ESPE Abstracts (2016) 86 P-P1-599

Validation of Prediction Models for Near Final Adult Height in Children with Idiopathic Growth Hormone Deficiency Treated with Growth Hormone for 1 Year

Saartje Straetemansa,b, Jean De Schepperb,c, Muriel Thomasb, Franciska Verlindeb & Raoul Roomanb


aMaastricht University Medical Center, Maastricht, The Netherlands; bThe Belgian Society for Pediatric Endocrinology and Diabetology, Brussels, Belgium; cUniveristy Hospital Brussels, Brussels, Belgium


Background: An accurate prediction of final height after the first year of growth hormone (GH) treatment may help clinicians to give parents and children more realistic expectations.

Objective and hypotheses: To validate two prediction models (with and without max. GH peak) for near final adult height (nFAH) by Ranke et al.

Method: Height data of 142 (93 male) idiopathic GH deficient (iGHD) children, treated with GH for at least four consecutive years with at least one year before puberty, who attained nFAH, were retrieved from the database of the Belgian Society for Pediatric Endocrinology and Diabetology. Bland Altman (BA) plots were used for validation of prediction models. Clark error grid (CEG) analysis was performed to assess the clinical significance of the differences between observed and predicted nFAH.

Results: Mean duration of GH therapy was 9.6 years. Mean nFAH for boys was 169.1 cm (−1.76 SD), for girls 157.5 cm (−1.54 SD), mean nFAH minus midparental height SDS was −0.41, median total height gain was 1.73 S.D. In males, the Ranke nFAH predictions were significantly higher than the observed nFAH (difference: 0.29 S.D. ±0.66; 95% CI 0.14–0.43; P<0.001). In females, there was no significant difference. BA: the means of the differences between the observed and predicted nFAH are close, but not equal to zero. There is a proportional bias with an overprediction for the smaller heights and an underprediction for the taller heights. We propose a small correction equation. CEG: 56% of predicted nFAH are in zone A (<0.5 S.D. difference from observed nFAH), 28–31% are in zone B (0.5–1 S.D. difference), 13–16% are in zone C (>1 S.D. difference). In females, 38–40% of predicted nFAH are in zone A, 38–40% in zone B, 22% in zone C.

Conclusion: Ranke’s prediction models for nFAH after first-year GH therapy accurately predict nFAH in females and overpredict nFAH in males by about 2 cm. We propose a small correction equation that should be evaluated in a separate cohort. These prediction models are useful for predicting nFAH after first-year GH treatment in clinical practice.

Article tools

My recent searches

No recent searches.