We performed two experiments to describe the magnitude of delayed-onset muscle soreness (DOMS) associated with blood flow restriction (BFR) exercise and to determine the contribution of the concentric (CON) versus eccentric (ECC) actions of BFR exercise on DOMS. In experiment 1, nine subjects performed three sets of unilateral knee extension BFR exercise at 35% of maximal voluntary contraction (MVC) to failure with a thigh cuff inflated 30% above brachial systolic pressure. Subjects repeated the protocol with the contralateral limb without flow restriction. Resting soreness (0-10 scale) and algometry (pain-pressure threshold; PPT) were assessed before and 24, 48 and 96 h post-exercise. Additionally, MVC and vastus lateralis cross-sectional area (CSA) were measured as indices of exercise-induced muscle damage. At 24-h post-exercise, BFR exercise resulted in more soreness than exercise without BFR (2.8 +/- 0.3 vs 1.7 +/- 0.5) and greater reductions in PPT (15.2 +/- 1.7 vs. 20 +/- 2.3 N) and MVC (14.1 +/- 2.5% decrease vs. 1.5 +/- 4.5% decrease) (p <or= 0.05). In experiment 2, 15 different subjects performed three sets of unilateral BFR exercise at 35% MVC with one limb performing only the CON action and the contralateral performing the ECC action. The aforementioned indices of DOMS were assessed before exercise and 24, 48 and 96 h post-exercise. At 24 h post-exercise, CON BFR exercise resulted in more resting soreness than ECC BFR exercise (3.0 +/- 0.5 vs. 1.6 +/- 0.4), and a greater decrease in MVC (9.8 +/- 2.7% decrease vs. 3.4 +/- 2.5% decrease) (p <or= 0.05). These data suggest that knee extension BFR exercise induces mild DOMS and that BFR exercise elicits muscle damage under atypical conditions with low-tension concentric contractions.