LAUSR

Aims It is recognised that different infant growth references (GR) can alter apparent infant growth patterns. The influence of the choice of GR on apparent associations between early growth and later outcomes in research studies is unknown. Our aim was to examine whether the choice of GR influences (1) apparent infant growth patterns; and (2) associations with later anthropometry and body composition (BC). Methods Weight, height, BMI, waist circumference (WC)) and BC (four component model: fat mass (FM), fat free mass (FFM)) were measured in 209 healthy British children (mean (SD) 11.1 (3.5) yrs, 104 boys). Infant weight data were extracted from parent-held child records for birth, 3, 6 and 12wk, 6 and 12mo and converted to standard deviation scores (SDS) using British 1990 Growth Reference (UK90) and World Health Organisation (WHO) growth standards. Change in Weight SDS (ΔwtSDS) using the two GRs were compared and associations between ΔwtSDS 3–6wk and 0–12mo and later measurements were assessed, adjusting for confounders. Results ΔwtSDS calculated using UK90 and WHO were significantly different at all ages. UK90 ΔwtSDS 0–3wk was significantly <0 and ΔwtSDS 3–6wk significantly higher. WHO SDS were significantly <0 for ΔwtSDS 3–6wk and significantly higher for 12wk-6mo, 6–12mo and −12mo. In boys, ΔwtSDS 36wk was positively associated with later height, weight and FFM regardless of GR although associations were stronger for WHO than UK90 (e.g. 0.65 WHO (p=0.02), 0.57 UK90 (p=0.04)). In girls, ΔwtSDS 3–6wk was positively associated with later FM and fat-free mass index (FMI) using both UK90 and WHO, and with weight, BMI, WC only using UK90; all associations were stronger using UK90 than WHO. ΔwtSDS birth-12mo in boys and girls was significantly associated with all later outcomes except FMI, with similar effect sizes using UK90 and WHO. Conclusion This study showed that choice of GR does result in differences in infant growth patterns (1). This does in turn, influence apparent associations with later outcomes (2): effects may differ by gender and growth period. This should be considered by researchers.