Evolution of growth hormone neurosecretory disturbance after cranial irradiation for childhood brain tumours: a prospective study.

To determine the aetiopathology of post-irradiation growth hormone (GH) deficiency, we performed a mixed longitudinal analysis of 56 24 h serum GH concentration profiles and 45 paired insulin-induced hypoglycaemia tests (ITT) in 35 prepubertal children, aged 1.5-11.8 years, with brain tumours in the posterior fossa (n = 25) or cerebral hemispheres (n = 10). Assessments were made before (n = 16), 1 year (n = 25) and 2 to 5 years (n = 15) after a cranial irradiation (DXR) dose of at least 30 Gy. Fourier transforms, occupancy percentage, first-order derivatives (FOD) and mean concentrations were determined from the GH profiles taken after neurosurgery but before radiotherapy (n = 16) and in three treatment groups: Group 1: neurosurgery only without DXR (n = 9); Group 2: > or = 30 Gy DXR only (n = 22); Group 3: > or = 30 Gy DXR with additional chemotherapy (n = 9). Results were compared with those from 26 short normally growing (SN) children. Compared with SN children, children with brain tumours had faster GH pulse periodicities (200 min vs 140 min) and attenuated peak GH responses to ITT (24.55 (19.50-30.20) vs 8.32 (4.57-15.14) mU/l) after neurosurgery, before radiotherapy. However, spontaneous GH peaks (19.05 (15.49-23.44) vs 14.13 (9.12-21.38) mU/l), 24 h mean GH (5.01 (4.37-5.62) vs 3.98 (2.63-5.89) mU/l) and FODs (1.43 (1.17-1.69) vs 1.22 (0.88-1.56) mU/l per min) were similar. The abnormalities present before radiotherapy persisted in group 1 children at 1 year when 24 h mean GH (2.45 (1.17-5.01) mU/l) and FODs (0.73 (0.26-1.20) mU/l per min) were additionally suppressed, although partial recovery was evident by 2 years. With time from radiotherapy, there was a progressive increase in GH pulse periodicity (Group 2: 200 min at 1 year, 240 min at > or = 2 years; Group 3: 140 min at 1 year, 280 min at > or = 2 years) and a decrease in 24 h mean GH (Group 2 vs Group 3 at > or = 2 years: 2.45 (1.70-3.47) vs 1.86 (1.32-2.69) mU/l) and FODs (Group 2 vs Group 3 at > or = 2 years; 0.56 (0.44-0.69) vs 0.44 (0.27-0.61) mU/l per min). Initial discrepancies between measures of spontaneous and stimulated (ITT) GH peaks were lost by 2 or more years (spontaneous vs ITT; Group 2: 7.76 (5.89-9.77) vs 3.80 (0.91-15.84) mU/l; Group 3: 6.03 (4.27-8.32) vs 3.80 (0.31-46.77) mU/l). After cranial irradiation, a number of changes evolved within the GH axis: faster GH pulse periodicities and discordance between physiological and pharmacological tests of GH secretion before irradiation gave way to a slow GH pulse periodicity, decreased GH pulse amplitude and rate of GH change (FOD) and, with time, eventual concordance between physiological and pharmacological measures. The evolution of these disturbances may well reflect differential pathology affecting hypothalamic GH-releasing hormone and somatostatin.