Changes in cognition during antiretroviral therapy: comparison of 2 different ranking systems to measure antiretroviral drug efficacy on HIV-associated neurocognitive disorders

Abstract

Objective: Although HIV-associated neurocognitive disorders should be treated with highly active antiretroviral treatment (HAART) regimens with good central nervous system (CNS) penetration, the definition of neuroactive HAART remains controversial. We compared 2 ranking systems to measure HAART neuroeffectiveness.

Methods: Patients with (n = 93) or at risk for (n = 92) HIV-associated neurocognitive disorders underwent neuropsychological (NP) test batteries before HAART initiation and at follow-up. Changes in normatively adjusted summary NP test z scores were calculated for each subject. Two neuropenetration scores were calculated: the central nervous system penetration reference score (number of drugs in the combination among zidovudine, abacavir, stavudine, lamivudine, efavirenz, nevirapine, indinavir, and lopinavir-ritonavir) and the CNS penetration-effectiveness (CPE) score: a summary score of 1 (high: penetration: [corrected] zidovudine, abacavir, delavirdine, [corrected] nevirapine, amprenavir-ritonavir, fosamprenavir-ritonavir, [corrected] indinavir-ritonavir, and lopinavir-ritonavir), 0.5 (intermediate penetration: [corrected] stavudine, lamivudine, emtricitabine, efavirenz, amprenavir, fosamprenavir, [corrected] atazanavir-ritonavir, atazanavir, [corrected] and indinavir), and 0 (low penetration: remaining ARVs) [corrected] for each drug in the combination. Main outcome measures were changes in global NPZ scores and in summary z scores on 5 domains.

Results: At regression analyses, higher CPE scores correlated with greater improvements in NPZ-4 (P = 0.0283), NPZ-8 (P = 0.0071), concentration and speed of mental processing (P = 0.0046), and mental flexibility (P = 0.0262) summary z scores. The correlation was stronger among NP-impaired patients. By contrast, higher estimates of neuroeffectiveness with the alternative system showed no correlation. No association was seen between CD4 and plasma viral load changes with both scores.

Conclusions: The CPE score represents a step forward toward the identification of a clinically useful approach to estimating HAART ability to improve cognition.