At present, no established practical and reliable endocrine test exists for differentiating between constitutional delay of puberty (CDP) and hypogonadotropic hypogonadism (HH). The most discriminating results have been reported by measuring LH and testosterone (T) responses to gonadotropin-relasing hormone (GnRH) analogues. In this study, 23 prepubertal boys aged 14 to 16.5 years underwent a modified triptorelin (a GnRH analogue) stimulation test, and they were followed clinically for up 24 to months. Sixteen subjects developed spontaneous puberty during the follow-up period and thus were diagnosed with CDP, and the remaining seven were diagnosed with HH. Retrospective evaluation of their LH, FSH and T responses revealed significant differences without any overlaps in serum LH levels at 4h (CDP: 33.2 +/- 9.3 vs. HH: 3.3 +/- 2.6 mIU/ml, p < 0.0002) and in serum T levels at 24 h (CDP: 369.3 +/- 128.1 vs. HH: 61.4 +/- 22.6 ng/dl p < 0.0002). We conclude that CDP can be clearly differentiated from HH by the LH response at 4 h and/or T response at 24 h after a single-dose triptorelin administration.