Background: An important aim of AACTG studies A5001 (ALLRT) and A362 is to assess the impact of highly active combination antiretroviral therapies (HAART) on neurologic disease in HIV. To address this objective, the NeuroScreen, comprising brief assessments of neuropsychological (NP) function and distal sensory polyneuropathy (DSPN), is being administered to subjects in these studies. NARC007 evaluated the NeuroScreen’s performance characteristics (e.g., sensitivity, specificity) by comparison to a comprehensive diagnostic evaluation.

Methods:  A subset of subjects (N=301) enrolled in A5001/A362 at 15 AACTG sites received both the NeuroScreen and a comprehensive neurodiagnostic evaluation. The cognitive NeuroScreen consisted of the Trailmaking test (parts A and B), and the Digit Symbol task, which together assess speed of information processing and mental flexibility. Cognitive impairment points (CIP) were assigned by comparing subjects' raw scores to age- and education-adjusted norms from an HIV-negative reference sample. Raw scores (RS) were converted to CIP as follows: RS 1-2 standard deviations (SD) below norms =1 CIP; RS > 2 SD below norms = 2 CIP; otherwise = 0 CIP.  Cases of DSPN at screening had both distal vibratory sensation loss and diminished ankle reflexes.  For the comprehensive neurodiagnostic evaluation, a single clinical rater (KR) classified subjects as NP impaired or unimpaired based on a review of normatively adjusted test scores and demographic information. Neurologists made DSPN diagnoses.

Results:  For the cognitive NeuroScreen, a conventional cut-off of 2 CIP yielded high specificity (98%), but low sensitivity (24%).  By comparison, a cut-off of 1 CIP retained good specificity (75%), but showed better sensitivity (54%) (overall correct classification rate 70%). NeuroScreen DSPN diagnoses yielded comparably high specificity (91%) and sensitivity (46%).

Conclusions:  The AACTG NeuroScreen is a brief, simple and low-cost global screening tool that provides acceptable diagnostic efficiency in HIV neurological disease.  Serial NeuroScreen assessments in A5001 and A362 will permit estimation of the prevalence and course of neurocognitive impairment and DSPN in over 3000 individuals.  Because detailed information on disease status and ART is also available for these subjects, the NeuroScreen will yield important information about the impact of HAART on the burden of neurological disease in HIV.

Keywords: Cognitive; neuropathy; HIV