Background:  Despite achieving adequate suppression of plasma HIV- RNA, incomplete immune recovery (IIR) occurs in 5-15% of patients on HAART. CVE and NAM may be related to immunodeficiency. Therefore, we studied the relation between IIR on HAART and the occurrence of cardiovascular events (CVE) and nucleoside analog mutation (NAM).

Methods:  Patients treated with HAART for at least 2 years with virological suppression (< 500 HIV-RNA copies/mL) for at least the previous 6 months were selected from the ATHENA cohort. Patients were grouped based on immune recovery after 2 years (baseline): CD4 cell count < 200 (A), 200 to 350 (B), 350 to 500 (C) and > 500 CD4 cells/µL (D). Kaplan Meier (KM) estimates and Cox proportional hazard models were used to analyze primary outcome measures: time from baseline to occurrence of CVE, time to NAM and time to occurrence of any of CVE, malignancy, AIDS, or death (‘event’).

Results:  In this study, 3071 patients were included and were allocated to groups A, B, C and D, respectively: 200 (6.5%), 646 (21.0%), 1414 (46.0%) and 811 (26.4%). 83% were male and 57% infected through MSM. Median age at baseline was significantly higher in the groups with IIR: 45.1 (A), 43.1 (B), 41.3 (C) and 41.8 (D) (overall p<0.0001). Mean CD4 cell count at baseline was 142 (A), 255 (B), 431 (C) and 728 (D) cells/µL. During follow up 198 events occurred. In cohort A significantly more events occurred (KM estimate of the percentage of patients with an event within 5 years from baseline 16.7%, compared with 9.9% (B), 9.3% (C) and 6.6% (D); overall log rank, P ≤0.0001). Hazard ratios (HR) adjusted for age, sex and transmission mode of groups B, C and D compared with A were 0.67 (95% Confidence Interval (CI) 0.42 to 1.07), 0.62 (0.40 to 0.96), and 0.47 (0.29 to 0.76). During the study, 53 CVE and 42 NAM occurred. There was a non-significant trend for shorter time to NAM in group A (KM estimate within 5 years from baseline 5.7% compared with 2.2% (B), 1.5% (C) and 1.8% (D), overall log rank, P =0.27). Significantly more CVE occurred in group A (KM estimate of the percentage of patients with a CVE within 5 years from baseline 4.7% compared with 1.9% (B), 2.4% (C) and 2.0% (D), overall log rank, P =0.02). Adjusted for age HR of groups B, C and D compared with A were 0.41 (95%CI 0.16 to 1.05), 0.71 (0.33 to 1.56), and 0.60 (0.25 to 1.41).

Conclusions:  Incomplete immune recovery (<200 CD4 cells/µL) after 2 years of HAART was associated with older age, more events, more CVE, and a trend for more NAM. New strategies to improve immune recovery in patients with incomplete immune recovery are warranted.