We read with interest the article by Shennan et al.,1 which reported the results of a trial of the treatment with metronidazole of women who were at high risk of preterm delivery as predicted by a positive test for fetal fibronectin. We would like to comment on their conclusion that metronidazole should not be used for the prenatal prevention of preterm birth.

The trial recruited women already at high risk of preterm birth, who were then found to be positive for fetal fibronectin at 24–27 weeks, and randomised them to treatment with metronidazole or placebo. While fetal fibronectin is the best predictor of preterm birth available at present,2 it is not a reliable test for the presence of bacterial vaginosis (BV). This is supported by the findings that the majority of the women randomised to metronidazole or placebo did not have BV (89 and 84%, respectively).

The association between BV and preterm birth, although not strong, is well established. Numerous clinical trials have been reported of screening both high-risk and low-risk women and treating those found to have BV with metronidazole. The results of these trials are heterogeneous with some demonstrating a reduction in the incidence of preterm birth, others demonstrating no benefit. Only one of these studies, in which women were treated with 2 days of oral metronidazole, reported an increased risk of preterm birth associated with treatment. However, in addition, a small number of studies have examined metronidazole treatment of Trichomonas vaginalis in pregnancy, or metronidazole treatment of women with a positive fibronectin test, and demonstrated an increase in the risk of preterm birth.

We would suggest that an alternative explanation for the release of fibronectin into the vaginal secretions is physiological breakdown of the fetal membranes over the lower uterine pole, as evidenced by the generation of the ‘zone of altered morphology’ in the fetal membranes.3 Remodelling of the lower uterine segment and cervix and associated alterations in the fetal membrane morphology may make women more susceptible to ascending infection. The characteristic loss of the decidual layer and the resident macrophage population may contribute to this.

Since the majority of women in the study of Shennan et al. did not have evidence of BV, it is plausible to speculate that the antibiotics in the treatment arm changed the uterovaginal flora, thereby facilitating pathogenic colonisation, which resulted in infection in some women and consequent poorer outcomes.

While it is wise to recommend caution in the use of any antibiotics in pregnancy, this advice may be particularly pertinent to those women who have tested positive for fetal fibronectin but have no evidence of infection. However, more evidence is required about the role of metronidazole and other antibiotics in women at risk of preterm birth, where there is clinical evidence of BV or other microbiologically pathogenic states.


  1. Top of page
  2. References
  • 1
    Shennan A, Crawshaw S, Briley A, Hawken J, Seed P, Jones G, et al. A randomised controlled trial of metronidazole for the prevention of preterm birth in women positive for cervicovaginal fetal fibronectin: the PREMET Study. BJOG 2006;113:6574.
  • 2
    Honest H, Bachmann LM, Gupta JK, Kleijnen J, Khan KS. Accuracy of cervicovaginal fetal fibronectin test in predicting risk of spontaneous preterm birth: systematic review. BMJ 2002;325:301.
  • 3
    McParland PC, Bell SC. The fetal membranes and mechanisms underlying their labour associated and prelabour rupture during pregnancy. Fetal Mat Med Rev 2004;15:136.