Objective: To determine if mortality is increased in individuals with human immunodeficiency virus type 1 (HIV-1)-associated neurocognitive disorders less severe than frank dementia.
Design: A prospective cohort study; median duration of follow-up was 2.4 years. Kaplan-Meier analysis and Cox proportional hazards models were used to compare survival times according to neurocognitive classification.
Setting: University-based research unit.
Participants: A volunteer sample of 414 individuals seropositive for HIV-1. Subjects were classified at their baseline evaluation as neuropsychologically (NP) normal or abnormal (impaired in > or = 2 NP test domains). A subgroup of NP abnormal subjects met operational criteria for HIV-associated minor cognitive motor disorder; the remaining subjects were designated NP impaired. Subjects with frank dementia were excluded.
Main outcome measure: Mortality.
Results: At the baseline evaluation, 256 (62%) of 414 subjects were designated normal; 109 (26%). NP impaired; and 49 (12%), minor cognitive motor disorder. One hundred six participants (26%) died during follow-up. Compared with the NP normal group, the unadjusted relative risk (RR) of death for all NP abnormal subjects (minor cognitive motor disorder and NP impaired) was significantly increased (RR, 1.7; 95% confidence interval [CI], 1.2-2.6; P < .005). After adjusting for concurrently measured predictors of survival (CD4 lymphocyte counts, Centers for Disease Control and Prevention HIV disease classification, hemoglobin concentration, and serum beta 2-microglobulin) in proportional hazards models, mortality for all NP abnormal subjects remained elevated (RR, 1.8; 95% CI, 1.2-2.8; P < .01). The elevation in mortality risk for subjects with minor cognitive motor disorder was statistically significant (RR, 2.2; 95% CI, 1.2-3.8; P < .01); for NP impaired subjects it was marginally significant (RR, 1.6; 95% CI, 1.0-2.8; P = .06).
Conclusions: The HIV-infected individuals with NP impairment had a higher risk of dying than those without impairment. This was particularly true for those meeting syndromic diagnostic criteria.