Transmyocardial laser revascularisation (TMLR) has emerged as an efficient treatment for angina refractory to conventional therapies. Moreover, some studies have suggested improvement of cardiac function. Its mechanism of action, initially attributed to blood flow from the ventricular cavity through the channels toward the ischaemic myocardium, remains to be defined. 21 pigs were randomised to TMLR of the left lateral wall (TMLR group; n = 7), to myocardial infarction (MI group; n = 7), or to both (TMLR-MI group; n = 7). In each group one animal was sacrificed at postoperative day 0, 7, 14 and 21, and 3 animals at postoperative day 28. The lateral wall of the left ventricle was fixed in 4% formaldehyde and sliced in a plane perpendicular to the channels for histological analysis. The evolution of the cross-section area of the lesions as well as their vascular densities at postoperative day 28, expressed as number vascular structures per mm2, were studied. All the channels were occluded by a clot at postoperative day 0 and by scar tissue at postoperative day 28. The cross-section area of the lesions diminished from 8.1 +/- 1.3 mm2 at postoperative day 0 to 2.4 +/- 0.2 mm2 at postoperative day 28. At postoperative day 28, the vascular density of scar tissue of the channels in the TMLR group was significantly increased in comparison with that of myocardial infarction in the MI group (47.1 +/- 10.2 vs. 23.8 +/- 9.1; p < 0.01). In the TMLR-MI group, channels scar became indistinguishable from surrounding myocardial infarction scar. In this model, laser channels are occluded early and do not salvage acute ischaemic myocardium. Mechanisms which could explain clinical improvement, as suggested by the results, are intense laser-induced neovascularisation and channel scar retraction (Laplace's law).