Endocrine Abstracts

Introduction

Growth hormone deficiency (GHD) is a rare cause of statural retardation in children. The diagnosis is evoked in front of a bundle of anamnestic, clinical and radiological arguments. Treatment consists of substitution by recombinant GH with the aim of restoring normal size. There is interindividual variability in the response to treatment. We therefore proposed to study the response to treatment as well as the predictive factors of this response in the Tunisian population.

Methods

This is a retrospective study including 34 patients who were followed at the National Institute of Nutrition in Tunis for congenital GH deficiency and who completed their treatment.

Results

The assessment of statural response was based on total statural gain at the end of treatment and final height, thus investigating factors predictive of good response to treatment for each of these two parameters. The mean total statural gain was 31.62 ±19 cm with extremes ranging from 3 to 75 cm corresponding to a mean statural gain of 1.17 ±0.8DS. Age of discovery was negatively correlated with total statural gain (r = –0.64; P <0.001).Total statural gain was 28.6±19.6 cm for boys versus 36.66±19.96 cm for girls with no significant difference (P = 0.150). Height, BMI at baseline (P = 0.04; P <0.001 respectively) were correlated with total statural gain, and baseline IGF1 (SD) and thyroid insufficiency were correlated with statural gain (P = 0.025; P = 0.017 respectively). Bone age at baseline was negatively correlated with statural gain(cm)(r = –0.67, P <0.001). For the standard deviation (SD) assessment of statural gain, only general disease history was correlated with this parameter(P = 0.005). Final height was 158.3±7.09 cm in boys and 144.79±14 cm in girls with a significant difference (P = 0.001). The parameters correlated to the final height were: general disease history(P = 0.002), combined deficit (0.015), existence of corticotropic insufficiency(P = 0.027), thyroid insufficiency (P = 0.048), number of deficits greater than 1(P = 0.027) and pathological aspect on hypothalamic-pituitary MRI (P = 0.01).

Discussion/Conclusion

The response to treatment in our population was close to the results of the literature series. Factors predictive of a good response to treatment vary from one individual to another and from one population to another. These factors can be integrated into a true predictive model allowing for an adaptation of treatment.