Predicting the growth response to growth hormone (GH) treatment in prepubertal and pubertal children with isolated GH deficiency--model validation in an observational setting (GeNeSIS).

AIM

Validation of a preexisting mathematical model for predicting height velocity in the first year of growth hormone (GH) treatment, using an independent cohort of pediatric patients with confirmed GH deficiency (GHD).

DESIGN

210 prepubertal and pubertal patients with isolated GHD were enrolled within the 'growth prediction module' of a prospective international postmarketing research program (GeNeSIS).

METHODS

The patients were grouped as 'prepubertal', at 'start of puberty' and 'pubertal', and the performance of the model was validated by comparing predicted and observed height velocity (HV) in the first treatment year for each group. The operational characteristics of the model, when used as a screening method to predict poor and good growth response to GH, were calculated using the first year change in height standard deviation score as a cut-off for treatment response.

RESULTS

Depending on pubertal stage, the growth prediction model explained 53-72% of the variability of the first year HV. The model fulfilled the statistical prerequisites to identify patients with poor and good responses to GH with sufficient reliability in individual patients.

CONCLUSIONS

The growth prediction model was successfully validated in a large cohort of prepubertal and pubertal pediatric patients under field conditions.