Background:  Atazanavir (ATV) is a well tolerated protease inhibitor (PI), but concern exists about its potency. ATV plasma levels might be useful to predict its efficacy and toxicity.

Methods:  Performance of ATV was examined in PI-experienced patients included in the expanded access program in Madrid. ATV plasma levels (C_12h; estimated C_min) were measured (HPLC-UV) at week 12. The genotypic inhibitory quotient was defined as C_12h/number of protease resistance mutations.

Results:  Data from 92 and 52 patients followed for 12 and 24 weeks, respectively, were analyzed:  median age, 42 years; male, 68%; time on prior HAART, 72 months; in patients with baseline viral load >50 copies/mL (65%), the median plasma HIV-RNA was 3.9 log₁₀; median CD4 count, 395 cells/mL; HCV co-infection, 45%; ATV was prescribed as rescue (45%), simplification (11%), or to ameliorate dyslipemia (24%) or other toxicities (16%); concomitant tenofovir (TDF), 78%; none received ritonavir boosting. At 24 weeks, virologic response  (≥1 log drop in HIV-RNA or to <50 copies/mL) was recorded in 64% (OT-analysis) or 57% (ITT). In patients with baseline detectable viral load, median HIV-RNA drop was 0.7 log. Median CD4 gain was 38 cells/ mL. Median ATV C_min was 0.12 μg/mL (IQR, 0.05-0.22 μg/mL). Immunologic and virologic responses did not correlate with ATV C_min. Neither the efficacy of ATV nor C_min significantly varied according to TDF use. The number of protease resistance mutations (IAS-USA list) tended to be associated with the VR: median of 5 protease resistance mutations in failing patients vs 1 in responders (p = 0.07). In subjects with baseline viral load >50 copies/mL, a higher viral load drop was associated with a higher genotype inhibitory quotient (p = 0.02; β -0.6; 95%CI -0.9, -0.2). At 24 weeks, only 4 patients (4%) discontinued treatment due to ATV-related toxicity (1 hyperbilirubinemia). Although bilirubin increased overall in up to 66% of subjects, grade 4 only developed in 10%. Increases in bilirubin at week 12 were associated with higher ATV plasma levels (p = 0.01; β 0.4; 95%CI 0.3 to 1.1). No significant liver enzyme elevations were observed. There were no significant differences in bilirubin nor transaminases according to HCV status. Hypertriglyceridemia and hypercholesterolemia significantly decreased.

Conclusions:  ATV-based regimens provide a significant virologic response in PI-experienced patients, particularly in subjects lacking protease resistance mutations. The genotype inhibitory quotient predicts accurately the virologic response to ATV. Hyperbilirubinemia is associated with ATV plasma levels.

Keywords: atazanavir; GIQ; hyperbilirubinemia