Background Dolutegravir is a once-daily integrase strand transfer inhibitor with no need for pharmacokinetic boosting that is approved for the treatment of HIV-1 infection. Because women are often under-represented in HIV clinical trials, we addressed the safety and efficacy of dolutegravir in women with HIV-1. Methods The ARIA study is a randomised, open-label, multicentre, active-controlled, parallel-group, non-inferiority phase 3b study done in 86 hospital and university infectious disease clinics, local health clinics, and private infectious disease clinics in 12 countries and one US territory, in North America, South America, Europe, Africa, and Asia. Eligible participants were women aged 18 years or older who had HIV-1 RNA viral loads of 500 copies per mL or greater, had received 10 days or less of previous antiretroviral therapy, and had tested negative for the HLA-B*5701 allele. Pregnant women were excluded. Eligible women were randomly assigned (1:1) to receive either a single-tablet regimen of dolutegravir plus abacavir and lamivudine once a day (dolutegravir group) or a three-tablet combination of ritonavir-boosted atazanavir plus coformulated tenofovir disoproxil fumarate and emtricitabine once a day (atazanavir group). Random treatment group assignment was stratified by plasma HIV-1 RNA viral loads and CD4 cell count at baseline. The primary endpoint was the proportion of participants with HIV-1 RNA viral loads of less than 50 copies per mL at week 48 in all participants who received at least one dose of study medication (intention-to-treat exposed population). We used a non-inferiority margin of −12%. Investigators monitored adverse events to assess safety. This study is registered with ClinicalTrials.gov, number NCT01910402. Findings Between Aug 22, 2013, and Sept 22, 2015, of 705 women assessed, 499 were randomly assigned to either the dolutegravir group (n=250) or the atazanavir group (n=249); two participants from each group were randomised to treatment but did not receive study medication. At week 48, 203 (82%) of 248 participants in the dolutegravir group compared with 176 (71%) of 247 in the atazanavir group had HIV-1 RNA viral loads of less than 50 copies per mL (mean difference 10·5%, 95% CI 3·1–17·8, p=0·005). One participant in the atazanavir group had nucleoside reverse transcriptase inhibitor–associated resistance that led to reduced emtricitabine susceptibility. Adverse events were similar between the dolutegravir and atazanavir groups; the most common were nausea (46 [19%] of 248 in the dolutegravir group vs 49 [20%] of 247 in the atazanavir group) and headache (28 [11%] vs 32 [13%]). Fewer participants in the dolutegravir group than the atazanavir group reported drug-related adverse events (83 [33%] vs 121 [49%]) or adverse events that led to discontinuation (ten [4%] vs 17 [7%]). One death was reported in each treatment group, but neither was considered related to the study medications. Interpretation The non-inferior efficacy and similar safety profile of the dolutegravir combined regimen compared with the atazanavir regimen support the use of dolutegravir for HIV-1 infection in treatment-naive women. Funding ViiV Healthcare.