482 TRANSIENT GROWTH DECELERATION IN NORMAL SHORT CHILDREN

Abstract

Several studies have been reported, in which children with growth retardation have been used as their own growth velocity (GV) controls in evaluating therapeutic interventions. In order to judge the validity of such an approach, the natural history of GV before and after a timepoint at which child is likely to be enrolled in such a study must be known. We have examined the GV of 24 children who underwent growth hormone (GH) testing because of concern about their growth (Height <3d centile, GV <5cm/yr), but were found not to be GH deficient. They were 17 boys and 8 girls, ranging in age from 2 to 16 yrs. (mean 10.05). Mean GV before testing was 3.8 cm/yr. Following testing, and without any specific treatment, GV was greater in all but 2 of them, with a mean of 5.1cm/yr. The difference (mean±SEM) was 1.3±0.298, p=0.0016) Thus, a transient growth deceleration appears to occur in some non- GH-deficient children, during which they are most likely to be referred for testing and subsequently be enrolled in a therapeutic trial. This creates an ascertainment bias that may produce spurious evidence of effectiveness of the intervention under evaluation. Matched or randomized controls are essential in such studies.