Usefulness of IGF-I assay for the diagnosis of GH deficiency in adults.

IGF-I is the best marker of GH secretory status but it also depends on the nutritional status and peripheral hormones such as insulin, glucocorticoids, thyroid hormones and gonadal steroids. Though monitoring IGF-I levels is the best way for evaluating appropriate GH replacement, the usefulness of IGF-I assay in the diagnosis of adult GH deficiency (GHD) is still matter of debate. To clarify this point in a large population of GHD adults (no. = 135, 61 women and 74 men; age, mean +/- SE: 43.8 +/- 1.4 yr, range 20-80 yr) we studied IGF-I levels, their reproducibility and association to peak GH response to GHRH + arginine (GHRH + ARG) test and insulin tolerance test (ITT). The results in GHD were compared with those in a large population of normal subjects (no. = 336, 233 women and 103 men, aged 20-80 yr). Mean IGF-I levels in GHD (77.8 +/- 4.9 micrograms/l) were clearly lower (p < 0.001) than those in normal subjects (170.2 +/- 4.7 micrograms/l). In Childhood Onset GHD (CO-GHD; no. = 40; age, mean +/- SE: 27.8 +/- 1.5 yr) IGF-I levels were lower than those in Adult Onset GHD (AO-GHD; no. = 95, age, mean +/- SE: 50.7 +/- 1.4 yr) (56.6 +/- 9.7 vs 87.1 +/- 5.4 micrograms/l, p < 0.0003). In both GHD and normal subjects IGF-I levels showed good, reproducibility (r = 0.92, p < 0.00001 and r = 0.62, p < 0.00001, respectively). In GHD, but not in normal subjects, IGF-I levels were positively associated to peak GH responses to GHRH + ARG (r = 0.57, p < 0.00001); on the other hand, the GH peak after ITT was not associated to IGF-I in GHD. In normal subjects, but not in GHD, IGF-I levels were negatively associated to age (r = -0.60, p < 0.00001). Considering individual IGF-I levels there was a clear overlap between GHD and normal subjects. However, this overlap was strongly dependent on age. In fact, in the third and fourth decade of life 83.6% of GHD had IGF-I levels below the 3rd centile of normal values; on the other hand, in the fifth-sixth decade and in ageing 47% and only 12% of GHD, respectively, had IGF-I levels low for age. In conclusion, our results demonstrate that IGF-I levels represent a reproducible marker of GH status and are reduced more in CO-GHD than in AO-GHD adults. An overlap exists between GHD and normal subjects, however this is small up to the 4th decade of life. Thus, though normal IGF-I levels do not rule out the existence of GHD, up to 40 yr low IGF-I levels strongly point to GHD if malnutrition and liver disease have been ruled out.