Burden of stroke attributable to selected lifestyle risk factors in rural South Africa

Abstract

Background: Rural South Africa (SA) is undergoing a rapid health transition characterized by increases in non-communicable diseases; stroke in particular. Knowledge of the relative contribution of modifiable risk factors on disease occurrence is needed for public health prevention efforts and community-oriented health promotion. Our aim was to estimate the burden of stroke in rural SA that is attributable to high blood pressure, excess weight and high blood glucose using World Health Organization's comparative risk assessment (CRA) framework.

Methods: We estimated current exposure distributions of the risk factors in rural SA using 2010 data from the Agincourt health and demographic surveillance system (HDSS). Relative risks of stroke per unit of exposure were obtained from the Global Burden of Disease Study 2010. We used data from the Agincourt HDSS to estimate age-, sex-, and stroke specific deaths and disability adjusted life years (DALYs). We estimated the proportion of the years of life lost (YLL) and DALY loss attributable to the risk factors and incorporate uncertainty intervals into these estimates.

Results: Overall, 38 % of the documented stroke burden was due to high blood pressure (12 % males; 26 % females). This translated to 520 YLL per year (95 % CI: 325-678) and 540 DALYs (CI: 343-717). Excess Body Mass Index (BMI) was calculated as responsible for 20 % of the stroke burden (3.5 % males; 16 % females). This translated to 260 YLLs (CI: 199-330) and 277 DALYs (CI: 211-350). Burden was disproportionately higher in young females when BMI was assessed.

Conclusions: High blood pressure and excess weight, which both have effective interventions, are responsible for a significant proportion of the stroke burden in rural SA; the burden varies across age and sex sub-groups. The most effective way forward to reduce the stroke burden requires both population wide policies that have an impact across the age spectra and targeted (health promotion/disease prevention) interventions on women and young people.

Fig. 1

CRA methodology – comparison of population of interest with actual (factual) exposure distribution versus corresponding theoretical optimum distribution (the counterfactual) (a) and impacts on relative risk of disease occurrence (b)

Fig. 2

Graphical representation of comparative risk assessment methodology showing the actual distributions of systolic blood pressure (fig. 2 a) and body-mass index (fig. 2 b) in Agincourt sub-district, South Africa, 2010 compared with the targeted “counterfactual” distribution

Fig. 3

Distribution of PAFs for stroke due to SBP and BMI in adult males, Agincourt, South Africa, 2010

Fig. 4

Distribution of PAFs for stroke due to SBP and BMI in adult females, Agincourt, South Africa, 2010