Value and limits of pharmacological and physiological tests to diagnose growth hormone (GH) deficiency and predict therapy response: first and second retesting during replacement therapy of patients defined as GH deficient

There is currently a debate about the use of pharmacological and physiological tests to define GH deficiency and predict response to GH therapy. In addition, a good response to therapy has also been described in subjects without GH deficiency. For further information, we reevaluated GH secretion during replacement therapy in a group of children defined as GH deficient and examined response to therapy in the subjects subdivided according to secretion. One hundred eighty four children (113 boys and 71 girls) initially diagnosed with GH deficiency by means of pharmacological (peak < 8 micrograms/L after arginine and L-dopa tests) and physiological tests (mean nocturnal concentration < or = 3.3 micrograms/L during sleep test) underwent the same tests 2.8 +/- 1.1 yr after start of GH therapy. Sixty eight patients were retested 1.5 +/- 0.4 yr after first retesting. At diagnosis 122 subjects had pathological pharmacological and physiological tests (group A), 30 subjects normal sleep test with pathological pharmacological tests (group B), and 32 subjects pathological sleep test with normal pharmacological tests (group C). At diagnosis 140 subjects were prepubertal and 44 pubertal. To evaluate response to therapy in relation to GH secretion at diagnosis and at both retestings, a number of auxological parameters were calculated during treatment. At first retesting, 107 subjects (58.1%) changed initial group of diagnosis, 34 of whom (18.5%) presented normal secretion in both pharmacological and physiological tests (group D). At second retesting, 31 of the 68 subjects reexamined (45.6%) changed first test results, and 33 (48.5%) reverted to the initial group of diagnosis; none of the 6 subjects of group D maintained normal secretion. Although the percentage of normalized subjects was higher in pubertal subjects (36.4%; P = 0.0003) than prepubertal subjects (8.9%), puberty did not prevent a reduction of secretion in some subjects. Response treatment during the first year of therapy was similar in the various groups. GH secretion seems to change in both prepubertal and pubertal children diagnosed with GH deficiency when pharmacological and physiological tests are repeated over time. Moreover, such tests may not represent a reliable tool for predicting response to treatment. GH secretion normalization at retesting may not necessarily represent the end of a transient secretory defect.