Challenges in identification of HIV‐Associated Neurocognitive Disorders, a focus on low and middle income countries

Abstract

AbstractBackgroundHIV‐associated neurocognitive disorders (HAND) are highly prevalent amongst people living with HIV (PLWH), especially in Sub‐Saharan Africa (SSA) where the majority of PLWH live. Current HAND identification measures lack accuracy in older PLWH on treatment, with substantial heterogeneity observed in SSA prevalence meta‐analyses There is a need to better understand the current cognitive phenotype of this newly emergent population, to develop appropriate screening tools. We report two linked studies.Study 1 compared diagnostic accuracy of the International HIV dementia scale (IHDS), and Identification and Interventions for Dementia in Elderly Africans (IDEA) screen, for identification of HAND in older PLWH in TanzaniaStudy 2 aimed to identify cognitive domains significantly associated with symptomatic HAND (s‐HAND) in older PLWH in Tanzania, with potential utility in future screening measures.MethodA systematic sample of 253 PLWH aged ≥ 50 attending a Government clinic underwent comprehensive HAND evaluation by consensus American Academy of Neurology (AAN) criteria including locally‐normed low‐literacy clinical neuropsychological assessment. A diagnostic accuracy study of the IHDS and IDEA screen versus AAN criteria was completed with strict blinding. Participants were offered identical annual reassessment at Y1 Y2 and Y3 and the relationship of cognitive domains with s‐HAND in a logistic mixed effects modelResultHAND was highly prevalent (47%) at baseline, increasing to 57.9% at year 3.At baseline, both IHDS and IDEA had limited diagnostic accuracy for HAND (AUROC 0.639–0.667 IHDS, 0.647–0.713 IDEA). On secondary analysis of longitudinal data, tests of verbal learning and memory, categorical verbal fluency, visual memory and visuoconstruction had AUROC > 0.7 for s‐HAND (0.70 ‐ 0.72; p<0.001 for all). Tests of visual memory, verbal learning and memory (delayed recall and recognition memory), psychomotor speed, language comprehension and categorical verbal fluency were independently associated with s‐HAND (p<0.01 for all). Neuropsychological impairment pattern varied by education.ConclusionCurrent HAND screening tests no longer perform well in the new population of older PLWH in Africa. Tests most associated with s‐HAND include broad cortical and subcortical domains, some not included in existing HAND screening tests and vary by educational background. This should be considered in future HAND screening measure development.