Reduced bone mineral density (BMD) has been reported in patients with isolated GH deficiency (GHD) or with multiple pituitary hormone deficiencies (MPHD). To investigate whether the severity of GHD was correlated with the degree of bone mass and turnover impairment, we evaluated BMD at the lumbar spine and femoral neck; circulating insulin-like growth factor I (IGF-I), IGF-binding protein-3 (IGFBP-3), and osteocalcin levels, and urinary cross-linked N-telopeptides of type I collagen (Ntx) levels in 101 adult hypopituitary patients and 35 sex- and age-matched healthy subjects. On the basis of the GH response to arginine plus GHRH (ARG+/-GHRH), patients were subdivided into 4 groups: group 1 included 41 patients with a GH peak below 3 microg/L (0.9 +/- 0.08 microg/L), defined as very severe GHD; group 2 included 25 patients with a GH peak between 3.1-9 microg/L (4.7 +/- 0.4 microg/L), defined as severe GHD; group 3 included 18 patients with a GH peak between 9.1-16.5 microg/L (11.0 +/- 0.3 microg/L), defined as partial GHD; and group 4 included 17 patients with a GH peak above 16.5 microg/L (28.3 +/- 4.3 microg/L), defined as non-GHD. In all 35 controls (group 5), the GH response after ARG+/-GHRH was above 16.5 microg/L (40.7 +/- 2.2 microg/L). In patients in group 1, circulating IGF-I (P < 0.001), IGFBP-3 (P < 0.05), osteocalcin (P < 0.001), and urinary Ntx levels (P < 0.001) were lower than those in group 3-5, which were not different from each other; the t score at the lumbar spine (-1.99 +/- 0.2) and that at the femoral neck (-1.86 +/- 0.3) were lower than those in groups 3 (-0.5 +/- 0.7, P < 0.01 and -0.3 +/- 0.7, P < 0.01, respectively), 4 (-0.5 +/- 0.2, P < 0.01 and -0.3 +/- 0.7, P < 0.01, respectively), and 5 (-0.5 +/- 0.2, P < 0.001 and 0.0 +/- 0.02, P < 0.001, respectively). In patients in group 2, circulating IGF-I and IGFBP-3 levels were not different from those in group 1, whereas the t scores at the lumbar spine (-1.22 +/- 0.3) and femoral neck (-0.9 +/- 0.3) were significantly higher and lower, respectively, than those in groups 1 and 5 (P < 0.05) but not those in groups 3 and 4, and serum osteocalcin and urinary Ntx levels were significant higher than those in group 1 and lower than those in groups 3-5 (P < 0.001). To evaluate the effect of isolated GHD vs. MPHD, patients were subdivided according to the number of their hormonal deficits, such as panhypopituitarism with (10 patients) or without (31 patients) diabetes insipidus, GHD with 1 or more additional pituitary deficit(s) (36 patients), isolated GHD (7 patients), 1-2 pituitary hormone deficit(s) without GHD (10 patients), and normal anterior pituitary function (7 patients). The t score at the lumbar spine and femoral neck and the biochemical parameters of bone turnover were not significantly different among the different subgroups with similar GH secretions. A significant correlation was found between the GH peak after ARG+GHRH and IGF-I, osteocalcin, urinary Ntx levels, and the t score at the lumbar spine, but not that at the femoral neck level. A significant correlation was also found between plasma IGF-I levels and the t score at the lumbar spine and femoral neck, serum osteocalcin, and urinary Ntx. Multiple correlation analysis revealed that the t score at the lumbar spine, but not that at the femoral neck, was more strongly predicted by plasma IGF-I levels (t = 3.376; P < 0.005) than by the GH peak after ARG+GHRH (t = -0.968; P = 0.338). In conclusion, a significant reduction of BMD associated with abnormalities of bone turnover parameters was found only in patients with very severe or severe GHD, whereas normal BMD values were found in non-GHD hypopituitary patients. These abnormalities were consistently present in all patients with GHD regardless of the presence of additional hormone deficits, suggesting that GHD plays a central role in the development of osteopenia in hypopituitary patients.