|dc.description.abstract||Non-communicable diseases (NCD) are also known as chronic diseases of lifestyle and cause the greatest burden of disease globally, whether measured as morbidity or mortality. Although there is consensus that obesity is a risk factor for NCD, differences of opinion exist as to what anthropometric measure or index of adiposity is the most effective in identifying those individuals who are at greatest risk, due to ethnic differences. Hence, uncertainty remains in the black South African population regarding the validity of different anthropometric measures and indexes and cut-off values to predict health outcomes. Africa is currently experiencing one of the most rapid demographic and epidemiological transitions in world history. Black South Africans have moved away from rural areas to more urban areas, where they are exposed to more psychosocial stress, less physical activity and adopted a more westernised diet (high in fat and has less carbohydrates and fibre), which is associated with an increased risk for the development of NCD and the metabolic syndrome (MetS). MetS is composed of a cluster of metabolic disorders associated with increased risk of cardiovascular diseases (CVD). The risk factors to define MetS include various combinations of abdominal obesity, high blood pressure, high fasting plasma glucose, dyslipidemia (increased triglycerides, low HDL cholesterol or both) and insulin resistance. There are currently three common definitions of MetS: the World Health Organization (WHO) definition, the National Cholesterol Education Program, Adult Treatment Panel III (NCEP-ATP III) definition and the International Diabetes Federation (IDF) definition. These definitions are in general agreement on the essential components of the MetS but differ in their cut-off values and method of combining the individual components. A prominent feature of the IDF definition is that abdominal obesity is a prerequisite component of the MetS, with abdominal obesity defined according to ethnic specific values of waist circumference (WC). The IDF definition suggests that sub-Saharan Africans should use the same WC cut-off values as Caucasians, derived from European data, until more specific data are available. The currently accepted WC cut-off values, might underestimate risk for metabolic syndrome risk factors in
African populations, and a substantial proportion of those who would need health care advice
would be missed in health screenings.
To determine the most effective anthropometric measure to indicate the presence of NCD risk factors (measures of lipids, fasting glucose, C-reactive protein, blood pressure and obesity) in a black South African population;
To determine the ethnically appropriate WC cut-off values for abdominal obesity in black South African men and women, to predict increased risk of metabolic syndrome risk factors (raised triglyceride, reduced HDL-cholesterol, raised blood pressure and raised fasting blood glucose), or two or more of these risk factors;
To determine the prevalence of the metabolic syndrome (MetS) with urbanisation, using three definitions (NCEP-ATP III, IDF, IDF with local WC cut-off values);
To assess the association of metabolic risk factors with abdominal obesity and raised blood pressure in a sub-Sahara African population.
The study formed part of the baseline data of the South African leg of the Prospective Urban and Rural Epidemiological (PURE) study. This study had a cross-sectional design that included randomly selected participants older than 35 with no reported chronic diseases of lifestyle, Tuberculosis or HIV. A total of 2010 black Setswana speaking participants was included of which 746 were men and 1264 were women.
Data was collected by a specialised multidisciplinary team. The participants signed an informed consent form. Questionnaires were issued during individual interviews and were conducted by extensively trained fieldworkers in the language of the participant's choice. Blood pressure was measured in the sitting position on the right arm, after 5 minutes of rest, using the validated OMRON HEM-757 automatic digital blood pressure monitor. All anthropometric measurements were done using the guidelines of the International Society for the Advancement of Kinanthropometry (ISAK). Blood serum and plasma concentrations of fasting triglycerides, highdensity lipoprotein cholesterol and fasting glucose was auto-analysed with a Konelab ™ clinical analyser by making use of standardised enzymatic procedures. Low-density lipoprotein cholesterol was calculated.
Statistical analysis was performed using SPSS for Windows (version 16.0) performing nonparametric statistical analysis. Student T -tests, one-way ANOVA tests, ANCOVA tests and Pearsons Chi-square tests were used to compare means and percentages. Cross-tabulation was used to calculate odds ratios. Waist-circumference cut-off values were determined by ROC curve analysis. Regression models were also used to quantify the prediction of NCD risk factors
measures by all 4 anthropometric measures/indexes. Linear regression analysis with adjustments for covariates (age, urbanization, HIV, and gender) (Model 1) and (age, urbanisation, HIV. gender and BMI) (Model 2) were used to determine the association of the percentage variance, explained by risk factors, in addition to models containing covariates, with abdominal obesity (WC) and raised blood pressure. Fasting triglycerides were not normally distributed and therefore log transformed before any comparisons were made. The relationships were expressed as standardised beta (β) coefficients and the percent increase in the adjusted R2 change when abdominal obesity (WC) or raised blood pressure was added to the model in comparison to model 1 and model 2.
As assessed by the ability of the anthropometric indices, to 1) account for the variability in each risk factor and 2) correctly identify individuals with increased NCD risk factors, the predictive abilities of BMI, WC and WHtR were similar. WC was slightly better (0.01-0.08 higher R2 value, p < 0.05) in predicting concentrations of total cholesterol (TC), fasting glucose (FG), triglycerides (TG), LDL-cholesterol (LDL-C), systolic blood pressure (SBP) and diastolic blood pressure (OBP) in men. BMI was slightly better in identifying men with reduced HDL-cholesterol (HDL-C) (0.03 higher R2, p < 0.05), while WHtR was slightly better (0.01-0.11 higher R2 value, p < 0.05) in predicting concentrations of C-reactive protein (CRP), TG, SBP, OBP in women. WC was slightly better in identifying reduced concentrations of HDL-C (0.08 higher R2, p < 0.05) in women. On the basis of two or more metabolic risk factors, WC and BMI were equal in their predictive ability of NCD risk factors according to the receiver operating characteristic (ROC) curve analysis (AUC=0.65) in men while WHtR and WC were equal in their predictive ability (AUC=0.65) in women. For men and women combined, WHtR was the most frequent predictor for NCD risk factors (5 out of 8 risk factors) and WC (3 out of 8 risk factors) as determined by AUC. Based on the receiver operating characteristic (ROC) curve analysis the WC value for predicting metabolic risk factors in this black African population was about 80 cm for men and women. The AUC for men was 0.653 (0.611-0.695 CI) and for women it was 0.643 (0.613-0.674 CI). According to the locally determined WC criteria, the prevalence of abdominal obesity was
28.1 % in the men and 52.4% in the women. The prevalence of MetS varied according to the definition used. The IDF definition with sub-Saharan Africa proposed waist circumference (WC) cut-off values (≥ 94 cm men; ≥ 80 cm women) and the IDF definition with proposed local WC cut-off values (≥ 80 cm for men and women) indicated a higher prevalence of MetS compared to the NCEP-ATP III definition. The highest prevalence of the MetS was obtained with the IDF definition with local WC cut-off values in the men (rural: 16.4%; urban: 17.6%) and women (rural: 28.4%; urban 38.2%) and the lowest with the NCEP-ATP III definition in men (rural: 2.3%;
urban: 0.8%) and women (rural: 16.3%; urban: 22.0%). Raised blood pressure (≥ 130/85 mmHg) was the most prevalent metabolic risk factor in both the rural (60.3-60.6%) and urban (72.0-72.6%) men and women. In the men no significant difference in the prevalence of the individual metabolic risk factors between the urban and rural groups could be found. The prevalence of most of the individual metabolic risk factors of the urban women was significantly higher compared to the rural women. The additional percentage of variance of individual continuous metabolic risk measures explained by either abdominal obesity (WC) or raised blood pressure (≥ 130/85 mmHg) were determined after adjusting for age, urbanisation, HIV status and gender. The standardised beta (β) coefficients and R2 change were higher for abdominal obesity (WC) rather than for raised blood pressure, except in the case of SBP, DBP and HDL-cholesterol. The association with the metabolic risk factors (fasting glucose, triglycerides, total cholesterol and HDL-C) was slightly stronger for abdominal obesity (WC) rather than for raised blood pressure (the association of SBP and DBP will always be high with raised blood pressure). After adjusting for BMI as presented in Model 2, the associations of fasting glucose
with raised blood pressure was no longer significant (p = 0.18).
WC and WHtR are the strongest anthropometric indicators of non-communicable diseases in this black South African population.
The optimal WC cut-off values (≥ 80 cm for men and women) to predict increased risk of metabolic syndrome risk factors (raised triglyceride, reduced HDL-cholesterol, raised blood pressure and raised fasting blood glucose), or two or more of these risk factors in this black African population are lower for the men and similar for the women than the WC cut-off value proposed by the lDF (≥ 94 cm for men and ≥ 80 cm for women).
The prevalence of the metabolic syndrome was significantly higher in the urban groups compared to the rural groups, across the metabolic syndrome definitions. Urbanisation significantly increased the prevalence of the metabolic risk factors and metabolic syndrome (IDF definition) in the women but not in the men.
Both, raised blood pressure and abdominal obesity have a significant association with metabolic risk factors in this Setswana speaking South African population. The association with the metabolic risk factors was slightly stronger for abdominal obesity (WC) rather than for raised blood pressure in this population.||