The previous therapeutic regimen did not influence the choice of boosted or unboosted ATV. In both groups, the main reason for switching therapy to ATV was virological failure; treatment simplification was the reason for 14.5% of switches to boosted ATV and 22.3% of switches to unboosted ATV. More patients on boosted ATV had switched because of lipid alterations and hepatotoxicity. No differences in backbone therapy were detected between the two groups; in particular, there was no

selleck difference in the use of TDF plus another nucleoside reverse transcriptase inhibitor (NRTI) (Fig. 1). Reasons for using unboosted ATV were: low RTV tolerance (42.3%), nonavailability of the 150 mg ATV formulation (12.3%), lower pill burden (9.2%), better expected compliance (6.2%), impaired liver function (6.2%), hyperlipidaemia (2.3%), other (16.2%) and unknown (5.3%). Therapy outcomes are reported in Table 2. The mean overall observation

time was 23.9 months [standard deviation (SD)±14.8 months]; 24.4 months (SD±14.4 months) for the boosted ATV group and 22.5 months (SD±15.9 months) for patients receiving unboosted ATV. Safety outcomes confirmed the results of several previous studies: hyperbilirubinaemia was the main grade 3–4 AE causing ATV interruption, more frequently in patients taking ATV/r [11 (2.9%) vs. 2 (1.5%)]. No treatment interruptions were reported for grade 3–4 hypertriglyceridaemia. At the BYL719 nmr end of follow-up, similar proportions of patients remained on ATV: 58.5% on unboosted and 58.1% on boosted; respectively, 27.7% and 30.3% had stopped the therapy and 13.9% and 11.6% were lost to follow-up. Data were not available regarding whether patients who interrupted ATV remained without any treatment or switched to another regimen. The mean time

to stopping ATV was 12.6 months in the unboosted ATV group and 14.9 months in the boosted ATV group; survival analysis found no difference in treatment times between the two groups, including patients taking ATV with TDF (Fig. 1; data truncated at 50 months because fewer than 20 patients remained at risk). No differences PIK3C2G were observed in the efficacy of ATV between the formulations or among the single causes of therapy interruption, which were virological failure, death, AEs, patient’s decision, or other reasons, after adjustment for multiple comparison. Regarding the causes of death, one patient died of sudden coronary death, one of nonspecified polyserositis, one of overdose and one for unknown reasons; the other deaths were related to existing terminal diseases: wasting syndrome (one patient), chronic respiratory failure (one), nonspecified cancer (two), hepatic cirrhosis (four) and lymphoma (two).