Refereed journal article or data article (A1)
Predictive utility of childhood anthropometric measures on adult glucose homeostasis measures: a 20-year cohort study
List of Authors: Wu Feitong, Ho Valentina, Fraser Brooklyn J., Schmidt Michael D., Dwyer Terence, Venn Alison J., Magnussen Costan G.
Publisher: Nature Publishing Group
Publication year: 2018
Journal: International Journal of Obesity
Journal name in source: International Journal of Obesity
Volume number: 42
Issue number: 10
Start page: 1762
End page: 1770
Number of pages: 9
ISSN: 0307-0565
eISSN: 1476-5497
DOI: http://dx.doi.org/10.1038/s41366-018-0177-z
Background/objectives
Childhood body mass index (BMI) predicts adult glucose homeostasis measures and type 2 diabetes mellitus, but little is known about the predictive utility of other anthropometric measures in childhood. We aimed to identify the anthropometric measure in childhood that best predicts adult glucose homeostasis measures and examine if the combination of additional anthropometric measures further improves predictive utility.
Methods
A 20-year follow-up of children participating in the Childhood Determinants of Adult Health Study (n = 2345, aged 7–15 years at baseline). Baseline anthropometric measures were waist circumference (WC), WC adjusted for height, weight adjusted for height, hip circumference, waist-hip-ratio, waist-height-ratio, BMI, conicity index, abdominal volume index (AVI), body adiposity index, and a body shape index. Fasting glucose and insulin levels measured at follow-up were used to define insulin resistance (HOMA2-IR), low beta-cell function (HOMA2-β), high fasting insulin, and impaired fasting glucose (IFG).
Results
All child anthropometric measures were significantly associated with HOMA2-IR, HOMA2-β, and high fasting insulin (relative risk = 1.12–1.55), but not IFG. AVI had the largest area under receiver-operating curve (AUC) in predicting adult HOMA2-IR (AUC, 95% confidence interval: 0.610, 0.584–0.637), HOMA2-β (0.615, 0.588–0.642) and high fasting insulin (0.613, 0.587–0.639). Combining each additional anthropometric measure with AVI did not appreciably increase predictive utility (an increase of 0.001–0.002 in AUC, p > 0.05 for all).
Conclusions
Anthropometric measures from a single time-point in childhood are associated with insulin-related outcomes 20-year later in adulthood. However, overall predictive utility was low and was not substantially enhanced by combining multiple different child anthropometric measures.