Body composition during early infancy : pre- and postnatal determinants and assessment approaches

Background, research gaps and aims Obesity has reached pandemic levels among adults worldwide and the increasing prevalence among children and adolescents is alarming. There is strong evidence that prenatal and early postnatal growth and developmental plasticity play a pivotal role in determining the risk of obesity and co-morbidities across the life span. In humans, developmental plasticity is heightened during the so-called first 1000 days (from conception to two years of age), such that a stimulus or an insult during this sensitive period may cause permanent alterations in the growth trajectory with potential for lifelong consequences. In this context, optimising growth during early life through appropriate nutrition and environmental exposures is of paramount importance. Clinical assessment of growth in infants has long been based on anthropometric measures, with birthweight being the most widely used marker of foetal nutrition and other intrauterine exposures. Both extremities of the birthweight spectrum, i.e., low birthweight (LBW) and high birthweight (HBW), have shown associations with obesity during childhood and adulthood. Hence, a better understanding of secular trends in birthweight and associated maternal factors at the population level may help in planning preventative strategies to reduce LBW and HBW rates and improve population health. Thus far, secular trends in birthweight and associations with maternal factors have not been studied in Tasmania, the Australian state with the highest rates of obesity in children and adults. As birthweight is only a crude indicator of the nutritional status of a newborn, body composition assessment is a valuable addition providing information of differential growth in fat mass (FM) and fat-free mass (FFM). Mounting evidence indicates that relatively greater gains in adiposity during early development is associated with later life obesity. Several pre- and postnatal risk factors for increased adiposity during infancy have been identified; however, there are inconsistencies and contradictions in findings. Differences in the age of infants, the technique used to estimate body composition, and the index used to elucidate variability of adiposity may explain some of these discrepancies. Further, in utero exposure to gestational diabetes mellitus (GDM) has been identified as one of the main risk factors for increased adiposity at birth. An increasing awareness of short- and long-term adverse effects of GDM to both mother and infant has led to substantial improvements in perinatal care and provision of treatments for women with GDM during recent years. However, evidence on whether treatments for GDM can normalise adiposity in newborns is still inconsistent. Among the various body composition techniques available, air displacement plethysmography (ADP using PEA POD) is one of the most ‚ÄövÑv¿practical‚ÄövÑvp approaches for use in early infancy. ADP allows precise estimations of body composition rapidly and non-invasively. However, the PEA POD only accommodates infants up to ~6 months of age (<10 kg body weight). Therefore, alternative body composition techniques should be used in longitudinal studies that follow-up infants beyond 6 months of age. The deuterium dilution (DD) technique is another body composition approach suitable for use in infants and has good reliability and validity. It can be used in infants from birth, is comparatively inexpensive, and collected body fluids can be stored until analysed. The DD technique may be a suitable alternative when infants can no longer be accommodated in the PEA POD. Thus far, only one study has reported that body composition measurements obtained using PEA POD and DD are in agreement. Infants who participated in this study were predominantly Asian, and their ages varied from 0.4-24.4 weeks. Thus, additional research is needed to test the agreeability of the two techniques in infants of other ethnicities and in larger samples of the same age. The four studies undertaken for this PhD thesis aimed to improve the knowledge and understanding of determinants and measures of infant body composition. The first study explored secular birthweight trends and associated maternal factors in Tasmania. The second study identified pre- and postnatal determinants of adiposity from birth to 6 months of life in a sample of healthy Tasmanian infants. The third study systematically reviewed the literature reporting different adiposity measures in newborns of mothers with GDM controlled with therapeutic interventions compared to those with normal glucose tolerance (NGT). The fourth and final study appraised the agreement of body composition measures assessed via PEA POD in relation to DD technique in 6-month-old infants. Methods and results Perinatal data of all live-born singletons and their mothers, linked by the Tasmanian Data Linkage Unit (n = 81700), were used to investigate the secular trends (from 2005 to 2018) in birthweight and associations with pre-pregnancy and pregnancy maternal factors in Tasmania. Over the 14 years, mean birthweight (3425 g to 3359 g) and the proportion of HBW (14.2% to 11.0%) decreased, while the proportion of LBW increased (4.8% to 6.5%). However, as of 2018, the rate of HBW (1 out of 9 babies) was still higher than the rate of LBW (1 out of 15 babies). A downward shift in gestation length distribution, increased rates of mothers with caesarean delivery, hypertensive disorders, age >35 years, and changes in ethnic demographics with an increased number of indigenous or immigrant mothers, may have contributed to this trend towards smaller babies. Although the rates of pre-pregnancy obesity and GDM (well-known risk factors of HBW) have risen, and maternal smoking (a major risk factor for LBW) has markedly decreased over the period, the impact of these changes in mothers was not apparent in the infant birthweight trend. A prospective longitudinal cohort study design was used to investigate associations between pre- and postnatal factors and infant adiposity measures from birth to 6 months. This study was conducted at the Launceston General Hospital, Tasmania, from September 2017 to October 2019. Information on pre- and postnatal exposure variables was obtained from mothers through an interviewer-administered questionnaire, and infant body composition was measured using PEA POD. Linear mixed-effects modelling with backward stepwise regression was used to assess longitudinal associations between pre- and postnatal factors and infant adiposity. To test whether the discrepancies in predictors of infant adiposity in the literature stemmed from using various indices to measure adiposity, a range of adiposity measures, i.e., FM, percent fat mass (%FM), fat mass index (FMI) and log-log index (FM/FFMP), were used. The body composition of 322 infants was assessed within 72 hours of birth, and of those, 174 and 109 were followed up at 3 and 6 months, respectively. Positive associations were observed between gestation length and infant FM, parity and infant %FM and FMI, and pre-pregnancy body mass index (ppBMI) and infant %FM at birth. Male infant sex and formula feeding were negatively associated with all adiposity indices at 6 months. Surprisingly, maternal intake of iron supplements during pregnancy was negatively associated with infant FM, %FM and FMI at 3 months and FM/FFMp at 6 months; however, this finding should be interpreted with caution as our analysis lacked information on doses and duration of prenatal supplements. These results suggest that some of the inconsistency in the literature regarding pre- and postnatal impacts on infant adiposity is potentially due to the use of different measures of adiposity in various studies. We systematically reviewed the literature reporting FM, %FM and skinfold thickness (ST) in infants of mothers with GDM controlled with therapeutic interventions (IGDMtr) following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. In total, 25 studies were included in the systematic review, of which 17 were included in the meta-analysis. Regardless of the type of treatment (insulin, metformin, glyburide), treating GDM lowered FM in newborns compared to no treatment. A meta-analysis of all studies showed that, compared to infants exposed to normal glucose tolerance (INGT), IGDMtr had higher overall adiposity (mean difference, 95% confidence interval (CI)) as measured by FM (68.46 g, 29.91 to 107.01) and %FM (1.98%, 0.54 to 3.42). In contrast, subcutaneous adiposity measured with ST did not differ between the two infant groups. However, a meta-analysis of a subgroup of more recent studies (data collection occurred during or after 2010) showed there were no significant differences in FM, %FM and ST between INGT and IGDMtr. It is possible that more intensive management of blood glucose levels in mothers with GDM during recent years has normalised the adiposity in their infants. We used Bland-Altman analysis to evaluate the agreement of body composition measures (FM, %FM and FMI) assessed via PEA POD and the DD technique in a sample (n = 72) of 6-month-old infants. The differences between the two methods were not constant (FM: bias = 25.26, 95%CI = -65.92 to 116.45; %FM: 0.33; -0.93 to 1.60; FMI: 0.06; -0.15 to 0.27); however, the limits of agreement (LOA) were wide and significant proportional bias was identified with DD technique underestimating infant adiposity at lower values and overestimating infant adiposity at higher values, in comparison to PEA POD. Further analyses were performed to investigate whether the adiposity values at lower or upper extremities significantly affected the bias. When the analysis was conducted with mean values above the first quartile (n = 53), LOA was somewhat narrower (FM: -667.84 to 519.91; %FM: -9.15 to 7.96; FMI: -1.58 to 1.27), and no proportional bias was detected (p > 0.1 for all). Wide LOA and significant...