Vincenzo Berghella, Mark Klebanoff, Cora McPherson, J.Christopher Carey, John C. Hauth, J.M. Ernest, R.Phillip Heine, Ronald J. Wapner, Wayne Trout, Atef Moawad, Kenneth J. Leveno, Menachem Miodovnik, Baha M. Sibai, J.Peter Van Dorsten, Mitchell P. Dombrowski, Mary J. O'Sullivan, Michael Varner, Oded Langer, for the National Institute for Child Health and Development Maternal Fetal Medicine Units Network

The purpose of this study was to determine whether sexual intercourse was associated with the treatment efficacy or the incidence of preterm birth in two large randomized trials in which metronidazole treatment of bacterial vaginosis or Trichomonas vaginalis did not reduce preterm birth. Secondary analysis of two multicenter, double-blind, placebo-controlled trials in which women with asymptomatic bacterial vaginosis on Gram stain or asymptomatic T vaginalis on culture were randomized at 16 to 23 weeks of gestation to metronidazole or placebo. In both studies, women took 2 g of metronidazole or placebo in the presence of a nurse (first dose) and were given a second dose to take 48 hours later. This regimen was repeated (third and fourth doses) at 24 to 29 weeks. At the time of the third dose, bacterial vaginosis and T vaginalis specimens were collected again. Patients who were randomly selected to receive metronidazole were analyzed for bacterial vaginosis and T vaginalis at 24 to 29 weeks and for preterm birth of <37 weeks of gestation, according to intercourse between first and second doses and between the second and third doses. Continuous variables were compared with the use of the Wilcoxon rank-sum test; categoric variables were compared with the use of the chi(2 ) test, Fisher exact test, or the Mantel-Haenzel test of trend. Sexual intercourse between the first and second doses or between the second and third doses did not influence the incidence of bacterial vaginosis (18% vs 24%; relative risk, 0.7; 95% CI, 0.5-1.1; and 23% vs 20%; relative risk, 1.2; 95% CI, 0.9-1.6, respectively) or T vaginalis (4% vs 8%; relative risk, 0.5; 95% CI, 0.1-3.6; and 5% vs 10%; relative risk, 0.5; 95% CI, 0.2-1.1; respectively) at 24 to 29 weeks of gestation compared with no intercourse. In the T vaginalis trial, sexual intercourse between the first and second doses or between the second and third doses did not influence the incidence of preterm birth (13% vs 17%; relative risk, 0.8; 95% CI, 0.3-2.1; and 16% vs 17%; relative risk, 1.0; 95% CI, 0.6-1.6; respectively) compared with no intercourse. In the bacterial vaginosis trial, although sexual intercourse between the first and second doses did not influence the incidence of preterm birth (11% vs 12%; relative risk, 0.9; 95 % CI, 0.6-1.5), sexual intercourse between the second and third doses was associated with a reduction in the incidence of preterm birth (10% vs 16%; relative risk, 0.6; 95% CI, 0.4-0.9) compared with no intercourse. Sexual intercourse was associated with neither the efficacy of metronidazole treatment of bacterial vaginosis or T vaginalis nor with the incidence of preterm birth. In the bacterial vaginosis study, intercourse between the second and third doses had a negative association with preterm birth.