We investigated the effect of indomethacin, a widely used nonsteroidal antiinflammatory drug, on the healing of myocardial infarction (MI). Experimental MI was produced in anesthetized, open-chest dogs by occluding the left anterior descending coronary artery. Ten dogs received indomethacin, 10 mg/kg i.v., and 11 received saline, 15 minutes and 3 hours after occlusion. After 6 weeks, the dogs were killed and their hearts were subjected to morphologic and biochemical analysis. The average thickness of the transmural scar and the noninfarcted left ventricular wall was measured at multiple sites in formalin-fixed left ventricular slices and the ratio of the thickness of the transmural scar to the noninfarcted wall determined. The average thickness of the noninfarcted wall was 8.80 +/- 0.19 mm (mean +/- SEM) in the control group and 8.44 +/- 0.26 mm in the indomethacin group (NS). The scar thickness was 7.24 +/- 0.64 mm in the control group and 3.56 +/- 0.40 mm in the indomethacin group (p less than 0.001). The ratio of scar to noninfarcted wall thickness was 0.83 +/- 0.07 in the control group and 0.43 +/- 0.04 in the indomethacin group (p less than 0.001). Scars in treated dogs did not differ from controls either by light microscopic histologic analysis or by analysis of hydroxyproline content per unit weight. We conclude that indomethacin results in marked scar thinning when given early after experimental MI.