Introduction: Black African populations present with low insulin sensitivity (SI) and hyperinsulinemia, the latter due to high insulin secretion and reduced clearance. In addition, they exhibit lower levels of central and ectopic fat, compared to their white counterparts, contradicting the known correlates of SI typically reported in white populations. Moreover, in black African women hyperinsulinemia is in excess of the level needed to compensate for low SI, with a corresponding high disposition index (DI), a marker of insulin response that accounts for the level of SI. Further, obese, black African women have a high risk for type 2 diabetes (T2D), but the correlates of hyperinsulinemia have not been fully elucidated, especially the role of ectopic fat and body fat distribution. Exercise training is beneficial to improve SI and DI, however, whether these effects are mediated by changes in ectopic fat in skeletal muscle, liver and pancreatic depots is unknown. Accordingly, exercise training can be used as a model to assess the correlates of hyperinsulinemia and SI in cohorts at high risk for developing T2D, such as obese black African women. This thesis therefore aims to describe the correlates of hyperinsulinemia and SI and to evaluate the effect of exercise training on these components with emphasis on the role of body fat distribution and ectopic fat in mediating these changes. Methods: Firstly, a cross-sectional analysis of 45 obese (BMI 30-40 kg/m2 ) black South African women (age 20-35 years) without T2D was conducted. Thereafter the women were block randomized into an exercise training (n=23) or no exercise (control, n=22) group. The exercise training group participated in a 12-week combined aerobic and resistance training programme (40-60 min session, 4 days/week) supervised by a biokineticist. Pre and post-intervention testing included assessment of acute insulin response to glucose (AIRg), SI, DI (AIRg x SI), insulin secretion rate (ISR), hepatic insulin extraction (HIE) and peripheral insulin clearance (CLp) (frequently sampled intravenous glucose tolerance test); body fat mass and regional adiposity (dual-energy X-ray absorptiometry); hepatic, pancreatic and skeletal muscle fat and abdominal subcutaneous (aSAT) and visceral adipose tissue (VAT) (magnetic resonance imaging); intramyocellular (IMCL) and extramyocellular fat content (EMCL) (magnetic resonance spectroscopy). Results: The baseline results showed that a high DI was associated with low VAT (r0.565, p< 0.001), pancreatic fat, soleus IMCL and EMCL with VAT explaining most of the variance in DI (32%). SI was inversely associated with VAT (rho -0.417, p=0.007) and AIRg was inversely and HIE was positively associated with VAT-aSAT ratio (rho - 0.345, p=0.029 and rho 0.510, p=0.011, respectively). DI was positively associated with CLp (rho 0.528, p=0.006), while its components (SI and AIRg) were not. Results from the intervention showed that exercise training increased DI (median (interquartile range): 6.1 (3.6-7.1) to 6.5 (5.6-9.2) x103 arbitrary units, p=0.028), SI (2.0 (1.2-2.8) to 2.2 (1.5-3.7) (mU/l) -1 min -1 , p=0.005) and VO2peak (mean ± standard deviation: 24.9±2.42 to 27.6±3.39 ml/kg/min, p< 0.001), with no changes in control group. Exercise training decreased body weight (84.1±8.7 to 83.3±.9.7 kg, p=0.038) and gynoid fat mass (18.5±1.7 to 18.2±1.6%, p< 0.001). AIRg, ISR, HIE, CLp, aSAT, VAT and ectopic fat were unchanged after exercise training. However, the control group increased body weight and aSAT. The increase in SI and DI were not associated with changes in body composition, body fat distribution or ectopic fat. Conclusion: Novel results from our cross-sectional analysis showed that, in obese black South African women, DI was positively associated with peripheral insulin clearance, probably due to higher SI of peripheral tissue. Moreover, the most important correlate of a high DI was low VAT independent of ectopic fat accumulation in other sites. Further, we showed that low AIRg and high HIE correlated with a high VAT-aSAT ratio, while low SI was associated with high VAT. These associations require further exploration to determine direction of causality. Findings from our exercise intervention study extend on previous research by showing that moderate-to-high intensity combined aerobic and resistance exercise training increased SI and improved cardiovascular fitness, but insulin secretion, hepatic insulin clearance, ectopic and central fat depots did not change. Our results suggest that hyperinsulinemia may not occur solely as a compensatory mechanism for low SI and that ectopic and central fat might not be the primary correlates of insulin resistance in this cohort. Rather, intrinsic factors within muscle and adipose tissue may be putative mediators for observed improvements in the metabolic outcomes but will require further elucidation. Further research is required to confirm the causal role of VAT on low DI and to determine whether a long-term exercise training program and/or a low carbohydrate/glycemic index diet will reduce AIRg in those with hyperinsulinemia.

The high prevalence of running related injury, particularly in novice runners has prompted the investigation into interventions to mitigate the risk of injury. This dissertation set out to investigate the effects of a progressive 12-week running intervention in novice runners wearing footwear with reduced cushioning. The aim was to understand how intrinsic characteristics of novice runners, namely body composition, strength and flexibility, influence running biomechanics and ultimately injury outcome, and whether footwear structure effects this relationship. Prior to the intervention, participants were assigned to wearing footwear with reduced cushioning (RC) or footwear with traditional cushioning (TC). Three-dimensional running biomechanics were collected during over ground running at 3.0 meters per second in their prescribed footwear. Other measured variables included lower limb strength, by means of an isokinetic dynamometer, lower limb flexibility, full body composition by means of Duel Energy X-Ray Absorptiometry and lower leg bone oedema by means of magnetic resonance imaging. Throughout the intervention, pain or discomfort was assessed. All variables were reassessed after the intervention. The programme used in this thesis resulted in a 11.1 % prevalence of injury, which is considerably lower than other studies. No differences in injury incidence, bone oedema or pain or discomfort prevalence were found between footwear groups, however the RC group experienced pain or discomfort more frequently. Footwear with reduced cushioning was found to promote kinematic strategies, including a lower foot strike angle (FSA), more flexed knee angle at foot strike and reduced knee range of motion during stance phase to compensate for the lack of cushioning. Whilst most novice runners adopted a rear foot strike pattern throughout the intervention, the RC group were four times more likely to reduce FSA. Intrinsic characteristics of novice runners may not be indicative of injury, however the intervention resulted in changes to these variables. These included improvements in movement-specific strength, increased passive hip flexor flexibility and weight loss. Footwear had no effect on these variables. Greater mass characteristics resulted in kinematic adaptations in the knee. This dissertation highlighted the importance of a conservative training structure to mitigate injury risk in novice runners. Additionally, footwear has limited effect on injury risk and thus should not be prescribed to promote biomechanical change, but rather to compliment a runner's current biomechanics and intrinsic characteristics.

Background: Developing tackle and ruck technique is important to improve performance in matches and reduce the risk of injury. Little is known regarding valid tools to assess tackle and ruck technique in rugby union. The aims of this thesis were (1) to assess the validity and representativeness of the contact assessment tool, and (2) to identify factors which may affect the degree to which contact technique developed in training transfers to matches. Methods: Tackle, ball-carry and ruck technique of players competing at different levels of play were assessed in a two-on-two training drill using standardised technical criteria. Technique scores between levels of play were compared to assess the validity of the contact assessment tool, and contact technique scores assessed in training and matches were compared to assess the representativeness of the tool. Physical qualities and questionnaire data on the importance of technique to improve performance and reduce injuries were compared to contact technique scores assessed in training, to determine the effect of physical conditioning and player's knowledge on contact technique proficiency. Finally, tackle and ruck technique scores assessed in training and matches were compared to measures of match performance and contact related injuries. Results: Senior players scored significantly higher in the tackle, ball-carrier and ruck assessment than academy 1st and 2nd level players, demonstrating the good construct validity of the assessment tool. Contact technique scores were associated with performance outcomes in training and in matches, although technique scores in matches were lower than technique scores in training. There were no significant relationships between player's knowledge of the importance of contact technique and their contact technique proficiency. There were moderate to large associations between various physical qualities and tackle, ball-carry and ruck technique scores assessed in training. Players with better contact technique in matches performed better in matches, however, there were no meaningful correlations between contact technique in training and match performance or match related contact injuries. Conclusions: These findings demonstrate the validity of a tool to assess contact technique in rugby union with good representative learning design, however progressing the drill into less structured environments is recommended to further improve the representativeness of the assessment environment. Furthermore, the findings highlight the importance of contact skill training and physical conditioning to ensure skills developed in training are transferred to match performance.

Bibliography: leaves 93-106.

COL5A1 encodes the α1 chain of type V collagen, a minor fibrillar collagen that is an important regulator of collagen fibril assembly. A polymorphism (rs12722, C/T) within the 3'-untranslated region (UTR) of COL5A1 is associated with chronic Achilles tendinopathy (TEN) and other soft tissue injuries as well as exercise-related phenotypes. These phenotypes are directly or indirectly associated with the mechanical properties of musculoskeletal soft tissue. It has therefore been hypothesised that variants in the COL5A1 gene, specifically the 3'-UTR, regulate synthesis of the α1(V) chain and type V collagen production. Type V collagen levels in turn regulate fibril architecture and structure and, thereby, mechanical properties of musculoskeletal soft tissues. Although the 3'-UTR of many eukaryotic genes have been shown to play an important regulatory role, the function of the COL5A1 3'-UTR is currently unknown. Aim. The primary aim of this thesis was therefore to determine whether the COL5A1 3'-UTR was functional and to identify functional differences between the COL5A1 3'-UTR cloned from participants with TEN and healthy asymptomatic control individuals. The secondary aim was to start mapping the functional regions within the 3'-UTR, focusing on regions which are potentially responsible for contributing to the tendinopathic phenotype.

INTRODUCTION: Chronic diseases of lifestyle (CDL) are a major cause of global morbidity and mortality. Although CDL are largely preventable and treatable through adopting and maintaining healthy lifestyle behaviours, CDL rehabilitation programmes remain an underutilised resource. Behaviour modification is thus complex, and requires a collaborative approach between psychologists and medical clinicians involved in the management of CDL. This thesis examined the role of psychological factors in the management of patients with CDL who participated in a comprehensive lifestyle intervention (CLI) programme. METHODS: An explanatory mixed methods design was used to describe the CLI experience. These included an initial clinical audit of 308 patients commencing and completing a twelve week CLI programme to test associations of psychological, demographic, medical and diagnostic factors with physical outcomes. Two qualitative studies were subsequently conducted to further understand patient experiences of CDL and CLI programmes. The first involved interviews of 14 patients at programme commencement and completion. The second consisted of a case study of a patient participating in the programme using human centred design principles as well as ethnography.

Includes bibliographical references / The prevalence of childhood overweight and obesity has increased substantially globally with a concurrent decline in both children's physical activity and fitness levels. The socio-ecological model proposes that health behaviour, such as physical activity, is influenced by multiple factors, at an individual, social and environmental level. However, there seems to be a lack of consensus in the current literature on the factors influencing physical activity in different settings (e.g. school, neighbourhood), and across a wide range of socio-economic conditions. To our knowledge, there are no data available on the role of different environmental factors (within the school and the neighbourhood environment) in relation to children's physical activity in a South African setting. Therefore, one of the key aims of this thesis was to assess the effectiveness of a novel intervention on fitness, measures of physical activity and factors that influence physical activity. Specifically, this thesis assessed the effectiveness of a school-based, curriculum-grounded, educator-focused intervention designed to increase physical activity and healthy eating in South African primary school students in low income settings, on fitness levels and physical activity related knowledge, attitudes and behaviour (Chapter 2). Additionally, this thesis examined factors within the school environment associated with observed physical activity in children during in-school break time, self-report activity, and moderate- to -vigorous, in-school objectively-measured activity (Chapter 3-5). Furthermore, this thesis assessed the influences of parental perceptions and the neighbourhood environment on children's physical activity (Chapter 6). The overarching aim of this thesis is to examine the association between environmental constructs (physical spatial and built environment, social environment, and policy environment) and children's physical activity.

Background: Global levels of overweight and obesity in preschool-aged children have increased dramatically in the last two decades, with most overweight and obese children younger than five years living in low- and middle-income countries (LMICs). Statistics from the 2013 South African National Health and Nutrition Examination Survey (SANHANES-1) confirm that levels of overweight and obesity are high in South African preschool-aged children, with prevalence rates of overweight and obesity up to 18.2% and 4.7%, respectively. This increasing problem of overweight and obesity in South African preschoolaged children highlights the need for intervening in this age group. Overweight and obesity interventions in preschool children typically include one or more of the following behaviours: physical activity, sedentary behaviour and screen time. Aim and objectives: The aim of this study was to characterise the preschool environment in rural South Africa, and to explore physical activity, gross motor skill proficiency, sedentary behaviour and screen time in rural South African preschool-aged children. Additionally, aims of this study were to explore the associations between gross motor skills, body composition and physical activity; and to assess compliance with current physical activity and sedentary behaviour guidelines. Methods: Preschool-aged children (3-5 years old, n=131) were recruited from three Preschools and two Grade R (reception year) settings in Agincourt, a rural village in north eastern South Africa. In order to gain an understanding of the Preschool and Grade R settings, an observation of the preschool environments was conducted using a tool adapted from the Outdoor Play Environmental Categories scoring tool, Environmental and Policy Assessment and Observation instrument, and the Early Learning Environments for Physical Activity and Nutrition Environments Telephone Survey. Each child’s height and weight was measured. Physical activity and sedentary behaviour were measured objectively using a hip-worn ActiGraph GT3X+ accelerometer for 7 days (24 hours, only removed for water-based activities). Gross motor skills were assessed using the Test for Gross Motor Development–Version 2 (TGMD-2). Physical activity and sedentary behaviour, including the contextual information for these behaviours, during the preschool day (08h00 until ±12h00) were measured using the Observational System for Recording Physical Activity in Children (Preschool Version). A separate sample of parents/caregivers were recruited (n=143) to complete a questionnaire that was adapted from the Healthy Active Preschool Years questionnaire and Preschool Physical Activity Questionnaire. Parents reported on their child’s screen time, and on factors within the home and community contexts in which physical activity and sedentary behaviours occur. Results: In terms of the environment, the Preschools and Grade R settings differed in that fixed play equipment only featured in the Preschool settings. Grade R settings had more open space in which to play. All Preschool and Grade R settings provided children with limited portable play equipment, and none of the schools had access to screens. Although all children recruited for the study were preschool-aged, the Grade R children were significantly older than the Preschool children (5.6±0.3years vs. 4.4±0.4 years, p <0.05). According to IOTF cut-offs, the prevalence of overweight/obesity was low (5.0%) in the sample, and 68.1% of children were classified as normal weight. On average, children spent 477.2±77.3 minutes in light- to vigorous-intensity physical activity (LMVPA) per day, and 93.7±52.3 minutes in moderate- to vigorous-intensity physical activity (MVPA). In terms of the new current guidelines (180min/day LMVPA, including 60min of MVPA, described as ‘energetic play’), and using average daily average of LMVPA and MVPA, 78.2% met current guidelines. Observed and objectively measured sedentary behaviour results revealed that children were more sedentary during preschool time (between 08:00 to 12:00) compared to the afternoons. Overall, boys were significantly more physically active than girls; and Preschool children did more physical activity during preschool time than Grade R children (all p< 0.05). Over 90% of the sample achieved an ‘average’ or better ranking for gross motor skill proficiency. The Grade R children were significantly more proficient than the Preschool children for all gross motor skill components (raw scores and standardised scores). Overall, boys achieved significantly better object control raw scores than the girls, and displayed greater proficiency than the girls in the strike (p=0.003), stationary dribble (p< 0.001) and kick (p< 0.001). None of the preschool or Grade R settings had access to screens such as televisions or iPads, and parent-reported screen time was low for the total sample (0.5±0.3hr/day). The majority of the sample (97.9%) met current screen time guidelines (<1 hour per day). Parents (82.5%) reported that they believed that their child did sufficient PA for their health, but 81.8% also reported believing that television time would not affect their child’s health. Parent responses revealed neighbourhood safety as a potential barrier to being physically active in the community. Conclusions: Rural preschool-aged children in South Africa appear to be engaged in adequate amounts of physical activity, particularly LMVPA, and are adequately proficient in gross motor skills. The children did not engage in excessive amounts of screen time. Overweight and obesity were not prevalent in this sample of rural preschool-aged children, and therefore it would appear that an intervention to reduce or prevent obesity by increasing physical activity, improving gross motor skills and reducing screen time is unnecessary. Rather, interventions that facilitate the increase in levels of MVPA in order to meet current physical activity guidelines are warranted. Additionally, it is essential that the high levels of physical activity (LMVPA) and good foundation of gross motor skills observed in this sample are promoted in an effort to maintain them throughout childhood. Future research may want to determine whether these activities (high levels of LMVPA, low levels of screen time) track throughout childhood and into adolescence.

Includes abstract. / Includes bibliographical references. / The HIMS test, which consists of controlled exercise at increasing workloads, has been developed to monitor changes in training status and accumulative fatigue in athletes. As the workload can influence the day-to-day variation in heart rate, the exercise intensity which is associated with the highest sensitivity needs to be established with the goal of refining the interpretability of these heart rate measurements. The aim of the study was to determine the within subject day-to-day variation of submaximal and recovery heart rate in subjects who reached different exercise intensities.

Obesity treatment requires approaches that target the reduction of body weight and fat mass. The improvement of cardiorespiratory fitness (CRF), metabolic flexibility and insulin sensitivity also contribute towards reducing obesity-associated risk factors. While energy restriction alone results in significant weight loss, exercise-only interventions provide small amounts of weight loss and prevent weight gain, while also improving many of the other variables targeted in obesity treatment. Once achieved however, successful weight-loss maintenance is challenging, with many individuals subsequently experiencing weight regain. The main objectives of this thesis were to explore the role of metabolic rate and substrate utilization in influencing body weight, body composition and insulin sensitivity. This twopart thesis hypothesised that: 1) exercise training, without dietary intervention, will improve metabolic rate and substrate utilization in a sedentary obese population, and that this would be associated with improved body composition, insulin sensitivity and CRF; and 2) metabolic rate, substrate utilization and insulin sensitivity are altered through weight loss/regain, predisposing these individuals to weight regain and impairing successful weight-loss maintenance. In Part 1 of this thesis a 12 week exercise intervention in sedentary, obese (BMI 30-40kg.m-2 ) black South African (SA) women (aged 22, IQR 21-24 years) was completed. Previous studies have shown that black SA women present with very low CRF, a key indicator of increased risk for non-communicable disease (NCD), and have a high prevalence of obesity and insulin resistance (IR). They are thus at increased risk for developing type 2 diabetes (T2DM). Furthermore, physiologically black SA women have also been shown to have less visceral adipose tissue (VAT) and more peripheral gluteal fat mass (FM) compared to their white counterparts, but are paradoxically more IR. Despite this presentation, to date there has been no supervised exercise intervention studies undertaken in this very high risk population group. The first study of this thesis (chapter 2) aimed to assess the effects of the exercise intervention on changes in CRF, energy expenditure (EE) and substrate utilization, both at rest and during steady-state exercise compared to non-exercising controls. It also assessed baseline and changes in these measurements in relation to changes in body composition. Black SA women (BMI 30-40 kg.m-2 , 20-35 y) were recruited and randomized into control (CTL, n=15), or exercise (EXE, n=20) groups. The CTL was instructed to maintain usual activity while the EXE completed 12 weeks of combined resistance and aerobic exercise training (4d.wk-1 , 40-60min.d-1 @ >70% peak heart rate (HRpeak)). At pre-intervention, a treadmill-based CRF test, measuring peak volume of oxygen consumption (VO2peak), was carried out. Thereafter resting and steady state exercise (50% VO2peak) energy expenditure (EE) and respiratory exchange (RER) were measured along with body composition (dualenergy X-ray absorptiometry (DXA)). A frequently sampled intravenous glucose tolerance test (FSIGT) was also carried out to determine changes in insulin sensitivity. These tests were repeated at post-intervention testing with steady state testing being carried out both at the same relative intensity (50% post-testing VO2peak) and the same workload (treadmill speed and gradient) as used for pre-testing. Dietary intake (4d diary) and daily step-count (ActivPAL) data was collected at pre-testing, 4, 8 &amp; 12 weeks. Results showed that all participants had very low baseline CRF, falling below the 20th percentile previously shown in African American women. In response to exercise training, CRF increased by ≈11% and rates of fat oxidation during steady-state exercise were improved, while in controls these remained unchanged. Compared to CTL, EXE also showed small but significant reductions, in weight, as well as BMI, waist (WC) and hip (HC) circumferences. In contrast weight, BMI and WC increased in non-exercising controls. Gynoid FM (absolute FM and as a proportion of total FM), rather than visceral adipose tissue (VAT), was reduced in exercise participants. Within the exercise group higher baseline fat oxidation rates during steady state exercise and lower resting carbohydrate oxidation rates explained 61.6% (p< 0.001) of the variability in changes in gynoid FM in response to 12 weeks of exercise training in this group. In conclusion, exercise training improved CRF and fat oxidation rates during submaximal exercise in sedentary, obese black SA women. Higher fat oxidation rates during steady state exercise and lower resting carbohydrate oxidation rates at baseline were associated with the mobilization of gynoid FM in response to exercise training, rather than VAT as is typically shown in exercise interventions. This novel finding potentially represents an ethnic/gender specific response to exercise training. Further studies are needed to confirm this. Similar exercise training programs, that are sustainable over the long term, would therefore be beneficial in achieving meaningful increases in CRF while also supporting weight management and body composition improvements in this high risk population group. Using data from the exercise intervention in obese black SA women, the second study of this thesis (chapter 3) investigated inter-individual variability in the CRF response (∆VO2peak) to exercise training. The study specifically aimed to compare changes in EE and substrate utilization at rest and during steady state exercise, body weight and composition and insulin sensitivity between high and low CRF responders to the 12 week intervention. Furthermore it aimed to explore associations between baseline metabolic rate, EE and substrate utilization and subsequent changes in CRF in response to exercise training, to determine if baseline variability in these measures contributed to inter-individual variability in the CRF outcome. Within the exercise group, high inter-individual variability in CRF response to exercise training was identified. Based on a median split in ∆VO2peak, high responders (HRS, n=10) increased CRF by 21.7 ±10.0% (p< 0.001) compared to no change in both low responders (LRS, n=10; +0.6 ±6.3%, p=0.748) and CTL (-3.2 ±10.8%, p=0.195). This occurred despite all groups having similar baseline VO2peak and the exercise group receiving the same exercise dose (number of exercise sessions attended and average intensity of the exercise sessions). At baseline, HRS derived ≈62% of energy expenditure from fat oxidation during steady-state exercise compared to just 41% in LRS, who relied to a greater extent on carbohydrate oxidation. Furthermore, HRS were ≈11 kg lighter than LRS. There was also a positive association between BMI and RER such that individuals with higher BMI showed lower fat utilization (i.e., higher RER). HRS reduced gynoid FM whereas in LRS this remained unchanged. This is in line with the findings of the previous chapter which showed that exercise-related reduction in gynoid FM was associated with greater baseline fat oxidation. LRS showed improvements in insulin sensitivity compared to CTL and HRS. Using regression analysis including the exercising participants, greater baseline carbohydrate oxidation rates both at rest and during steady state exercise predicted a poorer CRF to exercise training, explaining 37.5% of the variability in ∆VO2peak. To the best of my knowledge, this is the first study to show that baseline variability in substrate utilization among sedentary obese individuals contributes towards explaining the variability in the CRF response to exercise training. However, further studies are required to confirm these results. Together, these studies show that higher fat oxidation rates are necessary for FM mobilization, while correspondingly reduced reliance on carbohydrate oxidation both at rest and during exercise supports improvements in CRF in response to exercise training. These findings add to a growing body of research aimed at explaining inter-individual variability in exercise intervention outcomes and may contribute to individualizing the exercise prescription. Part 2 of this thesis used a cross-sectional approach and investigated firstly whether there was evidence for metabolic adaptation to weight loss/regain in response to long term weight maintenance, potentially predisposing individuals to future weight gain/regain. Secondly, I investigated whether insulin sensitivity is altered as a result of prior weight loss history, or whether successful weight loss restores insulin sensitivity to levels that are comparable to phenotypically similar controls with no weight loss history. Weight stable, BMI-matched South African women aged 20-45 years with or without a history of prior weight loss were screened and recruited. Four groups were defined as follows: Weight Reduced (RED, n=15) - lost at least 15% of body weight &amp; maintained a reduced weight (BMI ≤ 27kg.m-2 ) for over 12 months (15% of body weight), but relapsed back to overweight or obese (BMI ≤ 27kg.m-2 ); and overweight or obese (BMI ≤ 27kg.m-2 ) stable-weight controls (OSW, n=11) - no history of significant weight loss. The first study in Part 2 (Chapter 4) compared metabolic rate and substrate utilization in RED and REL to their respective BMI-matched controls with no weight loss history, both at rest and in response to a high fat meal challenge. Metabolic rate and substrate utilization were measured both at rest, immediately after consumption of the high fat test-meal and every hour thereafter for three hours. Dietary intake (3 x 24h recalls) and physical activity (ACTi Graph GT3X accelerometer worn for 7 days) data was collected and body composition was measured (bioelectrical impedance, BIA). Questionnaires were also completed covering weight history, socio-economic status and eating behaviour. Results showed that there was no difference in either resting EE or substrate utilization between the RED and REL compared to the respective BMI-matched controls, after accounting for fat free mass (FFM). The TEF, postprandial EE (absolute and per kg FFM), post-prandial energy balance, RER, fat oxidation rate and post-prandial fat balance were similar between RED and REL compared to their respective controls, indicating that there was no evidence of metabolic adaption to weight loss. However, successful weight-loss maintainers did show behavioural strategies that may have counteracted weight-loss associated adaptive thermogenesis and supported weight-loss maintenance. These individuals had manipulated macronutrient intake (increasing protein and reducing carbohydrate intake), were more physically active, exhibiting less sedentary behaviour and increased moderate and vigorous activity, and had greater fat free soft tissue mass (FFSTM). While the presence of adaptive thermogenesis is not disputed in these results, the distinct physiological and behavioural differences together observed in the RED may have been instrumental in attenuating weight-loss associated declines in EE, shown to persist into weight-loss maintenance. Together with these lifestyle strategies, weight reduced individuals also reported greater dietary restraint in comparison to controls. This is surprising after such a significant period of weight-loss maintenance (median weight-loss maintenance: 30 months) and highlights the ongoing challenges to maintain reduced weight. These findings contribute to the relatively smaller body of research into the longer-term persistence of weight-loss associated adaptive responses in comparison to that covering the acute weight loss phase. It also highlights strategies that may be effective in counteracting metabolic adaption to weight loss. As such, these strategies may warrant inclusion as part of weight-loss maintenance programs as they potentially help to reduce the risk for weight regain as a result of weight-loss associated adaptive thermogenesis. The next study in Part 2 of the thesis (Chapter 5) aimed to examine the impact of successfully maintained weight loss and weight-loss relapse on insulin sensitivity compared to BMImatched controls without a weight loss history. Predictors of variability in insulin sensitivity were also explored. Following the measurement of resting metabolic rate and substrate utilization a 75g oral glucose tolerance test was used to determine fasting and 2hr plasma glucose and insulin. The Homeostatic Model Assessment (HOMA-IR) and insulin sensitivity index (ISI(0,120)) were used to assess insulin sensitivity. A novel finding of this study was that successfully maintained, weight-reduced individuals displayed enhanced measures of insulin sensitivity (lower HOMA-IR and higher ISI(0,120) measurements), compared to all other groups, including BMI-matched controls with no weight loss history. Previously studies have investigated changes in insulin sensitivity in response to weight loss and in weight-loss maintenance, but not necessarily in comparison to individuals without a weight loss history as defined by this study protocol. With weight regain however, insulin sensitivity measures for REL were not different compared to either LSW or OSW, showing that enhanced insulin sensitivity accompanying weight loss is likely reversed with weight regain. Prior weight history, fasting substrate utilization, measures of body weight and composition, protein intake per kilogram, physical activity and CRF were all associated with measures of insulin sensitivity. Using these variables in regression models, ≈60% of the variability in insulin sensitivity in both HOMA-IR and ISI(0,120). Weight loss and weight regain history followed by fasting RER were the most significant independent predictors of insulin sensitivity. In conclusion, a novel finding was that successfully weight-reduced individuals are more insulin sensitive than their BMI-matched controls with no weight loss history, independent of dietary intake and physical activity. This remains evident even after significant periods of maintaining the reduced weight. Weight loss maintenance programs are essential to retaining metabolic benefits acquired through weight loss. Remaining physically active by reducing sedentary behaviour and in particular including small amounts of vigorous physical activity significantly predicts improved insulin sensitivity. This thesis includes a number of novel findings. In Part 1, we showed that in response to exercise training gynoid FM, rather than VAT, was reduced in sedentary obese black SA women undergoing a 12 week exercise intervention, which may represent an important ethnic/gender specific response. We also showed that substrate utilization plays an important role in altering body composition and CRF in response to an exercise intervention. Greater fat oxidative capacity at the outset resulted in an enhanced ability to reduce gynoid FM in response to exercise training. Furthermore, a greater reliance on carbohydrate rather than fat oxidation during baseline testing predicted a poorer CRF response. Identification of individuals with a lower capacity for fat oxidation at the outset of an exercise intervention may therefore allow for a more targeted exercise prescription, which may in turn improve outcomes of exercise interventions. The lack of clinically significant weight loss suggests that future exercise interventions should prescribe exercise EE of sufficient magnitude to achieve weight loss and emphasize adherence to this prescription or include some dietary restriction. Education around the possible adaptive responses to increased EE and the imposed energy deficit, highlighting the strategies employed by weight reduced individuals from Part 2 of this thesis, may help to attenuate potential metabolic adaption to increased EE and further improve the weight loss outcomes of exercise-only interventions. It may also help to inform weight-loss maintenance programs to assist individuals to maintain the reduced weight following weight loss. The enhanced insulin sensitivity in weight reduced individuals as shown in Part 2, may potentially represent an ongoing and persistent adaptive response to weight loss that may in itself increase the risk for weight-loss relapse. Education around the physiological adaption to significant weight loss and emphasizing strategies that may counteract this metabolic adaptation may improve the efficacy of both weight-loss and weight-loss maintenance programs.