The Early and Late Phases of Ischemic Preconditioning
A Comparative Analysis of Their Effects on Infarct Size, Myocardial Stunning, and Arrhythmias in Conscious Pigs Undergoing a 40-Minute Coronary Occlusion
Abstract The effectiveness of the late phase of ischemic preconditioning (PC) in protecting against myocardial infarction and the concomitant contractile dysfunction after sustained ischemia remains unclear. The early and late phases of PC have not been compared using the same protocol in the same experimental model; furthermore, the late phase of PC has not been assessed in the conscious state in a large animal preparation. The goal of this study was to directly compare the effects of early and late PC on myocardial infarct size and postischemic dysfunction in chronically instrumented, conscious pigs. Four groups of pigs were subjected to a 40-minute coronary occlusion followed by 3 days of reperfusion. Group 1 (n=7) served as control. Group 2 (n=6) was subjected to ten 2-minute occlusion/2-minute reperfusion cycles 25 minutes before the 40-minute occlusion (early PC). Groups 3 (n=7) and 4 (n=4) were subjected to 10 and 25 cycles, respectively, of 2-minute occlusion/2-minute reperfusion 24 hours before the 40-minute occlusion (late PC). Infarct size averaged 45.1±5.9% of the region at risk in control pigs, was reduced by 79% (to 9.4±3.2%) in group 2, but did not differ in groups 3 (33.3±4.8%) and 4 (38.8±8.2%) versus group 1. Power analysis demonstrated that there was an 80% probability of detecting a 40% decrease in infarct size in groups 3 and 4 versus group 1. The recovery of systolic wall thickening (measured with ultrasonic crystals) after the 40-minute occlusion was poor in groups 1, 3, and 4 but markedly enhanced in group 2 throughout the 3 days of reperfusion; this beneficial effect could have been due to limitation of infarct size, alleviation of stunning, or both. Thus, a series of ten 2-minute coronary occlusions had a profound (≈80%) early infarct–limiting effect, which was associated with a marked functional benefit. This protection, however, disappeared 24 hours later and could not be reinstituted by increasing the number of PC coronary occlusions to 25. The incidence and duration of ventricular tachycardia after reperfusion was not changed by either early or late PC; no conclusions could be drawn regarding ventricular fibrillation or ischemia-induced ventricular tachycardia, since these arrhythmias did not occur in control animals. Taken together, the present results demonstrate striking differences between the early and late effects of PC: In conscious swine subjected to a sustained coronary occlusion, a PC protocol that induces powerful protection during the early phase of PC fails to induce any protection during the late phase, indicating either that a late protective effect of PC does not exist or that, if it exists, it must be weaker than the early protective effect.
Ischemic PC is the phenomenon whereby brief ischemic episodes render the heart more resistant to subsequent ischemic insults.1 2 3 4 5 6 It has recently become apparent that ischemic PC consists of two phases: an early phase, which occurs within minutes and disappears within 2 to 4 hours from the PC ischemia,7 8 9 and a late phase, which becomes manifest 24 hours later.10 11 12 13 14 15 16 17 While the early phase of PC has been clearly shown to protect against infarction and, in some models, against arrhythmias,1 7 8 9 18 19 20 21 22 23 24 25 26 the effects of the late phase of PC on infarct size and arrhythmias after a sustained coronary occlusion are controversial. Kuzuya et al,11 Marber et al,10 and Baxter and colleagues13 15 have shown that ischemic PC reduces infarct size 24 hours later in open-chest canine11 and rabbit10 13 15 models. Yang et al14 have recently reported a reduction in infarct size 24 hours after ischemic PC in conscious rabbits. In contrast, using a protocol similar to that used in the above-mentioned studies, Tanaka et al27 were unable to demonstrate a protective effect of late PC against infarction in open-chest rabbits.10 13 15 Similarly, Jagasia et al28 failed to observe any infarct-limiting effect of late PC in open-chest rats. With regard to the effect of PC on myocardial stunning after a sustained ischemic insult resulting in subendocardial infarction, studies in open-chest animal models23 29 have reported that early PC does not protect; no data are available regarding late PC. Evaluation of this form of myocardial stunning is particularly difficult because the reperfused region contains a complex admixture of necrotic subendocardium and stunned subendocardium and because the relative proportions of these two components in different parts of the reperfused region can be highly variable.30 As to the effects of ischemic PC on arrhythmias, only a few studies dealing with the late phase of protection have been published. Shiki and Hearse19 reported that open-chest rats were protected against reperfusion-induced VT and VF within 60 minutes after ischemic PC, but this effect disappeared 24 hours later. Yang et al,14 however, showed that in conscious rabbits ischemic PC reduced ischemia-induced VF 24 hours later.
An assessment of the effects of late PC on infarct size and arrhythmias is complicated by the fact that all published studies,10 11 13 15 27 28 except the report by Yang et al,14 have been performed in open-chest animals that were recovering from a recent thoracotomy. It is possible that either the postoperative state or the abnormal conditions associated with the open-chest state (such as anesthesia, trauma, elevated catecholamine levels, fluctuations in body temperature, etc) may interfere with the phenomenon of late PC or with the end points examined (infarct size, postischemic contractile performance, and incidence and severity of arrhythmias).31 32 For example, both the severity of myocardial stunning and the magnitude of the attendant free radical generation are greatly exaggerated in open-chest compared with conscious dogs.33 34 Ischemia- and reperfusion-induced arrhythmias differ in open-chest compared with conscious animals.25 35 For a given level of collateral flow and region at risk, infarct size has been found to be greater in anesthetized versus conscious dogs,36 although this is not the case in rabbits.3 5 The size of a myocardial infarction is affected to a major extent by myocardial temperature,37 38 which is unstable in open-chest models.34 Finally, the duration of early PC differs in conscious versus open-chest rabbits,9 25 and several investigators have found that PC is affected by the specific type of anesthesia used. All of these problems would be obviated by evaluating the effects of late PC in conscious animals.
Accordingly, we sought to analyze late PC in a conscious porcine model. We have recently found in conscious pigs that a sequence of ten 2-minute coronary occlusion/2-minute reperfusion cycles attenuates myocardial stunning after reversible ischemia 24 hours later12 16 17 ; specifically, when the same sequence of ten 2-minute occlusions is repeated on 2 consecutive days, the duration and degree of postischemic dysfunction are significantly less on the second day compared with the first day,12 16 17 indicating the development of a late PC effect against stunning. This protective effect is powerful, resulting in an ≈50% decrease in the severity of myocardial stunning and is consistent, being observed in every animal studied.12 16 17
In view of the conflicting reports regarding the effects of late PC on infarct size and arrhythmias, in view of the lack of any data on the effect of late PC on the recovery of contractile function after a subendocardial infarction, and in view of the fact that almost all of the studies of late PC have been performed in open-chest preparations, we investigated in our conscious pig model whether late PC protects against a partly irreversible ischemic insult resulting in subendocardial infarction. Specifically, in the present study, we sought to determine whether a sequence of ten 2-minute coronary occlusion/2-minute reperfusion cycles (which is highly effective in inducing late PC against the myocardial stunning caused by another sequence of 2-minute coronary occlusions12 16 17 ) also confers protection against the infarction, contractile dysfunction, and arrhythmias associated with a 40-minute coronary occlusion performed 24 hours later. We compared the effects observed 24 hours after the sequence of coronary occlusions (late PC) with those observed 25 minutes after the sequence (early PC) in order to gain insight into the relative potency of early and late PC. Having found that a sequence of ten 2-minute ischemic episodes fails to reduce infarct size 24 hours later, we then sought to determine whether increasing the number of PC cycles from 10 to 25 would induce a protective effect. Because of the difficulties inherent in distinguishing reversible from irreversible contractile abnormalities in the setting of a subendocardial infarction (due to the admixture of viable and necrotic tissue),30 the primary purpose of measuring systolic wall thickening in the present study was to determine, in general terms, whether PC enhances the recovery of contractile function in the reperfused region rather than to specifically dissect the functional effects of PC resulting from attenuation of stunning from those resulting from reduction in infarct size.
Materials and Methods
In order to allow comparisons between the present results and our prior data, the same conscious pig model previously used to study late PC against myocardial stunning after reversible ischemia12 16 17 39 was used in the present study. The preparation and techniques have been detailed previously.12 16 17 39 The study was performed in accordance with the guidelines of the Committee on Animals of Baylor College of Medicine and with the Guide for the Care and Use of Laboratory Animals (Department of Health and Human Services, Publication No. [NIH] 86-23).
Domestic pigs of either sex (weight on the day of the study, 28.2±3.3 kg) were premedicated with ketamine hydrochloride (20 mg/kg IM) and atropine sulfate (0.04 mg/kg IM). Sixty minutes later, anesthesia was induced with methohexital sodium (4 to 8 mg/kg IV), after which the animals were intubated. Anesthesia was maintained with 0.5% to 1.0% methoxyflurane. A left thoracotomy was performed in the 5th intercostal space under sterile conditions. Tygon catheters were placed in the left atrium and right ventricle, and an additional catheter was introduced into the femoral artery and advanced to the thoracic aorta. A hydraulic occluder and a Doppler flow velocity probe were implanted around the mid LAD. Two insulated copper wires were sutured to the right ventricle to record the electrocardiogram. To measure LV wall thickening, 10-MHz pulsed Doppler ultrasonic crystals33 were sutured to the epicardial surface, two in the center of the region to be rendered ischemic and another in an area remote from it (posterior LV wall); each probe was sutured with four 5-0 prolene stitches penetrating 0.5 to 1.0 mm into the myocardium, thus producing minimal trauma. All wires and catheters were tunneled under the skin and exteriorized through small incisions on the back. The chest was closed in layers, and a small tube was left in the thorax to evacuate air and fluids after surgery. Antibiotics were administered intravenously before surgery and daily thereafter (cefazolin, 30 mg/kg BID; gentamicin, 0.7 mg/kg BID). Arterial blood gases, hematocrit, rectal temperature, and heart rate were measured daily after instrumentation to ensure that the animals had fully recovered from the surgical procedure. The catheters were flushed daily until the end of the protocol. All pigs were allowed to recover for a minimum of 10 days after surgery and were trained for at least 6 days to lie quietly in a specially designed cage. The cage was constructed of wood and could be adjusted (length, 90 to 115 cm; width, 55 to 73 cm) to match the size of the pig.
Throughout the experiment, pigs were studied as they lay quietly in a cage in a quiet, dimly lit room. Aortic pressure was measured with Statham P23Db pressure transducers. All measured variables (aortic pressure, LAD blood flow velocity, LV wall thickening, and the electrocardiogram) were recorded simultaneously on an eight-channel direct-writing oscillograph (Gould Brush System 200).
The experimental protocol is illustrated in Fig 1⇓. Pigs were assigned to a control group or to one of three preconditioned groups. In the control group (group 1, n=7), pigs were subjected to a 40-minute LAD occlusion followed by 3 days of reperfusion. To assess the protective effects of the early phase of PC, pigs in group 2 underwent a sequence of ten 2-minute LAD occlusions, each separated by 2 minutes of reperfusion, followed 25 minutes later by a 40-minute LAD occlusion and 3 days of reperfusion. This protocol was selected as a result of pilot studies in which 3 pigs underwent a sequence of ten 2-minute LAD occlusions, followed 15 minutes later by a 40-minute LAD occlusion. Because all 3 pigs developed VF upon reperfusion, in group 2 we lengthened the interval between the PC ischemia and the 40-minute coronary occlusion to 25 minutes. To assess the protective effects of the late phase of PC, pigs in group 3 (n=7) underwent a sequence of ten 2-minute LAD occlusions, each separated by 2 minutes of reperfusion, followed 24 hours later by a 40-minute LAD occlusion and 3 days of reperfusion. To determine whether increasing the number of PC cycles brings about a protective effect during the late phase of PC, pigs in group 4 (n=4) were subjected to 25 (instead of 10) cycles of 2-minute LAD occlusion/2-minute reperfusion and, 24 hours later, a 40-minute LAD occlusion followed by 3 days of reperfusion.
In all groups, pigs were sedated with diazepam on 2 consecutive days, ie, on the day of the 40-minute LAD occlusion and on the antecedent day. On each day, diazepam was given in doses sufficient to maintain sedation for 4 to 5 hours12 16 17 39 (the initial dose was 1.5 to 2.5 mg/kg IV over 60 minutes; subsequent additional doses were given as needed). Thus, the dosage of diazepam was identical in all groups. Hemodynamic variables were recorded at the following time points: before administration of diazepam (baseline); 5 minutes before the 40-minute LAD occlusion (preocclusion); 20 and 40 minutes into the LAD occlusion; and 30 minutes and 1, 2, 3, 24, and 72 hours after reperfusion following the 40-minute LAD occlusion. The electrocardiogram was recorded continuously throughout the 40 minutes of LAD occlusion and the initial 15 minutes of reperfusion. LV wall thickening was recorded at the following time points: baseline (before diazepam); before the first 2-minute LAD occlusion or the 40-minute occlusion, depending on the protocol (preocclusion); 1 minute after each reperfusion during the sequence of 10 occlusions and reperfusions; at 20 minutes into the 40-minute LAD occlusion; and at 15 minutes and 1, 3, 24, and 72 hours after reperfusion following the 40-minute LAD occlusion. To measure regional myocardial blood flow, radioactive microspheres were injected at 30 minutes into the 40-minute LAD occlusion, as previously described.12 16 17 39 The following measurements of arrhythmic activity were obtained: PVCs, incidence and duration of VT (defined as four or more consecutive ventricular ectopic beats), total number of ectopic beats (number of PVCs plus beats of VT), and incidence of VF.
Measurement of Regional Myocardial Function
Regional myocardial function was assessed as systolic wall thickening by using a pulsed Doppler probe, as previously described.12 16 17 39 40 The beginning of systole was determined from the peak of the QRS complex on the right ventricular electrogram, and the end of systole was determined from the onset of the rapid rise in LAD blood flow velocity after systole.12 16 17 39 40 Percentage systolic thickening fraction was calculated as the ratio of net systolic thickening to end-diastolic wall thickness, multiplied by 100.40 In all animals, measurements from at least 10 beats were averaged at baseline, and from at least 5 beats at all subsequent time points. As indicated above, two thickening Doppler probes were implanted in the potentially ischemic region. The measurements used for this study are those derived from the probe that gave the lowest values of wall thickening (ie, the most severe degree of myocardial stunning) after reperfusion.
Measurement of Infarct Size and Regional Blood Flow
At the end of the study, the pigs were heparinized (5000 U IV), anesthetized (sodium pentobarbital, 35 mg/kg IV), and killed with an intravenous bolus of KCl. The hearts were excised, and the LAD was cannulated at the site of the previous occlusion. To delineate the occluded vascular bed, the aortic root was perfused for 2 minutes with a 0.5% solution of Monastral blue dye in 6% dextran 70 in normal saline at a pressure of 100 mm Hg,12 16 17 39 and the previously occluded LAD was simultaneously perfused with colorless fluid (6% dextran 70 in normal saline) at the same pressure. As a result of this procedure, the portion of the LV supplied by the previously occluded coronary artery (ischemic bed) was identified by the absence of blue dye.12 16 17 39 The heart was then cut into 1.0-cm-thick slices in a plane parallel to the atrioventricular groove, and all atrial, valvular, and right ventricular tissues were excised. To delineate infarcted from viable myocardium, the slices were incubated in a 1% solution of triphenyltetrazolium chloride (pH 7.4) at 37°C for 20 minutes. All slices were then weighed, photographed, and fixed in 10% formaldehyde solution. Transparencies were projected onto a paper sheet at a 10-fold magnification and traced to delineate infarcted, ischemic/reperfused, and nonischemic regions. The corresponding areas were measured by videoplanimetry, and from these measurements, infarct size was calculated as a percentage of the occluded bed.41 Five transmural specimens (1 to 2 g) were then obtained from both the ischemic and the nonischemic regions (to avoid admixture of ischemic and nonischemic tissue, ischemic specimens were obtained at least 0.5 cm inside the boundaries of the occluded bed). Each specimen was divided into endocardial and epicardial halves, weighed, and placed in scintillation vials containing 10% neutral buffered formalin. Regional myocardial blood flow was calculated by standard methods.40
Data are reported as mean±SEM. Hemodynamic variables and wall thickening were evaluated by a two-way repeated measures ANOVA (group and time) to determine whether there was a main effect of group, a main effect of time, or a group-time interaction. If the global tests showed a significant main effect or interaction, post hoc contrasts between different time points in the same group or between different groups at the same time point were performed with paired or unpaired Student’s t tests, as appropriate, using the Bonferroni correction.42 Occluded bed size, infarct size, and regional blood flow were analyzed with a one-way ANOVA followed by unpaired Student’s t tests with the Bonferroni correction. Because they did not follow a normal distribution, the total number of PVCs and ectopic beats and the duration of VT were analyzed using nonparametric methods (Kruskal-Wallis one-way ANOVA). The distribution of PVCs in each 4-minute interval from 0 to 40 minutes of LAD was compared among groups with the Mann-Whitney U test. The relation between infarct size and postischemic wall thickening was assessed by linear regression analysis using the least-squares method; the slopes of the repression lines were compared by ANCOVA. Since the data in the control group (group 1) and in the two late PC groups (groups 3 and 4) were found to lie along the same regression line, a comparison was performed between the regression line for the early preconditioned group (group 2) and the regression line for groups 1, 3, and 4 combined. All statistical analyses were performed using the SAS software system.43 Two-way repeated-measures ANOVA was performed using the procedure GLM (General Linear Models).43
A total of 29 pigs were entered into the study. Three pigs (1 in group 1, 1 in group 2, and 1 in group 4) were excluded because of failure of the balloon occluder during either the multiple 2-minute LAD occlusions or the 40-minute occlusion. Two pigs in group 4 died of ischemia-induced VF during PC with 25 LAD occlusions. The remaining 24 pigs form the basis of the present study. Of these, 7 were assigned to a control group (group 1), 6 to the early PC group (group 2), and 7 and 4 to the late PC groups (groups 3 and 4, respectively).
Arterial Blood Gases, Hematocrit, Temperature, and Diazepam Dose
On the day of the 40-minute LAD occlusion, arterial pH, Po2, hematocrit, and rectal temperature were within physiological limits in all groups (Table 1⇓). The doses of diazepam were similar among the four groups on the day of the 40-minute LAD occlusion (2.96±0.21, 2.85±0.24, 2.65±0.19, and 2.95±0.30 mg/kg in groups 1, 2, 3, and 4, respectively; P=NS) as well as on the day before the 40-minute LAD occlusion (2.86±0.22, 2.81±0.26, 2.65±0.39, and 2.75±0.31 mg/kg in groups 1, 2, 3, and 4, respectively; P=NS).
The administration of diazepam did not produce significant hemodynamic alterations, as indicated by the comparison of baseline (before diazepam) and preocclusion (after diazepam) measurements (Table 2⇓). All measured variables (heart rate, systolic arterial pressure, rate-pressure product, and LAD blood flow) remained stable within each day. There were no significant differences in hemodynamic variables among the four groups at any time point (Table 2⇓).
Regional Myocardial Blood Flow
The measurements of regional myocardial blood flow are summarized in Table 3⇓. Thirty minutes into the 40-minute LAD occlusion, blood flow to the ischemic region was virtually nil in all groups, both in the subepicardial and in the subendocardial layers of the LV wall. There was no statistically significant difference among the four study groups with respect to epicardial, endocardial, or mean transmural flow to the nonischemic zone.
Myocardial Infarct Size
Table 4⇓ summarizes the weights of the LV, region at risk, and infarct, as well as the risk region–to–LV and infarct–to–risk region ratios. The individual values of infarct size, expressed as a percentage of the region at risk, are illustrated in Fig 2⇓. The size of the risk region was similar among the four groups (Table 4⇓). In the control group, the 40-minute LAD occlusion resulted in infarct sizes that averaged 45.1±5.9% of the region at risk (Fig 2⇓). PC with ten 2-minute LAD occlusion/2-minute reperfusion cycles followed by 25 minutes of reperfusion before the prolonged (40-minute) LAD occlusion (group 2) reduced infarct size to 9.4±3.2% of the region at risk (P<.05 versus the control group) (Fig 2⇓), indicating a marked protective effect of the early phase of PC against infarction. However, the same PC protocol had no significant effect on infarct size 24 hours later (group 3): infarct size averaged 33.3±4.8% of the region at risk, which was not statistically different from the control group (Fig 2⇓). Increasing the number of PC occlusions from 10 to 25 (group 4) did not result in a late PC effect against infarction (38.8±8.2% of the risk region, P=NS versus the control group) (Fig 2⇓).
A statistical analysis was performed to measure the probability that a reduction of infarct size during the late phase of PC could have been missed because of insufficient sample size (type II error).42 Because the purpose of this investigation was to evaluate the ability of late PC to reduce infarct size and because an increase of tissue necrosis by PC has never been suggested, we measured only the probability of demonstrating a decrease in infarct size (one-tailed analysis). Fig 3⇓ illustrates the power of the study as a function of the hypothetical reduction of infarct size by late PC (α=.05). In the group in which late PC was induced with 10 LAD occlusions (group 3), there was an ≈80% probability of demonstrating a 45% reduction in infarct size compared with the control group (group 1), if such a reduction did indeed occur (Fig 3⇓, upper panel). The probability of detecting a 79% reduction of infarct size (which corresponds to the reduction effected by early PC in group 2 versus group 1) was almost 100%. Thus, if late PC were as effective as early PC in limiting cell death, there would have been virtually no chance that such an effect would have been missed in this study.
Since infarct size did not differ appreciably in the two groups that underwent late PC (groups 3 and 4) (Table 4⇑, Fig 2⇑), we repeated the power analysis after pooling the data from all of the pigs in groups 3 and 4, so as to measure the probability that a reduction of infarct size by late PC (induced with either 10 or 25 LAD occlusions) was missed. The results of this analysis show that there was an 80% probability of demonstrating a 40% reduction in infarct size (Fig 3⇑, lower panel). In previous studies showing a protective effect of early PC1 7 8 9 18 22 24 or late PC10 11 13 against infarction, the average reduction of infarct size was ≥40%. Therefore, it is unlikely that a biologically important effect of late PC, if present, remained undetected in this study.
Regional Myocardial Function
Because of Doppler probe malfunction, measurements of wall thickening could be obtained only in 5 of the 7 pigs in group 1, 5 of the 6 pigs in group 2, and 6 of the 7 pigs in group 3. Wall thickening measurements were obtained in all 4 pigs in group 4. In the nonischemic region, the systolic thickening fraction did not differ significantly among the control group (group 1) and the preconditioned groups (groups 2, 3, and 4) before and during the 40 minutes of LAD occlusion and during the 3 days of reperfusion (Table 2⇑).
In the LAD-dependent region, baseline systolic thickening fraction was similar in all groups (29.6±2.3%, 25.2±3.1%, 21.3±1.6%, and 25.0±5.0% in groups 1, 2, 3, and 4, respectively; P=NS). The preocclusion measurements of thickening fraction were similar to the baseline measurements in groups 1, 3, and 4 but were markedly decreased in group 2, as a result of the PC protocol (Fig 4⇓). During the 40-minute LAD occlusion, all groups exhibited similar degrees of dyskinesis (Fig 4⇓). After reperfusion, control pigs (group 1) exhibited essentially no recovery of wall thickening, even at 3 days after occlusion release (Fig 4⇓), indicating severe stunning of the viable myocardium within the risk region. In contrast, pigs in the early PC group (group 2) showed a marked enhancement in the recovery of wall thickening, which was evident as early as 15 minutes after reperfusion and persisted up to 3 days (Fig 4⇓). In the late PC groups (groups 3 and 4), the recovery of wall thickening was indistinguishable from that in the control group (Fig 4⇓), suggesting that late PC had no effect on myocardial stunning.
The enhanced recovery of contractile function in group 2 could have been caused by infarct size limitation or by a protective effect of early PC against myocardial stunning. In an attempt to distinguish between these two possibilities, the measurements of thickening fraction at 1, 3, 24, and 72 hours of reperfusion were plotted as a function of infarct size (Fig 5⇓). ANCOVA demonstrated that at 1, 3, 24, and 72 hours after reperfusion, the slope of the regression line for the early PC group (group 2) was significantly different from that of the regression line for groups 1, 3, and 4 combined (P<.01 at all time points). This would suggest that for similar infarct sizes, the severity of stunning was reduced in group 2, ie, that early PC alleviated the severity of myocardial stunning independent of the reduction of infarct size. The results of this analysis, however, cannot be considered conclusive because the infarct sizes in group 2 were clustered in the low infarct/risk region range (as a result of the infarct-sparing effect of early PC); only 1 pig in group 2 had an infarct size comparable to that in groups 1, 3, and 4. Since linearity in each experimental group over the total range of infarct sizes could not be demonstrated, a parabolic relationship between infarct size and wall thickening cannot be excluded. Thus, it remains possible that the functional improvement observed in group 2 was due to the smaller infarct sizes.
Fig 5⇑ also illustrates the fact that in the late PC groups (groups 3 and 4), all points lie around the regression line of the control group (group 1), which would support the concept that late PC did not have any effect on myocardial stunning.
The total number of PVCs, the total number of ectopic beats, and the incidence of VT and VF during the 40-minute LAD occlusion are shown in Table 5⇓. The number of PVCs in each 4-minute interval from 0 to 40 minutes of LAD occlusion is illustrated in Figs 6⇓ and 7⇓. No VT or VF was observed in any group, which may reflect the relatively small region at risk. In the control group, the occurrence of PVCs showed a biphasic profile, with a first peak developing at 13 to 16 minutes after LAD occlusion, followed by a period of remission and then by a second, smaller peak at 37 to 40 minutes after LAD occlusion (Fig 6⇓). This profile of occurrence of PVCs was modified in the early phase of ischemic PC. As shown in Fig 6⇓, a sequence of ten 2-minute LAD occlusions (ending 25 minutes before the 40-minute occlusion [group 2]) delayed the occurrence of PVCs during the 40-minute occlusion (an effect that was statistically significant [P<.05, Mann-Whitney U test]), although it did not significantly decrease the total number of PVCs (Table 5⇓). (Similar results were obtained in pilot studies in 3 pigs subjected to ten 2-minute LAD occlusions and then to a 40-minute occlusion after a 15-minute interval [data not shown].) On the other hand, late PC with either 10 or 25 LAD occlusions performed 24 hours earlier produced no consistent changes in the distribution of PVCs during the 40-minute occlusion (Fig 7⇓) and had no significant effect on the total number of PVCs (Table 5⇓).
The total number of PVCs and ectopic beats, the incidence and duration of VT, and the incidence of VF during the first 15 minutes of reperfusion following the 40-minute LAD occlusion are summarized in Table 5⇑. No pig developed VF in the control group. As mentioned in “Materials and Methods,” in the pilot studies in which the interval between the PC ischemia and the 40-minute LAD occlusion was 15 minutes, 3 of 3 pigs developed VF. When the interval between the PC ischemia and the 40-minute LAD occlusion was prolonged to 25 minutes (group 2), the incidence of VF fell to 14% (1 of 6 pigs). Although group 2 exhibited fewer PVCs than did group 1 (control group), there were no significant differences between group 2 and group 1 with respect to the total number of ectopic beats or the incidence or duration of VT. PC performed 24 hours earlier with 10 cycles of LAD occlusion/reperfusion (group 3) did not have any significant effect on reperfusion-induced arrhythmias. PC performed 24 hours earlier with 25 LAD occlusion/reperfusion cycles (group 4) resulted in a significant increase in the duration of reperfusion-induced VT and in the total number of ectopic beats compared with the control group values (Table 5⇑).
The goal of the present study was to perform a comprehensive comparison of the effects of early and late PC on infarct size and postischemic dysfunction under conditions that are as physiological as possible. The salient findings can be summarized as follows: (1) In conscious pigs, a sequence of ten 2-minute coronary occlusions markedly reduces (by ≈80%) the size of the infarction resulting from a 40-minute occlusion produced 25 minutes later, indicating a powerful early PC effect against cell death. (2) This protective effect is associated with a marked and sustained enhancement of the recovery of contractile function. (3) Despite these powerful early protective effects, the same PC protocol (ten 2-minute occlusions) fails to reduce the size of the infarction and the severity of the postischemic dysfunction resulting from a 40-minute occlusion produced 24 hours later. (4) Increasing the number of PC coronary occlusions from 10 to 25 does not bring about a protective effect 24 hours later. (5) Although no conclusions can be drawn regarding the incidence of VF, PC with 10 or 25 occlusions fails to decrease the number of PVCs during ischemia or reperfusion and the incidence and duration of VT during reperfusion 24 hours later.
Previous studies10 11 13 14 15 27 have examined the effects of late PC on infarct size in rabbits and dogs, mostly in open-chest preparations.10 11 13 15 27 To our knowledge, this is the first study (1) to examine early PC against infarction and arrhythmias in conscious pigs, (2) to assess late PC against infarction and arrhythmias in pigs, (3) to assess PC (early or late) against postischemic dysfunction in the setting of a subendocardial infarction in a conscious animal model, and (4) to directly compare the effects of the early and late phases of PC in the same animal model using the same experimental protocol. Taken together, the present results demonstrate a dichotomy between the early and late effects of PC. We found that a PC protocol that was highly effective in providing protection against cell death during the early phase of PC failed to induce any protection during the late phase of PC, indicating that in conscious swine late PC against infarction either does not exist or, if it exists, is weaker than early PC. Similarly, the ten 2-minute occlusion protocol was highly effective in enhancing recovery of contractile function after a partly reversible ischemic insult during the early phase of PC but failed to do so during the late phase of PC. The fact that the same protocol induces early but not late protection against infarction and contractile dysfunction after a 40-minute coronary occlusion suggests important qualitative or quantitative differences between the mechanism of early and late PC.
Effect of the Early Phase of PC on Infarction
Numerous previous investigations have established the early infarct–sparing effects of ischemic PC in open-chest models.2 3 4 5 However, only a few studies9 25 have documented the ability of early PC to reduce infarct size in a conscious animal model (rabbits). All previous studies in larger species (dogs and pigs) have used anesthetized preparations (eg, References 1, 7, 22, and 241 7 22 24 ; reviewed in Reference 33 ). Thus, to date, no published report has demonstrated the ability of ischemic PC to reduce infarct size during the early phase of protection in a large mammalian heart in the conscious state. Because there are differences between open-chest and conscious preparations with respect to stunning,34 arrhythmias,25 35 infarction,36 and PC9 25 (possibly as a result of the confounding effects of anesthesia, surgical trauma, abnormal hemodynamics, fluctuations in body temperature,34 37 38 exaggerated free radical production,33 etc, associated with the open-chest state) and because these factors could affect PC, it is important to verify whether ischemic PC is equally protective in awake animals. The present study demonstrates that ischemic PC is remarkably effective in limiting cell death during the early phase of protection in conscious pigs: infarct size was reduced from an average of 45.1±5.9% of the risk region to 9.4±3.2% (a 79% reduction). The present study also demonstrates that ischemic episodes need not last 3 to 5 minutes in order to induce PC; ischemic episodes <3 minutes can be highly effective, if they recur in clusters.
In contrast to conscious dogs, in which brief coronary occlusions can be performed without sedation,33 34 35 40 in our experience conscious pigs require sedation with diazepam because of restlessness, which continues to occur even after several days of adaptive training to the laboratory environment.12 16 17 39 However, in accordance with previous studies,12 16 17 39 in this investigation the administration of diazepam had no effect on hemodynamic variables (Table 2⇑) or wall thickening (Fig 4⇑). The animals were still conscious and responsive to stimuli but remained calm throughout the protocol; as a result, hemodynamics were stable during ischemia and early reperfusion (Table 2⇑).
Effect of the Late Phase of PC on Infarction
In contrast to the early phase of PC, the efficacy of the late phase of PC in reducing infarct size is controversial. Kuzuya et al11 have demonstrated in open-chest dogs that a sequence of four 5-minute coronary occlusions reduces the size of the infarction produced by a 90-minute coronary occlusion applied 24 hours later. Using open-chest rabbits, Marber et al10 have shown that a sequence of four 5-minute coronary occlusions significantly reduces the size of the infarction resulting from a 30-minute coronary occlusion applied 24 hours after the PC ischemia. Similar results have been subsequently reported by Baxter and colleagues13 15 in a similar model. Yang et al14 have recently reported that PC with four cycles of 5-minute coronary occlusion/10-minute reperfusion 24 hours before the 30-minute ischemia reduces infarct size in conscious rabbits. On the other hand, using an open-chest rabbit preparation and an experimental protocol similar to that of Marber et al10 and Baxter and colleagues,13 15 Tanaka et al27 failed to show a significant infarct-limiting effect either 24 or 48 hours after the PC ischemia. Furthermore, Jagasia et al28 did not observe any reduction in infarct size in open-chest rats subjected to a 45-minute coronary occlusion 24 hours after PC with either three 3-minute occlusions or one 15-minute occlusion. The reason for these conflicting results is unknown.
The present findings indicate that in spite of a profound early protection against infarction, PC with 10 or 25 cycles of 2-minute LAD occlusion/2-minute reperfusion did not have a significant infarct-limiting effect 24 hours later in conscious pigs. We cannot exclude the possibility that a late protective effect could have been observed using different PC protocols. However, our study demonstrates that a PC protocol that is highly effective in reducing infarct size during the early phase of PC fails to reduce infarct size during the late phase of PC. This indicates that in conscious pigs, a late protective effect of ischemic PC against cell death either does not exist or that, if it exists, it must be weaker than the early protective effect.
The apparent discrepancy between our results and those of previous authors10 11 13 14 15 may be explained by differences in species, experimental preparations, and/or protocols. The failure of the present study to detect a late phase of protection against infarction cannot be ascribed to variations in the major determinants of infarct size. The size of the infarct was normalized to the size of the risk region, which was relatively consistent among pigs. Collateral flow during the 40-minute LAD occlusion was virtually nil in all animals. Body temperature (another major determinant of infarct size37 38 ) was virtually indistinguishable among the various groups (Table 1⇑). Finally, hemodynamic variables were similar in all groups (Table 2⇑). It is unlikely that the reason for the negative results was an insufficient PC ischemic stimulus. In group 4, a total of 50 minutes of ischemia (twenty-five 2-minute LAD occlusions) was applied as the PC stimulus, which is far in excess of the total ischemic burden (20 minutes) previously found to induce late PC against infarction in rabbits10 13 14 15 and dogs.11 Furthermore, in group 2, a PC protocol of ten 2-minute LAD occlusions (total ischemic burden of 20 minutes) was sufficient to trigger a “classic” early protective response, with a 79% decrease in infarct size—an effect comparable to that observed in most studies of early PC against infarction.1 2 3 4 5 7 8 9 18 22 23 24 29 Thus, judging from the ability to limit necrosis, our PC protocol was as powerful as that used in previous investigations.1 2 3 4 5 7 8 9 18 22 23 24 29
Our inability to detect a late phase of protection against infarction cannot be attributed to the use of an excessively severe ischemic insult as the test for protection. The severity of an ischemic insult is determined not only by the duration of the coronary occlusion but also by the level of collateral perfusion, the myocardial oxygen demands, the myocardial temperature, and probably other factors. The net result of the aggregate influences of all of these factors is the amount of tissue necrosis, which can therefore be considered as an indicator of the overall severity of the ischemic insult. Although the duration of coronary occlusion that we used (40 minutes) is longer than that (30 minutes) used in prior studies of late PC in rabbits,10 13 14 15 the amount of necrosis resulting from the 40-minute occlusion in our control animals is comparable to that noted in the control animals in previous studies in which late PC was observed (45.1±5.9% of the risk region in our study versus 56.9±6.5% in the study by Marber et al,10 53.6±5.7% in the study by Baxter et al,13 and 35.7±2.3% in the study by Yang et al14 ). Furthermore, the duration of the sustained occlusion in our study was much shorter than that (90 minutes) used by Kuzuya et al.11 The infarct size in our control group is also comparable to that reported in the control groups of most previous studies of early PC against infarction.1 2 3 4 5 7 8 9 18 22 23 24 29 It is possible that a late PC effect could be detected using shorter coronary occlusions resulting in smaller infarcts. However, if the protection afforded by late PC were limited only to occlusions shorter than 40 minutes and to infarcts <45% of the risk region, such protection would clearly be less powerful than that afforded by early PC.
The results obtained in group 2 provide a “positive control,” in the sense that they demonstrate the ability of our experimental model to detect a reduction in infarct size. It must be stressed that the purpose of this study was to determine whether late PC induces a reduction in infarct size comparable to that effected by early PC. As detailed in “Results,” the power of our study to demonstrate a late protective effect similar to that observed during the early phase of PC (ie, a 79% reduction in infarct size) was almost 100%. Thus, there is virtually no chance that a decrease in infarct size similar to that seen in the early phase was missed in the late phase. Furthermore, our study had sufficient (80%) power to detect a 40% reduction in infarct size by late PC, which is less than the reductions noted by Marber et al10 (45% reduction) and Kuzuya et al11 (46% reduction) in their studies of late PC.
Effects of the Early and Late Phases of PC on Stunning
Little is known regarding the ability of ischemic PC to attenuate the severity of myocardial stunning in the tissue salvaged by reperfusion after a sustained coronary occlusion associated with subendocardial infarction. The present results demonstrate for the first time that the early phase of ischemic PC brings about a marked improvement in the recovery of contractile function in the tissue that survives within the risk region. This improvement is very rapid, becoming manifest as early as 15 minutes after reperfusion (Fig 4⇑), and is sustained for at least 3 days. As elaborated in “Results,” however, the analysis of the relation between wall thickening and infarct size does not enable us to distinguish whether the functional improvement was due to reduction of infarct size, attenuation of stunning, or both. Because the infarct sizes in group 2 were clustered in the low range of the infarct–to–risk region ratios and because there was almost no overlap between group 2 and the other groups (Fig 5⇑), linearity in each experimental group over the total range of infarct sizes could not be demonstrated. Consequently, it is not possible to determine whether pigs with comparable infarct size had different functional recovery in group 2 vis-à-vis the other groups, ie, whether the enhanced recovery of function was independent of the reduction in infarct size.
In contrast to the early phase of PC, our study demonstrates that the late phase of PC had no effect on the recovery of function after a 40-minute coronary occlusion. Here, the analysis of the wall thickening–infarct size relation (Fig 5⇑) would suggest that late PC failed to alleviate the severity of stunning in the surviving myocardium, since the ranges of infarct size were comparable in the control group and in the late PC groups (groups 3 and 4) and since all points lay along the same regression line.
A limitation of the present study that applies to both the early and the late phases of PC is that in the setting of a subendocardial infarction, in which the transmural distribution of necrotic tissue within the risk region is highly irregular, measurements obtained with a single ultrasonic crystal may not be representative of the overall contractile function of the reperfused region.30 Techniques that provide an integrated assessment of function in the entire risk region (eg, two-dimensional echocardiography) may be better suited to examine the effect of PC on myocardial stunning after a subendocardial infarction.
Effect of the Early Phase of PC on Arrhythmias
Most previous studies in open-chest rats or in isolated rat hearts have reported a protective effect of early PC against VT and VF.19 20 21 26 In large animal species, the effects of PC on arrhythmias have been variable.21 44 Our present results indicate that early PC does not limit reperfusion-induced VT, but they do not enable us to assess the effect of early PC on VF or ischemia-induced VT, because the incidence of these arrhythmias in control animals was nil (Table 5⇑). Early PC reduced the number of PVCs after reperfusion (Table 5⇑). It was interesting that when we performed the 40-minute occlusion 15 minutes after the end of the PC ischemia, there appeared to be an exacerbation of reperfusion-induced VF (3 of 3 pigs developed VF in our pilot studies compared to 0 of 6 pigs in the control group). The fact that the incidence of reperfusion-induced VF decreased when the interval between the PC ischemia and the sustained 40-minute occlusion was lengthened to 25 minutes (1 of 6 pigs, group 2) suggests that the incidence of arrhythmias after PC may be influenced by the interval between the PC stimulus and the subsequent ischemic insult. These apparently paradoxical effects of ischemic PC on arrhythmias may reflect activation of the ATP-sensitive K+ channels after brief episodes of ischemia.45 46 Activation of ATP-sensitive K+ channels has been suggested to have both proarrhythmic and antiarrhythmic effects during myocardial ischemia, depending on the experimental circumstances.47 48
Effect of the Late Phase of PC on Arrhythmias
To date, only two studies14 19 have reported the effects of the late phase of ischemic PC against arrhythmias, with conflicting results. Shiki and Hearse19 demonstrated in open-chest rats that a 5-minute coronary occlusion attenuated the ventricular arrhythmias induced by a second coronary occlusion performed 10 to 60 minutes later; however, the protective effect against arrhythmias disappeared 24 hours later. In contrast, Yang et al14 reported in conscious rabbits that a sequence of four cycles of 5-minute coronary occlusion/10-minute reperfusion reduced the incidence of ischemia-induced VF during a 30-minute coronary occlusion performed 24 hours later. In the present study, late PC with ten 2-minute coronary occlusions (group 3) had no significant effect on the total number of PVCs and on the incidence or duration of VT during reperfusion (Table 5⇑). When a sequence of twenty-five 2-minute occlusions was used (group 4), the duration of VT after reperfusion was significantly increased compared with that in control pigs (Table 5⇑), suggesting that this PC protocol may have a detrimental effect. Our data, however, do not enable us to draw conclusions regarding ischemia- or reperfusion-induced VF and ischemia-induced VT, since these arrhythmias did not occur in groups 1, 3, or 4. Consequently, if late PC protects against VF or ischemia-induced VT, this effect would not have been detected in our study. In view of these considerations, our results are not in conflict with those of Yang et al.14 Because of the differences in experimental models, comparison of our data with those of Shiki and Hearse19 is difficult.
In summary, the present study is the first direct comparison of the cardioprotective effects of the early and late phases of ischemic PC in the same preparation and with the same protocol. The results reveal striking differences. We found that repetitive 2-minute coronary occlusions have a profound early infarct–limiting effect, which is associated with a marked enhancement of the recovery of contractile function. This protection, however, disappears 24 hours later and cannot be reinstituted by increasing the number of ischemic episodes during the PC protocol. The inability of late PC to protect against cell death was observed despite the use of an animal model of moderate infarct size (averaging only 45% of the risk region). Neither early nor late PC affects VT during reperfusion, but no conclusions can be drawn regarding VF or ischemia-induced VT, since our experimental model was not primarily designed to assess arrhythmias. In investigations of PC, the number of possible variations in the experimental preparation or protocol is extremely large. Therefore, it is impossible for any single study to rule out the existence of late PC against sustained ischemia, no matter how many groups or protocols are used. Nevertheless, our data clearly indicate that if a late protective effect of ischemic PC exists in conscious swine, it must be weaker than the earlier PC effect.
Selected Abbreviations and Acronyms
|LAD||=||left anterior descending coronary artery|
|LV||=||left ventricle (ventricular)|
|PVC||=||premature ventricular contraction|
This study was supported in part by National Institutes of Health grants R01 HL-43151 and HL-55757 (Dr Bolli) and by American Heart Association, Kentucky Affiliate, Inc, grants KY-96-GB-31 (Dr Tang) and KY-96-GB-32 (Dr Qiu). We gratefully acknowledge Jennifer S. Pocius and Alejandro Tumang for expert technical assistance, Dr Jane Goldsmith for advice on statistical analysis, and Sandy Dunaway and Kimberly Mayes for secretarial assistance.
- Received August 19, 1996.
- Accepted December 24, 1996.
- © 1997 American Heart Association, Inc.
Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation. 1986;74:1124-1136.
Walker DM, Yellon DM. Ischaemic preconditioning: from mechanisms to exploitation. Cardiovasc Res. 1992;26:734-739.
Murry CE, Richard VJ, Jennings RB, Reimer KA. Myocardial protection is lost before contractile function recovers from ischemic preconditioning. Am J Physiol. 1991;260:H796-H804.
Van Winkle DM, Thornton JD, Downey DM, Downey JM. The natural history of preconditioning: cardioprotection depends on duration of transient ischemia and time to subsequent ischemia. Coron Artery Dis. 1991;2:613-619.
Marber MS, Latchman DS, Walker JM, Yellon DM. Cardiac stress protein elevation 24 h after brief ischemia or heat stress is associated with resistance to myocardial infarction. Circulation. 1993;88:1264-1272.
Kuzuya T, Hoshida S, Yamashita N, Fuji H, Oe H, Hori M, Kamada T, Tada M. Delayed effects of sublethal ischemia on the acquisition of tolerance to ischemia. Circ Res. 1993;72:1293-1299.
Sun JZ, Tang XL, Knowlton AA, Park SW, Qiu Y, Bolli R. Late preconditioning against myocardial stunning: an endogenous protective mechanism that confers resistance to postischemic dysfunction 24 h after brief ischemia in conscious pigs. J Clin Invest. 1995;95:388-403.
Baxter GF, Marber MS, Patel VC, Yellon DM. Adenosine receptor involvement in a delayed phase of myocardial protection 24 hours after ischemic preconditioning. Circulation. 1994;90:2993-3000.
Yang XM, Baxter GF, Yellon DM, Fletcher JR, Downey JM, Cohen MV. Second window of protection in conscious rabbits. J Mol Cell Cardiol. 1995;27:A26. Abstract.
Baxter GF, Goma FM, Yellon DM. Temporal characterization of the ‘second window of protection’: duration of the anti-infarct effect after ischemic preconditioning. Circulation. 1995;92(suppl I):I-389. Abstract.
Tang X-L, Qiu Y, Park S-W, Sun J-Z, Park SW, Kalya A, Bolli R. Time-course of late preconditioning against myocardial stunning in conscious pigs. Circ Res. 1996;79:424-434.
Liu GS, Thornton J, Van Winkle DM, Stanley AWH, Olsson RA, Downey JM. Protection against infarction afforded by preconditioning is mediated by A1-adenosine receptors in the rabbit heart. Circulation. 1991;84:350-356.
Shiki K, Hearse DJ. Preconditioning of ischemic myocardium: reperfusion-induced arrhythmias. Am J Physiol. 1987;253:H1470-H1476.
Hagar JM, Hale SL, Kloner RA. Effect of preconditioning ischemia on reperfusion arrhythmias after coronary artery occlusion and reperfusion in the rat. Circ Res. 1991;68:61-68.
Schott RJ, Rohmann S, Braun ER, Schaper W. Ischemic preconditioning reduces infarct size in swine myocardium. Circ Res. 1990;66:1133-1142.
Cohen MV, Liu GS, Downey JM. Preconditioning causes improved wall motion as well as smaller infarcts after transient coronary occlusion in rabbits. Circulation. 1991;84:341-349.
Cohen MV, Yang X-M, Downey JM. Conscious rabbits become tolerant to multiple episodes of ischemic preconditioning. Circ Res. 1994;74:998-1004.
Osada M, Sato T, Komori S, Tamura K. Protective effect of preconditioning on reperfusion induced ventricular arrhythmias of isolated rat hearts. Cardiovasc Res. 1991;25:441-444.
Tanaka M, Fujiwara H, Yamasaki K, Miyame M, Yokota R, Hasegawa K, Fujiwara F, Sasayama S. Ischemic preconditioning elevates cardiac stress protein but does not limit infarct size 24 or 48 h later in rabbits. Am J Physiol. 1994;267:H1476-H1482.
Jagasia D, Whiting JM, McNulty PH. Ischemic preconditioning fails to produce a second window of protection 24 hrs later in the rat. Circulation. 1996;94(suppl I):I-184. Abstract.
Ovize M, Kloner RA, Przyklenk K. Effect of preconditioning on regional contractile function: combined experience from three protocols. J Mol Cell Cardiol. 1992;24(suppl III):S13. Abstract.
Bolli R. Mechanism of myocardial ‘stunning.’ Circulation. 1990;82:723-738.
Vatner SF, Braunwald E. Cardiovascular control mechanisms in the conscious state. N Engl J Med. 1975;29:970-976.
Ning XH, Zweng TN, Gallagher KP. Ejection- and isovolumic contraction-phase wall thickening in nonischemic myocardium during coronary occlusion. Am J Physiol. 1990;258:H490-H499.
Li XY, McCay PB, Zughaib M, Jeroudi MO, Triana JF, Bolli R. Demonstration of free radical generation in the ‘stunned’ myocardium in the conscious dog and identification of major differences between conscious and open-chest dogs. J Clin Invest. 1993;92:1025-1041.
Triana JF, Li X-Y, Jamaluddin U, Thornby JI, Bolli R. Postischemic myocardial ‘stunning’: identification of major differences between the open-chest and the conscious dog and evaluation of the oxygen radical hypothesis in the conscious dog. Circ Res. 1991;69:731-747.
Duncker DJ, Klassen CL, Ishibashi Y, Herrlinger SH, Pavek TJ, Bache RJ. Effect of temperature on myocardial infarction in swine. Am J Physiol. 1996;270:H1189-H1199.
Verbinski SG, Schwartz LM, Vander Heide RS, Reimer KA. Minor variation in pericardial temperature is a major predictor of infarct size in dogs. Circulation. 1995;92(suppl I):I-457. Abstract.
Park SW, Tang XL, Qiu Y, Sun JZ, Bolli R. Nisoldipine attenuates myocardial stunning induced by multiple coronary occlusions in conscious pigs and this effect is independent of changes in hemodynamics or coronary blood flow. J Mol Cell Cardiol. 1995;28:655-666.
Bolli R, Zhu WX, Thornby JI, O’Neill PG, Roberts R. Time-course and determinants of recovery of function after reversible ischemia in conscious dogs. Am J Physiol. 1988;254:H102-H114.
Patel BS, Jeroudi MO, O’Neill PG, Bolli R. Effect of human recombinant superoxide dismutase on canine myocardial infarction. Am J Physiol. 1990;258:H369-H380.
Wallenstein S, Zucker CL, Fleiss JL. Some statistical methods useful in circulation research. Circ Res. 1980;47:1-9.
SAS Institute Inc. SAS/STAT User’s Guide, Release 6.03 Edition. Cary, NC: SAS Institute Inc; 1988.
Reimer KA, Murry CE, Jennings RB. Cardiac adaptation to ischemia: ischemic preconditioning increases myocardial tolerance to subsequent ischemic episodes. Circulation. 1990;82:2266-2268.