Abstract
INTRODUCTION: The proportion of patients with low GH response to provocative tests increases with the number of other pituitary hormone deficiencies, reason why in panhypopituitary patients GH stimulation tests may be unnecessary to diagnose GH deficiency (GHD) PURPOSE: To re-evaluate the diagnostic cut-offs of GH response to GHRH + arginine (ARG) test related to BMI, considering the patients' pituitary function as the gold standard for the diagnosis of GHD. METHODS: The GH responses to GHRH + ARG were studied in 358 patients with history of hypothalamic-pituitary disease. GHD was defined by the presence of at least 3 other pituitary deficits (n = 223), while a preserved somatotropic function was defined by the lack of other pituitary deficits and an IGF-I SDS ≥ 0 (n = 135). The cut-off with the best sensitivity (SE) and specificity (SP), was identified for each BMI category using the ROC curve analysis. To avoid over-diagnosis of GHD we subsequently searched for the cut-offs with a SP ≥ 95%. RESULTS: The best GH cut-off was 8.0 μg/l (SE 95%, SP 100%) in lean, 7.0 μg/l (SE 97.3%, SP 82.8%) in overweight, and 2.8 μg/l (SE 84.3%, SP 91.7%) in obese subjects. The cut-off with a SP ≥ 95% was 2.6 μg/l (SE 68.5%, SP 96.6%) in overweight and 1.75 μg/l (SE 70.0%, SP 97.2%) in obese subjects. CONCLUSIONS: This is the first study that evaluates the diagnostic cut-offs of GH response to GHRH + ARG related to BMI using a clinical definition of GHD as gold standard. Our results suggest that with this new approach, the GHRH + ARG cut-offs should be revised to avoid GHD over-diagnosis.