A Thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in ful lment of the requirements for the degree of Doctor of Philosophy in Medicine Johannesburg, South Africa 2017. / Background: The persistent burden of undernutrition, with increasing prevalence of obesity and metabolic disease risk among children and adolescents, has become a global public health problem. Research has shown that risk factors established in childhood and adolescence may contribute to the development of non-communicable diseases (NCDs) in adulthood. This is of particular concern in South Africa, given its rapid socio-economic, political and epidemiological transitions. Research into the trends of nutrition transition in rural children and adolescents, whose particular health needs have been under-served and poorly delineated in the past, provides a unique opportunity to study the e ects of rapid health transitions on development. Aim: To determine the association of nutrition, body composition and metabolic disease risk in rural South African children and adolescents. Study design: Three cross-sectional studies were undertaken to address the overall aim of this research. The speci c objectives of each study were: (1) to determine the association of nutrition on body composition and metabolic disease risk in children and adolescents; (2) to examine the associations between body mass index (BMI), disordered eating attitude and body dissatisfaction in female adolescents, and descriptive attributes assigned to silhouettes of di ering body habitus in male and female adolescents; and (3) to investigate associations between diet and cardiovascular disease (CVD) risk factors in adolescents. Method: One cross-sectional study, 3 analyses were nested within the Agincourt Health and Socio-demographic Surveillance System (HDSS) site, in the Bushbuckridge subdistrict, Mpumalanga Province, South Africa. In 2009, a random sample of 600 children and adolescents, from age groups 7 to 8 years, 11 to 12 years and 14 to 15 years, were selected from 3489 children who had participated in a 2007 growth survey. These children and adolescents had to have lived in Agincourt at least 80% of the time since birth or since 1992, when enrolment into the Agincourt Health and Socio-Demographic Surveillance System (HDSS) began. Height and weight were measured to determine BMI. Age and sex-speci c cut-o s for underweight and overweight/obesity were determined using those of the International Obesity Task Force. Body image satisfaction using Feel-Ideal Discrepancy (FID) scores, Eating Attitudes Test-26 (EAT-26), perceptual female silhouettes and pubertal assessment were collected through self-administered questionnaires. Blood pressure (systolic (SBP) and diastolic (DBP)) was measured, fasting blood samv ples were collected for the determination of glucose and lipids. Waist to hip ratio cut-o s of (WHR) >0.85 for females, >0.90 for males, waist to height ratio (WHtR) of >0.5 for both sexes, and waist circumference (WC) of >80 cm for females and >94 cm for males were used to determine the risk of adiposity. For abnormal lipids: high density lipoprotein cholesterol (HDL-C) cut-o s of >1.03 mmol/l, low density lipoprotein-cholesterol (LDL-C) of >2.59 mmol/l, triglycerides (TGs) of >1.7 mmol/l and total cholesterol (TC) of >5.17 mmol/l were used. Pre-hypertension prevalence was computed using the average of 2 readings of SBP or DBP, being >90th but <95th percentile for age, sex and height. Dietary intake was assessed using semi-quantitative food frequency questionnaire. T-test and ANOVAs for normally distributed data and Wilcoxon-Mann- Whitney test was used to determine signi cant di erences by sex and by pubertal stages for EAT-26 and EAT-26 sub-scores. Chi square tests were done to determine signi cant associations between the categorical variables. Bivariate linear regression was employed to test associations and signi cant tests were set at the p<0.05 level. Results: Study component (1): Stunting levels were higher in the boys than in the girls in mid to late childhood and combined overweight and obesity prevalence was higher in girls than in boys. The girls' BMI was signi cantly greater at ages 11 and 12 years than that of the boys [girls: 18 3.4, 95% con dence interval (CI): 17.33- 18.69; boys: 17 2, 95% CI: 16.46-17.25; p-value 0.004] and at ages 14-15 years (girls: 22 4.1, 95% CI: 20.82-22.47; boys: 19 2.4, 95% CI: 18.39-19.38; p-value < 0.001). Prehypertension (de ned as < 90th centile for age, sex and height) was higher in girls (15%) than boys (10%). Further, impaired fasting glucose was detected in 5.3% of girls and 5% of boys. High-density lipoprotein cholesterol (>1.03 mmol/l) concentrations were observed in 12% of the girls and 0.7% of the boys, which is indicative of cardiometabolic risk. Study component (2): The prevalence of overweight and obesity was higher in girls than boys in early and mid to post pubertal stages. The majority (83.5%) of the girls reported body image dissatisfaction (a desire to be thinner or fatter). The girls who wanted to be fatter had a signi cantly higher BMI than the girls who wanted to be thinner (p=0.001). There were no di erences in EAT-26 score between pubertal groups, or between boys and girls within the two pubertal groups. The majority of the boys and the girls in both pubertal groups perceived the underweight silhouettes to be \unhappy" and \weak" and the majority of girls in both pubertal groups perceived the normal silhouettes to be the \best". Study component (3): Added sugar and sweets contributed 10% and maize meal and vi bread contributed 7.2% to the total number of food items consumed respectively. Girls had higher intakes of total fat, saturated fat and cholesterol after adjusting for dietary energy intake and age (all p<0.001). The prevalence of combined overweight and obesity was 13.8% in girls and 3.1% in boys (p<0.001). In addition, indicators of adiposity were higher in females, abnormal waist circumference (WC) (6.7%), waist to hip ratio (WHR) (22.0%) and waist to height ratio (WHtR) (18.0%), compared to males, (0%), (3.1%) and (6.2%) respectively (all p<0.001). Girls had higher low-density lipoprotein (LDL) (12(9.3%) vs. 3(2.3%), p=0.01), total cholesterol (17(12.7%) vs. 5(3.5%), p<0.001) and were more pre-hypertensive (28(15.3%) vs.15(8.4%), p=0.04) than the boys. Furthermore, the bivariate associations between dietary intakes (total energy, total carbohydrate (CHO), total dietary fat and saturated fat) and anthropometric indices (BMI and WC) showed that body mass index (BMI) was associated with total energy (p=0.05) and BMI and WC were associated with total fat (p=0.01, p=0.03) and saturated fat (p<0.001, p=0.02) in females respectively. Conclusions: In conclusion, this thesis highlights that girls in rural South Africa had a higher prevalence of combined overweight and obesity than did boys, stunting was more prevalent amongst boys than girls in mid to late childhood and metabolic risk factors that were associated with adiposity, and linked to diet, were higher in girls than in boys. This study has provided useful information for targeting critical health promotion intervention programmes to optimise child nutrition as part of a noncommunicable disease preventative strategy, especially, in remote areas in rapidly transitioning South Africa. / LG2018