Preterm birth is a significant cause of infant morbidity and mortality. Considerable evidence suggests that infections play a key role in causing women to have preterm labor and delivery. The specific bacteria found in amniotic fluid and the placenta in association with premature birth are thought to come from the vagina. This is especially true in women with bacterial vaginosis. This infection is estimated to affect about 800,000 pregnant women in the United States every year. Women with bacterial vaginosis have a higher incidence of preterm delivery. Several studies have shown that metronidazole (with or without erythromycin) reduces the risk of preterm delivery in women with bacterial vaginosis. Carey and associates conducted a multicenter clinical trial to determine if screening for and treating bacterial vaginosis with antibiotics would reduce the risk of preterm delivery in asymptomatic, pregnant women.

Women who were between eight weeks and 22 weeks, six days of gestation were initially screened for bacterial vaginosis and infection with Trichomonas vaginalis. Exclusion criteria for screening included an increased vaginal discharge, itching, burning, odor, allergy to metronidazole, antibiotic use within the previous 14 days, multifetal gestations or any medical illness that required regular or intermittent drug therapy. Vaginal specimens were obtained from the junction of the upper one third and lower two thirds of the lateral vaginal wall, placed on a glass slide and tested with a pH stick. If the vaginal pH was more than 4.4, the slide was sent to a reference laboratory for Gram staining. Results were interpreted according to the Nugent criteria, where a score of seven or more combined with a vaginal pH of more than 4.4 is considered diagnostic for bacterial vaginosis. The vaginal specimens were also examined for T. vaginalis. Women who tested positive for T. vaginalis as an isolated or coinfection organism were excluded from the trial.

Women with a diagnosis of bacterial vaginosis whose pregnancies were between 16 weeks and 23 weeks, six days of gestation were randomized to receive eight capsules of metronidazole (250 mg each) or placebo. The first dose was given at the study site, and a further dosage of eight capsules was to be taken at home 48 hours later.

A follow-up visit was conducted between 24 weeks and 29 weeks, six days of gestation, which was always at least 14 days after the initial treatment. Screening for bacterial vaginosis was repeated, and all participants were treated a second time with the same two-dose regimen regardless of the results of the Gram stain.

All women enrolled received routine pre-natal care. Outcomes assessed included preterm labor, admissions and visits to the hospital, the use of tocolytic therapies, premature rupture of membranes, intra-amniotic infection, neonatal sepsis, neonatal intensive care unit admissions, neonatal deaths and preterm delivery. Preterm delivery was defined as a birth occurring before 37 weeks of gestation.

Of the 21,965 women screened, 6,540 were diagnosed with bacterial vaginosis. The mean age was 23 years. After exclusion criteria were applied and patient withdrawals, 1,953 women were randomized to receive metronidazole or placebo. Of this total, 90 percent returned for the follow-up visit; however, outcome data were available for only 98.3 percent of the women. Bacterial vaginosis was still present in 22.2 percent of the women in the metronidazole-treatment group and in 62.6 percent of the placebo group at the follow-up visits.

The frequency of preterm delivery was 12.2 percent in the metronidazole-treatment group compared with 12.5 percent in the group receiving placebo. The frequency of delivery before 37 weeks of gestation did not differ significantly between the metronidazole group and the placebo group. In addition, there was no significant difference between the two groups in the rate of deliveries before 35 weeks and 32 weeks of gestation. Other outcomes for which no significant differences were observed included preterm delivery related to spontaneous labor or rupture of membranes. Treatment with metronidazole did not reduce the rates of hospital admission for preterm labor, receipt of tocolytic drugs, vaginal infections, neonatal intensive care unit admission or neonatal death.

The authors conclude from this study that screening asymptomatic women for bacterial vaginosis and initiating subsequent treatment with oral metronidazole does not decrease the risk of preterm labor or delivery. This appears to be so despite proof that metronidazole is effective in treating women with bacterial vaginosis (cure rate of approximately 78 percent in this study). These results conflict with those reported in earlier studies that showed antibiotics (specifically metronidazole and erythromycin) did prevent preterm birth.

editor's note: The results of this study are significant and may change the way many family physicians and obstetricians approach this problem. In our own prenatal clinic, we have been screening all patients for bacterial vaginosis during the first trimester and treating those who test positive. This practice will need to be reevaluated on the basis of this new evidence.—j.t.k.