Predicting the response to growth hormone treatment in short children with chronic kidney disease.

CONTEXT: Short stature in children with chronic kidney disease (CKD) is due to various underlying congenital or acquired renal disorders resulting in variable impairment of renal function and variable response to GH treatment.
OBJECTIVE: It was the aim to develop a mathematical model that allows the prediction of the individual growth response and to identify nonresponders.
DESIGN: Data from 208 prepubertal children on conservative or dialysis treatment in a large pharmaco-epidemiological survey, the KIGS (Pfizer International Growth Database), were used for the model and data from 67 similar CKD patients registered at the Dutch Growth Research Foundation for validation.
RESULTS: Annualized height velocity (centimeters per year) during the first year of GH treatment was best predicted by age at start, weight sd score, underlying renal disorder (hereditary kidney disorder), glomerular filtration rate (at baseline), and GH dosage. Using these parameters, the final model explained 37% of the overall variability of growth response. Standard error of the estimates was 1.6 cm. Age was the most important predictor of growth response (20.3% of variability) followed by weight sd score at start, and 27.2% of the variability of the second-year response could be predicted by the first-year response and glomerular filtration rate. Nonresponders of the validation group could be correctly identified.
CONCLUSION: Based on simple clinical variables, a robust prediction model was developed that provides realistic expectations of individual growth response to GH in short children with CKD. The model will help in identifying nonresponders and to tailor treatment strategies.