Abstract
INTRODUCTION: In childhood, growth hormone (GH) deficiency (GHD) diagnosis is based on auxological assessment and biochemical provocative tests, whose reliability remains disputed. Recently, several papers have been published on standardising the duration of some tests. The aim of our study was to analyse the possible length reduction of the L-DOPA provocative test. METHODS: We retrospectively investigated the response of GH to L-DOPA in 256 children, analysing 267 tests (some patients were retested over time for the persistence of severe auxopathy). We studied the same data considering GH peak threshold both at 8 ng/mL (Italian GHD cut-off) and at 10 ng/mL (international cut-off). Based on stimulation tests, patients were divided into two groups: GHD and no-GHD short children. We described the results in the whole population and then clustering for gender and pubertal stage. We termed as index the test stopped at 90 min. RESULTS: The GH peak after L-DOPA mostly occurred at 60 min. The sensitivity of the index test was the highest, while the specificity was slightly higher using the 8 ng/mL threshold (specificity = 0.68; 95% CI 0.60-0.76) then using the 10 ng/mL threshold (specificity = 0.56; 95% CI 0.47-0.65) at 90 min. The two ROC curves showed moderate performance of the test at 90 min. While the negative predictive value was 100% in both tests, the positive predictive value was slightly better with 10 ng/mL cut-off. Considering the two groups established by GHD definition and placing a GH threshold at 10 ng/mL, stopping L-DOPA test time at 90 min would have changed the test result and subsequentially patient's classification in 3/267 of the analysed tests (1.1%), while with the Italian GH threshold value at 8 ng/mL in 7/267 of the tests (2.6%). CONCLUSIONS: Our research shows that omitting 120-min time reduces L-DOPA test specificity, especially with GHD cut-off at 10 ng/mL.