Pregnancy Outcomes Stayed Reassuring After Rifapentine-Based TB Exposure
A secondary analysis from a phase 3 tuberculosis trial brings a rare, practical data point to a field long forced to make decisions with thin evidence. Among women who became pregnant during tuberculosis treatment in the S31/A5349 study, investigators observed no elevated rate of fetal loss or congenital anomalies after exposure to the 4-month rifapentine/moxifloxacin regimen.
The findings matter because pregnancy and active TB collide in a setting where maternal infection itself carries serious risk, yet drug safety data remain limited. Global estimates have placed active TB in pregnancy at more than 200,000 cases annually, and untreated disease can lead to pregnancy complications, nonobstetric maternal death, and infant death. Even when treatment is started, clinicians still face a narrow evidence base for the safety of several standard TB drugs during gestation.
The trial context makes the analysis especially useful. S31/A5349 was a multicenter randomized phase 3 noninferiority study in nonpregnant participants 12 years of age and older with drug-susceptible pulmonary TB. Investigators compared two treatment-shortening rifapentine-containing regimens with the standard 6-month control regimen. One investigational arm used rifapentine, moxifloxacin, and isoniazid for 4 months, with pyrazinamide during the first 2 months. The other used rifapentine plus isoniazid for 4 months, with pyrazinamide and ethambutol during the first 2 months. The 4-month rifapentine/moxifloxacin regimen was previously shown to be noninferior to the control regimen and was described as safe and well tolerated, while the rifapentine-only arm failed to meet noninferiority criteria.
Pregnancy had not been an enrollment category for the parent trial. Women who were pregnant or breastfeeding were excluded because of uncertainty around rifapentine, moxifloxacin, and pyrazinamide in those settings. Yet some participants conceived during the study, creating an unplanned but clinically important opportunity to examine outcomes after actual in-trial drug exposure.
Who Became Pregnant During Treatment
Among 740 women in the trial, 97 became pregnant, representing 13.1% of the female cohort. Those 97 participants accounted for 102 pregnancies. Of those pregnancies, 30, or 29.4%, involved exposure to study drugs. In practical terms, conception occurred on or before the last study dose date in those cases, based on estimated conception dating used by the investigators.
The analysis separated pregnancies by treatment arm and tracked fetal outcomes. Fetal loss occurred in 3 of 13 pregnancies in the control regimen group, 1 of 9 in the rifapentine/moxifloxacin group, and 1 of 8 in the rifapentine regimen group. Among 21 live births in exposed pregnancies, seven in each study arm, 1 infant with a congenital anomaly was reported in a participant who had received the rifapentine regimen.
One infant anomaly across 21 live births does not create a signal of major teratogenic concern. It does, however, illustrate the level of caution still required when extrapolating from adult TB efficacy trials to pregnancy care. The sample is too small to rule out uncommon harms, but it is large enough to offer some reassurance that inadvertent early exposure did not produce an obvious excess of fetal loss or structural abnormality in this cohort.
Why the Finding Matters at the Bedside
For clinicians treating TB in women of reproductive potential, this analysis addresses a real-world problem: pregnancies happen during therapy, even when the trial protocol asks participants to avoid them. When pregnancy is discovered, treatment is usually stopped and TB care is transferred to local standard practice. The central question is what has already happened to the fetus before that point.
The answer here is more reassuring than restrictive. Among women who became pregnant while assigned to the short rifapentine/moxifloxacin regimen, there was no elevated rate of fetal losses or congenital anomalies. Because this regimen is now recommended by the World Health Organization and the CDC for drug-susceptible TB in nonpregnant patients 12 years of age and older, the pregnancy data help fill a gap created by the original exclusion of pregnant women from enrollment.
Rifapentine-based shortening of TB therapy has obvious appeal in pregnancy care. A shorter regimen can reduce the duration of maternal illness, the burden of adherence, and the window in which active TB itself can destabilize pregnancy. The biological logic is straightforward: durable TB control lowers ongoing inflammatory and nutritional stress on the mother, which may indirectly reduce obstetric risk. The remaining uncertainty is drug-specific fetal safety, especially for regimens containing agents not historically studied in pregnant populations at scale.
The Limits Remain Built Into the Design
The investigators were careful about how they handled pregnancy detection and exposure assignment. Women of childbearing potential had to have a negative pregnancy test at screening unless they were surgically sterilized or postmenopausal, and they were required to use contraception or abstain during study treatment. Pregnancy tests were not mandated during follow-up, though clinicians could perform them at their discretion. When pregnancy was reported or detected during treatment, study drugs were permanently stopped and women continued with follow-up until the pregnancy outcome was known.
Those procedures make the dataset more credible, but they also define its limitations. This was a secondary analysis, not a pregnancy-focused trial. The total number of exposed pregnancies was small, and the analysis did not collect nonstudy TB treatment outcomes after transfer to local care. That leaves unanswered questions about the comparative maternal course after treatment interruption and switch to local regimens.
The exposure window also deserves attention. Some pregnancies occurred during study treatment, while others arose during the follow-up period. Only pregnancies with estimated conception dates on or before the last study dose date were considered exposed, and that estimation relied on site-level judgment and available clinical information. Even with real-time safety review by the CDC Clinical Research Branch, conception dating is not as exact as a randomized drug exposure schedule.
Still, the practical message is difficult to ignore. In a cohort of women who conceived during a major TB treatment trial, short-course rifapentine/moxifloxacin exposure did not show an obvious reproductive safety penalty. For a disease in which treatment choices during pregnancy have often been constrained more by absence of data than by proven harm, that matters at the point of care and in the design of future TB studies.