Short-term changes in growth and urinary growth hormone, insulin-like growth factor-I and markers of bone turnover excretion in healthy prepubertal children

Childhood growth is a non-linear process. To assess whether there is a biochemical correlate of non-linear growth, we have measured free pyridinoline (fPYR) and deoxypyridinoline (fDPYR) excretion in seven healthy prepubertal children, aged 6.1-7.7 years. To examine the link between short-term growth and hormone output, urinary growth hormone (uGH) and insulin-like growth factor-I (uIGF-I) were also measured. Height and weight were measured and a timed overnight urine was collected three times per week from September to July, with results expressed as a weekly change in height (Dheight(w)) or weight (Dweight(w)), and as weekly average hormone or bone marker excretion (uGH(w), uIGF-I(w), fPYR(w), fDPYR(w)). Subject specific SD scores (SDS) were derived for each variable.Dheight(w)and Dweight(w)did not correlate to uGH(w), uIGF-I(w), fPYR(w)or fDPYR(w). Dheight(w)SDS was weakly but significantly correlated to fPYR(w)SDS (r = +0.16;P<0.05) and fDPYR(w)SDS (r = +0.15;P<0.05). The percentage of high frequency (2-4 weeks) variation in uGH(w)excretion, as defined by time series analysis, was correlated with the mean uIGF-I(w)(r = +0.81;P<0.05), which in turn was significantly reduced (92 +/- 38 vs 120 +/- 47 ng;P<0.001) during periods of slow growth (Dheight(w)< 0.05 cm/week). We conclude that in normal children the amount of urinary fPYR, fDPYR, GH and IGF-I does not provide a direct biochemical correlate of growth from week to week. However good growth is associated with a relative increase in fPYR and fDPYR excretion, while poor growth is associated with reduced IGF-I excretion, which in turn is influenced by the temporal secretory pattern of GH over 2-4 weeks.

The cyclical nature of prepubertal growth

We have examined the complete longitudinal height velocity (HV) data of 135 (80 male and 55 female) chromosomally normal children from the Edinburgh Longitudinal Growth Study who were measured six-monthly between age 3 years and the onset of the adolescent growth spurt. Individual HV curves appeared to consist of a regular series of accelerations and decelerations in a cyclical fashion. After excluding variations due to measurement error, and basing the analysis on the pattern of acceleration, we were able to identify a number of spurts of regular occurrence. The mid-childhood spurt was clearly identified at ages 7.0 in boys and 6.7 years in girls, and could be identified in all children except one girl. Other spurts were also apparent; a pre-school spurt at ages 4.8 and 4.6, a late-childhood spurt at ages 9.2 and 8.6, and in children with an average-to-late onset of puberty, a prepubertal spurt at ages 10.8 and 10.0 years in boys and girls respectively. Synchronization at peak HV was performed for each spurt as described by Shuttleworth (1937) for the adolescent growth spurt. The overall pattern of growth appeared to be cyclical with a mean peak interval of 2.2 years in boys and 2.1 years in girls, the cycles appearing to continue until interrupted by the onset of the adolescent growth spurt. Sitting height velocity (SHV) and leg length velocity (LLV) curves also showed a cyclical pattern, but each varied independently. The magnitude of the HV spurts in an individual was dependent on the synchrony between the phases of SHV and LLV spurts. The cyclical pattern of prepubertal growth with its rapid changes in HV should be taken into account when assessing the growth of any child and in the response to any treatment offered.