Mechanical loading is known to cause increases in connective tissue blood flow of human tendon and in a local release of vasodilatory substances like prostanoids and bradykinin (Langberg et al. 1999; Boushel et al. 2000). The present study evaluated the importance of prostaglandin for exercise-related blood flow increases in connective tissue. The study was approved by the Ethical Committee of Copenhagen. Healthy, young volunteers (n = 24, aged 23-31 years) underwent after giving informed consent, 30 min of intermittent, isometric, plantar flexion with both calf muscles (2 s contraction/2 s relaxation, torque output: 900 ± 50 N m) either without (n = 6, Control) or with blockade of the cyclo-oxygenase (Cox)-mediated formation of prostaglandins (Kurumbail et al. 1996; Vane et al. 1998), either Cox-2 specific (n = 10, Celecoxib 200 mg X 2 per day for 3 days) or Cox unspecific (n = 8, acetylsalicylic acid 500 mg per day for 3 days and indomethacin 100 mg 12 h and 1 h before the experiment). Prostaglandin E2 (PGE2) concentration in peritendinous tissue was determined by microdialysis using labelled [15-3H(N)]PGE2 in the perfusate (relative recovery 44 ± 5 % (mean ± S.E.M.) at rest and 47 ± 5 % during muscular contractions). Blood flow was measured by ¹³³Xenon wash-out. Statistical analysis with Mann-Whitney or Wilcoxon’s rank sum tests were used. In control, interstitial PGE2 concentrations rose in response to exercise (0.8 ± 0.2 ng ml^-1 (rest) to 1.4 ± 0.5 (exercise), P < 0.05), whereas during unspecific Cox inhibition tissue PGE2 was completely inhibited and no increase in PGE2 was observed during exercise. Cox-2 specific blockade did not inhibit tissue PGE2 at rest but totally abolished the exercise-induced increase (0.8 ± 0.1 ng ml^-1 (rest) to 0.7 ± 0.2 (exercise)). Blood flow was similar in the three experimental groups at rest, whereas during exercise the increase in flow was significantly reduced in the Cox-2 (to 7.7 ± 1.2 ml (100 g tissue)^-1 min^-1) and Cox unspecific inhibited group (to 8.2 ± 1.3 ml (100 g tissue)^-1 min^-1), respectively, compared with control (to 11.0 ± 1.8 ml (100 g tissue)^-1 min^-1). The findings suggest that Cox-2 specific mechanisms are responsible for the exercise-induced increase in prostaglandin synthesis, and that increase in tissue PGE2 plays a role in the increased tissue flow in peritendinous connective tissue observed during muscular contractions and physical loading of tendons.