Myocardial water content and infarct size were studied in 39 pigs randomly assigned to a nonintervention group, a group with an intracoronary infusion of a control solution, and a group with a hyperosmotic infusion to 450 mosM by the addition of D-mannitol. The intracoronary solutions were selectively infused into the left anterior descending coronary artery just distal to the occlusion site starting 48 min after occlusion. Reperfusion was performed 3 min later and the infusion rate progressively tapered off over the following 33 min. Multiple myocardial fragments were then obtained in nine pigs, from endocardial, mesocardial, and epicardial regions of the ischemic and control myocardium. Water content measured after 48 h of dessication was significantly greater in the reperfused [530 +/- 7 ml/100 (mean +/- SE) g dry wt] compared with control myocardium (374 +/- 3; P less than 0.0001) and similar in reperfused control and isotonic infusion groups (556 +/- 7 and 543 +/- 8 ml/100 g dry wt); it was 491 +/- 11 with intracoronary D-mannitol infusion, representing 35% less increase (P less than 0.001). In the 30 remaining pigs, area at risk and infarct size were measured 24 h later by in vivo fluorescein and in vitro triphenyltetrazolium chloride. Infarct size was similar in control and in the isotonic reperfused hearts, 6.80 +/- 1.05 and 6.22 +/- 0.76% of ventricular weight, and smaller with D-mannitol, 4.46 +/- 0.46 (P less than 0.05). The ratio of infarct size to area at risk was also smaller [0.415 +/- 0.029 vs. 0.543 +/- 0.052 and 0.547 +/- 0.045 (P less than 0.02)].(ABSTRACT TRUNCATED AT 250 WORDS)