Objective: To review the available data on risk factors for cardiovascular disease (CVD), the influence of urbanisation of Africans on these risk factors, and to examine why stroke emerges as a higher risk than ischaemic heart disease (IHD) in the health transition of black South Africans. Design: A review of published data on mortality from and risk factors of CVD in South Africans. Setting: South Africa. Subjects: South African population groups and communities. Methods: The available data on the contribution of stroke and IHD to CVD mortality in South Africa are briefly reviewed, followed by a comparison of published data on the prevalence and/or levels of CVD risk factors in the different South African population groups. The impact of urbanisation of black South Africans on these risk factors is assessed by comparing rural and urban Africans who participated in the Transition and Health during Urbanisation of South Africans (THUSA) study. Results and conclusions:The mortality rates from CVD confirmed that stroke is a major public health problem amongst black South Africans, possibly because of an increase in hypertension, obesity, smoking habit and hyperfibrinogenaemia during various stages of urbanisation. The available data further suggest that black South Africans may be protected against IHD because of favourable serum lipid profiles (low cholesterol and high ratios of high-density lipoprotein cholesterol) and low homocysteine values. However, increases in total fat and animal protein intake of affluent black South Africans, who can afford Western diets, are associated with increases in body mass indices of men and women and in total serum cholesterol. These exposures may increase IHD risk in the future.

Objective: To determine the relative validity of the culture-sensitive quantitative food frequency questionnaire (QFFQ) developed for the Transition, Health and Urbanisation in South Africa (THUSA) study by 7-day weighed food records, urinary nitrogen excretion and basal metabolic rate (BMR). Design: A cross-sectional study. Setting: A community-based study in a population strati®ed according to level of urbanization. Subjects: Residents of the North West Province, South Africa, aged between 15 and 65 years. The weighed food record study comprised 74 participants while 104 participants collected 24-hour urine samples. Methods: All participants were interviewed using the QFFQ. For the weighed food record study, participants kept detailed weighed food diaries for seven consecutive days. For the urinary nitrogen study, participants made one 24-hour urine collection. Completeness of the urine collections was checked against 240 mg para-aminobenzoic acid. BMR was estimated by the Scho®eld equations. Results: Spearman rank correlation coef®cients between the QFFQ and weighed food record ranged between 0.14 (®bre) and 0.59 (vitamin C). The QFFQ tended to underestimate intakes compared with the weighed records. Quintile distributions were similar for both methods. The correlation between urinary nitrogen excretion and dietary intake was poor. Possible underreporting was identi®ed for 43% of the participants with the QFFQ and 28% with the weighed food record. Conclusions: The QFFQ appeared to be a relatively valid instrument for the assessment of dietary intakes of the population of the North West Province. The use of biomarkers in this population was dif®cult and needs further investigation.

Objective: To develop and assess the reproducibility of a quantitative food frequency questionnaire (QFFQ) sensitive to the culture of the African population of the North West Province, South Africa. Design: A cross-sectional study. Setting: A community-based ®eld study in a population strati®ed according to level of urbanization from deep rural to urban upper class. Subjects: A total of 144 (99 women and 45 men) residents of the North West Province, aged between 15 and 65 years, participated in the study. Methods: A culture-sensitive, 145-item interviewer-administered QFFQ was designed to cover the whole diet. Portion sizes were estimated from a food portion photograph book (FPPB) showing foods in three portion sizes. The QFFQ was administered twice, 6±12 weeks apart. Results: Spearman rank correlation coef®cients between the two administrations varied from 0.14 for calcium to 0.75 for alcohol. The mean percentage difference between intakes was 8.5 (standard deviation = 9.9). Energy, protein, carbohydrate and calcium gave differences within 10%. Few signi®cant differences among correlation coef®cients or percentage difference for gender, age group or strata of urbanization were present. Bland±Altman plots showed signi®cant proportional bias for protein, ®bre and vitamin C. More than 70% of the participants were classi®ed into adjacent quintiles for all nutrients. For food groups, correlation coef®cients ranged from 0.25 for milk to 0.45 for vegetable and maize meal groups and 80% of participants were classi®ed into adjacent quintiles. Conclusions: The QFFQ appeared to be a reproducible dietary intake assessment instrument.

Thirty-one men (47 (SD 14) years) and thirty-four women (35 (SD 13) years) took part in a 4-week randomized cross-over trial to compare the effect of six mugs of black tea daily v. placebo (water, caffeine, milk and sugar) on blood lipids, bowel habit and blood pressure, measured during a run-in period and at the end of weeks 2,3 and 4 of the test periods. Compliance was established by adding a known amount of p-aminobenzoic acid (PABA) to selected tea bags, and then measuring its excretion in urine. Mean serum cholesterol values during run-in, placebo and on tea drinking were 5-67 (SD 1-05), 5-76 (SD 1-11) and 5-69 (SD 1-09)mmoYI (P=0-16). There were also no significant changes in diet, LDL-cholesterol, HDL-cholesterol, triacylglycerols, and blood pressure in the tea intervention period compared with placebo. Compared with placebo, stool consistency was softened with tea (P = 0-04), and no other differences were found in bowel habit. Results were unchanged when fifteen 'non-compliers', whose PABA excretion indicated that fewer than six tea bags had been used, were excluded from the analysis, and when differences between run-in and tea periods were considered separately for those who were given tea first or second.

The objective of the present study was to assess the relationship between anthropometric measurements and risk factors for non-communicable diseases (NCD) in South African black women. A cross-sectional sample of 1040 apparently healthy black female volunteers, 15 -70 years old, was recruited from thirty-seven randomly selected sites in the North West Province, stratified according to level of urbanisation. We analysed the association between BMI, waist:hip (WHR), waist circumference (WC) and skinfold measurements and the following risk factors for NCD: blood pressure, serum lipids, fasting serum glucose and insulin and plasma fibrinogen, by using age-adjusted correlation analyses and stepwise regression analysis. Of the subjects, 28 : 6 % were obese (BMI.30). After adjustment for age and smoking status, BMI correlated significantly with diastolic blood pressure (r 0 : 21, P¼0 : 037), serum triacylglycerols (TG) (r 0 : 30, P¼0 : 003), fasting glucose (r 0 : 29, P¼0 : 005) and log fasting insulin (r 0 : 24, P¼0 : 02). There was a significant negative correlation between BMI and HDL-cholesterol (r 20 : 38, P,0 : 001). Similar but stronger correlations were found between both WC and WHR and these risk factors. Together with age, WC was a significant predictor of TG, HDL-cholesterol and fasting glucose in regression analysis, while subscapular skinfold was a significant predictor of diastolic blood pressure and fasting glucose concentration. Triceps skinfold was a significant predictor of total serum cholesterol, LDL-cholesterol, plasma fibrinogen and the insulin sensitivity index. Measures of obesity, particularly WC, are associated with the risk for NCD in black South African women, in which a high rate of obesity has been found. Obesity: Black women: Non-communicable disease risk factorsThere is controversy in the literature, and a lack of evidence regarding the detrimental effects of obesity on risk factors of non-communicable disease (NCD) in African women. 'Healthy obesity' was first described by Walker et al. (1989) in rural black women, when one or more of the adverse sequelae of obesity could be found in only 12 : 5 % of the obese women studied and 87 : 5 % were free of any sign of the metabolic syndrome.