Background:  The association between ART drug resistance mutations and clinical progression in patients with triple-drug class virological failure has not been reported previously. We aimed to describe the pattern of mutations and examine these mutations as predictors of death within the group of patients with triple-drug class virological failure.

Methods:  All patients with triple-drug class virological failure during 1996-2004 were identified in the Danish HIV Cohort Study. Virological failure was defined as a viral load of >1000 copies/mL for a total of 120 days while receiving treatment with a given drug class. We performed genotypic resistance tests on virus from the closest available plasma sample within the period 1 month before to 6 months after the date of triple-drug class virological failure. We computed time to death from date of triple-drug class virological failure and constructed Kaplan-Meier survival tables. The influence of resistance mutations (International AIDS Society-USA, 2005 criteria) on mortality was assessed using Cox regression analysis.

Results:  Among 2797 patients initiating HAART, 179 developed triple-drug class virological failure and were followed for a median of 4.3 years (IQR 2.8 to 5.5) thereafter. The 133 patients with an available resistance test had a mean of 6.99 mutations, hereof 3.86 nucleoside reverse transcriptase inhibitor (NRTI) mutations, 1.78 non-NRTI (NNRTI) mutations, and 1.33 primary protease inhibitor (PI) mutations; 61% harbored resistance mutations toward all 3 major drug classes, 80% toward at least 2 drug classes, and 89% toward at least 1 drug class. Overall mortality rate was 74 (95%CI 54 to 101) per 1000 person-years. Each additional mutation conferred an increased risk of death of 9%, crude mortality rate ratio (MRR) = 1.09 (1.01 to 1.18), adjusted MRR  = 1.08 (1.00 to 1.17), adjusted for CD4 count and HIV RNA at time of triple-drug class virological failure, and date of triple-drug class virological failure. Thus, patients in the upper half, harboring 9 mutations or more, had increased mortality compared with patients harboring 8 mutations or fewer, adjusted MRR = 2.27 (1.15 to 4.47). Individual mutations significantly associated with increased mortality were: A62V, MRR = 24.02 (4.92 to 117.29); G48V, MRR = 8.63 (3.14 to 23.77); G190A/S, MRR = 3.00 (1.56 to 5.76); E44D, MRR = 2.95 (1.50 to 5.81); T215Y (but not T215F), MRR = 2.75 (1.42 to 5.33); V118I, MRR = 2.31 (1.15 to 4.66); M41L, MRR = 2.24 (1.14 to 4,42); L210W, MRR = 2.03 (1.08 to 3.79); V82F/A/T/S, MRR = 2.03 (1.09 to 3.78); and Y181C/I, MRR = 2.01 (1.06 to 3.81). No mutations were significantly associated with decreased mortality.

Conclusions:  In patients with triple-drug class virological failure, genotypic drug-resistance mutations were associated with increased risk of death, independent of CD4 count, HIV RNA, and date of triple-drug class virological failure.