Prediction of response to growth hormone treatment in pre-pubertal children with growth hormone deficiency.

Background: GH therapy response varies substantially among patients. Several models were developed to predict the efficacy of GH therapy in children.

Aim: To evaluate the accuracy of a growth prediction model using data from an Italian pediatric GH deficiency (GHD) cohort (GeNeSIS, Growth Prediction Sub-study).

Methods: Open-label, multicenter study in 22 Italian pre-pubertal GH treatment- naïve patients with GHD (8 female, 14 male, 0.5 to 12.2 yr), 18 isolated GHD, 4 multiple pituitary hormone deficiency given recombinat human GH therapy (0.025-0.035 mg/kg/day) for 12 months. Growth prediction was performed, after 3 months of treatment, using baseline data [bone age (BA) and IGF-I], a urinary marker of bone turnover [deoxypyridinoline crosslinks (DPD)] at 4 weeks, and height velocity (HV) at 3 months. Results were expressed as 1st-yr HV using the following equation: 1-yr HV (cm) = 3.543 - (2.337 × BA) - (0.010 × IGF-I) + (0.100 × DPD) + (0.299 × 3-month HV). Predictions were compared to the 1st-yr HV and accuracy was calculated as percentage of the difference between mean calculated HV and the real 1st-yr HV.

Results: For females predicted HV was 12.98 ± 4.82 cm/yr and actually was 13.05 ± 3.91 cm/yr after the 1st year; for males predicted HV was 13.95 ± 5.39 cm/yr and actually was 12.93 ± 5.02 cm/yr.

Conclusions: In this paediatric Italian cohort with GHD, a growth prediction model seems to be a valid tool to assess 1st-yr response to GH treatment in Italian children.