Gender difference in HIV-1 RNA viral loads

Objectives To test and characterize the dependence of viral load on gender in different countries and racial groups as a function of CD4 T-cell count. Methods Plasma viral load data were analysed for > 30 000 HIV-infected patients attending clinics in the USA [HIV Insight (TM) (Cerner Corporation, Vienna, VA, USA) and Plum Data Mining LLC (East Meadow, NY, USA) databases] and the Netherlands (ATHENA database; HIV Monitoring Foundation, Amsterdam, Netherlands). Log-normal regression models were used to test for an effect of gender on viral load while adjusting for covariates and allowing the effect to depend on CD4 T-cell count. Sensitivity analyses were performed to test the robustness of conclusions to assumptions regarding viral loads below the lower limit of quantification (LLOQ). Results After adjusting for covariates, women had (nonsignificantly) lower viral loads than men (HIV Insight (TM): - 0.053 log(10) HIV-1 RNA copies/mL, P = 0.202; ATHENA: - 0.005 log(10) copies/mL, P = 0.667; Plum: - 0.072 log(10) copies/mL, P = 0.273). However, further investigation revealed that the gender effect depended on CD4 T-cell count. Women had consistently higher viral loads than men when CD4 T-cell counts were at most 50 cells/mu L, and consistently lower viral loads than men when CD4 T-cell counts were greater than 3 50 cells/mu L. These effects were remarkably consistent when estimated independently for the racial groups with sufficient data available in the HIV Insight (TM) and Plum databases. Conclusions The consistent relationship between gender-related differences in viral load and CD4 T-cell count demonstrated here explains the diverse findings previously published.