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.