Mechanical dysfunction is the strongest predictor of sudden cardiac death due to arrhythmia. Contraction-excitation feedback whereby changes in myocardial length/tension influence the time course of repolarization and excitability would provide a possible mechanism. Such a relationship has been shown in animals but has yet to be demonstrated in man. A useful model for studying this relationship is provided by the process of weaning off cardiopulmonary bypass after routine coronary artery surgery. During this weaning period of approximately 1 min, the heart is converted from being partially empty and flaccid (i.e., a "nonworking" state) to being filled and stretched to support the circulation (i.e., a "working" state). Monophasic action potentials (MAPs) were recorded from the left ventricular epicardium as a measure of repolarization time in 16 patients at discontinuation of cardiopulmonary bypass. Systolic pressure was recorded from the radial artery line. Measurements were made at three stages that related to different dynamic states of the heart: (1) starting to come off bypass ("minimally working"), defined as the time of first appearance of an inflection on the arterial pressure trace indicating the start of left ventricular ejection and valve opening, when arterial pressures represent left ventricular pressure, (2) half off bypass ("partially working"), and (3) off bypass ("wholly working"). During the process of discontinuing bypass MAP duration shortened, while systolic pressure increased. MAP duration at 90% and 60% repolarization (MAP D90, MAP D60) decreased from 288.0 -+ 29.5 msec (mean ± SEM) and 235.0 ± 27.9 msec in the minimally working heart to 274.5 ± 30.2 msec and 224.2 ± 27.3 msec in the partially working heart (p < .001), with a subsequent decrease to 261.0 ± 28.8 and 214.0 ± 28.7 when the heart was wholly working (p < .001). Systolic pressure increased from 54.1 + 9.3 mm Hg in the minimally working heart to 65.9 + 13.8 mm Hg in the partially working heart (p < .001) and subsequently increased to 75.5 ± 13.3 mm Hg when the heart was wholly working (p < .001). Mean heart rates did not change significantly. A strong correlation was obtained between absolute MAP duration and systolic pressure. Regression analysis revealed: MAP D90 vs systolic pressure (p < .001) and MAP D60 vs systolic pressure (p < .01). An impressive correlation was also obtained between the change in MAP duration and change in systolic pressure, with p < .001 for AMAP D90 vs Asystolic pressure and p < .02 for AMAP D60 vs lAsystolic pressure. When changes in AMAP D90 and AMAP D60 from the minimally working to the partially working state were analyzed separately both failed to reach significance. However, the change from the partially working to wholly working state for AMAP D90 vs Asystolic pressure was significant (p < .02), although that for AMAP D60 just failed to reach significance (p < .1). Our results provide support for contraction-excitation feedback in man whereby changes in myocardial force/tension relation...