Background Stretch-induced arrhythmias (SIAs) can be elicited in normal canine left ventricles by transient diastolic dilatation. Since clinically important ventricular arrhythmias arise most commonly in failing and dilated ventricles, we hypothesized that the arrhythmogenic effect of transient diastolic stretch would be enhanced in chronically dilated failing canine hearts.Methods and Results Heart failure was induced in seven dogs by right ventricular pacing at 250 min sup -1 for 20.2+-1.6 days. Left ventricular (LV) mechanical properties were measured in vivo with serial echocardiograms in these seven dogs with the dogs awake and tranquilized to confirm the development of LV dilation and failure. By the third week of pacing, average short-axis area ejection fraction decreased by 64.3% (P<.001) as end-diastolic and end-systolic diameters increased by 25.9% and 50.7%, respectively (P<.001). After heart failure was established, the hearts were harvested and in vitro data were obtained as an isolated, blood-perfused ventricle preparation. A computerized servo pump system connected to an LV intracavitary balloon was used to measure and control LV volume. Results were compared with in vitro data obtained from eight ventricles not subjected to pacing (controls). LV contractility, quantitated in vitro as the slope of the peak isovolumic pressure-volume relation (Emax) normalized to LV cavity size, was much lower in the heart failure group than in controls (182+-18 versus 365+-38 mm Hg, P<.001). In all isolated hearts, SIAs were induced using an electromechanical stimulation protocol in which eight paced beats at 2 Hz were followed by a transient increase in LV volume during early diastole. Prestretch volume (Vi) was selected to yield end-diastolic pressures of 4 to 8 mm Hg in all hearts. The fractional increase in LV volume (Delta V) that produced SIAs 50% of the time (Delta V sub 50/Vi) was smaller in failing hearts than in controls (0.78+-0.04 versus 1.18+-0.17, P=.009), indicating an increased sensitivity to SIAs in the failing hearts. Although ventricular pairs were occasionally induced in both groups, the great majority of the arrhythmias induced in both groups were single extrasystoles, and nonsustained runs of ventricular tachycardia were never elicited in either group. LV end-diastolic and peak stretch pressures were similar in the two groups, but LV end-diastolic wall stress was higher by 35.7% (P=.029) in the dilated failing ventricles because LV hypertrophy, which tends to normalize wall stress as the heart dilates, did not occur during the 3 weeks of pacing. For stretch stimuli of comparable arrhythmogenic effectiveness, peak LV wall stress during stretch was similar in the two groups, whereas the fractional increase in volume was significantly smaller in the heart failure group, indicating impaired viscoelastic properties in the failing ventricles. In five control ventricles, acute exposure to 0.5 micromole/L dobutamine increased ventricular sensitivity to the induction of SIAs, as shown by a decrease in Delta V50/Vifrom 1.27+-0.16 to 1.06+-0.11 (P=.04).Conclusions Altered mechanical properties and/or neurohumoral adaptations associated with chronic dilation and failure predispose the ventricle to induction of ventricular extrasystoles by transient LV diastolic stretch. (Circulation. 1994;90:2022-2031.)