It has been proved that to analyze the factors that determine responsiveness to rhGH and to develop growth prediction models can help doctors to individualize the treatment and maximize the effect.
Objectives: To set up and validate the predictive models of growth responses to rhGH treatment in the first year in prepubertal short stature children with various GH secretary statuses.
Methods: Growth responses to rhGH treatment in the first year, height velocities (HV) and increases in height SDS (DeltaHtSDS), in 62 prepubertal short stature children with various GH secretary statuses were analyzed retrospectively. There were 27 patients with complete growth hormone deficiency (cGHD), 23 with partial GHD (pGHD) and 12 with idiopathic short stature (ISS) in the model group. According to the peak GH value in GH provocative test, the group of pGHD was divided into pGHD-1 (5 - 6.9 microg/L, 12 patients) and pGHD-2 (7 - 9.9 microg/L, 11 patients). All the cases in model group were used for setting up Model-total and the cases of growth hormone deficiency for Model-GHD. Predictive models, including Model-GHD and Model-total, to HV and DeltaHtSDS were set up by the way of multiple regression analysis, based on the results of simple correlation analysis. Other 14 children were included according to the same criteria with the model group, the validation group. The validation group was analyzed prospectively. The actual growth responses were compared with the predicted values calculated by different models so that the predictive models could be validated.
Results: The simple correlation analysis showed that HV and DeltaHtSDS in the first year were negatively correlated with the same group factors at baseline: chronological age, bone age, height SDS, differences between the height SDS and the target height SDS, peak value in GH provocative test and IGF-1SDS. All the 4 predictive models were found to be significant at a level of P < 0.05, R(2) ranged from 0.244 to 0.519. The two models predicted HV and Model-GHD for DeltaHtSDS were proved to be validated. The observed and predicted responses positively and significantly correlated with each other, r value ranged from 0.753 to 0.996. And there was no significant difference between them when tested by paired t test.
Conclusions: The availability of the predictive model will help to individualize the growth hormone treatment in prepubertal short stature children with various growth hormone secretary status.