Growth Hormone Provocation Tests and the Response to Treatment
© 2005 Prime Health Consultants, Inc.
Cole and colleagues explored the relationships among a variety of anthropometric, demographic, and clinical variables in an attempt to identify those that contribute significantly to the variance in growth response to exogenous growth hormone (GH). The authors utilized the UK Pharmacia KIGS database; 337 children were identified with the diagnosis of isolated GH deficiency and met the inclusion criteria (age <10 years, Tanner pubic hair stage 1, GH provocation test result <20mU/L, follow-up at least 1 year). Children with cranial tumors or history of radiation therapy were excluded. The primary outcome variable was linear growth in the first year of therapy, while the secondary outcome variable was growth response during the second year. Covariates included height and weight SDS at the initiation of treatment, age, sex, birthweight SDS, gestation, midparental height SDS, GH dose, injection frequency, growth velocity in the year prior to treatment, and maximal GH provocation test result. Multiple regression analysis was used to identify factors predicting growth response. Ten factors predicted 42% of the variance in the first year’s growth response: time between visits (years), age at baseline (years), age2 at baseline (year2), height SDS at baseline, weight SDS at baseline, midparental height SDS, birth weight SDS, maximum GH peak, GH dose, and number of GH injections/week. The maximum GH response to provocative stimulation was the most predictive factor, accounting for 9.9% of the variance. Only 3 factors predicted the growth response in the second year: maximum GH response to stimulation, GH dose, and injection rate; these 3 factors explained 6.8% of the variance. However, when the first year’s response to GH and baseline height SDS were added to the model, 21.6% of the variance was accounted for and the results of the provocative test no longer contributed significantly. The authors discussed the importance of the GH provocative test results as a predictor of the first year’s growth response to exogenous GH and the correlation between the first year’s response and the growth rate during the second year (r=0.45).They also discussed previously published information regarding the high sensitivity but low specificity of GH provocative testing in detecting GH deficiency. They encouraged more standardization of GH provocative testing procedures and of serum GH analyses as one way of reducing the different findings presented in various forums as to the relevance of such testing to the diagnosis of GH deficiency. They conclude that the GH response to 2 provocative tests (<20mU/l) should be used to decide whether or not a trial of GH should be initiated, but that the result of the first year’s response should be the criterion for determining whether or not to continue such therapy.
Editor’s Comment: GH stimulation tests are used by most pediatric endocrinologists in the United States to determine, at least initially, which short children will receive exogenous GH. Indeed, the results of such testing remain the sole criteria used by most third-party payors in deciding whether or not GH therapy will be a covered prescription drug benefit. Only recently have criteria been loosened for some very specific clinical conditions, ie, chronic renal failure, Turner syndrome, children with intrauterine growth retardation, etc. Most short children do not fit into these specific categories for which stimulation testing is not a requirement, and the decision as to whether or not to prescribe GH remains problematic. The results presented in the paper by Cole and colleagues seem to justify using the “gold standard” as an important criterion in making that decision. It should be noted, however, that all of the children studied had been classified in the registry as GH deficient and thus, children with other causes of short stature may have been excluded from the analyses. That such children might benefit from GH therapy, ie, increase their predicted adult height regardless of their response to provocative testing, is not explored in this paper. Perhaps the rigor of standardized testing and serum GH analyses will increase our ability to make these decisions, but such testing should never be the sole criteria for medication trials.
William L. Clarke, MD