The objective of this study was to evaluate virological and pharmacological determinants of a 24-week virological response to unboosted atazanavir (ATV) in highly drug-experienced HIV-infected patients. Among patients enrolled in the ATV Expanded Access Program, those with HIV-RNA >1000 copies/mL, a genotype performed within three months from the baseline (BL), and who completed 24 weeks of treatment, were included. They received at least three antiretrovirals, including ATV 400 mg once daily without boosting. ATV plasma levels were evaluated after four weeks of treatment by high performance liquid chromatography (HPLC). ATV genotypic inhibitory quotient (GIQ) was calculated as the ratio between ATV Ctrough and the number of the BL ATV-related protease resistance mutations (among the following: 10I/V/F, 20R/M/I, 24I, 331/F/V, 36I/L/V, 46I/L, 48V, 54V/L, 63P, 71V/T/I, 73C/S/T/A, 82A/F/S/T, 84V, and 90M). Thirty-five subjects were included. At baseline, median (interquartile range) CD4+ T-lymphocytes, HIV-RNA, and ATV resistance mutations were 232.5 (106-303)/microL, 4.7 (4.2-5.1) log10 copies/mL, 2 (1-6), respectively. Thirteen (37.1%) subjects were off-therapy and 11 (31.4%) showed no PI mutation at baseline. Median steady-state ATV Ctrough was 230 ng/mL (87-520), for an ATV GIQ of 86.5 (25.5-165.5). Median HIV-RNA changes from baseline at weeks 4, 12 and 24 were -1.76 (from -0.44 to -2.12), -1.41 (from -0.41 to -2.81) and -1.44 (from -0.42 to -2.71) log10, respectively. The HIV-RNA changes were correlated to the number of ATV resistance mutations at each time point (P < 0.05), whereas no correlation was found between ATV Ctrough or ATV GIQ and HIV-RNA changes. In conclusion, the number of ATV resistance mutations is the only correlate to virological response through 24 weeks of treatment with unboosted atazanavir 400 mg once daily.