To assess the performance of growth hormone stimulation tests (GHSTs) in the evaluation of short stature. It was a single-centre retrospective study carried out in children evaluated for short stature between January 2005 to March 2020. The clonidine stimulation test (CST) and glucagon stimulation test (GST) were used to assess growth hormone (GH) reserve (GST was performed only when peak GH levels were between 5 to ≤10 ng/mL on CST). A GH level of <5 ng/mL on CST or ≤10 ng/ml on both was used to corroborate GH deficiency. A total of 556 children were eligible for this study. The mean (SD) age was 12.9 (3.5) years, and 66.3% were male. The peak GH level [median (IQR)] was 5.50 ng/ml (1.90 - 7.50) on CST (at 60 minutes) and 7.45 ng/ml (2.15 - 10.77) on GST (at 120 minutes). On restricting sampling to two time points, the false positive rate was 13.6% on CST (60, 90 minutes) and 11.5% on GST (120, 150 minutes). Similarly, restricting to three time points was associated with a false positive rate of 8.5% on CST (60, 90, 120 minutes) and 3.8% on GST (90, 120, 150 minutes). Using the treating clinician-determined diagnosis of GHD as a reference standard, the optimal cut-off of peak GH on CST was 7.79 ng/ml (sensitivity: 83.8%; specificity: 89.4%). Restricting the GH sampling to fewer time points is associated with an increase in the false positivity rate (FPR).
Read full abstract