Gammaglutamyltranspeptidase (gamma-GT) levels before treatment are predictive of Tipranavir interruption in ART multiexperienced HIV-1-infected patients

The Italian Tipranavir EUP/EAP study was an observational study involving nine Italian centres that enrolled three class drug-experienced patients who consecutively entered the EUP/EAP TPV programs. The Cox model performed to assess the risk correlates of interrupting TPV included sex, age, HIV risk factors, HBV/HCV status, use of T20, CD4, HIV-RNA, ALT, γ-GT, cholesterol, triglycerides, glucose levels determination before starting TPV. The study enrolled 175 patients followed up for a median time of 30 weeks (range 3-68). TPV was interrupted by 46 patients (16 for intolerance, 14 for immuno-virological failure, four for disease progression – including two deceased – 12 for patient decision). The factors independently associated with treatment interruption for any cause were previous ART duration (OR 1.18, 95%CI 1.03-1.35, p=0.016 per each additional year) and γ-GT BL (grade 2 vs 0: OR 6.31, 95%CI 2.49-16.4, p< 0.0001). The γ-GT BL median level in the 16 patients who interrupted TPV for intolerance was 122 IU/L (range: 11-352). Both γ-GT and ALT were significantly increased at interruption compared to BL (p=0.041 and p=0.016 respectively, Wilcoxon test). A transient protective effect against γ-GT and TG increase was observed in patients receiving T20 with TPV (p=0.034 and p=0.027 respectively at week 12 in a Cox model assessing the risk of increasing one toxicity grade). γ-GT level seems to be relevant and more sensitive than ALT/AST level or HBV/HCV status in predicting the risk of TPV interruption. Unreported alcohol abuse could be considered a potential cofactor limiting the effectiveness and safety of TPV/RTV treatment.