Why not to include M-cells in ventricular computer models

Cardiomyocytes show considerable heterogeneity in action potential duration (APD) throughout the ventricles. Based on canine experiments, it has been proposed that a population of cells exhibiting extremely long APDs, the M-cells, is present in the midmyocardium. This cell type continues to be used in simulation studies. In this review, however, we argue against including them in computer models of the ventricles. Our argument is based on experimental findings reported by other research groups who have looked for M-cells in several species. Other than in the canine wedge model, M-cells are neither found consistently in the same locations nor as a large band, but rather as sparse, small islands. Using computer simulations, we demonstrate that M-cells are not required to explain the electrocardiogram and that their inclusion may result in non-physiological behaviour. Finally, as initially proposed, M-cells only manifest prolonged APD at extremely slow pacing rates, below sinus rhythm, and would not be seen in arrhythmias. Therefore, we conclude that there is no proof of a large enough population of M-cells to affect transmural APD, nor is it necessary, so M-cells should not be included in computer models.