Viral tropism by geno2pheno as a tool for predicting CD4 decrease in HIV-1-infected naive patients with high CD4 counts

OBJECTIVES To investigate the value of tropism (determined by genotypic testing) to predict CD4 depletion in HIV-infected antiretroviral-naive patients with high CD4 counts. METHODS Viral tropism was determined by geno2pheno (false positive rate = 10%) in 223 HIV-infected subjects naive to antiretrovirals with CD4 count ≥350 cells/μL and HIV-RNA >500 copies/mL enrolled in the ICONA Foundation Study for whom a stored plasma sample (baseline) was retrospectively tested. We monitored CD4 cell count and identified predictors of decline before antiretroviral therapy initiation, applying a mixed linear model with covariates (age, gender, tropism, HIV risk factor, calendar year of HIV infection, months from HIV diagnosis to baseline, hepatitis C virus status, CD4 and HIV-RNA at sample collection and duration of follow-up). RESULTS Two hundred and twenty-three subjects met the eligibility criteria; 137 (61%) were male and the median age was 35 (31-40) years. Median follow-up was 16.4 (3.2-37.2) months. Median CD4 decrease during follow-up was -157 (-278 to -13) cells/μL. At baseline, 192 (86%) subjects were defined as harbouring R5 virus and 31 (14%) non-R5. Median CD4 count was 571 (458-729) cells/μL and median HIV-RNA was 4.08 (3.57-4.55) log(10) copies/mL. At multivariable analysis, a greater mean CD4 decrease was associated with non-R5 viral tropism (-159.9 ± 12.22, P = 0.0002) at baseline. Other significant covariates were female gender, older age, intravenous drug use, longer duration of follow-up, and higher CD4 cell count and higher HIV-RNA at sample collection. CONCLUSIONS In patients with CD4 counts ≥350 cells/μL, non-R5 viral tropism by geno2pheno is predictive of CD4 decrease independent of their viral set point and CD4 counts.