Timing of the Infancy-Childhood Growth Transition in Rural Gambia

Abstract

The Karlberg model of human growth describes the infancy, childhood, and puberty (ICP) stages as continuous and overlapping, and defined by transitions driven by sequential additional effects of several endocrine factors that shape the growth trajectory and resultant adult size. Previous research has suggested that a delayed transition from the infancy to the childhood growth stage contributes to sub-optimal growth outcomes. A new method developed to analyze the structure of centile crossing in early life has emerged as a potential tool for identifying the infancy-childhood transition (ICT), through quantifying patterns of adjacent monthly weight-for-age z-score (WAZ) deviation correlations. Using this method, the infancy-childhood transition was identified as taking place at around 12 months of age in two cohorts of UK infants. Here, we apply this method to data collected as part of a longitudinal growth study in rural Gambia [the Hormonal and Epigenetic Regulators of Growth, or HERO-G study, N = 212 (F = 99, M = 113)], in order to identify the ICT and assess whether timing of this transition differs across groups based on sex or birth seasonality. We calculated Pearson correlation coefficients for adjacent monthly WAZ score deviations. Based on the patterns of change in the correlation structure over time, our results suggest that the infancy-childhood transition occurs at around 9 months of age in rural Gambian infants. This points to an accelerated ICT compared to UK infants, rather than a delayed ICT. A comparatively later transition, seen in UK infants, allows maximal extension of the high rates of growth during the infancy stage; an earlier transition as seen in Gambian infants cuts short this period of rapid growth, potentially impacting on growth outcomes in childhood while diverting energy into other processes critical to responses to acute infectious challenges. Growth in later developmental stages in this population offers an extended window for catch-up.

Keywords: childhood; growth; hormonal growth regulation; infancy; infancy-childhood transition.

Figure 1

(A) Weight for age z-scores (WAZ) by age. WAZ means and standard deviations given in Table 1. Blue circles: HERO-G infants; Green squares: Widdowson cohort; Red triangles: CIGS cohort. Widdowson and CIGS data from Cole et al. (28). (B) Correlations by age between adjacent pairs of WAZ deviations. Correlations were computed using data from every available subject. The sample sizes for each correlation are given in Table 2. Blue circles: HERO-G infants; Green squares: Widdowson cohort; Red triangles: CIGS cohort. Widdowson and CIGS data from Cole et al. (28). (C) Correlations by age between adjacent pairs of WAZ deviations. The sample sizes for each correlation are given in Table 2. Orange circles: female HERO-G infants; Purple squares: male HERO-G infants. (D) Correlations by age between adjacent pairs of WAZ deviations. Orange circles: female HERO-G infants; Purple squares: male HERO-G infants; Green squares: Widdowson cohort; Red triangles: CIGS cohort. Widdowson and CIGS data from Cole et al. (28). (E) Correlations by age between adjacent pairs of WAZ deviations. The sample sizes for each correlation are given in Table 2. Brown circles: HERO-G infants born in the dry season; Green circles: HERO-G infants born in the wet season. (F) Correlations by age between adjacent pairs of WAZ deviations. Brown circles: HERO-G infants born in the dry season; Green circles: HERO-G infants born in the wet season; Green squares: Widdowson cohort; Red triangles: CIGS cohort. Widdowson and CIGS data from Cole et al. (28).