Objectives: The AtLaS-M randomized trial showed that in patients with HIV-1 RNA <50 copies/mL on atazanavir/ ritonavir + two NRTIs, switching to a dual therapy with atazanavir/ritonavir+lamivudine had superior efficacy as compared with continuing the previous triple therapy. This substudy was designed to evaluate at 48 weeks the impact of the dual therapy versus the three-drug atazanavir/ritonavir-based therapy on the HIV-1 cellular reservoir as reflected by the quantification of blood-associated HIV-1 DNA levels. Methods: In a representative subset of 201 of 266 randomized patients (104 in the dual-therapy arm and 97 in the triple-therapy arm) total HIV-1 DNA levels in whole blood at baseline and after 48 weeks and factors associated with the HIV-1 DNA levels were evaluated. Results: The mean baseline HIV-1 DNA levels (2.47 log 10 copies/10 6 leucocytes) were comparable between arms. A significant mean decrease between baseline and week 48 was observed: -0.069 log 10 copies/10 6 leucocytes in the dual-therapy arm (P=0.046) and -0.078 in the triple-therapy arm (P=0.011); the mean difference between arms was -0.009 (P=0.842). Nadir CD4 count was inversely correlated with baseline HIV-1 DNA (P=0.009); longer duration of ART and lower nadir CD4 correlated with a less prominent HIV-1 DNA decrease (both P <0.005). Higher baseline HIV-1 DNA was associated with residual viraemia at week 48 (P=0.031). Conclusions: When compared with continuing three-drug therapy, atazanavir/ritonavir+lamivudine dual therapy resulted in a similar decline in HIV-1 DNA levels in patients with sustained virological suppression. These data support the safety of this simplified treatment strategy in terms of its effect on the cellular HIV-1 reservoir. © The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.