Prediction of the outcome of growth hormone provocative testing in short children by measurement of serum levels of insulin-like growth factor I and insulin-like growth factor binding protein 3.

Serum levels of insulin-like growth factor I (IGF-I) and insulin-like growth factor binding protein 3 (IGFBP-3) reflect the secretion of endogenous growth hormone (GH) in healthy children and exhibit little diurnal variation, which makes them potential candidates for screening of GH deficiency (GHD). We evaluated serum IGF-I and IGFBP-3 levels in relation to the outcome of GH provocative testing in 203 children and adolescents (111 boys and 92 girls) in whom GHD was suspected. A total of 1030 children served as control subjects. In children less than 10 years of age, IGF-I levels were below the cutoff limit in 8 of 15 children with GHD (sensitivity 53.3%) and above the cutoff limit in 47 of 48 children with a normal GH response (specificity 97.9%). Similarly, IGFBP-3 levels were below the cutoff limit in 9 of 15 children with GHD (sensitivity 60%) and above the cutoff limit in 47 of 48 children with a normal GH response (specificity 97.9%). Consequently the predictive value of a positive test result in prepubertal children was 88.8% for IGF-I and 90% for IGFBP-3. In children and adolescents between 10 and 20 years of age, IGF-I levels were below the cutoff limit in 34 of 46 children with GHD (sensitivity 73.9%) and above the cutoff limit in 63 of 94 children with normal GH response (specificity 67%). IGFBP-3 levels were below the cutoff limit in 26 of 46 children with GHD (sensitivity 56.5%) and above the cutoff limit in 74 of 94 children with a normal GH response (specificity 78.7%). Accordingly the positive predictive value and in 10- to 20-year-old children was 52.3% for IGF-I and 56.5% for IGFBP-3. Combination use of IGF-I and IGFBP-3 gave additional diagnostic information. We conclude that a subnormal IGF-I level, and especially a subnormal IGFBP-3 level, are highly predictive of a subnormal GH response to a provocative test in prepubertal children in whom GHD is suspected. On the other hand, a normal IGF-I or IGFBP-3 level does not exclude GHD. The predictive value of IGF-I and IGFBP-3 in pubertal children is diminished in comparison with that in prepubertal children. We believe that IGF-I and IGFBP-3 are valuable tools in the evaluation of childhood GHD.