In the first 48 weeks, new AIDS events were observed in 2.7% of late presenters, 0.8% of late starters and 0.9% Selleck Sirolimus of ideal starters (P=0.0001; χ2 test). In contrast, among those who remained alive beyond week 48 and were still under follow-up, the rate of new AIDS events between weeks 48 and 96 was similar in the three groups at 1.3, 1.0 and 0.5%, respectively (P=0.11; χ2 test). Deaths were more frequent in late presenters in the first 48 weeks (2.0%vs. 1.0 and 0.5% in late and ideal starters, respectively; P=0.0003; χ2 test) but among those surviving the first 48 weeks, death rates between weeks 48 and 96 were similar in the three groups (1.0%vs. 1.1 and 1.0% in late and ideal starters, respectively; P=0.96; χ2 test). Overall,

clinical progression (new AIDS events or death) occurred in 4.6% of late presenters, 1.8% of late starters and

1.4% of ideal starters in the first 48 weeks (P=0.0001). The rates of new clinical progression after a year of HAART (i.e. among those who remained alive AG 14699 and under follow-up beyond week 48) were 2.2, 1.9 and 1.3% in late presenters, late starters and ideal starters, respectively (P=0.21). Multivariable analyses (Table 2) suggested that late presenters were at increased risk of a new AIDS event or death in the first 48 weeks (P=0.01) compared with late starters, but that this excess risk was lost if patients survived beyond week 48 (P=0.83). In contrast, clinical progression rates in late starters and ideal starters did not differ significantly, either at 48 (P=0.64) or 96 (P=0.40) Phosphoglycerate kinase weeks. The differences in virological and immunological endpoints between late presenters and late starters were unchanged after additionally controlling for the pre-HAART CD4 cell count and viral load. However, the difference in clinical progression rate in late presenters compared with late starters at 48 weeks was reduced and nonsignificant [adjusted odds ratio (OR) 1.69; 95% confidence interval (CI) 0.93, 3.06; P=0.09]. When we assumed that all individuals who were lost to follow-up or who had missing viral load values at week 48 had not achieved

virological suppression (a missing equals failure approach), virological suppression rates were 55.3, 58.6 and 56.1% in late presenters, late starters and ideal starters, respectively. Differences between late presenters (adjusted OR 1.08; 95% CI 0.91, 1.28; P=0.37) and late starters, and between ideal starters (adjusted OR 0.94; 95% CI 0.79, 1.12; P=0.47) and late starters were not significant in multivariable analyses. Similar results were obtained at 96 weeks. When similar analyses were performed for the clinical outcome, in which patients who were lost to follow-up over the first 48 weeks (and between week 48 and week 96 for the 96-week outcome) were assumed to have experienced clinical failure, failure rates were 16.2, 12.9 and 20.5% in late presenters, late starters and ideal starters, respectively, at week 48 (P=0.0001), and 18.8, 17.7 and 18.8%, respectively (P=0.