Long-term response to recombinant human growth hormone treatment: a new predictive mathematical method

• Published in: Journal of endocrinological investigation Vol. 41; no. 7; pp. 839 - 848
• Main Authors: Migliaretti, G., Ditaranto, S., Guiot, C., Vannelli, S., Matarazzo, P., Cappello, N., Stura, I., Cavallo, F.
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.07.2018; Springer Nature B.V
• Subjects: Body height; Children; Endocrinology; Growth hormones; Mathematical models; Medicine; Medicine & Public Health; Metabolic Diseases; Original Article; Patients; Sensory neurons; GH deficit; Gompertzian function; Growth hormone therapy; Particle swarm optimization (PSO) model; Height gain; Final height prediction; Radial basis functions (RBF)
• ISSN: 1720-8386; 0391-4097; 1720-8386
• DOI: 10.1007/s40618-017-0816-6

Introduction Recombinant GH has been offered to GH-deficient (GHD) subjects for more than 30 years, in order to improve height and growth velocity in children and to enhance metabolic effects in adults. Aim The aim of our work is to describe the long-term effect of rhGH treatment in GHD pediatric patients, suggesting a growth prediction model. Material and methods A homogeneous database is defined for diagnosis and treatment modalities, based on GHD patients afferent to Hospital Regina Margherita in Turin (Italy). In this study, 232 GHD patients are selected (204 idiopathic GHD and 28 organic GHD). Each measure is shown in terms of mean with relative standard deviations (SD) and 95% confidence interval (95% CI). To estimate the final height of each patient on the basis of few measures, a mathematical growth prediction model [based on Gompertzian function and a mixed method based on the radial basis functions (RBFs) and the particle swarm optimization (PSO) models] was performed. Results The results seem to highlight the benefits of an early start of treatment, further confirming what is suggested by the literature. Generally, the RBF–PSO method shows a good reliability in the prediction of the final height. Indeed, RMSE is always lower than 4, i.e., in average the forecast will differ at most of 4 cm to the real value. Conclusions In conclusion, the large and accurate database of Italian GHD patients allowed us to assess the rhGH treatment efficacy and compare the results with those obtained in other Countries. Moreover, we proposed and validated a new mathematical model forecasting the expected final height after therapy which was validated on our cohort.