Antibiotic treatment of bacterial vaginosis to prevent preterm delivery: Systematic review and individual participant data meta-analysis

Background: Bacterial vaginosis (BV) increases preterm delivery (PTD) risk, but treatment trials showed mixed results in preventing PTD. Objectives: Determine, using individual participant data (IPD), whether BV treatment during pregnancy reduced PTD or prolonged time-to-delivery. Data Sources: Cochrane Systematic Review (2013), MEDLINE, EMBASE, journal searches, and searches (January 2013–September 2022) (“bacterial vaginosis AND pregnancy”) of (i) clinicaltrials.gov; (ii) Cochrane Central Register of Controlled Trials; (iii) World Health Organization International Clinical Trials Registry Platform Portal; and (iv) Web of Science (“bacterial vaginosis”). Study Selection and Data Extraction: Studies randomising asymptomatic pregnant individuals with BV to antibiotics or control, measuring delivery gestation. Extraction was from original data files. Bias risk was assessed using the Cochrane tool. Analysis used “one-step” logistic and Cox random effect models, adjusting gestation at randomisation and PTD history; heterogeneity by I2. Subgroup analysis tested interactions with treatment. In sensitivity analyses, studies not providing IPD were incorporated by “multiple random-donor hot-deck” imputation, using IPD studies as donors. Results: There were 121 references (96 studies) with 23 eligible trials (11,979 participants); 13 studies (6915 participants) provided IPD; 12 (6115) were incorporated. Results from 9 (4887 participants) not providing IPD were imputed. Odds ratios for PTD for metronidazole and clindamycin versus placebo were 1.00 (95% CI 0.84, 1.17), I2 = 62%, and 0.59 (95% CI 0.42, 0.82), I2 = 0 before; and 0.95 (95% CI 0.81, 1.11), I2 = 59%, and 0.90 (95% CI: 0.72, 1.12), I2 = 0, after imputation. Time-to-delivery did not differ from null with either treatment. Including imputed IPD, there was no evidence that either drug was more effective when administered earlier, or among those with a PTD history. Conclusions: Clindamycin, but not metronidazole, was beneficial in studies providing IPD, but after imputing data from missing IPD studies, treatment of BV during pregnancy did not reduce PTD, nor prolong pregnancy, in any subgroup or when started earlier in gestation.

McDonald randomized women with singleton pregnancies who were at least 17 years old to metronidazole (n=429) or placebo (n=428). They were screened around 19 weeks, but randomized around 24 weeks. Donors were selected who were randomized between 20 and 28 weeks. In the MZ group, 22 women with previous preterm births had two preterm births and 407 women without previous preterm births had 29 preterm births. In the control group, 24 women with previous preterm births had ten preterm births and 404 women without previous preterm births had 22 preterm births. The imputed data sets have exactly the same number of preterm and term births by treatment group and previous preterm status.
Moniri randomized 120 women to metronidazole (n=60) or control (n=60). Randomization occurred between 20 and 34 weeks of gestational age. In both groups there were 2 premature births. The imputed data sets have exactly the same number of preterm and term births by treatment group.

Clindamycin studies
Subtil et al randomized 1904 women to clindamycin and 956 to placebo, at less than 15 weeks' gestation. Four of the five studies with data using clindamycin randomized women at this early gestational range. The Kiss study was not used to create imputation donors because gestational age at delivery was not provided in individual weeks. An imputed data set was created by randomly selecting individuals from the four comparable studies so that counts match published treatment and outcome results. Five separate data sets were created to allow for multiple imputation inference (Rubin 1987). The summary characteristics of the imputed data sets exactly match the known summaries of the actual data. Joesoef et al (1995) provides counts of number of subjects receiving clindamycin (n=340) and placebo (n=341), number very preterm (less than 32 weeks), number preterm (less than 37 weeks), and a fact about randomization (14-26 weeks). Information on rates of low birth weights (of current pregnancies) and mother's age (in ranges) also are available. An imputed data set was created by randomly selecting individuals from the four comparable studies so that counts match published treatment and outcome results. Sampling weights analogous to post stratification sampling weights were created to give donors relative frequencies to match the mother's age distribution. Five data sets were created. The summary characteristics of the imputed data sets exactly match the known summaries of the actual data or are not statistically significantly different in terms of low birth weight and mother's age categories.
Vermeulen and Bruinse (1999) included twenty-two women with BV, eleven each in clindamycin and placebo groups. One in each group had a delivery before 34 weeks.
Randomization occurred before 26 weeks of gestation. Women had singleton pregnancies. An imputed data set was created by randomly selecting individuals from the four comparable studies so that counts match published treatment and outcome results.
Kekki (2001) provides counts of subjects receiving clindamycin (n=187) and placebo (n=188), number of preterm (less than 37 weeks), a fact about randomization (12-19 weeks), and age ranges of mothers. An imputed data set was created by randomly selecting individuals from the four comparable studies so that counts match published treatment and outcome results. All mothers had singleton pregnancies. Sampling weights based on empirical likelihood methods were created to give donors relative frequencies to resemble the reported mean mother's age in both groups. Five data sets were created. The summary characteristics of the imputed data sets exactly match the known summaries of the actual data. The average mother's age in the imputed data sets is not statistically different from the reported mother's age in either the treatment or placebo groups. Guaschino (2003) randomized 112 women to clindamycin or no treatment; 55 to clindamycin and 57 to control. Six in each group were lost to follow up. The study provides counts of number of preterm (less than 37 weeks), a fact about randomization (14-25 weeks), and some other information, including the average gestational age at randomization by group. An imputed data set was created by randomly selecting individuals from the four comparable studies so that counts match published treatment and outcome results. All mothers had singleton pregnancies. Sampling weights based on empirical likelihood methods were created to give donors relative frequencies to resemble the reported mean gestational age at randomization in both groups. The imputed data sets have exactly the same number of preterm and term births by treatment group. The average gestational age at randomization is not statistically different from that reported in either the treatment groups. Giuffrida and Mangiacasale (2006) randomized sixty BV positive women to either clindamycin (n=30) or hydrogen peroxide (n=30). The women were at 13-16 weeks of gestation at the time of randomization. There were no preterm birth events. An imputed data set was created by randomly selecting individuals from the four comparable studies so that counts match published treatment and outcome results.
Kiss randomized women to treatment with clindamycin (n=188) or control (n=187). The exact gestational age at delivery was not available, but ranges were reported as less than 33 weeks, 33 to 36 weeks, and 37 weeks or more. Women were randomized at either 16 or 21 weeks, with each value likely representing a range of gestational ages at randomization. For each woman in the Kiss database, donors were selected from other studies using CM by matching on twin status, previous preterm status, nulliparous status, gestational age at randomization within two-and-a-half weeks, and range of gestational age at delivery. Indicators of extremely early, very early, and premature birth were updated when needed to be consistent with the imputed gestational age at delivery.