Metabolic, methodological and developmental aspects of body composition : Studies in women and children with special reference to early life mechanisms behind childhood obesity

Sammanfattning: In recent decades the number of children with overweight has increased worldwide. To understand the mechanisms behind this development, knowledge regarding metabolism and physiology in relation to the nutritional situation in early life is of importance. In particular, information about body composition development during early childhood is relevant. This thesis presents three studies in this area. In the pregnancy study serum samples, collected from 23 women before, during and after pregnancy, were analysed for serum levels of leptin, adiponectin and resistin and used to assess insulin resistance (HOMA-IR) in relation to the total body fat (TBF) content of the women. TBF (%) and leptin were significantly correlated with HOMA-IR before and during pregnancy. When HOMA-IR was regressed on TBF (%) the slope of the regression line was 0.111 in gestational week 32 and significantly (p<0.05) higher than the value before pregnancy, 0.046, indicating that healthy pregnancy enhances the relationship between body fatness and insulin resistance. In the HF-study hydration of fat-free mass (hydration factor, HF) was assessed in 12 newborns using the doubly labelled water (DLW) method and air displacement plethysmography (PeaPod). HF was 80.9% with a low biological variability (0.81% of average HF). In the longitudinal study the body density of 108 healthy fullterm infants (53 girls, 55 boys) was measured at one and 12 weeks of age using PeaPod. Body composition was calculated using two models (Fomon’s and Butte’s). BMI values for the mothers of the infants were assessed before pregnancy. Body composition and total energy expenditure using the DLW-method were assessed in 20 of these children at the age of 1.5 years, when their sleeping metabolic rate was measured using indirect calorimetry and their resting energy metabolism was calculated using prediction equations. Butte´s model gave significantly (p<0.05) lower values for TBF than Fomon´s model, and invalid results for five newborns. Using Fomon´s model, at one week of age girls contained 13.4 ± 3.7 % and boys contained 12.5 ± 4.0 % TBF. The corresponding figures at 12 weeks were 26.3 ± 4.2 % and 26.4 ± 5.1 %. The mothers’ BMI values before pregnancy were correlated with the body weight but not with the TBF (g,%) or fat-free mass (g) of their infants at one week of age. At 1.5 years of age girls (n=9) contained 28.0±2.8 % and boys (n=11) 28.3±3.7 % TBF. Between one and 12 weeks of age all infants increased their TBF content, while 13 children increased and seven children decreased their TBF content between the ages of 12 weeks and 1.5 years. The results demonstrated that predicting rather than measuring resting energy metabolism involves a risk for spurious correlations between TBF and physical activity level. The level of physical activity (x), was negatively correlated with [TBF (%) at 1.5 years minus TBF (%) at 12 weeks] (y), r=-0.52, p=0.02. In conclusion, the results suggest that the body fat content of a woman has a stimulating effect on the growth, rather than on the fat retention, of her foetus. They also show that the Fomon model is the best available model when calculating the body composition of infants from body density. Finally, the results indicate that physical activity at the age of 1.5 years is important regarding the rate at which the high level of body fat, typical of infancy, decreases in early childhood.