A Research Report submitted to the School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, in fulfillment of the requirements for the Degree of Masters in Epidemiology and Biostatistics. Johannesburg, February 2014 / Background: Overweight continues to be a major indicator of adulthood non-communicable diseases (NCDs) risk and can be linked to early childhood and adolescent dietary lifestyles. Research has shown that the burden of childhood and adolescent overweight in sub-Saharan Africa (SSA) and especially in South Africa is rising. There is therefore an urgent need to address this burden in the context of overweight risk related to nutrition transition. The current study used Birth to Twenty (BT20) cohort study data to examine the association between dietary intake and overweight status of 17-year old adolescents residing in Soweto, South Africa in 2007/2008. Methods: A secondary data analysis was conducted including 227 seventeen-year old adolescents (43.6% boys) from the BT20 cohort study. A modified quantitative Food Frequency Questionnaire (FFQ) was used to estimate dietary intake. Height and weight were measured by trained research assistants and overweight was categorized using the International Obesity Task Force (IOBTF) cut-offs. A multinomial logistic regression model was used to examine the association between overweight category (BMI) and dietary intake, socio-demographic factors and other covariates. The base reference used for Body Mass Index (BMI) category was normal weight. Results: The overall overweight prevalence was 14.5% (33/227), and prevalence of underweight was 23.3% (53/227) among the 17-year old Soweto adolescents. The prevalence of overweight was significantly higher in girls (19.3%) than boys (8.1%). The prevalence of underweight was 33.3% for boys and 15.6% for girls. The univariate analysis found the following macronutrient intake positively associated with boys’ BMI category: fat 189.4g (IQR=119.9-239.8, P<0.001), proteins 154.6g (IQR = 98.7-198.2, P<0.001) and carbohydrate 673.1g (IQR = 461-794.2, P<0.001). The following macronutrients: fat 176.4g (IQR = 99.5-220.1, P<0.001), proteins 126.1g (IQR = 70.4-145, P<0.001.) and carbohydrates 523.1g (IQR = 326.6-640.9, P<0.001) were positively associated with girls’ BMI category. The socio-demographic factors found associated with girls’ BMI were ethnicity (p=0.042) and maternal education (p=0.05). Also factors such as washing machine ownership (p=0.046), and parents having a car (p=0.048) were positively associated with boys BMI category. The multinomial logistic regression showed no significant differences in dietary intake when overweight boys were compared to normal weight boys: fat intake (RRR=0.99, 95% CI=0.95-1.01), carbohydrate (RRR=0.99, 95% CI=.96-1.01) and energy intake (RRR=1.01, 95% CI=0.99-1.01). Similar non significant results were observed for girls: fat intake (RRR=1.0, 95% CI =0.96-1.01), carbohydrate (RRR=1.0, 95% CI =0.99-1.02) and energy intake (RRR=1.0, 95% CI=0.99-1.01). Girls from the coloured community were more likely to be underweight than black African girls (RRR= 2.8, 95% CI=0.89-8.57). The results also showed that girls from mixed ancestral community (RRR= 0.25, 95% CI=0.05-1.20) were less likely to be overweight than black African girls. Discussion and Conclusion: The survey indicated high prevalence of both underweight and overweight among 17 year old adolescents residing in Soweto. This underscores the urgent need for both underweight and overweight prevention interventions and also highlights the need for an integrated surveillance system for both underweight and overweight among South African adolescents.

Obesity is common in adolescence and associated with medical complications. As maladaptive emotional processing is assumed to influence obesity, this research investigated emotional openness (EO), a general model of emotional processing, in normal-weight versus overweight adolescents, with consideration of gender differences, in order to determine whether a particular EO profile is characteristic of overweight adolescents. This research also explored the psychometric characteristics of the Dimensions of Openness to Emotions Questionnaire (DOE-20), a comprehensive multifactorial instrument that assesses emotional processing. The DOE-20 and the Positive and Negative Affect Schedule (PANAS) short form were completed by 160 adolescents (mean age: 14.36 ± 0.61 years), 39 of which were overweight. A multivariate analysis of variance showed no difference with respect to overweight status, but better ability to recognize and regulate emotions in boys than girls. The original five-factor structure of the DOE-20 was confirmed by confirmatory factor analysis; however, internal consistency was modest. As this research did not replicate the EO profile previously found in obese adults, we discuss the adequacy of the DOE-20 to evaluate EO in adolescents as well as the ability of the construct of EO to assess emotional processing in obesity. Further research should examine the links between EO, problematic eating behaviors, and obesity.