Intravenous magnesium lowers mortality in patients with suspected myocardial infarction. We tested the hypothesis that the protective effect may be due to a direct, local influence of magnesium on myocardial reperfusion injury in a dog model of ischaemia/reperfusion. Ten anaesthetized open chest dogs underwent 1 h of left anterior descending artery (LAD) occlusion and 6 h of reperfusion. The animals received intracoronary (i.c.) magnesium aspartate (Mg, n = 5) or vehicle infusion (n = 5) for the first hour of reperfusion. Mg infusion was adapted to actual LAD flow (ultrasonic flow probe) to increase regional plasma concentration by 4 mmol L-1. Regional myocardial function was measured as percent systolic wall thickening (sWTh, sonomicrometry). Intracoronary Mg increased LAD flow during application (at 15 min reperfusion; Mg, 194 +/- 44 (mean +/- SD); control, 116 +/- 41 mL min-1 100 g-1, P < 0.01). sWTh decreased during coronary occlusion from 14.3 +/- 7.1% to -4.7 +/- 2.7% in the control group and from 14.8 +/- 2.5% to -4.1 +/- 3.1% in the Mg group. Throughout the reperfusion period wall function remained depressed in both groups to a similar degree (control, -3.5 +/- 1.8%; Mg, -3.0 +/- 1.9% at 6 h reperfusion). Global haemodynamics were not different. Infarct size after 6 h reperfusion (TTC staining) was similar in both groups (Mg, 20.6 +/- 5.0; control, 24.4 +/- 8.7% of area at risk). Regional magnesium application (i.c.) to post-ischaemic reperfused myocardium had no influence on infarct size or post-ischaemic regional wall function in this model.(ABSTRACT TRUNCATED AT 250 WORDS)