Incident cognitive impairment in a longitudinal cohort of older adults in rural South Africa, 2014‐19

Abstract

BackgroundWe aimed to determine the incidence of cognitive impairment and its key sociodemographic, social, and health‐related predictors at the first longitudinal follow‐up of the population‐representative “Health and Aging in Africa: A Longitudinal Study of an INDEPTH Community in South Africa” (HAALSI) cohort in South Africa.MethodData were from 3,861 adults aged ≥40 in rural Agincourt sub‐district, South Africa from 2014‐19, who were free from cognitive impairment at baseline. Cognitive impairment was defined as scoring ≥1.5 SD below the baseline mean composite time orientation and episodic memory score, or requiring a proxy interview with “fair” or “poor” proxy‐reported memory. Limitations to activities of daily living (ADLs) were compared according to incident cognitive impairment status. Incidence rates (IRs), incidence rate ratios (IRRs), and 95% confidence intervals (CIs) for cognitive impairment were estimated according to sociodemographic, social, and health‐related characteristics using modified Poisson regression. IRs and IRRs were weighted to account for mortality over the follow‐up. IRRs were adjusted for age, sex/gender, and country of birth.ResultOver a mean follow‐up of 3.7 years, 309/3,861 at‐risk participants newly developed cognitive impairment (IR=24.0 per 1000 person‐years (PY); 95% CI: 21.6‐26.8). Incidence increased steadily with age, from IR=9.1 per 1000 PY (95% CI: 5.5‐16.1) among those aged 40‐44 years at baseline to IR=76.5 per 1000 PY (95% CI: 63.2‐93.4) among those aged 80+. At least one ADL limitation was prevalent at follow‐up in 39% of those with an incident cognitive impairment, compared to 7% of non‐cognitively impaired participants. The incidence of cognitive impairment did not vary by sex/gender, HIV infection status, or cardiovascular risk factors, but was strongly graded according to education, literacy, household assets, employment, marital status, and frequency of alcohol consumption. For example, IRR=1.13 (95% CI: 0.91‐1.40) for female vs. male, IRR=1.05 (95% CI: 0.77‐1.43) for HIV‐positive vs. HIV‐negative, IRR=2.40 (95% CI: 1.81‐3.17) for illiterate vs. literate.ConclusionThis study presents one of the first incidence rate estimates for cognitive impairment in sub‐Saharan Africa, where populations are beginning to rapidly age. Social and socioeconomic disparities in incident cognitive impairment rates were apparent in a similar pattern as in many high‐income countries.