From the 12-month time-point, 36 of 2052 patients died (5566 person-years). Overall, 52.0% of deaths after 8 months (26 of 50) and 50.0% of deaths after 12 months (18 of 36) were

in discordant responders. In an unadjusted analysis, the risk of an AIDS event after either 8 or 12 months was not significantly different for discordant and concordant responders. GSK269962 cell line However, the risk of death was higher for discordant responders at both 8 months (IRR 2.27, 95% CI 1.30–3.95, P=0.004) and 12 months (IRR 3.19, 95% CI 1.66–6.14, P<0.001). After adjusting for age, baseline viral load and CD4 cell count, and having an AIDS event prior to the follow-up at 8 and 12 months, the risk of death was still higher for discordant responders at 8 months (IRR 2.08, 95% CI 1.19–3.64, P=0.01) and 12 months (IRR 3.35, 95% CI 1.73–6.47, P<0.001) (Table 6). At 8 months, the

risk of death was also slightly higher in those who were older (IRR 1.03 per additional year, 95% CI 1.00–1.06, P=0.048); however, baseline viral load, CD4 cell count and having had an AIDS event prior to the point of determining discordancy were not significantly associated with death. At 12 months, older age was again associated with an increased risk of death (IRR 1.03, 95% CI 1.00–1.07, P=0.050), with a higher baseline CD4 count being associated with a reduced risk (IRR 0.63 per 100 cells/μL increase, 95% CI 0.44–0.90, P=0.012). The risk of an AIDS event in the adjusted analysis was only significantly NVP-BGJ398 associated with baseline viral load when discordancy was categorized at 8 months (IRR 1.82, 95% CI 1.14–2.88, P=0.011). Despite the efficacy of HAART in suppressing HIV viral replication, a rather large proportion CYTH4 of individuals experienced a limited increase in CD4 cell count, or no increase, by around 8 or 12 months. Such responses, assessed at 12 months and, to a lesser extent, at 8 months, were associated with poorer outcome. In many patients (35% of those evaluable)

the discordant response was transient, on the definition used here, with a >100 cells/μL increase by 12 months, even though this was not achieved earlier. Changing treatment was not associated with a change in status between 8 and 12 months. This suggests that the later improvement in CD4 cell count seen in some patients categorized early as having a suboptimal CD4 response was a consequence of a continued, albeit slow, recovery of immune function on HAART, rather than a result of a change of regimen to one with greater potency with respect to restoration of immune function. The incidence of a discordant response in this study was 32% at 8 months and 24% at 12 months. These rates need to be seen in the particular context of the inclusion criteria for the study, which were intended to select a homogeneous group of patients with respect to an early virological response, and to ensure the availability of follow-up data.