The relative effects of nicotinic acid (NA) and nitroglycerin (NT) added to cold high K+ cardioplegia were studied, to represent the two moieties of the adenosine triphosphate-sensitive potassium channel (KATP) activator nicorandil (N). In addition, we made a pooled analysis of a large series of experiments performed in our Laboratory to investigate the effects of KATP activation by N, or block (by glibenclamide, G), on postcardioplegic myocardial dysfunction. In both studies, reversibility from myocardial dysfunction (stunning) was assessed by the positive inotropic agent dobutamine. Guinea pig papillary muscle preparations were immersed in Tyrode's solution (O2 content 16 ml/l, 37 degrees C), then hypoxic (O2 content 5 ml/l) superfusion with hypothermic (20 degrees C) cardioplegic Saint Thomas' Hospital solution (STHS) was performed for 120 min. We investigated: A) 5 groups based on treatments added to STHS: 1) saline (Control (C)); 2) N = 1 mmol/L; 3) G = 1 mumol/L (also given for 15 min in Tyrode's solution); 4) NA = 1 mmol/L; 5) NT = 100 mumol/L; B) 76 consecutive experiments and we defined, independent of whether just before or during STHS: 1) KATP activation (by N, in the concentration range 1 mumol/L to 1 mmol/L, n = 36); 2) KATP block (by G 1 mumol/L, either alone or just before N, n = 20); 3) controls (n = 20) (either saline, n = 12, or saline plus dimethyl sulfoxide, as vehicle, at the ratio 100 to 1, n = 8). Absolute isometric contractility variables were evaluated along with percent changes of baseline values: 1) at 30 s of STHS, 2) after 60 min of reoxygenation with Tyrode's solution and 3) following further 15 min of dobutamine 10 mumol/L. In all preparations, developed tension (DT), time to peak tension (TPT), DT/TPT and time to arrest (TTA) were measured. In study A): TTA was significantly abbreviated (intergroup F = 5.79, p < 0.001) in N (49 +/- 11 s, mean +/- SD) p < 0.01 vs C and NA). At 30 s of STHS %DT/TPT was unchanged among groups. By contrast, after 60 min of reoxygenation %DT/TPT in N (118 +/- 35%, p < 0.05 vs C, p < 0.01 vs G) was improved (intergroup F = 5.48, p < 0.002). G, NA and NT showed recovery of contractility similar to C. However, after dobutamine the poorest %DT/TPT were seen in G (p < 0.01 vs C, p < 0.05 vs N). In study B): using the multivariate logistic model, with KATP activation, the odds of normal contractile response, respectively at 60 min of reoxygenation (t = 2.81) and after dobutamine (t = 3.22), were 29.8 and 8.86 of controls, whereas TTA (t = -1.59) was inversely related. Moreover, with KATP block the odds after dobutamine was 0.204 of controls. The relative operating characteristic plots showed areas under the curve greater than 0.7, which is evidence for accurate assessment of the predictive rules adopted. This is the first report where a probabilistic approach to cardioplegia-related experiments showed high accuracy in predicting the recovery of post-hypoxic contractile function (stunning). The results indicate that on postcardioplegic stunning: 1) KATP activation by N and KATP block by G (both given prior to or contemporary with hypoxia) have opposite effects; 2) the favorable effects of nicorandil seem unrelated to its nicotinamide or nitrose moieties.