Background:  The association between the accumulation of resistance mutations and mortality remains unclear. Here, we present an updated analysis of the combined impact of adherence and accumulated antiretroviral (ARV) resistance mutations on non-accidental HIV-related mortality in an ARV-naïve cohort of 2506 individuals beginning HAART between 1996 and 2004 in British Columbia, Canada.

Methods:  Of the 2506 people in the study, HIV drug-resistance genotypes were available for 1472 (58.7%) individuals. Resistance events were determined via genotyping of patient-derived plasma samples, and interpreted using a modified International AIDS Society-USA (IAS-USA) list (n = 6073 resistance genotypes). The accumulation of drug-resistance classes was described for 3 adherence strata (defined as: <50%, 50 to 95%, and ≥95%) using a mean cumulative function analysis over a median of 5.6 years of follow-up. Resistance categories were defined as lamivudine (3TC), NRTI (except 3TC), NNRTI, and protease inhibitors (PI). Analysis of the causes of death was performed using ICD-9 and ICD-10 codes.

Results:  Within 5.6 years of initiating HAART, 247 of the 1472 participants (16.8%) died and 1188 survived (80.1%). A total of 37 (3.1%) were not on therapy for at least 5.6 years. The mean accumulated drug class resistance mutations (mean cumulative function score) was 0.81 (95%CI = 0.75 to 0.88) for the survivors, and 1.33 (1.09 to 1.56) for the non-survivors. In the <50% adherence stratum, the overall mean cumulative function score was 0.72 (0.59 to 0.85) (survivors = 0.63; 0.50 to 0.77; non-survivors = 1.14; 0.71 to 1.58). At 50 to 95% adherence, the overall mean cumulative function score was 1.09 (0.98 to 1.21) (survivors = 1.00; 0.88 to 1.13; non-survivors = 1.58; 1.18 to 1.98). At ≥95% adherence, the overall mean cumulative function score was 0.82 (0.75 to 0.88) (survivors = 0.77; 0.68 to 0.87; non-survivors = 1.28; 0.87 to 1.69). In agreement with previous results, mean accumulated HIV drug-class resistance mutations were greatest in the intermediate adherence group. Results were similar when all-cause mortality instead of HIV-related mortality was considered.

Conclusions:  Mean accumulated HIV drug-class resistance mutations are greater in non-survivors than survivors, across all adherence strata, but in all cases a median of fewer than 2 resistance categories had accumulated. This suggests that there is very little accumulated resistance to the 4 drug categories across all adherence strata in both survivors and non-survivors, but that resistance is a risk factor for mortality.