The objective of this study was to find predictive factors of lopinavir/ritonavir (LPV/r) discontinuation for drug-related toxicities in highly pre-treated human immunodeficiency virus (HIV)-infected subjects. The study was an observational study of HIV patients starting LPV/r with HIV RNA > 3log10 copies/mL and a follow-up > or = 6 months. Parameters studied were HIV RNA, CD4+ cell counts, metabolic parameters and drug-related adverse events. Acquired immune deficiency syndrome (AIDS) events and deaths were recorded. The Kaplan-Meier (KM) model was used to estimate time-dependent probability, and the multivariable Cox model to identify predictors of LPV/r discontinuation for adverse events. The study evaluated 416 HIV-infected patients. Seventy-seven patients (18.5%) discontinued LPV/r for toxicities. Adverse events leading to LPV/r discontinuation were gastrointestinal symptoms in 40 cases, hyperlipidaemia in 27 and increase of aspartate aminotransferase (AST)/alanine aminotransferase (ALT) in 10 patients. Nineteen patients (4.6%) developed an AIDS event during observation and 15 (3.6%) died. The KM probability of LPV/r discontinuation for toxicities was 5.3% (range 3.1-7.5%) at month 12 and 15.7% (range 12.1-19.3%) at month 24. Subjects with hepatitis C virus (HCV)-HIV co-infection (odds ratio (OR) 7.40; 95% confidence interval (CI) 3.73-14.66 versus HCV-negative; P = 0.001) and receiving LPV/r plus nucleoside reverse transcriptase inhibitors (NRTIs) and protease inhibitor (PI)/non-nucleoside reverse transcriptase inhibitor (NNRTI) (OR 1.74; 95% CI 1.04-2.91 versus LPV/r plus only NRTIs; P = 0.04) showed a higher risk of LPV/r discontinuation by a Cox analysis, whereas non-intravenous drug abusers (IVDUs) (OR 0.40; 95% CI 0.24-0.67 versus IVDUs; P = 0.001) had a lower risk. The rate of discontinuation for toxicity decreased by 17% for each additional month of LPV/r exposure (OR 0.83; 95% CI 0.80-0.86 for each additional month; P < 0.001). LPV/r was substantially well tolerated. Diarrhoea was the most frequent adverse event leading to discontinuation. HCV-HIV co-infected patients and patients with a short exposure to LPV/r have a higher risk of discontinuing LPV/r and should be strictly monitored.