Twenty untrained male volunteers were required to run downhill for 45 minutes on a motor driven treadmill to induce muscle soreness. The volunteers took diclofenac or placebo before and for 72 hours after two runs 10 weeks apart, in a randomised double blind crossover design. Subjective soreness was assessed before and at intervals up to 72 hours after each run; venous blood samples, collected at the same time intervals, were used to estimate serum activities of creatine kinase, lactate dehydrogenase and aspartate aminotransferase and serum concentrations of creatinine and urea. Subjective soreness and the biochemical parameters increased after both runs, although the serum enzyme response to the second run was reduced. Diclofenac had no influence on the serum biochemical response to downhill running. Although overall soreness was not affected by diclofenac, individual soreness measurements were reduced by diclofenac at the first period of the study. These results suggest that diclofenac does not influence muscle damage, but may slightly reduce the associated soreness.