Is Metronidazole Safe to Take During Pregnancy?

Metronidazole is generally considered safe during pregnancy based on decades of research involving hundreds of thousands of pregnancies. Major clinical guidelines, including those from the CDC, recommend it for treating bacterial vaginosis and other infections in pregnant women. But the evidence is not perfectly clean, and a few findings deserve attention, particularly around first-trimester exposure and specific infection types where the drug’s benefits get murky.

What the Largest Studies Actually Found

The most reassuring evidence comes from studies that tracked real-world outcomes in pregnant women who were prescribed metronidazole for infections during pregnancy. A study of nearly 2,829 mother-infant pairs found that about a third of the mothers had been treated with the drug, and after adjusting for other factors, there was no link between metronidazole and preterm birth, low birth weight, or birth defects. The odds ratios were essentially 1.0 across the board, meaning treated women fared no differently than untreated women on those outcomes.¹

An earlier meta-analysis pooling seven studies reached the same conclusion: the overall odds ratio for birth defects after first-trimester exposure was 0.93, with a confidence interval that comfortably straddled 1.0. The authors stated plainly that metronidazole does not appear to carry an increased teratogenic risk.²

These are the kinds of findings that have kept metronidazole on the standard list of pregnancy-compatible antibiotics for years. But the picture gets a bit more complicated when you zoom into specific study designs and specific outcomes.

The First-Trimester Question

The first trimester is when organs form, so any medication taken during those weeks gets extra scrutiny. Here the data is mostly reassuring but not unanimously so. A 2021 systematic review and meta-analysis that pooled data from over 411,000 pregnant women found a slightly elevated odds ratio for major malformations after first-trimester exposure, at 1.15, but this did not reach statistical significance. The result was driven largely by case-control studies, which found a statistically significant increase, while higher-quality prospective and retrospective cohort studies did not.³

That same meta-analysis turned up one specific malformation worth flagging: congenital hydrocephalus showed a statistically significant increase with an odds ratio of about 4.0, though this was based on only two studies and the absolute number of affected pregnancies was very small. The review also found a pooled odds ratio of roughly 1.7 for spontaneous abortion, which persisted even after excluding studies at high risk of bias.³

The spontaneous abortion signal is hard to interpret. Women who take metronidazole in early pregnancy are taking it because they have an active infection, and infections themselves can raise miscarriage risk. Disentangling the drug’s effect from the infection’s effect is a persistent challenge in this area of research, and the authors acknowledged this. Still, it is the kind of finding that makes clinicians weigh the decision more carefully in the first trimester than later on.

The contrast with the older meta-analysis, which found an odds ratio of 0.93 for first-trimester defects, is worth noting. That analysis used a smaller pool of studies with different designs. The disagreement probably reflects differences in which studies were included and how rigorously confounders were controlled rather than any fundamental contradiction in the biology. The overall takeaway from both analyses is that metronidazole is unlikely to be a strong teratogen, but a small risk from first-trimester exposure has not been completely ruled out.

The Drug Does Cross the Placenta

Metronidazole freely crosses the placenta. In studies measuring drug concentrations in fetal tissue after a single intravenous dose of 500 mg, levels in the fetus reached about two-thirds of the level found in the mother’s blood.⁴ Oral dosing showed similar patterns: after a 400 mg tablet taken an hour before first-trimester pregnancy termination, the drug was measurable in fetal tissue, with concentrations ranging up to about 3 micrograms per gram.⁵

This might sound alarming, but placental transfer alone does not equal harm. Many medications considered safe in pregnancy cross the placenta. The relevant question is whether the fetal exposure leads to damage, and on that question the outcome studies discussed above provide the real evidence. The pharmacokinetic data mainly matters to researchers designing dosing regimens and to clinicians choosing between systemic and topical routes.

Bacterial Vaginosis and Why Metronidazole Is Prescribed So Often in Pregnancy

Bacterial vaginosis is the single most common reason pregnant women end up on metronidazole. It is a disruption of the normal vaginal bacterial community that can cause discharge, odor, and discomfort, and it has been linked to preterm delivery. CDC guidelines recommend metronidazole as a first-line treatment for BV in pregnancy.⁶

Here is where an important distinction comes in. Metronidazole is effective at clearing the infection, with cure rates around 70% regardless of whether you take it orally or as a vaginal gel.⁷ But clearing the infection does not appear to reduce the risk of preterm birth. A large trial published in the New England Journal of Medicine found that preterm delivery occurred at essentially the same rate in women treated with metronidazole and those given a placebo, about 12% in both groups. The drug also did not reduce rates of intraamniotic infections, postpartum infections, or neonatal intensive care admissions.⁸

An individual participant data meta-analysis confirmed this pattern: metronidazole for BV during pregnancy produced an odds ratio of 1.00 for preterm delivery, which is about as null a result as you can get. Clindamycin, the main alternative antibiotic for BV, initially showed a benefit in studies that provided individual patient data, but that advantage disappeared after accounting for missing data from other studies.⁹

So the rationale for treating BV in pregnancy with metronidazole is to resolve the infection itself and relieve symptoms, not to prevent preterm birth. That is a worthwhile goal, but it changes the risk-benefit calculation. If a woman has symptomatic BV, treatment makes sense because the infection is bothersome and could progress. If the BV is asymptomatic and discovered incidentally, the case for treatment is weaker, because clearing it does not seem to change pregnancy outcomes.

Trichomoniasis Tells a Different Story

Trichomoniasis is a sexually transmitted parasitic infection, and metronidazole is the standard treatment outside of pregnancy. The drug clears the parasite effectively in pregnant women too, with about 90% of women testing negative after treatment. But one large U.S. trial that treated asymptomatic trichomoniasis between 16 and 23 weeks of pregnancy was stopped early because treated women actually had a higher rate of preterm birth, not a lower one. The risk ratio was 1.78, meaning treated women were nearly twice as likely to deliver preterm as those who received a placebo.¹⁰

That finding was unexpected and alarming, but its interpretation has been debated. One theory is that killing trichomoniasis organisms releases inflammatory products that trigger preterm labor. An observational study looking at the same question found no increased preterm birth risk with metronidazole treatment of trichomoniasis, suggesting the randomized trial’s finding might not tell the whole story.¹¹

The practical result is that guidelines now recommend treating symptomatic trichomoniasis in pregnancy with metronidazole, because untreated symptomatic infection also carries risks, but screening and treating asymptomatic trichomoniasis during pregnancy is not routinely recommended. The lesson from the stopped trial is not that metronidazole itself is dangerous but that treating an asymptomatic parasitic infection during pregnancy can paradoxically cause more harm than leaving it alone.

Oral Versus Vaginal, and How the Route Matters

If you are prescribed metronidazole during pregnancy, your provider may offer a choice between oral tablets and vaginal gel or suppositories. For BV, the two routes work equally well: cure rates of about 70% have been documented for both.⁷ The vaginal route delivers the drug locally and results in lower blood levels, which theoretically means less fetal exposure. This can be appealing during the first trimester when concerns about systemic exposure run highest.

From a microbiome perspective, both routes affect vaginal bacteria in broadly similar ways. Oral treatment was somewhat better at reducing certain BV-associated organisms, but the difference was not statistically significant in a head-to-head comparison. Reassuringly, beneficial Lactobacillus bacteria in the vagina did not decrease with either route.¹²

The downside of vaginal treatment is that it cannot treat infections elsewhere in the body. For trichomoniasis or infections in the gut, oral dosing is necessary. Side effects also differ: oral metronidazole commonly causes nausea, a metallic taste, and stomach upset, while the vaginal form largely avoids those but can cause local irritation. For pregnant women already dealing with morning sickness, the vaginal option for BV may be more tolerable.

Metronidazole Compared to Clindamycin

Clindamycin is the main alternative to metronidazole for BV during pregnancy, and the two drugs come up in almost every discussion of antibiotic choice. A trial specifically comparing them in pregnancies at risk for preterm labor found no significant difference in BV eradication rates. About 60% of women in each group were cured after first-line treatment, and after a second round the rates climbed to around 77-79%. Rates of preterm delivery and late miscarriage were also similar between groups. Side effects were reported by roughly 30% of women on metronidazole and about a quarter of those on clindamycin, and only a small fraction in either group stopped treatment because of them.¹³

The practical takeaway is that the two antibiotics are roughly interchangeable for BV in pregnancy. If one causes unacceptable side effects, switching to the other is reasonable. Clindamycin carries its own concerns, including a risk of Clostridioides difficile infection with oral use, so it is not inherently “safer” just because it is not metronidazole.

When Metronidazole Does Not Work

Metronidazole resistance is rare but does occur, particularly with trichomoniasis. When a pregnant woman has a trichomoniasis infection that does not respond to standard dosing, clinicians face a genuine dilemma. Various alternative treatment regimens exist for non-pregnant patients, including higher-dose metronidazole, tinidazole, and combination approaches, but none of these alternatives have been evaluated for safety during pregnancy.¹⁴ This is an area where the evidence base is genuinely thin, and management decisions end up being made case by case in consultation with infectious disease specialists.

The Animal Cancer Data and What It Means for You

You may come across warnings that metronidazole causes cancer in laboratory animals. This is true: high-dose, long-term administration has produced tumors in rodents. The drug is also a proven mutagen in bacterial test systems and shows some evidence of being able to damage DNA in human cells. Based on these findings, some researchers have argued it should be considered a potential human carcinogen.¹⁵

However, the epidemiological studies that have looked for a cancer link in people who took metronidazole have not found convincing evidence of one. The doses given to laboratory animals are far higher than what a pregnant woman would receive during a typical seven-day course. The mutagenicity data is a reason the drug is not prescribed casually, but it has not translated into observable cancer risk in humans at therapeutic doses. For a short course during pregnancy, this concern sits well below the more immediate questions about birth outcomes and infection control.

Practical Considerations for Each Trimester

The risk-benefit calculation shifts somewhat as pregnancy progresses. In the first trimester, when organ formation is happening and the spontaneous-abortion signal from the meta-analysis adds a layer of uncertainty, most clinicians will prescribe metronidazole only when the infection clearly needs treatment. Deferring treatment to the second trimester, if the infection allows it, is a common strategy. Some providers prefer vaginal metronidazole for BV in early pregnancy to reduce systemic exposure, though no trial has directly proven this makes a difference in outcomes.

In the second and third trimesters, the safety profile looks more straightforward. The large outcome studies show no increase in preterm birth or low birth weight with metronidazole use during these periods.¹ The exception, as discussed, is the use of the drug for asymptomatic trichomoniasis, where the benefit of treatment is questionable regardless of trimester.

One thing to keep in mind is that untreated infections carry their own pregnancy risks. BV left untreated has been associated with preterm birth, chorioamnionitis, and postpartum endometritis. Trichomoniasis has been linked to premature rupture of membranes and low birth weight. The decision to treat is always a balance between the known risks of the infection and the largely reassuring but not perfectly clean safety data on the medication. In most cases where a woman has a symptomatic infection that metronidazole can treat, the evidence favors treatment.