Isolated childhood growth hormone deficiency: a 30-year experience on final height and a new prediction model

Abstract

Purpose: We aimed to evaluate the near-final height (nFHt) in a large cohort of pediatricpatients with growth hormone deficiency (GHD) and to elaborate a new predictive method of nFHt.

Methods: We recruited GHD patients diagnosed between 1987 and 2014 and followed-up until nFHt. To predict the values of nFHt, each predictor was run in a univariable spline.

Results: We enrolled 1051 patients. Pre-treatment height was -2.43 SDS, lower than parental height (THt) (-1.09 SDS, p < 0.001). The dose of recombinant human GH (rhGH) was 0.21mg/kg/week at start of treatment. nFHt was -1.08 SDS (height gain 1.27 SDS), higher than pre-treatment height (p < 0.001) and comparable to THt. 1.6% of the patients were shorter than -2 SDS from THt. The rhGH dose at nFHt was 0.19 mg/kg/week, lower than at the start (p < 0.001). The polynomial regression showed that nFHt was affected by gender, THt, age at puberty, height at puberty, age at the end of treatment (F = 325.37, p < 0.0001, R² 87.2%).

Conclusion: This large national study shows that GHD children can reach their THt. The rhGH/kg/day dose significantly decreased from the start to the end of the treatment. Our model suggests the importance of a timely diagnosis, possibly before puberty, the beneficial effect of long-term treatment with rhGH, and the key-role of THt. Our prediction model has a very acceptable error compared to the majority of other published studies.

Keywords: Final height; Growth; Growth hormone deficiency; Growth hormone retesting; Insulin-like growth factor 1; LMG method; Prediction.