The impact of prior antiretroviral experience on adherence and clinical outcomes in patients initiating a new highly active antiretroviral therapy (HAART) regimen is not well defined. We performed an observational cohort analysis of antiretroviral-experienced or -naive HIV-infected patients prescribed a new HAART regimen (3538 patients) and enrolled in Kaiser Permanente Northern California from 1997 through 2002. Outcomes evaluated were HAART adherence and changes in HIV RNA level and CD4 T-cell counts over 12 and 24 months. The antiretroviral-naive group had a significantly greater odds of achieving >/=95% adherence to the HAART regimen over 12 months (adjusted odds ratio [OR] = 1.88, p < 0.001) and over 24 months (OR = 1.66, p < 0.001). The odds of achieving HIV RNA levels below limits of quantification also was higher among patients who were antiretroviral naive; over 24 months OR = 6.88, p < 0.001. Adjusted change in CD4 T-cell count was also greater among naive patients over 24 months (+67 cells/muL, p < 0.001), compared to antiretroviral-experienced patients. Years of antiretroviral experience did not affect adherence or any outcome measure. Adjusting for HAART adherence did not significantly affect HIV RNA results but did lessen CD4 T-cell count differences between antiretroviral-naive and -experienced patients. Thus, antiretroviral naive patients have improved HAART adherence, HIV RNA control, and CD4 T-cell count increases compared to antiretroviral-experienced patients regardless of years of antiretroviral experience. These findings should help direct HAART adherence efforts in HIV care clinics.