Improved screening for growth hormone deficiency using logical analysis data.

BACKGROUND: The growth hormone (GH) Research Society proposed criteria based on auxological parameters (height, difference between target and actual heights, growth rate) to predict GH deficiency (GHD) followed by GH stimulation tests. These criteria are very sensitive but not specific enough. The objective of this study was to improve screening for GHD, so reducing the need for stimulation tests. MATERIAL/METHODS: The patients were divided into three groups: Group 1 (n=54) with GHD and pituitary stalk interruption syndrome; Group 2 (n=104) with a normal GH peak after pharmacological and sleep tests; Group 3 (n=19) with transient GHD. The GHD diagnosis was based on the parameters proposed by the GH Research Society to evaluate the GH secretion, first alone and then plus plasma insulin-like growth factor (IGF) I concentrations. Data were analysed using the logical analysis of data framework. RESULTS: Screening can be achieved using a simple graph based on IGF I and the growth rate (GR): (1) all patients below a given line (GR < or =-7.3 - 1.3xIGF (SDS)) had GHD; (2) all but two patients above another given line (GR >-4.5 + 6.4/(IGF + 4.5) (SDS)) did not have GHD; (3) in-between, patients in a "gray area" cannot be diagnosed using only growth rate and IGF I. Most of the transient GHD were in or near the "gray area". CONCLUSIONS: Short children can be screened for GHD using growth rate and IGF I. Magnetic resonance imaging can then be used to detect pituitary stalk interruption syndrome or a tumor.