Background:  HIV genotypic and phenotypic susceptibility testing (G/PT) can optimize ART drug selection, HIV suppression, and CD4 cell responses in ART-experienced patients. We evaluated the influence of G/PT use on mortality rates in a large cohort of HAART-treated patients followed multiple years.

Methods:  We analyzed data from HAART-experienced participants in the HOPS, a multicenter, prospective, observational study of U.S. HIV-infected persons. We studied 2729 HAART-experienced patients from January 1, 1999 to June 30, 2005, whose post-HAART initiation HIV RNA was >1000 copies/cm³. We used logistic regression to assess factors associated with G/PT use and Cox proportional hazards models to compare mortality rates among those who did vs did not undergo G/PT. We adjusted for significant baseline demographics (age, insurance, HIV risk), and baseline CD4 cell counts.

Results:  Of 2729 HAART-experienced patients (median follow-up 2.77 years), 1030 (38%) underwent G/PT, of whom 613 were triple-ART-class experienced; and 1699 (62%) HAART-experienced patients did not undergo G/PT. The median number of HAART regimens before the first G/PT was 2. Patients with nadir CD4 <350 cells/cm³ or aged >40 years were more likely to undergo G/PT (each p <0.001). Mortality rates for those with ≥1 G/PT compared with those with no G/PT were 30.6 deaths/1000 person-years vs 40.9 deaths/1000 person-years, respectively (p = 0.001). In a Cox proportional hazards analysis that adjusted for baseline CD4, age at first viral load test (>40 vs ≤40 years), insurance type (public vs private), and HIV risk (injecting drug user [IDU] vs non-IDU), patients with ≥1 G/PT had a lower hazard for death (hazard ratio [HR] = 0.64, p = 0.001). Among triple-ART-class experienced participants, having had ≥1 G/PT was associated with an even lower hazard for death (HR = 0.23, p <0.001). Comparison of 398 patients who died to 398 survivors, matched on factors adjusted for in the Cox PH model, showed reduced mortality associated with having had ³1G/PT (RR = 0.56, p <0.001).

Conclusions:  In the HOPS, the use of HIV susceptibility testing was strongly associated with improved survival after adjusting for stage of HIV disease, demographics, age, and calendar year. Presumably, this survival benefit reflects long-term improvements in HIV suppression and CD4 counts achievable with G/PT-guided HAART selection. However, G/PT use may also reflect improvements in other aspects of care.