Antibiotic treatment of bacterial vaginosis does not help prevent premature birth, according to a study supported by the US National Institute of Child Health and Human Development (NICHD) and reported in the New England Journal of Medicine (2000;342:534). Commenting on the results of the study, Dr Duane Alexander (director, NICHD) said: "What was thought to be a promising experimental treatment is, in fact, ineffective for preventing preterm birth in pregnant women who have bacterial vaginosis but do not have symptoms."
In the trial, 953 pregnant women were treated with metronidazole, and 966 women with a placebo. The women received a first treatment between their 16th and 23rd weeks of pregnancy, and a second treatment between their 24th and 29th weeks of pregnancy. Bacterial vaginosis was eliminated in 77.8 per cent of the women who received metronidazole, and 37.4 per cent of the women who received placebo.
Premature birth occurred at similar levels in both groups (12.2 per cent in the metronidazole group and 12.5 per cent in the placebo group). In addition, more than 10 per cent of the women whose vaginosis was eliminated by the antibiotic treatment gave birth prematurely.
One of the authors of the study, Dr Mark Klebanoff (director of NICHD's division of epidemiology, statistics and prevention research) said: "Although the literature consistently indicates that intrauterine infection and bacterial vaginosis are associated with preterm birth, the results of our study do not support metronidazole treatment of asymptomatic pregnant women with bacterial vaginosis regardless of whether they were otherwise considered at either high or low risk for preterm birth." Dr Klebanoff added that the study results did not completely rule out the effectiveness of metronidazole treatment under other circumstances and said that, for example, it was possible that a longer duration of treatment or treatment earlier in pregnancy might be needed.
Dr Peter Brocklehurst (National Perinatal Epidemiology Unit, Institute of Health Sciences, Oxford) told The Journal on February 23 that the NICHD trial was interesting and had been extremely carefully undertaken. He added that earlier, smaller trials had shown conflicting results and that this large, well run trial had been needed to confirm what many people suspected regarding the use of antibiotic treatment to prevent preterm births.
Also speaking to The Journal on February 23, Mr Ronald Lamont (consultant obstetrician, Northwick Park hospital, Middlesex and author of an accompanying editorial [ibid, p581]) said that "this paper adds weight to the negative side of the argument but the trial is not ideal and should not result in therapeutic nihilism". He emphasised that studies to date had used various antibiotics at different doses and recruited women at different levels of risk from premature delivery and so it was unsurprising that different results were seen. Mr Lamont added that it was well established that infection was a cause of premature labour and that more work was needed to establish the usefulness of antibiotics in its prevention.