Evaluation and application of a highly sensitive assay for serum growth hormone (GH) in the study of adult GH deficiency.

Previous work from our laboratory addressing the diagnosis of GH deficiency in adults showed that RIA measurement of the 24-h integrated GH concentration (IGHC) was unable to discriminate between hypopituitary and age-, sex-, and body mass index-matched normal subjects because of the occurrence of undetectable levels (< 200 ng/L) within both groups. In contrast, full separation was achieved using stimulation by the insulin tolerance test (ITT). The data showed no significant relationship between IGHC and insulin-like growth factor-I (IGF-I) within either group. To determine whether limited sensitivity obscured diagnostic and physiological information, we assessed and modified a commercially available enzyme-linked immunosorbent assay (Elegance GH ELISA, Bioclone Australia) to achieve a high sensitivity (1 ng/L) and applied it to the study of IGHC and the relationship to IGF-I in a study group of 30 normal and 19 subjects with severe organic hypopituitarism. Using this assay, the IGHCs from all subjects were detectable and correlated significantly with detectable values obtained by RIA (n = 24; r = 0.80; P = 0.0001). Mean IGHC in normal subjects was significantly higher than that in hypopituitary subjects (852 +/- 131 vs. 97 +/- 28 ng/L), but the IGHCs from the two groups were not completely separate. Twenty-six percent of hypopituitary subjects had IGHC values within the normal range (111-3454 ng/L). IGHC decreased with age in normal subjects. Age stratification improved the separation, but an overlap remained in the young (< 50 yr old) and old (> 50 yr old) groups. Measurement of 12-h nocturnal IGHC levels improved the separation between hypopituitary and normal subjects in the young subjects only. IGHC was significantly related to IGF-I in hypopituitary (r = 0.59; P = 0.0084) and normal subjects (r = 0.55; P = 0.0017) and in the combined groups (r = 0.64; P = 0.0001). The data show that a sensitive ELISA reliably quantifies IGHC in normal and hypopituitary subjects. IGHCs in hypopituitary patients are lower, but not clearly separated from values in normal counterparts despite their having unequivocally impaired GH responses to ITT. We conclude that 1) IGHC in normal subjects can be reliably defined by sensitive ELISAs; 2) the diagnostic utility of the IGHC does not match the reliability or simplicity of an ITT, and 3) GH is a significant regulator of IGF-I in both normal and reduced states of GH secretion.