Physical activity, change in blood pressure and predictors of mortality in older South Africans - a 2-year follow-up study

Objective. A 5-year follow-up study of a cohort of 200 historically disadvantaged older South Africans was conducted to: (i) characterise current levels of habitual physical activity; (ii) relate physical activity to current risk factors for chronic disease; and (iii) identify risk factors associated with 2-year mortality. The baseline sample, drawn in 1993, was found to have a high prevalence of hypertension (71.7%). Research design. Retrospective cohort study. Methods. A baseline sample of 200 persons aged greater than or equal to 65 years, resident in the Cape Peninsula, was randomly drawn by means of a two-stage cluster design. Baseline measurements included: anthropometry, waist/hip ratio, systolic and diastolic blood pressure, body mass index (BMI), serum albumin, serum ferritin, haemoglobin and fasting plasma glucose levels, plasma lipid profiles, oral glucose tolerance test and self-reported health status. Subjects were revisited after 2 years, at which time an adapted version of the Yale Physical Activity Survey was administered and measurements of blood pressure and anthropometry were repeated. Statistical analyses. Spearman's rank-order correlations were used to describe relationships between various current risk factors and physical activity. Logistic regression was used to determine predictors of 2-year mortality from baseline data. Results. At follow-up, 142 of the subjects (66 men, 76 women) were traced and measurements collected. Thirty-two subjects were reported to have died by relatives living in the same household (22 men, 10 women). Levels of reported physical activity in the survivors were two-thirds lower than those reported in a sample of North Americans of similar age. There was an inverse association between age and physical activity (r = -0.31; P < 0.0005) and a positive association between BMI and physical activity (r = 0.29; P < 0.005). There was, however, no association between physical activity and systolic or diastolic blood pressure. In men, BMI in the lower tertile (P = 0.07) and serum ferritin levels were positively associated with increased mortality. Serum albumin levels were protective over the 2-year follow-up period (OR = 0.85; P < 0.05). In women, being diabetic (OR = 4.88; P = 0.06) and having a waist/hip ratio in the upper tertile (OR = 3.26; P = 0.06) were associated with mortality. Conclusions. Physical activity levels in this sample of older historically disadvantaged South Africans were habitually low. Simple anthropometric assessments incorporating weight and waist/hip ratio, together with serum albumin measurements, may be useful to screen general health risk for older adults at primary care level and provide indications for social or medical intervention. Further, strategies for earlier detection and effective management of diabetes, particularly in older women, may reduce premature mortality in this population.