In the NNRTI group, eight events of hepatotoxicity in 122 PYT were observed in the first year of therapy (6.6%), while for the whole period beyond 1 year 16 episodes in 569 PYT were found (2.8%; P = 0.04). Thus, the risk of developing hepatotoxicity was significantly higher in the first year after NNRTI treatment initiation. All hepatotoxic events in our DNA Synthesis inhibitor population occurred in 18 patients; four of them (22.2%) accounted for multiple LEEs over the years. All of these patients continued their NNRTI use

despite these multiple events. Five patients (4.1%) accounted for the five events of severe hepatotoxicity; none of them discontinued therapy because of this severe event, as the LEE had either resolved spontaneously or was attributed to other medication which was adjusted or stopped. One hundred and four patients (85.2%) did not show any clinically relevant hepatotoxicity. This retrospective cohort analysis shows that prolonged use of NNRTIs (≥ 3 years) is not accompanied by an increasing incidence of hepatotoxicity compared with the first year of NNRTI use. We did not find a difference in the risk for developing hepatotoxicity between patients using either EFV or NVP for ≥ 3 years. HCV coinfection was independently associated with the development of LEEs during NNRTI treatment. The incidence of hepatotoxicity did not differ significantly between

the NNRTI and PI groups. To date, a few studies have reported on the liver safety of long-term PFT�� order use of NVP and EFV [6, 9-11]. Most of these studies gave rates of discontinuation because of hepatotoxicity, but did not give the exact number of hepatotoxic events or describe HSP90 the time course. The significantly higher risk of liver toxicity in patients with an HCV/HIV coinfection using NNRTI has been reported before [1, 12]. The intriguing question is whether the occurrence of LEEs in these patients is indeed a marker of drug toxicity or the result of liver enzyme fluctuations in the context of chronic viral hepatitis infection [13]. It is remarkable that, although a higher proportion of patients in the PI group were HIV/HCV-coinfected,

there was no difference with the NNRTI group in terms of the number of hepatotoxic events. We observed a distinct pattern in the incidence of hepatotoxic events over the years of therapy. The number of hepatotoxic events in the first year of NNRTI therapy was significantly higher than in the period that followed. It seems that the number of events declined over the years, even in patients who had already experienced moderate to severe hepatotoxicity in the first year. This observation suggests that it is safe to continue NNRTI-based HAART, even in case of (asymptomatic) hepatotoxicity in the first year of therapy. The debate regarding the pathogenesis of NNRTI-induced hepatotoxicity is ongoing.