Differences in reproducibility and peak growth hormone responses to repeated testing with various stimulators in healthy adults.

In healthy adults, GH responses to provocative testing are variable between subjects. Information on the intra-subject variability is limited, despite the importance attached to GH stimulation tests in the diagnosis of GH deficiency. We have investigated and compared the variability of different GH stimulation tests in a group of healthy control subjects. In 16 healthy non-obese adults, two insulin tolerance tests (ITT) (0.15 IU/kg body weight i.v. and a fall in blood glucose < or = 2.2 mmol/l) two GHRH tests (1 microgram/kg body weight i.v.), and two clonidine (CLO) (300 micrograms p.o.) + GHRH (60 min later) tests were performed in the morning after an overnight fast. A pyridostigmine (PD) (120 mg p.o. 60 min before GHRH) + GHRH test was performed twice in an extended group of 31 healthy adult subjects. A wide range of GH responses was observed. Both during the ITT and the GHRH test, low values in the range generally recognized to reflect impairment of GH secretory status were encountered. The median (range) peak GH responses in tests 1 and 2 were: (a) ITT: 14.4 micrograms/l (4.1-71.1) and 14.0 micrograms/l (0.09-69.5), (b) GHRH test: 21.7 micrograms/l (0.71-56.2) and 18.4 micrograms/l (1.6-55.1); (c) CLO + GHRH test: 57.4 micrograms/l (22.9-209) and 65.8 micrograms/l (12.2-206); (d) PD + GHRH test: 36.5 micrograms/l (9.1-125) and 44.6 micrograms/l (6.3-101). The coefficients of variation (CV) were: 58% (ITT), 45% (GHRH), 46% (CLO + GHRH) and 26% (PD + GHRH). The peak GH responses were significantly different in all tests (CLO + GHRH > PD + GHRH > GHRH > ITT). In the individual subject, there was no systematic correlation between the peak GH responses in the different stimulation tests. In conclusion, we found that the stimulated GH responses were highly variable in all tests, and that the peak GH responses differed. Test results in patients should be evaluated against test-specific reference values, and caution is justified in the interpretation of low responses in a single test.