A longitudinal study of bacterial vaginosis during pregnancy

To determine the longitudinal changes in the incidence of vaginosis in pregnancy.

A prospective study of women during pregnancy.

A District General Hospital in North-West London.

Seven hundred and eighteen pregnant women attending antenatal clinics. At their first attendance and subsequently, Gram-stained vaginal smears were examined and Mycoplasma hominis and Gardnerella vaginalis were sought by culture.

Initially, 87 (12%) women had bacterial vaginosis diagnosed on Gram-stained reading of the vaginal smears. Examination of further smears, obtained from 176 women at 36 weeks of gestation, showed that those whose vaginal flora was normal initially, and who went to term, rarely developed vaginosis (three of 127, 2.4%). Samples were obtained at 36 weeks gestation from 32 women who had bacterial vaginosis initially, and went to term. In almost 50% (15 of 32) of these a normal lactobacillus-dominated flora had regenerated. Thirty-five women (5%) had initial vaginal smears graded as intermediate. From this group, six of the 17 (35%) women from whom samples were obtained at 36 weeks gestation still had flora of an intermediate pattern; 10(59%) now had normal flora and only one (6%) had developed bacterial vaginosis. Women with bacterial vaginosis were more likely to be culture-positive for M. hominis than those with normal flora (34/78 versus 10/563, odds ratio 42.73 (18.9 to 102.3) P < 0.001), or to be culture-positive for G. vaginalis than those with normal flora (35/78 versus 21/563, odds ratio 21.0 (10.75 to 41.2) P < 0.001).

Pregnant women do not commonly develop bacterial vaginosis after 16 weeks gestation, and if present, it remits spontaneously in approximately half of those who reach term. As bacterial vaginosis is associated with increased rates of second trimester miscarriage and preterm delivery, any treatment aimed at its eradication in pregnancy should be given no later than the beginning of the second trimester of pregnancy.