A cohort of 185 HIV-infected injection drug users (IDUs) and seronegative controls was followed with semiannual neuropsychological assessments for up to 4.5 years. Changes in cognitive performance over time were evaluated, and results of seronegative controls were used to adjust for level of education and practice effects. The effects of duration of follow-up, decline in CD4+ count, development of clinical symptoms, antiretroviral use, and diagnosis of AIDS on changes in neuropsychological performance over time were assessed with regression models using the generalized estimating equation approach. Improvement in performance over time, consistent with practice effects, was observed for all measures. The only subtest for which the magnitude of the practice effects was mildly attenuated relative to the seronegative controls was Grooved Pegboard, dominant hand. After adjusting for disease progression and antiretroviral therapy use, none of the time trends for the neuropsychological test scores were significant, suggesting no decline in performance of the seropositive patients relative to the seronegative controls. With development of clinical symptoms, there was a trend in the direction of declining performance. For subjects reporting two or more symptoms but not using antiretroviral therapy, the trend was not significant, whereas having two or more symptoms and using antiretroviral therapy was associated with significantly worse performance on tests of psychomotor speed and memory. With development of AIDS, a significant decline in performance was observed on measures of motor and psychomotor speed as well as memory. There is thus no evidence to suggest that HIV infection in the context of chronic drug and alcohol use significantly alters the frequency or rate of progression of cognitive symptoms. These findings suggest that the natural history of cognitive changes secondary to HIV infection is similar among HIV-infected IDUs and other risk groups such as homosexual/bisexual men.
