MULTIVARIATE PREDICTION OF SPONTANEOUS REPETITIVE RESPONSES IN VENTRICULAR MYOCARDIUM EXPOSED IN-VITRO TO SIMULATED ISCHEMIC CONDITIONS

Guinea-pig ventricular myocardium was partly exposed to normal Tyrode' s superfusion and partly to altered conditions (using modified Tyrode' s solution) set to simulate acute myocardial ischemia (PO2 80 +/- 10 m mHg; no glucose; pH 7.00 +/- 0.05; K+ 12 mM). Using a double-chamber t issue bath and standard microelectrode technique, the occurrence of sp ontaneous repetitive responses was investigated during simulated ische mia (occlusion) and after reperfusing the previously ischemic superfus ed tissue with normal Tyrode's solution (reperfusion). In 62 experimen ts (42 animals) the effects of: (1) duration of simulated ischemia (13 21 +/- 435 s), (2) stimulation rate (1002 +/- 549 ms) and (3) number o f successive simulated ischemic periods (occlusions) (1.58 +/- 0.92) o n: (1) resting membrane potential, (2) action potential amplitude, (3) duration of 50 and 90% action potentials and (4) maximal upstroke vel ocity of action potential were studied. All variables were considered as gradients (delta) between normal and ischemic tissue. Both during o cclusion and upon reperfusion, spontaneous repetitive responses were c oded as single, couplets, salves (three to nine and >10) or total spon taneous repetitive responses (coded present when at least one of the a bove-mentioned types was seen). The incidence of total spontaneous rep etitive responses was 31% (19/62) on occlusion and 85% (53/62) upon re perfusion. Cox's models (forced and stepwise) were used to predict mul tivariately the occurrence of arrhythmic events considered as both tot al spontaneous repetitive responses and as separate entities. These mo dels were applicable since continuous monitoring of the experiments en abled exact timing of spontaneous repetitive response onset during bot h occlusion and reperfusion. In predicting reperfusion spontaneous rep etitive responses, total spontaneous repetitive responses and blocks o bserved during the occlusion period were also considered. Total occlus ion spontaneous repetitive responses were predicted by: (1) longer del ta 50% action potential duration (t=2.68), (2) shorter delta 90% actio n potential duration (t=-2.17) and (3) fewer occlusive periods (t=-2.4 6). Total reperfusion spontaneous repetitive responses were predicted by a longer delta action potential amplitude (t=2.18). Due to few even ts during occlusion, prediction of individual arrhythmic entities was not possible. Upon reperfusion single spontaneous repetitive responses were predicted by longer delta maximal upstroke velocity of action po tential (t=2.59) and shorter delta 90% action potential duration (t=-2 .55); couplets were predicted by longer delta 50% action potential dur ation (t=3.26); longer delta action potential amplitude predicted salv es (>10) (t=3.26). This study demonstrates that different electrophysi ological variables, simultaneously measured in normal and simulated is chemic tissue, predict in vitro arrhythmic events seen during occlusio n and reperfusion. While occlusion spontaneous repetitive responses ar e mainly related to the difference in action potential duration betwee n normal and ischemic tissue, reperfusion spontaneous repetitive respo nses are associated with lower action potential amplitudes in the isch emic zone. Differences also exist among various arrhythmic entities up on reperfusion. Drug interventions might be tested in this preparation to see whether prevention of spontaneous repetitive responses is achi eved. Based on the evidence presented here, class III antiarrhythmics and ATP potassium-channel openers and/or blockers are potential candid ates.