Variable benefit in neuropsychological function in HIV- infected HAART- treated patients Abstract—The authors examined cognitive performance change in 101 indi- viduals with advanced HIV infection on highly active antiretroviral therapy (HAART), using standard neuropsychological testing in three visits, over a 27-month-period. Cognitive performance stabilized in a majority of HIV+ par- ticipants over time. A neuroactive HAART regimen was associated with neuro- psychological improvement. Decline occurred in a minority with lower nadir CD4. The current CD4 count and plasma viral load were not associated with cognitive change. NEUROLOGY 2006;66:1447–1450 Lucette A.J. Cysique, PhD; Paul Maruff, PhD; and Bruce J. Brew, MBBS, MD, FRACP In the highly active antiretroviral therapy (HAART) era, AIDS dementia complex (ADC) 1 still occurs, but prospective studies addressing illness activity and stability have had shortcomings 2-3 : entry has been restricted to those already with impairment, and the observation period has been relatively short. Therefore, we sought to prospectively examine the neuropsychological performance in individuals at risk of HIV-related cognitive change and treated with HAART for 5 years. 4 Methods. Subjects. One hundred HIV+ individuals with stage C3 HIV disease were randomly invited to participate from the outpatient clinics at St. Vincent’s Hospital. Eighty-one individuals came for a second visit at 6 months, 51 for a third visit at 15 months, and 38 for a fourth visit at 27 months. Demographic, clinical, and laboratory data are presented in table 1. Thirty sero- negative controls matched for age and education were recruited to develop norms for change and assessed at one follow-up session. 4 Procedure. The neuropsychological battery included the as- sessment of attention, learning, memory, motor coordination, com- plex attention/psychomotor speed, language, visuoconstruction (see Cysique et al. 4 for additional details). Data analysis. Decline was classified using the within SD (WSD)– based reliable change index (RCI). 5 The control group WSD was used as the statistical reference of stability over time to compute the RCI in all individuals in the short term (6 months) and long term (27 months). To identify HIV+ individuals with significant cognitive decline over the first 6 months, we defined it as an RCI of less than -1.96 on two or more of the 10 tests. 6 To identify which of the neuropsychological tests was most sensitive to HIV-related cognitive change, we compared performance on each of the 10 measures between the HIV+ with cognitive decline and HIV+ without (nondecliners) using t tests. Six tests sensitive to HIV-related cognitive decline over the short term were used to compute a composite change score (composite RCI) to identify decline over the long term. Standardized composite RCIs of -1.96 and -1.5 were classified as abnormal (i.e., significant decline). 6 Relationships between demographic, treatment-related, and clini- cal variables were explored with the composite RCI and multiple regression as well as Pearson correlations when appropriate. 7 To explore a potential attrition effect, HIV+ individuals who dropped out were compared at baseline with the patients who remained in the study on all variables, and no differences were found. The rate of decline between dropout cases and remaining participants was also explored and no differences were found. Results. Over the short term, the prevalence of neuro- psychological decline was higher in the HIV+ group than in controls (30% vs 13%, p  0.05). HIV+ decliners and nondecliners differed on six neuropsychological measures (effect sizes 0.4): domains of verbal learning, memory, motor coordination, psychomotor speed, and complex at- tention (these were retained to compute the composite RCI). When decline was defined as an RCI of less than -1.5 SD and based on the six measures to develop the composite RCI, we found that 13 individuals were classi- fied as decliners. When the 10 initial measures were used to develop a composite RCI, we also found that 13 individ- uals were classified as decliners. Therefore, fewer tests did not alter significantly the frequency of classification of de- cline. Moreover, of the 13 decliners, nine individuals were identical in both composite RCI representing a 70% agree- ment in the classification of decline. Based on the composite RCI, the prevalence of decline varied between 5.88% and 13.72% at session III and re- mained static at 5.26% at session IV The profile of change for individual cases is illustrated in the figure. Cognitive decline over the long term was related to lower nadir CD4 cell counts, past depressive episode, past HIV-related brain diseases, and initial number of AIDS- defining illnesses. Individuals with a longer disease dura- tion, older individuals, and individuals with lower level of education showed improvements in cognitive performance over time. Self-reported anxiety and depressive symptoms were also associated with decline in cognitive functioning. Complex attention was positively associated with the pres- ence of at least three neuroactive antiretrovirals compos- ing HAART (neuroactive was defined as a HAART regimen including at least three neuroactive antiretrovirals 8 ). Cur- rent CD4 cell count and plasma viral load were not associ- ated with cognitive performance overtime (table 2). Discussion. In this study, we examined the extent to which cognitive function changed overtime in HAART-treated individuals with advanced HIV in- fection. First, we found that 30% of HIV+ partici- pants showed reliable cognitive decline over the short term, but the majority improved over the long term, albeit with considerable variability. Second, change in cognitive performance was not related to From the Faculty of Medicine (L.A.J.C.), St. Vincent’s Clinical School, Uni- versity of New South Wales, Sydney, Australia; School of Psychology (P.M.), LaTrobe University, Melbourne, Australia; Departments of Neurology and HIV Medicine (B.J.B.), St. Vincent’s Hospital, Sydney, Australia. Disclosure: The authors report no conflicts of interest. Received October 18, 2005. Accepted in final form January 23, 2006. Address correspondence and reprint requests to Dr. Lucette Cysique, HNRC, 150 West Washington Street, 2nd Floor, San Diego, CA 92103; e-mail: lcysique@ucsd.edu Copyright © 2006 by AAN Enterprises, Inc. 1447