Objective: To investigate the value of antemortem cognitive functioning in predicting postmortem evidence of HIV encephalitis (HIVE).

Methods: Thirty-nine subjects were assessed during life with a comprehensive neuropsychological battery and went on to autopsy within 18 months of testing. Cognitive impairment was determined by blind clinical ratings, based on demographically corrected test scores. Presence of HIVE was based on postmortem immunocytochemical detection of the viral protein gp41 or by measurement of HIV RNA by PCR in multiple brain areas as well as by histopathologic evidence such as microgliosis, presence of multinucleated giant cells, and myelin pallor in several brain regions.

Results: The sensitivity and specificity of neurocognitive impairment in detecting the occurrence of HIVE were 67 and 92%. Eighteen of 19 subjects with antemortem neurocognitive impairment had evidence of HIV-related brain disease (positive predictive value = 95%).

Conclusion: Neuropsychological assessment can help select HIV-positive patients for treatment of CNS disease.