Background: The serum GH cut-off value for pharmacological tests of GH secretion (PhT GH) depends on the type of test and also on the method used for determining serum GH. Cut-off serum GH values as different as 5–10 ng/ml, have been reported, and have been validated biochemically. We have used the growth velocity (GV)-standard deviation score (SDS) during the first year of treatment with rhGH to validate these cut-offs on a biological basis. Methods: Fifty pre-pubertal patients with short stature (height ≤–2 SDS and GV ≤–1.2 SDS) were studied. GH deficiency (GHD) was diagnosed in 39 patients, on the basis of clinical and auxological parameters and on the serum concentration of IGF-1, and non-GHD in the other 11 patients. Two PhT GH (arginine and clonidine) were carried out in the 50 patients. Serum GH was determined by two different methods: one detecting most of serum GH isoforms, named Total GH (HGH Bio-Tech, MAIA Clone), and another one, only detecting the 22 kDa GH, named 22K GH (GH-22K IFMA, Wallac). Results: Basal data: all patients with GHD and with non-GHD had maximal serum GH response (MaxR) values below and above the cut-off, respectively, for the serum Total GH and 22K GH. The mean 22K GH/Total GH ratio was similar to previous publications. Post-rhGH treatment data: the two groups improved their height SDS during the first year of treatment, particularly patients with GHD. A receiver-operator curve was used to define the best threshold for post-treatment GV-SDS that separates GHD from non-GHD patients. This value was 1.91 GV-SDS. A negative correlation between first year treatment GV-SDS and pre-treatment serum GH MaxR was found for the two assays (p < 0.001). Then, the best cut-off GV-SDS, previously calculated with the receiver-operator curve (1.91 SDS) was used to interpolate the corresponding serum GH values, as determined by the two methods. For Total GH, the value was 10.8 ng/ml, and for 22K GH, it was 5.4 ng/ml. Conclusion: The cut-off values calculated by biological means to separate GHD from non-GHD were remarkably similar to those calculated biochemically (10.0 and 4.8 ng/ml, respectively) for Total and 22K GH. This is a biological validation for using different cut-off values, appropriate for each assay, to diagnose GHD.