Search for a command to run...
Highly active antiretroviral therapy (HAART) has substantially improved the immune function in many HIV-infected individuals. However, the occurrence of milder forms of neurocognitive dysfunction persists, possibly caused by the poor penetration of some antiretroviral agents into the central nervous system. This study reports on the changes in the incidence of mild neurocognitive impairment since the introduction of HAART in individuals initially found to be neurocognitively normal. Highly active antiretroviral therapy (HAART), the clinical strategy of combining antiretroviral medications from different classes, has restored immune function for many immunosuppressed individuals infected with HIV. Because some antiretroviral agents penetrate poorly into the central nervous system, patients may remain susceptible to HIV-associated neurocognitive disorders despite viral suppression in other tissues. This theoretical susceptibility is supported by observations of increased proportions of patients with HIV-associated dementia compared with those with other AIDS-defining illnesses [1], and in the proportion of new HIV-associated dementia cases occurring at higher CD4 cell counts [1,2]. Together, these observations support the theory that HAART treats the central nervous system sub-optimally. Milder forms of HIV-associated cognitive impairment also occur, and can impact social and occupational functioning. Two previous studies reported the benefits of HAART on neuropsychological performance [3,4]. However, to our knowledge, no published study has reported on the changes in the incidence of mild neurocognitive impairment since the introduction of HAART on individuals who initially presented as being neurocognitively normal. The San Diego HIV Neurobehavioral Research Center (HNRC) has performed repeated detailed neurological and neuropsychological evaluations of HIV-infected individuals at differing disease stages. In this analysis, we report on the incidence of neuropsychological impairment in subjects who were not initially impaired but were at high risk of neurocognitive complications because of clinical evidence of immunosuppression (stage C, 1993 CDC classification [5]). Subjects were divided into two groups, on the basis of when HAART became widely available. Those who had a baseline and at least one follow-up visit before 31 December 1995 were included in a ‘pre-HAART’ group. Only data from visits before the cut-off date were considered. Subjects in the ‘post-HAART’ group had both a baseline evaluation and at least one other visit after 1 January 1996. For all subjects, inclusion was restricted to those completing a detailed battery of neuropsychological tests at baseline and follow-up visits. These tests encompassed eight domains: abstraction/flexibility, attention/speed of information processing, learning, memory, verbal, perceptual-motor, motor, and sensory functioning. Clinical neuropsychological impairment ratings were assigned for each domain by a senior neuropsychologist. A diagnosis of neuropsychological impairment was made if impairment was present in at least two domains. Details of the parent HNRC study and methods have been reported previously [6–8]. Kaplan–Meier survival analyses estimated the distributions of time to an initial incidence of cognitive impairment, and log rank tests compared distributions. Covariate effects were tested using Cox proportional hazards modelling. Data from 46 subjects with AIDS were analysed, 22 in the pre-HAART group and 24 in the post-HAART group. Follow-up time was similar in both groups [mean (SD) range in years]: 1.4 (1.1), 0.25–4.6 for the pre-HAART group versus 1.7 (0.8), 0.5–3.1 for the post-HAART group (P > 0.2). Mean (SD), median baseline CD4 cell counts were 53 (74), 28 versus 174 (230), 97/μl for the pre- and post-HAART groups, respectively (Wilcoxon rank sum test P < 0.05). Twice as many subjects (eight out of 22, 36%) in the pre-HAART group developed new neuropsychological impairment than in the post-HAART group (four out of 24, 17%). Neuropsychological impairment occurred sooner in the pre-HAART cohort (one-sided P < 0.04). After accounting for censoring by Kaplan–Meier analysis, cognitive impairment occurred at a median of 1.5 years for the pre HAART group. Because only an estimated 17% of the post-HAART group became impaired during follow-up of 3.1 years’ duration, the median time was not calculable but is estimated to occur beyond 3.1 years (Fig. 1).Fig. 1.: Kaplan–Meier survival estimate for distribution of time to neurocognitive impairment (in years) on or before 31 December 1995 (pre-highly active antiretroviral therapy) versus on or after 1 January 1996 (post-highly active antiretroviral therapy). Tick marks on curves represent censored observations. HAART, Highly active antiretroviral therapy.The mean (SD), median CD4 cell counts (/μl) at the follow-up visit (defined as the visit when first assessed as neuropsychologically impaired or the last recorded visit if unimpaired) were 20 (26), eight for the pre- and 196 (179), 150 for the post-HAART groups (Wilcoxon rank sum test P = 0.0001). Of note was the decline in CD4 cell counts from baseline for subjects followed before the widespread availability of HAART, whereas the post-HAART group had higher CD4 cell counts since baseline. A Cox model to test for independent effects of the CD4 cell count, at the follow-up visit, on estimating time to neurocognitive impairment resulted in shorter times to impairment for lower CD4 cell counts (P < 0.01). A significant relationship between these variables withstood adjustment for treatment era (P = 0.03), but identification of the HAART group did not significantly improve the original model developed with the CD4 cell count alone (marginal effect P = 0.8). Our results suggest that HAART helps to maintain intact cognitive functioning in high-risk patients with relatively unrestricted access to HAART. This protection is predominantly mediated by a HAART-induced improvement in immune function reflected by the CD4 cell count. We are currently expanding this study to examine whether protection is afforded by specific classes of antiretroviral drugs. We plan to investigate the role of HIV-RNA viral load, not available in a sufficient number of subjects for the reported study, and to extend our observations to individuals in the medically asymptomatic stages of disease. Reena Deutscha Ronald J. Ellisa J. Allen McCutchana Thomas D. Marcottea Scott Letendrea Igor Grantab and the HNRC Groupa