No association was found between metronidazole treatment during pregnancy and preterm birth, low birth weight, or congenital anomalies among the 2,829 singleton/mother pairs in this cohort. Metronidazole is recommended for treatment of BV and T. vaginalis
). These conditions have been associated with preterm birth (10
). While it is widely recognized that symptomatic BV and T. vaginalis
should be treated in pregnant women, the American Academy of Pediatrics also notes that some experts recommend screening and oral treatment of BV in asymptomatic pregnant women at high risk for preterm delivery (9
). In this cohort, compared to women who were screened but untreated for BV/abnormal flora, women who were treated experienced reduced preterm delivery (19
Several trials have suggested that there may be an increased risk for preterm birth among women given metronidazole during pregnancy. In one study of 617 women, there were more preterm deliveries among women given two 2-g doses of oral metronidazole for asymptomatic T. vaginalis
infection than among women in the placebo group (RR, 1.8 [95% CI, 1.2 to 2.7]) (18
). In a separate study, asymptomatic pregnant women with a positive cervical or vaginal fetal fibronectin test were given metronidazole (250 mg 3 times daily) and erythromycin (250 mg orally 4 times daily) or placebo. Among the 76 women with a previous spontaneous preterm delivery, there was an increased risk of repeat preterm birth among women who were given metronidazole without a currently accepted clinical indication. In other U.S. trials and studies in which women were treated with metronidazole for a currently accepted clinical indication such as BV, no increased risk of preterm birth has been found (6
). Women in the cohort described in this study were not randomized to treatment; they also received metronidazole for laboratory-confirmed clinical indications.
Two trials in Africa have also shown an increased risk of adverse events among infants of mothers given metronidazole during pregnancy. In one South African trial, primigravida or women with a previous mid-trimester abortion or preterm delivery diagnosed with BV by clinical criteria (pH > 4.7, amine odor, gray discharge, greater than 20% clue cells, and fewer than 2+ lactobacilli by microscopic examination) who were treated with 400 mg oral metronidazole twice daily for 2 days had a significantly increased risk of preterm birth compared to women who took vitamin C (P
= 0.03) (22
). In a subanalysis of a trial of presumptive sexually transmitted disease (STD) treatment in Rakai, Uganda, women with T. vaginalis
who were randomized to receive orally 1 g azithromycin, 400 mg cefixime, and 2 g metronidazole had more low-birth-weight infants than women who took multivitamins (RR, 2.5 [95% CI, 1.1to 5.5]), but did not have significantly increased rates of preterm birth (RR, 1.3 [95% CI, 0.8 to 2.0]) (17
Although studies have not shown metronidazole to be carcinogenic in humans (2
), it has been documented as a bacterial mutagen and carcinogen in animals (14
). In humans, metronidazole crosses the placenta. Therefore, there is a theoretical risk to the fetus that prenatal care providers may consider when making therapeutic choices for pregnant women. A survey of obstetricians and gynecologists from the mid-1990s revealed that only 1.5% of providers prescribed oral metronidazole during the first trimester (21
). A 1999 survey found that 95% of prenatal care providers used oral metronidazole to treat nonpregnant patients with symptomatic BV, but 56% prescribed it for pregnant patients with symptomatic BV. Furthermore, only 12% of providers prescribed oral metronidazole or clindamycin for treatment of BV during the first trimester (4
). These studies suggest that providers still have theoretical concerns about oral metronidazole use. Providers now have the option of choosing topical metronidazole, although no surveys have been published recently to allow for comparison with current practices.
Multiple epidemiologic studies and meta-analyses have not shown an association between metronidazole treatment during pregnancy and congenital anomalies in infants. Case reports from the 1980s document midline facial defects (cleft lip/palate and holotelencephaly) in three infants whose mothers took metronidazole during pregnancy (5
). These defects usually arise in months 2 to 3 of gestation. A Hungarian case-control study of 30,000 pregnant women who had healthy infants and 17,000 women who had infants with congenital anomalies did not find increased odds of any congenital anomaly except cleft lip/palate after metronidazole treatment in the second or third month of pregnancy (OR, 8.54 [95% CI, 1.06 to 68.86]) (12
). The authors note that this represents an increase of 0.03 cases per 1,000 above the expected number (1 per 1,000) of cleft lip/palate cases in Hungary without metronidazole treatment. In this cohort, we did not find any infants with cleft lip/palate whose mothers were treated with metronidazole; two infants whose mothers were not treated with metronidazole had cleft lip/palate.
Two other large studies found no association between metronidazole exposure and congenital anomalies. A retrospective cohort study of Medicaid records of 104,339 Michigan births found 63 congenital anomalies among 1,083 infants of mothers with first trimester metronidazole treatment (RR, 0.92 [95% CI, 0.7 to 1.2]) (24
). A Danish population-based retrospective cohort study found no association with congenital anomalies (OR, 0.44 [95% CI, 0.11 to 1.81]) among infants of metronidazole-treated mothers (26
). Two smaller studies also found no association; a retrospective cohort study of 1,387 pregnant women from Tennessee found no excess of congenital anomalies (RR, 1.2 [95% CI, 0.9 to 1.6]) among infants of women who filled a prescription for metronidazole between 30 days before and 120 days after their last menstrual period, (23
) and a cohort study in Israel found 3 anomalies among 190 infants of mothers with first trimester metronidazole treatment and 8 anomalies among 575 infants whose mothers did not take metronidazole (RR, 1.13 [95% CI, 0.30 to 4.23]) (13
In addition to these studies, meta-analyses have evaluated first trimester metronidazole treatment. None of them found an association between first trimester metronidazole treatment and congenital anomalies (weighted OR, 0.93 [95% CI, 0.73 to 1.18] [3
] and OR, 1.08 [95% CI, 0.90 to 1.29] [7
This study has several limitations. First, instances of metronidazole treatment that were not recorded in the prenatal chart could have occurred, leading to misclassification. Pregnancy outcomes are not known for all women who could have potentially been included in this cohort. For example, data were not available on spontaneous or elective abortions or stillbirths that could have occurred subsequent to the development of a congenital anomaly. This study only reports congenital anomalies that were evident shortly after birth. Each of these limitations would lead to a bias toward the null. Finally, because of the small sample size, there is limited power to detect a change in the proportion of infants with congenital anomalies.
The results of this study in combination with others suggest that metronidazole treatment during pregnancy, given for currently accepted clinical indications at recommended doses, is not associated with an elevated risk of the adverse outcomes examined. There may still be a need to offer information to prenatal care providers about clear indications for metronidazole use in pregnancy. We suggest that providers follow the current recommendations of the CDC STD treatment guidelines regarding metronidazole use in pregnancy (8