VARIABILITY OF RESPONSES IN REPEATED GROW TH-HORMONE RESERVE TESTS INCHILDREN OF NORMAL STATURE

The presence of false negatives in the evaluation of patients with sho rt stature and the variability of 24 hour growth hormone (GH) physiolo gical studies in the normal population are well known. Therefore the r eproducibility of two widely used pharmacological tests was studied in normal children. Forty prepuberal children were evaluated (34 boys an d 6 girls), with chronological ages ranging from 2 years 11 months to 12 years 11 months (mean: 9 years 1 month), bone ages from 3 years 2 m onths to 12 years 6 months (mean: 8 years 4 months) and with normal st ature and growth velocity (SDS > -2) and normal body mass index (BMI < 25). Clonidine test was performed (100 mu g/m(2) body surface) measur ing GH (ng/ml) 0,60 and 90 min in 20 patients (Group I). Exercise-Prop ranolol test was performed (0,5 mg/kg weight) with basal and post-exer cise GH measurements in 20 patients (Group II). The tests were repeate d at one week intervals and each child was his own control. Group I sh owed (mean = SD): 1st test: B = 1.78 +/- 1.59, Max Resp = 13.16 +/- 8. 34; 2nd test: B = 1.17 +/- 0.51, Max Resp 15.12 +/- 8.09. Group II sho wed (mean +/- SD): 1st tes: B = 1.38 +/- 0.58, Max Resp 16.97 +/- 9.69 ; 2nd test: B = 1.54 +/- 1.16, Max Resp = 13.49 +/- 7.81. Wilcoxon' s test did not show significant differences when comparing B vs B and Ma x Resp vs Max Resp in both groups. The analysis of individual tests sh owed 39% of dissimilar responses in Group I and 18% in Group II when a value of GH greater than or equal to 10 ng/ml was considered as a pos itive response, When this value was set at 7 ng/ml, 22% of dissimilar responses were found in Group I and 24% in Group II. No correlation wa s found between the responses of the two tests for both groups, neithe r for the value of 7 ng/ml nor for that of 10 ng/ml (Mc Nemar' s test) . It can be concluded that: 1) no significant differences were found b etween each test and its repetition; 2) in the evaluated tests it was impossible to predict the result of their repetition in terms of posit ive or negative responses; 3) dissimilar individual responses, dependi ng on where the ''normality limit'' is set, indicate the necessity of establishing such a limit more accurately and on the basis of adequate population studies.