or seven days.11,13 Our regimen was similar in efficacy to that used in the other studies in treating bacterial vaginosis, but a longer course might be needed to eradicate organisms from the upper genital tract. Alternatively, an additional antibiotic that has antiinflammatory properties or a different spectrum of activity, such as erythromycin, might be required to reduce the risk of preterm birth.
The administration of therapy earlier or later in pregnancy might have produced different results, because the intrauterine infection associated with bacterial vaginosis may antedate the pregnancy.21 We chose to treat early in the second trimester to avoid fetal exposure to metronidazole in the first trimester and to repeat the regimen late in the second trimester or early in the third trimester so as to spread treatment over as wide a period as practical. There was no difference in the benefit of treatment between women treated before 20 weeks' gestation and those treated at or after 20 weeks, which is when the membranes seal the uterus closed,21 and there was no reduction in the occurrence of delivery at less than 32 weeks' gestation, which is closest to the time treatment was administered. These findings suggest that our timing of treatment was appropriate.
Our results show that screening pregnant women for asymptomatic bacterial vaginosis and treating the condition with a short course of orally administered metronidazole did not reduce the risk of preterm birth despite its effectiveness in eradicating bacterial vaginosis. Although the literature consistently indicates that intrauterine infection and bacterial vaginosis are associated with preterm delivery,3,4,5,6,7,8,9,10 the results of our study do not support the use of metronidazole to prevent preterm delivery among pregnant women with asymptomatic bacterial vaginosis, regardless of whether they are otherwise considered at either high or low risk for preterm delivery.
Supported by grants from the National Institute of Child Health and Human Development (U10 HD21410, U10 HD21414, U10 HD27869, U10 HD27917, U10 HD27905, U10 HD27860, U10 HD27861, U10 HD27883, U10 HD27889, U10 HD27915, U10 HD34122, U10 HD34116, U10 HD34210, U10 HD34208, and U10 HD34136) and the National Institute of Allergy and Infectious Diseases (AI 38514 and U01 HD36801).
We are indebted to P. Hitchcock, D.V.M., for contributing to the design of this study.
Source Information
From the Department of Obstetrics and Gynecology, University of Oklahoma, Oklahoma City (J.C.C.); the National Institute of Child Health and Human Development, Bethesda, Md. (M.A.K., R.P.N.); the Department of Obstetrics and Gynecology, University of Alabama at Birmingham, Birmingham (J.C.H.); the Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Pittsburgh, Pittsburgh (S.L.H., R.P.H.); the Biostatistics Center, George Washington University, Rockville, Md. (E.A.T., M.L.F.); the Department of Obstetrics and Gynecology, Wake Forest University School of Medicine, Winston-Salem, N.C. (J.M.E.); the Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center, Dallas (K.J.L.); the Department of Obstetrics and Gynecology, Thomas Jefferson University, Philadelphia (R.W.); and the Department of Obstetrics and Gynecology, University of Utah, Salt Lake City (M.V.). Other authors were Wayne Trout, M.D., Department of Obstetrics and Gynecology, Ohio State University, Columbus; Atef Moawad, M.D., Department of Obstetrics and Gynecology, University of Chicago, Chicago; Baha M. Sibai, M.D., Department of Obstetrics and Gynecology, University of Tennessee, Memphis; Menachem Miodovnik, M.D., Department of Obstetrics and Gynecology, University of Cincinnati, Cincinnati; Mitchell Dombrowski, M.D., Department of Obstetrics and Gynecology, Wayne State University, Detroit; Mary J. O'Sullivan, M.D., Department of Obstetrics and Gynecology, University of Miami, Miami; J. Peter VanDorsten, M.D., Department of Obstetrics and Gynecology, Medical University of South Carolina, Charleston; Oded Langer, M.D., Department of Obstetrics and Gynecology, University of Texas at San Antonio, San Antonio; and James Roberts, M.D., Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Pittsburgh, Pittsburgh.Presented at the annual meeting of the Society for Maternal–Fetal Medicine, San Francisco, January 21–23, 1999.
Address reprint requests to Dr. Klebanoff at NICHD, NIH, 6100 Executive Blvd., Rm. 7B03 MSC 7510, Bethesda, MD 20892-7510, or at mk90h{at}nih.gov.
References
Other members of the National Institute of Child Health and Human Development Network of Maternal–Fetal Medicine Units are as follows: University of Alabama at Birmingham: R. Copper, A. Northen, W. Andrews; University of Chicago: P. Jones, M. Lindheimer; University of Cincinnati: N. Elder, T. Siddiqi; George Washington University Biostatistics Center: C. MacPherson, S. Leindecker; Magee Women's Hospital: S. Caritis, M. Cotroneo, T. Camon; University of Miami: S. Beydoun, C. Alfonso; National Institute of Child Health and Human Development: C. Catz, D. McNellis, S. Yaffe; Ohio State University: J. Iams, F. Johnson, M. Landon; University of Oklahoma: G. Thurnau, A. Meier; Medical University of South Caroli-na: B. Collins, F. LeBoeuf, R. Newman; University of Tennessee: B. Mercer, R. Ramsey; University of Texas at San Antonio: M. Berkus, S. Nicholson; University of Texas Southwestern Medical Center: M. Sherman, S. Bloom; Thomas Jefferson University: M. DiVito, J. Tolosa; University of Utah: D. Dudley, L. Reynolds; Wake Forest University: P. Meis, E. Mueller-Heubach, M. Swain; and Wayne State University: S. Bottoms (deceased), G. Norman.
| |||||||||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
