BACKGROUND: A study from Botswana identified an increased risk of neural tube defects (NTDs) in infants of mothers with HIV who were treated with dolutegravir around the time of conception. We aimed to examine associations of dolutegravir use with NTDs and pregnancy loss using large health-care claims databases from the USA, a country with folic acid fortification of food. METHODS: In this cohort study, we analysed health-care claims data, recorded in the Merative MarketScan commercial database (MarketScan data) and Centers for Medicare & Medicaid Services Medicaid database (Medicaid data) from Jan 1, 2008, to Dec 31, 2020. We identified pregnancies with enrolment during their entire duration among women aged 15-49 years and we estimated time of conception. For each pregnancy, we determined HIV status and periconceptional exposure to dolutegravir or other antiretroviral agents. We estimated and compared the incidence rate of NTDs, stillbirths, and pregnancy loss (ie, spontaneous or induced abortions) by type of periconceptional antiretroviral exposure. We calculated adjusted risk ratios of the adverse outcomes using Poisson models adjusting for demographic and clinical factors. FINDINGS: Of 4 489 315 pregnancies in MarketScan data and 14 405 861 pregnancies in Medicaid data that had full enrolment, we identified 69 pregnancies in MarketScan data and 993 pregnancies in Medicaid data that were associated with HIV and periconceptional dolutegravir exposure. For women without HIV, the NTD rate was 4·1 per 10 000 live births (95% CI 3·9-4·3) in MarketScan and 5·7 per 10 000 live births (5·6-5·8) in Medicaid. No NTD cases were found among those with dolutegravir or non- dolutegravir antiretroviral drug exposure in the MarketScan data; only one NTD case was identified among women with dolutegravir, and three among women with non-dolutegravir antiretroviral exposure in Medicaid. After adjusting for covariates, there were no significant differences in risk ratios of NTD between groups with periconceptional dolutegravir or non-dolutegravir antiretroviral exposure and the group without HIV. However, compared with women without HIV, the risk of pregnancy loss was higher among women exposed to antiretroviral therapy: for dolutegravir exposure the adjusted risk ratio was 1·73 (95% CI 1·20-2·49) in MarketScan data and 1·41 (1·30-1·54) in Medicaid data; for non-dolutegravir antiretroviral exposure the adjusted risk ratio was 1·23 (1·10-1·37) in MarketScan data and 1·11 (1·07-1·15) in Medicaid data. INTERPRETATION: We studied the largest US cohort of women with periconceptional or early-pregnancy dolutegravir exposure. Our results do not show an increased risk of NTDs in exposed infants in the USA. Administrative databases can be used, with rigorous methodology, to study correlates of rare outcomes, such as NTDs, and to monitor for adverse pregnancy outcomes in women who receive antiretrovirals. FUNDING: US Centers for Disease Control and Prevention.