Key points are not available for this paper at this time.
The life expectancy of HIV-infected patients has dramatically changed in the last 2 years since the introduction of triple drug combinations, mainly those including protease inhibitors 1. However, issues on compliance and adverse effects have now became evident as a limiting cause of benefit in a substantial proportion of individuals 2. Amongst others, liver toxicity is now one of the leading reasons for withdrawn treatment 3–6. We analysed the first 187 patients who began highly active antiretroviral therapy (HAART) in a reference centre for HIV/AIDS located in Madrid. Examination of liver function was made at baseline and after 1 month of starting treatment. Criteria for liver injury were established when (i) both alanine and aspartate aminotransferase levels increased at least twofold with respect to baseline values; (ii) bilirubin levels were raised above 2.5 g/l; and (iii) development of clinical hepatic failure (ascites, encephalopathy, jaundice) occurred. The mean age of the study population was 37.5 ± 8 years, and 157 (84%) were men. Study subjects had acquired HIV infection through injecting drug use (48%), homosexual contacts (37.4%), and heterosexual relationships (14.5%). Most (84.1%) had received nucleoside analogues, and had detectable levels of HIV plasma viraemia or low CD4 cell counts when they began HAART. Protease inhibitors were administered at different rates: indinavir (73.8%), saquinavir (10.2%), and ritonavir (7.5%). Moreover, the combination of ritonavir plus saquinavir was administered in 8.6% of patients. The most frequent nucleoside combinations were as follows: stavudine plus lamivudine (66.3%), zidovudine plus lamivudine (11.8%), stavudine plus didanosine (9.6%), zidovudine plus zalcitabine (5.3%), and zidovudine plus didanosine (1.6%). A significant increase in transaminases (more than twofold) was recognized in 26 (13.9%) patients, whereas bilirubin increases above 2.5 g/l were seen in seven (3.7%) individuals. Eleven (5.9%) subjects needed to stop the medication because of either hepatic cytolysis (nine patient), or hyperbilirubinaemia (one patient), or both (one patient). Four (2.1%) individuals developed clinical hepatic decompensation, and one of them died. This patient was receiving stavudine, lamivudine plus indinavir, and he had been diagnosed with multiple chronic hepatitis (hepatitis B, C and D), although this was his first episode of clinical liver failure. In injecting drug users (IDU), baseline values of alanine and aspartate aminotransferases were significantly higher (P 0.5 g/l) occurred more frequently in IDU (31 out of 86; 36%) than in the other patients (25 out of 93; 26.9%; P < 0.001). Similarly, transaminase values raised more than twofold in 17 (19.8%) out of 86 IDU, but in only eight (8.6%) of the remaining 93 subjects (P = 0.08). Patients receiving indinavir developed hyperbilirubinaemia more frequently (56 out of 138; 40.6%) than subjects receiving other protease inhibitors (three out of 49; 6.1%; P < 0.001). Any nucleosides analogue was more hepatotoxic than the others. Chronic HCV infection was an independent predictor of hepatotoxicity: transaminase levels increased more than twofold in 16 (21.1%) out of 76 subjects with chronic HCV infection, but only in four (7.4%) out of 54 without chronic HCV (P = 0.03). In contrast, no association was found between development of significant hyperbilirubinaemia and chronic HCV (P = 0.2). In conclusion, hepatotoxicity is frequently seen in patients under HAART, and can force the withdrawal of antiviral treatment in a significant proportion of patients, occasionally resulting in fatal outcome. The development of cytolytic injury is particularly common in subjects with chronic HCV, and meanwhile cholestasis seems to be more frequent in patients under indinavir-containing regimens. Since most HIV-positive IDU are chronically infected with HCV, clinicians must be alert on the hepatic complications that can follow the introduction of HAART in these patients.
Rodrı́guez-Rosado et al. (Wed,) studied this question.