Growth hormone testing for the diagnosis of growth hormone deficiency in childhood: a population register-based study.

Evaluation of GH secretion using pharmacological GH stimulation tests (GHST) remains a current practice, although the reliability of GHST has been questioned, and many pitfalls have been pointed out. We have analyzed all of the 6373 GH stimulation tests that led to the initiation of GH therapy in 3233 children treated in France from 1973-1989. Tests and GH measurements were performed by individual centers and collected by the Association France-Hypophyse. GH deficiency (GHD) was due to craniospinal irradiation (11%), was due to organic causes or associated with multiple deficiencies (22%), or was considered idiopathic (65%); 2% of the patients were considered non-GHD. Eleven different pharmacological tests were used, and 62 of the 66 theoretical pairs of tests were used at least once. The most frequent combination of tests (ornithine in one instance and insulin in another) was used in 12.7% of patients. The reliability of the GH peak measured by comparing the results of 2 tests in the same patient was poor, as measured by intraclass correlation coefficients below 0.8. Multivariate analysis identified several parameters positively or negatively associated with peak plasma GH: calendar year of initiation of treatment, etiology of GHD, height SD score, bone age SD score, puberty, weight SD score, genetic target height SD score, and the nature of the pharmacological agent used. We believe that several of these factors (weight SD score, genetic target height SD score, and nature of the agent) identify biases in the diagnosis of GHD. We conclude that GHST should be performed with a very limited number of agents, interpreted after the establishment of reference values in age-matched normal children, and associated with other clinical and biochemical parameters for establishing the diagnosis of GHD.