5%) in the NNRTI find more group and one patient (1.9%) in the PI group had undetectable viral load at baseline, defined as HIV RNA < 400 HIV-1 RNA copies/mL.

Patients in the NNRTI group had a significantly higher CD4 count than those in the PI group (452 vs. 221 cells/μL, respectively; P < 0.01). These differences could be explained by the fact that many patients were switched from a PI-based regimen to an NNRTI-based regimen when these drugs became available. Regarding NVP users, 50% of female patients and 40% of male patients had CD4 counts < 250 and < 400 cells/μL, respectively, at the start of the treatment. In 2006, the new therapeutic strategy was implemented which restricted the use of NVP to patients with CD4 cell counts below these cut-off values, because higher CD4 cell counts were shown to be associated with an increased risk of hepatotoxicity [8]. The results of viral hepatitis coinfection (both HBV and HCV) evaluations were available for 92.6% of all patients. During NNRTI therapy, 14.8% of the study population experienced a > 2.5-fold elevation in serum ALT (grade ≥ 2) (Fig. 1). A total of 21 events of moderate and five events of severe liver toxicity

were observed during 691 person-years of therapy (PYT) with NNRTI (3.04 and 0.72 per 100 PYT, respectively). A subanalysis showed an equal risk for the development of hepatotoxicity in patients using NVP and those using EFV (16.7% vs. 13.8%, respectively; P = 0.51). Regarding the incidence of severe hepatotoxicity, two events in the EFV group PI3K inhibitor (0.47 per 100 PYT) and three events in the NVP group (1.1 per 100 PYT) were Oxalosuccinic acid observed (P = 0.37). The baseline CD4 counts in these three NVP-using patients with severe LEEs before the start of HAART were 508, 120 and 19 cells/μL, respectively. No significant difference in moderate hepatotoxicity between NVP and EFV was demonstrated

(1.8 vs. 3.3 per 100 PYT, respectively; P = 0.250). In the PI group, 10 patients (18.5%) showed at least grade 2 hepatotoxicity; 22 events of moderate and three events of severe hepatotoxicity were seen during the 468 PYT, with no significant difference in incidence between the NNRTI and PI groups (14.8% vs. 18.5%, respectively; P = 0.52). However, the two groups differed significantly in the baseline incidence of HCV coinfection, which is known to be associated with an increased risk of hepatotoxicity [1]. Excluding all HCV-positive patients from the analysis gave a cumulative incidence of 12.3% for NNRTI-using patients vs. 9.1% for those using PIs (P = 0.57). In the univariate analysis, only HCV coinfection was associated with the development of hepatotoxicity in the NNRTI group [odds ratio (OR) 1.83; 95% confidence interval (CI) 1.33-4.24; P < 0.01]. Hepatotoxicity was observed in 50% of coinfected patients compared with 12.3% in patients without HCV infection (P < 0.01).