Background: BMS-663068 is an oral prodrug of BMS-626529, an attachment inhibitor that binds to HIV-1 gp120, blocking viral attachment to host CD4 cells. AI438011 is an ongoing trial investigating the effi cacy, safety, and dose- response of BMS-663068 in treatment-experienced, HIV-1-infected patients. Herein we present the results of the primary analysis. Methods: AI438011 is a phase 2b, randomised, active-controlled trial, at 53 hospitals and outpatient clinics across ten countries in North and South America, Europe, and Africa. Individuals with an HIV-1 RNA viral load of at least 1000 copies per mL and a BMS-626529 half-maximum inhibitory concentration lower than 100 nmol/L were randomly assigned (1:1:1:1:1) to receive either BMS-663068 at 400 mg twice daily, 800 mg twice daily, 600 mg once daily, or 1200 mg once daily or ritonavir-boosted atazanavir (300 mg of atazanavir and 100 mg of ritonavir once daily), each with 400 mg of raltegravir twice daily and 300 mg of tenofovir disoproxil fumarate once daily as a backbone. The sponsor, participants, and investigators were masked for BMS-663068 dose but not for allocation. Primary endpoints were the proportion of patients with an HIV-1 RNA viral load less than 50 copies per mL (response rate) at week 24 and the frequency of serious adverse events and adverse events leading to discontinuation, up to the week 24 analysis. The primary analyses included all patients who received at least one dose of study drug (modifi ed intention-to-treat population). This study is registered at ClinicalTrials.gov, NCT01384734. Findings: Between July 26, 2011, and July 16, 2012, 581 participants were assessed for eligibility. Of these, 254 patients were randomly assigned to receive either BMS-663068 (n=52 for the 400 mg twice daily group, n=50 for the 800 mg twice daily group, n=51 for the 600 mg once daily group, and n=50 for the 1200 mg once daily group) or ritonavir-boosted atazanavir (n=51). 200 patients received at least one dose of BMS-663068, and 51 patients received at least one dose of ritonavirboosted atazanavir. At week 24, 40 (80%) of 50 patients in the BMS-663068 400 mg twice daily group, 34 (69%) of 49 patients in the 800 mg twice daily group, 39 (76%) of 51 patients in the 600 mg once daily group, and 36 (72%) of 50 patients in the 1200 mg once daily group had an HIV-1 RNA viral load less than 50 copies per mL, compared with 38 (75%) of 51 patients in the ritonavir-boosted atazanavir group. Serious adverse events were noted in 13 (7%) of 200 patients in the BMS-663068 groups and fi ve (10%) of the 51 patients in the ritonavir-boosted atazanavir group. Four (2%) of the 200 patients in the BMS-663068 groups and two (4%) of the 51 patients in the ritonavir-boosted atazanavir group discontinued because of adverse events. No serious adverse events or adverse events leading to discontinuation were BMS-663068-related. Grade 2-4 adverse events related to study drug(s) occurred in 17 (9%) of 200 patients across the BMS-663068 groups and 14 (27%) of 51 patients in the ritonavir-boosted atazanavir group. For the BMS-663068 groups these events were mostly single instances with no dose relation and for the ritonavir-boosted atazanavir group these were mostly gastrointestinal or hepatobiliary disorders associated with hyperbilirubinaemia. Interpretation: In a comparison with ritonavir-boosted atazanavir, effi cacy and safety of BMS-663068 up to the week 24 analysis support continued development of BMS-663068, which is being assessed in a phase 3 trial in heavily treatment-experienced individuals.