Objective: There is a large inter-individual variation in the growth hormone (GH) response to exercise, but within-individual variation is unknown. The purpose of this study was to determine the reproducibility of the GH response to a single 30s sprint on a cycle ergometer. Design: Eleven non-obese male volunteers completed two trials separated by at least seven days during which they completed a single all-out 30s sprint on a cycle ergometer followed by 60 min of rest. Blood samples were taken at rest and at regular intervals post-exercise. Results: No differences were found in mean power output during the sprint, or in peak blood lactate concentrations or lowest measured pH following the sprints. Re-test correlation was significant for both peak GH concentrations (r = 0.97, P < 0.05) and GH AUC (r = 0.97, P < 0.05). Within-subject error (change in mean typical error) of the peak GH concentrations and GH area under the curve (AUC) was 4.3 +/- 3.4 mug l(-1) and 2.9 +/- 54.3 min mug l(-1), respectively. Within-subject percentage error (percentage change in mean +/- typical percentage error) for peak GH concentration and GH AUC was 33.5 +/- 26.7% and 1.1 +/- 20.0%, respectively. Conclusion: Growth hormone AUC is a reproducible measure of the GH response to sprint exercise. (C) 2003 Elsevier Ltd. All rights reserved.