BACKGROUND:To investigate the association between efavirenz (EFV) use during conception or first trimester (T1) of pregnancy and the occurrence of birth defects. SETTING:Seven observational studies of pregnant HIV-positive women across 13 European countries and Thailand. METHODS:Individual-level data were pooled on singleton pregnancies included in participating cohorts in 2002-2015. Birth defects were coded according to ICD-10 and the EUROCAT classification. We performed mixed-effects logistic regression models to assess the association between EFV exposure in utero and likelihood of birth defects. RESULTS:We included 24,963 live births from 21,093 women. At conception, 30.2% (7537) women were on a non-EFV-based regimen, 4.8% (1200) on EFV, and 65% (16,226) were unexposed to antiretroviral therapy (ART). There were 412 infants with ≥1 birth defect, a prevalence of 1.65% (95% confidence interval: 1.50 to 1.82). Limb/musculoskeletal and congenital heart defects were the most common defects reported. Birth defects were present in 2.4%, 1.6%, and 1.3% of infants exposed to non-EFV, EFV, and unexposed to ART during conception/T1 (P = 0.135), respectively. The association between exposure to ART during conception/T1 and birth defects remained nonsignificant in adjusted analyses, as did exposure to EFV versus non-EFV (adjusted odds ratio 0.61; 95% confidence interval: 0.36 to 1.03, P = 0.067). Among the 21 birth defects in 19 infants on EFV, no neural tube defects were reported. CONCLUSIONS:Prevalence of birth defects after exposure to EFV-based compared with non-EFV-based ART in conception/T1 was not statistically different in this multicohort study, and even lower. EFV is at least as safe as other ART drugs currently recommended for antenatal use.