We performed a retrospective analysis of our experience with dual nucleoside regimens to look for predictors of long term benefit. The study evaluated a cohort of 68 HIV-infected patients treated at 3 Italian hospital-based facilities. The results were evaluated using univariate and multivariate statistical analysis. Fourty-three males and 25 females were treated for 22 +/- 14 months. Sixty three patients (92.6%) suffered no or low-grade side-effects. Thirty-four patients (50%) reached a viral load <400 copies/ml (undetectable). Viremia remained persistently undetectable in 9 cases (13.2%). Variable relapses of viremia were seen in 13 patients (19.1%) even though their therapys were not modified. Eight patients (11.8%) showed relapsing viremias persistently around or below 10,000 copies/ml. All patients reaching undetectable viremia but one showed increasing or stable CD4+ cell counts. Factors predicting favourable response were: pre-treatment CD4+ T-cells >150/microl, pre-treatment viremia <50,000 copies/ml, pre-treatment lymphocytes >1,500/microl, and no previous exposure to NRTI. Total lymphocyte counts and CD4+ T-cells showed a significant correlation. Dual NRTI regimens may be still considered for patients unable to tolerate HAART regimens and presenting with favourable predictors of response.