Impact of different cut-off limits of peak GH after GHRH-arginine stimulatory test, single IGF1 measurement, or their combination in identifying adult patients with GH deficiency

Objective: To evaluate the impact of different peak GH cut-off limits after GHRH-Arg test, IGF1 measurement, or their combination in identifying patients with GH deficit (GHD). Design and patients: Totally, 894 normal subjects (used for determining IGF1 normative limits) and 302 patients with suspected GHD were included. Different peak GH cut-off limits (used by European (depending on body mass index (BMI)) or North American (4.1 μg/l) Endocrine Societies, by HypoCCs (2.5 μg/l), or with 95% specificity (based on BMI), Method 1. 2, 3, or 4 respectively) and IGF1 were considered. Methods: Peak GH after GHRH-Arg and IGF1. Results: Different peak GH cut-off limits recognized different proportions of GHD (range. 24.8―62.9%). Methods 1 and 2 with high sensitivity recognized a higher proportion (95.5 and 92.5% respectively) of GHD among patients with three (T) pituitary hormone deficits (HD), whereas Method 4 (with high specificity) identified 96.7% normal subjects among those without pituitary HD; on the contrary, Method 4 identified only 75% GHD among patients with THD, whereas Method 1 recognized a high proportion (40%) of GHD among subjects without HD. Of the total patients, 82% with THD and 84.5% without HD were recognized as GHD or normal respectively by IGF1. Among the remaining patients with THD and normal IGF1. 75% was recognized as GHD by Method 1: among patients without HD and abnormal IGF1, 87.5% was identified as normal by Method 4. Overall, combination of IGF1 and Method 1 or Method 4 identified 9 5.5% GHD among patients with THD and 98.1% normal subjects among those without HD. Conclusions: Single peak GH cut-offs have limits to sharply differentiate GHD from normal subjects; IGF1 may be used for selecting patients to be submitted to the GHRH-Arg test; the peak GH cut-off limits to be used for identifying healthy or diseased patients depend mainly on the clinical context.