Obesity and its determinants in girls from five ethnic groups

In light of alarming rises in the prevalence of obesity worldwide, tackling the obesity ‘epidemic’ is now a national health priority in many countries. Increasingly, population measures that provide accurate estimates of body fatness in children are required. Body mass index (BMI), or weight standardised for height, remains the most cost-effective and practical tool in this regard. However, there is evidence that the association between BMI and body fatness is variable in children from different ethnic backgrounds. The primary aim of this thesis was to investigate the appropriateness of BMI thresholds for defining overweight and obesity in female children and adolescents from five diverse ethnic groups. Secondary objectives were to examine the associations between weight control practices and perceptions, and to compare objectively-measured physical activity levels with participation in active transport (AT). In order to achieve the primary aim stated above, it was necessary to obtain valid and reliable measures of body fat percentage (%BF) in a large sample of children. Bioelectrical impedance analysis (BIA) is well suited for this purpose, providing a portable and cost-effective means to estimate fat-free mass (and subsequently %BF). While equations exist for European, Maori, and Pacific Island children, findings from the preliminary study described in Chapter 2 demonstrate that there are no BIA equations appropriate for Chinese and Indian children. Given that these two groups are two of the fastest growing ethnicities in New Zealand, a new equation was developed that enables Asian girls to be included in future BIA research. The main study of this thesis involved a large-scale investigation of body composition in New Zealand’s five major ethnic groups (European, Maori, Pacific Island, East Asian, and South Asian). A total of 1,081 adolescent girls aged 11-16 years participated in the Girls’ Activity and Body Composition (ABC) Study. To extend the age range, data were combined with another study of 5-11-year-old New Zealand children (595 girls), coined the Body-Size and Steps in Children (BASIC) Study. Both studies measured BMI from height and weight, %BF from bio impedance measurements, and physical activity using sealed multiday memory pedometers over five consecutive days. A questionnaire was also administered to the adolescent-aged girls to gather data related to weight perceptions and practices. Initial analyses of the main dataset demonstrated that existing BMI definitions of overweight were relatively insensitive for predicting excess %BF in South and East Asian girls. Conversely, low specificity was observed for Pacific Island and Maori children. These findings provided the rationale for the second set of analyses: the development of BMI cut-off points that correspond to an equivalent level of %BF across all ethnicities. The adjusted BMI curves for overweight and obesity ranged from an average of 3.3 and 3.8 kg.m-2 (respectively) lower than international standards in South Asian girls to 1.5 and 1.9 kg.m-2 higher in Pacific Island girls. Clearly, the proposed changes will have significant effect on our estimates of overweight and obesity in this population group. Subsequent investigation revealed that many adolescent girls misclassify their weight status. However, the number of girls trying to lose weight exceeded those who perceived themselves as being overweight, with the magnitude of the difference dependent on ethnicity. It was concluded that interventions and educational campaigns that assist girls in recognising a state of excess body fat are a priority for all ethnic groups to ensure that behavioural changes necessary to combat widespread overweight and obesity are adopted. Finally, it was observed that the physical activity levels of the participants were significantly lower on weekends (9,528  4,407) than on weekdays (12,597  3,630). Furthermore, a consistent decline in daily step counts was observed with age: after adjustment for ethnicity and socioeconomic status (SES), girls in school years 9-10 achieved 2,469 (weekday) and 4,011 (weekend) fewer steps than girls in years 1-2. Daily step counts also varied by ethnicity, with Maori girls the most active and South Asian girls the least active. Overall, girls who used AT to and from school averaged 1,052 more weekday steps than those who did not use AT. These data suggest that adolescent-aged girls and girls of Asian descent are priority groups for future physical activity interventions, and that the promotion of AT in girls appears to be worthwhile