Height velocity and IGF-I assessment in the diagnosis of childhood onset GH insufficiency: do we still need a second GH stimulation test?

Abstract

The diagnosis of GH insufficiency (GHI) in childhood is not straightforward. Our aim was to test the sensitivity and specificity of height velocity (HV), IGF-I, IGFBP-3 and GH stimulation tests alone or in combination in the diagnosis of GHI. A retrospective review of patients with GHI and idiopathic short stature (ISS) diagnosed in our centre and followed up to the completion of linear growth. Thirty-three GHI children and 56 children with ISS were evaluated. GHI diagnosis was based on fulfilment of anthropometric, endocrine and neuroradiological criteria: stature < or = -2 z-score, delayed bone age (at least 1 year), GH peak response to at least two different provocative tests < 10 micro g/l (20 mU/l), brain MRI positive for hypothalamus-pituitary abnormalities, catch-up growth during the first year of GH replacement therapy > or = 75th centile, peak GH response to a third provocative test after growth completion < 10 micro g/l (20 mU/l). Children with anthropometry resembling that of GHI but with peak GH responses > 10 micro g/l (20 mU/l) were diagnosed as ISS. All subjects underwent standard anthropometry. GH secretory status was assessed by clonidine, arginine and GHRH plus arginine stimulation tests. IGF-I and IGFBP-3 circulating levels were measured by immunoradiometric assay (IRMA). The following cut-off values were chosen to discriminate between GHI and nonGHI short children: HV < 25th centile over the 6-12 months prior to the initiation of GH therapy, peak GH responses < 10 or < 7 micro g/l (< 20 or < 14 mU/l) and IGF-I and IGFBP-3-values < -1.9 z-score. Sensitivity (true positive ratio) and specificity (true negative ratio) were evaluated. Taking 10 micro g/l (20 mU/l) as the cut-off value, sensitivity was 100% and specificity 57% for GH provocative tests, whereas taking 7 as the cut-off value, sensitivity was 66% and specificity rose to 78%. Sensitivity was 73% for IGF-I and 30% for IGFBP-3 measurement, whilst specificity was 95% for IGF-I and 98% for IGFBP-3 evaluation. HV assessment revealed a sensitivity of 82% and a specificity of 43%. When HV and IGF-I evaluations were used in combination, sensitivity reached 95% and specificity 96%. When both HV and IGF-I are normal (26% of our subjects) GHI may be ruled out, whereas when both the indices are subnormal (23%) GHI is so highly likely that the child may undergo only one GH provocative test and brain MRI and, thereafter, may begin GH therapy without any further test. In case of discrepancy, when IGF-I is normal and HV < 25th centile (44% of children), due to the relatively low sensitivity of IGF-I assessment and low specificity of HV, the patient should undergo GH tests and brain MRI. Finally, in the rare case of HV > 25th centile and subnormal IGF-I-values (7%), due to the high specificity of IGF-I measurement, the child should undergo one provocative test and brain MRI for the high suspicion of GHI. Our results suggest that a simple assessment of HV and basal IGF-I may exclude or, in association with only one stimulation test, confirm the diagnosis of GH insufficiency in more than half of patients with short stature.