The purpose of the study was to investigate the effect of an increase in mechanical knee joint loading during running on the serum COMP level. On two different test days, 20 healthy men ran with knee orthoses for 30 min on a treadmill (v = 2.2 m/s). On day 1, the orthoses were passive, whereas on day 2 they were pneumatically driven (active) and thus increased the external knee flexion moments (+30.9 Nm) during stance phase. Lower-limb mechanics and serum COMP levels (baseline; 0, 0.5, 1, 2 h post running) were analyzed. COMP levels increased immediately after running with passive (+35%; pre: 7.5 U/l, 95%CI: 6.4, 8.7, post: 9.8 U/l, 95%CI: 8.8, 10.8, p < 0.001) and active orthoses (+45%; pre: 7.6 U/l; 95%CI: 6.4, 8.8, post: 10.3 U/l, 95%CI: 9.2, 11.5, p < 0.001), but they did not differ between interventions. While running with active orthoses, greater ankle dorsiflexion angles, knee flexion angles, and moments occurred (p < 0.05). Comparing both interventions, the Δ COMP pre-post, meaning the difference (Δ) between running with active and passive orthoses in pre to post COMP level change (=level after (post) running minus level before (pre) running), correlated negatively with Δ COMP baseline (difference between the baseline COMP level before running with active and passive orthoses, r = -0.616; p = 0.004), and with a positive tendence with the Δ maximum knee flexion (r = 0.388; p = 0.091). Therefore, changes in COMP concentration after physical activity seem to be highly influenced by the COMP baseline level. In addition, correlation analysis indicates that modifications in knee joint kinematics have a greater effect on cartilage metabolism than an increase in joint moments. © 2018 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 36:1937-1946, 2018.