Abstract
Objective. To determine the utility of growth hormone (GH) measurement with the insulin tolerance test (ITT), and to carry out a cost-effective analysis of the diagnosis of GH deficiency. Material and methods. Ninety-nine patients clinically suspected of GH deficiency were evaluated over a period of 14 months (January 2005 to April 2006). Post-insulin samples of GH and blood glucose (BG) samples were drawn at six different time-points. Serum GH levels of <10 µg L (prepubertal) and <6.1 µg L (adolescents) were taken as cut-off for the normal response. Results. Ninety-nine ITTs were carried out during the study period, and GH levels were found to be deficient in 47 subjects. Specificities at different time-points were 0 %, 54 %, 77 %, 62 %, 39 % and 23 % for 0, 30, 45, 60, 90 and 120 min, respectively, in the prepubertal group, and 5 %, 41 %, 80 %, 87 %, 77 % and 46 % at the same time-points for the adolescent group. Accuracy was highest at 45 and 60 min in both the prepubertal and adolescent groups. The receiver operating characteristic curve showed that the highest area under the curve was found in samples drawn at 45, 60 and 90 min in both the prepubertal and adolescent groups. Conclusion. Our data suggest that 0, 45, 60 and 90 min samples are sufficient for diagnosing GH deficiency, which could lead to potential cost reductions of up to 29.8 %.