Future directions: Growth prediction models

Background. One strategy for optimizing growth hormone (GH) treatment is to develop mathematical models based on clinical data from the large numbers of subjects in the KIGS (Pfizer International Growth Study Database) and to compare the observed versus predicted growth responses in subjects with short stature secondary to idiopathic GH deficiency (GHD), Turner syndrome, small birth size and idiopathic causes of short stature. Mathematical Models: Variables employed in derived regression equations include those related to birth status, genetic potential, current clinical status, laboratory data and GH treatment schedule. These models can provide an accurate estimate of potential growth on GH therapy and the tools to optimize and individualize GH therapy to obtain maximum height with the least risk and the lowest cost. Current prediction models explain around 58% of GH responsiveness in subjects with GHD, 46% in subjects with Turner syndrome and 52% in those born small for gestational age. Future Considerations: The predictive value of these models could be improved by the inclusion of extended anthropometric variables and biological parameters such as insulin-like growth factor I levels. However, recent reports that common polymorphisms of the GH receptor (GHR) gene may be associated with variations in response to GH suggest that, in the future, molecular genetics may provide an additional tool for refining growth prediction models. This possibility is being explored in a pilot study examining the effects of candidate genes in a targeted KIGS population to determine whether the GHR gene or other gene variants contribute to growth response over the first year of GH treatment. Copyright (c) 2007 S. Karger AG, Basel.