Search

In tandem with the ever-increasing aging population in low and middle-income countries, the burden of dementia is rising on the African continent. Dementia prevalence varies from 2.3% to 20.0% and incidence rates are 13.3 per 1000 person-years with increasing mortality in parts of rapidly transforming Africa. Differences in nutrition, cardiovascular factors, comorbidities, infections, mortality, and detection likely contribute to lower incidence. Alzheimer's disease, vascular dementia, and human immunodeficiency virus/acquired immunodeficiency syndrome-associated neurocognitive disorders are the most common dementia subtypes. Comprehensive longitudinal studies with robust methodology and regional coverage would provide more reliable information. The apolipoprotein E (APOE) ε4 allele is most studied but has shown differential effects within African ancestry compared to Caucasian. More candidate gene and genome-wide association studies are needed to relate to dementia phenotypes. Validated culture-sensitive cognitive tools not influenced by education and language differences are critically needed for implementation across multidisciplinary groupings such as the proposed African Dementia Consortium.

Background and Aims: South Africa is a middle-income country with high levels of income inequality and a rapidly aging population and increasing dementia prevalence. Little is known about which risk factors for dementia are important and how they differ by social determinants of health as well as key demographic characteristics such as sex and wealth. We sought to calculate the population attributable risks (PARs) for established potentially modifiable risk factors for dementia among these different groups. Methods: We obtained risk factor prevalence from population-based surveys for established dementia risk factors (diabetes, midlife hypertension, midlife obesity, physical inactivity, depression, smoking, low educational attainment, social isolation). We used relative risk estimates reported in previous meta-analyses and estimated PARs using Levin's formula and accounting for communality. We tested for one-way and two-way interactions by sex and wealth using Pearson's χ2. In stratified analyses, we performed tests for trend using logistic regression. Results: The prevalence of established risk factors for dementia ranged from 5% for depression to 64% for low education. After accounting for communality, the risk factors contributing the greatest PAR were low education (weighted PAR 12%, 95% CI 7% to 18%), physical inactivity (9, 5-14%), and midlife hypertension (6, 5-14%). Together, 45% of dementia cases may be attributable to modifiable risk factors (95% CI 25-59%). We found significant interactions (p < 0.005) between sex, wealth, or both (sex * wealth) and each risk factor except social isolation and physical activity. Low education was inversely associated with wealth in both male and female. The PAR for midlife hypertension, obesity, and diabetes was associated with increasing wealth, and was higher in female. In contrast, the PAR for smoking was higher in male (8% vs. 2%) and was associated with increasing wealth among female only. We found that either a strategy of large reductions in selected risk factors with the highest PAR (midlife hypertension, smoking, physical inactivity) or small reductions across all risk factors could potentially reduce dementia cases by as many as 250,000 by 2050. Discussions: The potential impact on dementia risk by decreasing exposure to established dementia risk factors is large and differs by sex and social determinants of health like wealth. Risk factor PAR should inform national and local health policy dementia initiatives in South Africa including which risk factors to target in the whole population and which to target in high-risk groups for maximum public health benefit.