Effect of Adenosine on Myocardial ‘Stunning’ in the Dog

Experimental Preparation and General Protocol

Pentobarbital-anesthetized dogs of either sex (13 to 34 kg) were instrumented with a snare and a Doppler flow probe around the mid left anterior descending coronary artery (LAD), a 27-gauge intracoronary needle just distal to the snare occluder, Doppler wall thickening probes, catheters in the aorta and left atrium, and a high-fidelity Millar pressure transducer in the left ventricular (LV) cavity. To prevent clotting, heparin was given immediately after insertion of the needle (3000 U IV) and thereafter (500 U/h). All dogs underwent a 15-minute LAD occlusion followed by 4 hours of reperfusion. Regional myocardial blood flow was determined by injecting radioactive microspheres, as previously described,²⁵ before coronary occlusion (baseline), 8 to 10 minutes after occlusion, 25 minutes after reperfusion (during adenosine infusion), and 4 hours after reperfusion. Systolic thickening fraction was assessed by using the pulsed Doppler epicardial probe.^{15 25 26 27 28 29 30 31 32} The total deficit of wall thickening after reperfusion was calculated by measuring the area between the wall thickening–time line and the baseline (100% line) during the first 4 hours of reflow.^{30 31} The size of the occluded coronary vascular bed was determined by a postmortem dual-perfusion technique.¹⁵

Adenosine Treatment

In phase A, dogs were assigned to one of three intracoronary treatments. Group I (control group) received vehicle (heparinized normal saline) at a rate of 2 mL/min beginning 30 minutes before occlusion and ending 60 minutes after reperfusion, whereas groups II and III received adenosine. Adenosine (Sigma Chemical Co) was dissolved in heparinized saline at a final concentration of 1 mg/mL; the pH of the solution (which was ≈2.8) was adjusted to 7.4 by using 0.1 mol/L NaOH. In group II, the infusion of adenosine was started 30 minutes before coronary occlusion and ended 60 minutes after reperfusion. In group III, the infusion of adenosine was started 2 minutes before reperfusion and ended 60 minutes after reperfusion. In both groups, adenosine was infused at a rate of 2 mL/min, corresponding to a dose of 2 mg/min. In pilot studies, this was found to be the highest dose that could be given intracoronarily without producing hemodynamic effects (ie, arterial hypotension). During the first 13 minutes of LAD occlusion, the rate of infusion was reduced to 10% of preocclusion values (0.2 mL/min) in all three groups to avoid washout of catabolites and in group II to prevent the local concentration of adenosine in the ischemic tissue from achieving very high levels.

In phase B, dogs were assigned to one of two intracoronary treatments: (1) group IV (control group) received vehicle for a duration of 30 minutes at a rate of 2 mL/min beginning 40 minutes before coronary occlusion and ending 10 minutes before occlusion, and (2) group V received a 4-mg/min infusion of adenosine for a duration of 30 minutes starting 40 minutes before coronary occlusion and ending 10 minutes before occlusion. Adenosine was dissolved as in phase A, except that the final concentration was 2 mg/mL.

Statistical Analysis

Data are reported as mean±SEM. Continuous variables were analyzed by a two-way repeated-measures ANOVA to determine whether there was a main effect of group, a main effect of time, or a time by group interaction. If the global tests showed a significant main effect or interaction, post hoc contrasts between groups at various time points or between time points within one group were performed with Student’s t tests for unpaired or paired data, as appropriate, and the resulting P values were adjusted according to the Bonferroni correction.³³ All analyses were performed using the SAS software system.³⁴ Two-way ANOVA was performed by using the SAS procedure glm.³⁴ The correlation between thickening fraction at 1 and 4 hours of reperfusion and transmural collateral blood flow was examined by using linear regression analysis, and comparisons between groups were made by ANCOVA, in which collateral flow was the covariate.

Phase A: Effect of Adenosine on Myocardial Stunning

Of the 72 dogs initially anesthetized in phase A, 37 (51%) were excluded for the reasons outlined in Table 1⇓. The final analysis included 14 dogs in group I (control), 10 dogs in group II, and 11 dogs in group III. Postmortem triphenyltetrazolium chloride (TTC) staining confirmed the absence of infarction in all animals. Among the dogs that underwent coronary occlusion, the incidence of ventricular fibrillation tended to be higher in group II (6 [38%] of 16) than in group I (3 [18%] of 17) and group III (2 [15%] of 13), but the differences were not statistically significant (P=.37 versus group I and P=.36 versus group III [χ² test with continuity correction]). It is important to point out that among the dogs that survived, there was no evidence that the severity of ischemia was less in group II than in groups I and III, since collateral flow and occluded bed size were similar in all three groups and the rate-pressure product during occlusion was actually higher in group II (see below).

Regional Myocardial Function

The systolic thickening fraction in the nonischemic (control) region did not differ significantly among the three groups at any time point during the protocol (Table 4⇑). Baseline systolic thickening fraction in the region to be rendered ischemic was 23.7±2.0%, 25.5±2.0%, and 28.3±2.1% in groups I, II, and III, respectively (P=NS). In group II, the thickening fraction decreased significantly during infusion of adenosine (to 84.1±3.6% of baseline, P<.01), indicating a negative inotropic effect of the nucleoside. During coronary occlusion, the extent of paradoxical systolic thinning in the ischemic region was similar in all three groups (Fig 1⇓). After reperfusion, control dogs (group I) exhibited little recovery of contractile function, and 4 hours after restoration of flow, the previously ischemic region was still dyskinetic, indicating severe myocardial stunning. Two-way ANOVA demonstrated that after reperfusion there was a significant time-by-group interaction, indicating that the changes in thickening fraction over time were different in the three groups. As Fig 1⇓ illustrates, in group II the recovery of function was enhanced throughout the 4-hour observation period; post hoc Student’s t tests with the Bonferroni correction confirmed that the thickening fraction was significantly greater than in the control group at 30 minutes after reflow and at all subsequent time points. In contrast, in group III post hoc tests demonstrated that the postischemic recovery of function did not differ significantly from that in the control group (Fig 1⇓). In comparison with the control group, the total deficit of wall thickening during the entire 4-hour period of reperfusion (an integrative assessment of postischemic dysfunction^{30 31} was reduced by 38% in group II (P<.01) and by only 14% in group III (P=NS).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Graph showing systolic thickening fraction in the ischemic-reperfused region 20 minutes into the infusion of adenosine (ADO) (ie, 10 minutes before coronary occlusion), 5 minutes after coronary occlusion (O), and at selected times after reperfusion in phase A. Group I indicates control dogs; group II, dogs treated with adenosine continuously before, during, and after ischemia; and group III, dogs treated with adenosine beginning at reperfusion. Thickening fraction is expressed as a percentage of baseline values. Data are mean±SEM.

As shown in Fig 2⇓, in control dogs wall thickening at 1 and 4 hours after reperfusion was directly related to collateral flow during coronary occlusion (r=.59 and r=.53, respectively). ANCOVA demonstrated that in group II this relation was altered, so that for the same level of collateral flow, thickening fraction after reperfusion was greater than in the control group by an average of 43% at 1 hour (P<.01) and 69% at 4 hours (P<.01) after reperfusion. ANCOVA demonstrated that the wall thickening–collateral flow relation was also significantly (P<.05) different between group II and group III. These results indicate that the enhanced recovery of wall thickening effected by adenosine infusion in group II was independent of any differences in blood flow during ischemia.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Scatterplots showing the relation between mean transmural collateral blood flow to the ischemic region during coronary occlusion (horizontal axis) and systolic wall thickening (vertical axis) at 1 and 4 hours after reperfusion in phase A. Group I indicates control dogs (n=14); group II, dogs treated with adenosine continuously before, during, and after ischemia (n=10); and group III, dogs treated with adenosine beginning at reperfusion (n=11). Collateral flow is expressed in milliliters per minute per gram, and thickening fraction is expressed as a percentage of baseline values (collateral flow is not expressed as a percentage of nonischemic zone flow, because nonischemic zone flow during coronary occlusion may have been altered by adenosine). Note that at any given collateral flow, group II exhibited greater wall thickening than did group I or group III. The linear regression equations for group I were as follows: 1 hour, y=−71.5+245.4x (r=.59, P<.05); 4 hours, y=−68.7+278.0x (r=.53, P=.05). The equations for group II were as follows: 1 hour, y=−11.8+143.8x (r=.61, P=.06); 4 hours, y=33.7+77.7x (r=.62, P=.06). The equations for group III were as follows: 1 hour, y=−59.0+304.7x (r=.78, P<.01); 4 hours, y=−45.6+233.1x (r=.45). ANCOVA demonstrated that the two regression lines for groups I and II were significantly different (P<.01) and that at comparable flows, thickening fraction in group II exceeded that in control dogs by an average of 43% at 1 hour and 69% at 4 hours. Furthermore, ANCOVA demonstrated that the regression lines for groups II and III were also different (P<.05).

To determine whether adenosine-induced hyperemia increases systolic wall thickening in the stunned myocardium (and thus may account for the increased wall thickening observed in group II at 30 minutes and 1 hour), at 4 hours of reperfusion adenosine was infused intracoronarily for 10 minutes at a rate of 2 mg/min, which did not produce any systemic hemodynamic effects, in nine dogs from groups I or III (Fig 3⇓). Despite an average 4.4-fold increase in LAD flow, the thickening fraction did not change significantly (−29.8±11.7% of baseline before adenosine versus −14.7±12.0% during adenosine, P=.14 [Fig 3⇓]).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Graph showing systolic thickening fraction in the ischemic-reperfused region 4 hours after reperfusion and during infusion of adenosine (ADO) (ie, 10 minutes later) in nine dogs from either group I or group III. LAD indicates left anterior descending coronary artery. Thickening fraction is expressed as a percentage of baseline values. Solid dots indicate mean±SEM.

Phase B: Effect of Adenosine on Preconditioning Against Myocardial Stunning

Of the 21 dogs initially anesthetized in phase B, 9 (43%) were excluded for the reasons outlined in Table 1⇑. Thus, the final analysis included 6 control dogs (group IV) and 6 dogs pretreated with adenosine (group V). As in phase A, postmortem TTC staining confirmed the absence of infarction in all animals.

Regional Myocardial Function

Systolic thickening fraction in the nonischemic (control) region did not differ significantly between the two groups (Table 6⇑). Baseline systolic thickening fraction in the region to be rendered ischemic was 23.6±2.8% in the control group and 24.5±2.1% in group V (P=NS). In group V, thickening fraction decreased transiently during the infusion of adenosine (81.4±4.8% of baseline values, P<.01) (Fig 4⇓), in analogy with the results obtained in group II. After discontinuation of adenosine (immediately before coronary occlusion), however, thickening fraction returned to values not significantly different from baseline (91.2±4.5% of baseline). During coronary occlusion, the degree of paradoxical systolic wall thinning was comparable in groups IV and V (Fig 4⇓). After reperfusion, there was no appreciable difference in the recovery of contractile function between the two groups, both of which exhibited severe stunning (Fig 4⇓). Two-way ANOVA failed to show a significant main effect for group or a group by time interaction. Fig 5⇓ illustrates the relation between postischemic wall thickening and collateral flow for these two groups. Note that there is overlap between groups IV and V at all levels of collateral flow. ANCOVA failed to reveal any difference between the two groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Graph showing systolic thickening fraction in the ischemic-reperfused region 20 minutes into the infusion of adenosine (ADO) (ie, 20 minutes before coronary occlusion), 1 minute before coronary occlusion (Pre-O) (ie, 9 minutes after the discontinuation of adenosine infusion), 5 minutes after coronary occlusion (O), and at selected times after reperfusion in phase B. Group IV indicates control dogs; group V, dogs treated with adenosine before ischemia. Thickening fraction is expressed as a percentage of baseline values. Data are mean±SEM.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Scatterplots showing the relation between mean transmural collateral blood flow to the ischemic region during coronary occlusion (horizontal axis) and systolic wall thickening (vertical axis) at 1 and 4 hours after reperfusion in phase B. Group IV indicates control dogs (n=6); group V, dogs treated with adenosine before ischemia (n=6). Collateral flow is expressed in milliliters per minute per gram, and thickening fraction is expressed as a percentage of baseline values (collateral flow is expressed in milliliters per minute per gram for consistency with Fig 2⇑). Note that there is considerable overlap between the two groups at all levels of collateral flow. The linear regression equations for group IV were as follows: 1 hour, y=−34.4+ 15.0x (r=.05, P=NS); 4 hours, y=−45.4+74.6x (r=.43, P=NS). The equations for group V were as follows: 1 hour, y=−45.2+102.8x (r=.41, P=NS); 4 hours, y=−18.7−10.6x (r=.04, P=NS). ANCOVA demonstrated that there was no significant difference between the two groups.

Effect of Adenosine on Myocardial Stunning

When adenosine was given continuously before, during, and after ischemia (group II), a significant and sustained attenuation of stunning was observed, which persisted throughout 4 hours of reperfusion despite the fact that adenosine infusion was discontinued after 1 hour of reperfusion. Since LAD blood flow returned to baseline values within minutes after stopping the infusion of adenosine, the improved recovery of function observed at 2, 3, and 4 hours of reperfusion cannot be attributed to increased myocardial perfusion (ie, to a Gregg phenomenon³⁵ ). Although at 30 minutes and 1 hour of reperfusion LAD flow was approximately fourfold higher in group II compared with group I (Table 4⇑), several considerations make it unlikely that this hyperemia accounted for the difference in wall thickening observed at these time points. First, in group II we found that when LAD flow returned to baseline values (≈10 minutes after adenosine was discontinued), the thickening fraction did not decrease significantly compared with the 1-hour values measured during hyperemia (Δ= −7.4±12.9%, P=NS). Second, if the enhanced wall thickening observed at 30 minutes and 1 hour in group II was due to the hyperemia, one would have expected a similar beneficial effect in group III, which exhibited a similar hyperemia at 30 minutes and 1 hour (Table 4⇑). Finally, infusion of adenosine at 4 hours failed to increase thickening fraction in the stunned myocardium despite a 4.4-fold increase in LAD flow (Fig 3⇑).

There were no other differences that could have accounted for the enhanced recovery in group II. This group was similar to groups I and III with respect to collateral flow, ischemic zone size, and most of the hemodynamic variables. For reasons that are unclear, group II had a higher heart rate and rate-pressure product at baseline and a higher rate-pressure product during occlusion (Table 4⇑), but these differences, if anything, should have exacerbated the severity of ischemia and thus should have led to an underestimation of the beneficial effects of adenosine. It should be noted that at 3 and 4 hours of reperfusion, the heart rate and rate-pressure product were similar in groups II and III (Table 4⇑), yet the recovery of function was significantly enhanced only in group II. Thus, we conclude that continuous infusion of adenosine before, during, and after ischemia reduces the severity of myocardial stunning by a mechanism that is independent of nonspecific factors, such as coronary flow or systemic hemodynamics.

Does Adenosine Attenuate Ischemic Injury or Reperfusion Injury?

In contrast to group II, group III exhibited no statistically significant improvement in the recovery of function when compared with the control group. The dose of adenosine used in group III (2 mg/min) was the same as that used in group II, and the fact that the infusion was started intracoronarily 2 minutes before the release of occlusion ensured that high local levels of adenosine were present in the LAD region at the very onset of reperfusion. Power analysis showed that there was a >90% probability that if adenosine given at reperfusion (group III) improved recovery of function to the same extent as adenosine given before, during, and after ischemia (group II), such an effect would have been detected in our study. With our sample sizes, we cannot rule out the possibility that adenosine given at reperfusion could exert a small protective effect (eg, an improvement of wall thickening of 15% of baseline values), but the physiological significance of such an effect would be questionable.

Thus, our results indicate that in the 15-minute occlusion model, administration of adenosine only during the reperfusion phase is either ineffective or much less effective than continuous administration of adenosine before ischemia, during ischemia, and during reperfusion. This implies that adenosine does not act primarily to decrease the component of “reperfusion injury” that contributes to the development of myocardial stunning.³⁶ (Such a component may still be decreased indirectly, however, as a result of the ability of adenosine to attenuate ischemic injury, since the severity of reperfusion injury in stunned myocardium appears to be modulated by the severity of the antecedent ischemic injury.^{26 27 28 29 31 36} On the other hand, the results of phase B clearly demonstrate that administration of adenosine only before ischemia fails to enhance the postischemic recovery of function, implying that the beneficial effects observed in group II were not due to the 30-minute infusion of adenosine given before coronary occlusion in this group. Taken together, the results of phases A and B indicate that in the 15-minute coronary occlusion model, the attenuation of myocardial stunning by adenosine is due to a beneficial effect that takes place mainly during the period of ischemia, not before ischemia or after reperfusion; ie, the primary mechanism of action of adenosine appears to be a reduction in the severity of the cellular perturbations that occur during ischemia (ischemic injury).

Does Adenosine Precondition Against Stunning?

In phase B, we sought to determine whether adenosine can precondition against stunning. The impetus to test this idea was provided by the demonstration that activation of adenosine A₁-receptors preconditions the myocardium against infarction^{11 13 19} and by recent reports^{20 21} suggesting that brief ischemia preconditions against stunning via an adenosine receptor–mediated mechanism. Our results clearly demonstrate that pretreatment with adenosine failed to protect the myocardium against subsequent stunning, despite the fact that adenosine was given for a relatively long time (30 minutes) and at a dose (4 mg/min) twofold greater than the dose (2 mg/min) found to be protective in group II. Thus, although adenosine preconditions against infarction,^{11 13 19} it does not precondition against the stunning induced by a 15-minute coronary occlusion. This finding further corroborates the concept that the pathogenesis and pathophysiology of myocardial stunning (a reversible form of injury) differ importantly from those of irreversible cellular injury (ie, myocardial infarction).³⁶

We elected to use a higher infusion rate in group V than in group II (4 versus 2 mg/min, respectively) to maximize the chances of detecting a preconditioning effect, if one truly existed. Since exogenous adenosine is rapidly taken up by red blood cells and endothelial cells, we reasoned that a higher infusion rate would be more likely to produce sufficient myocardial A₁-receptor activation to induce a preconditioning effect. The failure of the 4-mg/min infusion to precondition against stunning is unlikely to be due to toxicity because (1) the hemodynamic effects were very mild (only a transient decrease in diastolic arterial pressure with no change in systolic pressure or heart rate); (2) these effects resolved promptly after the infusion of adenosine was stopped, ie, before coronary occlusion; and (3) similar doses of adenosine (3.75 mg/min IC) have been shown to be protective in models of infarction.^{2 4}

Previous Studies of Whether Adenosine Treatment Attenuates Myocardial Stunning

In a previous study,³⁷ a bolus of adenosine given at the time of reperfusion was found to accelerate the rate of early contractile recovery (without affecting the final recovery) in stored or heterotopically transplanted rat hearts studied in vitro; this beneficial effect was associated with a sustained improvement in coronary flow. The difference between these findings and our present results may reflect differences in experimental preparations, species, and/or severity of ischemic injury.

Although many studies have suggested a beneficial effect of adenosine on postischemic myocardial dysfunction in a variety of experimental models,^{1 2 3 4 5 8 9 10 12 14 15} extrapolation of these results to the setting of regional myocardial stunning after reversible ischemia would be problematic. Many of these studies^{3 5 8 9 12 15} examined the effects of inhibitors of adenosine catabolism and transport, not the effect of exogenous adenosine. Some studies^{1 5 9} were conducted in isolated hearts, so that the results cannot necessarily be extrapolated to intact animals. Other studies^{2 4 10} involved models of prolonged (90- to 120-minute) ischemia resulting in subendocardial infarction, which differ from models of “pure” stunning caused by a brief (<20-minute), completely reversible ischemic insult.³⁶ Since the amelioration of postischemic dysfunction noted in these studies^{2 4 10} was associated with a decrease in infarct size, one cannot comment on the specific effects on myocardial stunning. Finally, two studies^{3 8} used a model of prolonged global ischemia with cardioplegic arrest and cardiopulmonary bypass, which again differs importantly from brief, regional ischemia.³⁶

To our knowledge, only three full-length reports^{12 14 15} have addressed the role of adenosine on myocardial stunning after reversible ischemia in vivo. Using an open-chest canine model of 15 minutes of coronary occlusion followed by 1 hour of reperfusion, Dorheim et al¹² observed a modest attenuation of postischemic dysfunction in dogs treated with erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA), an inhibitor of adenosine deaminase. Zughaib et al¹⁵ demonstrated that augmentation of endogenous adenosine with EHNA plus 6-(4-nitrobenzyl)mercaptopurine ribonucleoside (NBMPR), a nucleoside transport inhibitor, resulted in attenuation of myocardial stunning in dogs subjected to a 15-minute coronary occlusion followed by 4 hours of reperfusion. Unlike the present study, the purpose of these previous studies^{12 15} was to examine the effects of increasing the levels of endogenous adenosine by inhibiting its metabolism and transport. However, inhibitors of adenosine catabolism and transport may act not only by augmenting adenosine levels but also by preventing the generation of oxygen radicals from xanthine and hypoxanthine; therefore, it is not possible to discern whether the protective effects of EHNA and NBMPR in these studies^{12 15} were due to the former or to the latter mechanism. In open-chest dogs subjected to six 5-minute occlusion/10-minute reperfusion cycles followed by 2 hours of reperfusion, Yao and Gross¹⁴ found a significant beneficial effect when a selective adenosine A₁-receptor agonist was infused continuously before and during the six cycles and a significant exacerbation of stunning when the animals were pretreated with an A₁-receptor antagonist. Unlike the present investigation, their study¹⁴ examined a model of stunning induced by repetitive ischemia, which differs in several respects from models of stunning induced by a single coronary occlusion.³⁶ Our results expand the findings of Dorheim et al,¹² Zughaib et al,¹⁵ and Yao and Gross¹⁴ by demonstrating that (1) exogenous adenosine ameliorates myocardial stunning after a single ischemic insult, (2) this beneficial effect is sustained for 4 hours after reperfusion, (3) it requires the administration of adenosine during ischemia, and (4) adenosine does not precondition against the stunning induced by a 15-minute coronary occlusion. Since exogenous adenosine should, if anything, increase oxyradical production through the xanthine oxidase reaction, our results imply that the salutary effects observed with EHNA alone¹² and EHNA+NBMPR¹⁵ were indeed mediated, at least in part, by augmentation of adenosine.

Previous Studies of Whether A₁-Receptor Activation Preconditions Against Stunning

In a preliminary report in isolated working rat hearts, Cave et al³⁸ found that a 5-minute period of perfusion with adenosine (10, 50, and 100 μmol/L) followed by a 5-minute adenosine-free perfusion period failed to increase the recovery of function after a subsequent 20-minute period of global ischemia. Although it is unclear whether the dysfunction observed in this model represents solely stunning, the results of Cave et al are in agreement with our conclusion that adenosine pretreatment does not mitigate myocardial stunning.

The preconditioning effects of adenosine A₁-receptors against stunning in vivo are controversial. In a rabbit model of stunning induced by a 10-minute coronary occlusion, Urabe et al²¹ found that a prior 1-minute occlusion resulted in enhanced recovery of wall thickening and that the adenosine receptor antagonist 8-phenyltheophylline blocked this beneficial effect. However, other studies in dogs²² and pigs²³ have failed to demonstrate a preconditioning effect of brief (2- to 5- minute) coronary occlusions against the stunning induced by a subsequent 15-minute occlusion. In a rabbit model of stunning induced by four cycles of 5-minute coronary occlusion/10-minute reperfusion, Bunch et al²⁰ found that the adenosine receptor antagonist PD 115199 caused a progressive deterioration of function after each cycle, suggesting that the adenosine produced during the first ischemic episode preconditioned against the subsequent episodes. In contrast, using a canine model of stunning induced by six 5-minute occlusion/10-minute reperfusion cycles, Yao and Gross¹⁴ found that the adenosine receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine decreased the recovery of function after the first cycle but caused no additional deterioration after the next five cycles, suggesting that the adenosine released during the first ischemic episode was protective during that episode but did not precondition against the stunning induced by subsequent episodes.

The reason(s) for the apparent discrepancy among these previous studies^{14 20 21 22 23} is unclear but may relate to differences in experimental models and/or species. It should be noted that no previous study has examined the ability of adenosine (or adenosine agonists) to precondition against the stunning induced by a 15-minute coronary occlusion in the dog.

Mechanism of Adenosine-Mediated Protection

Adenosine exerts a number of actions that could beneficially affect myocardial stunning (reviewed in References 17, 39, and 40^{17 39 40} ). Theoretically, adenosine may reduce free radical formation by inhibiting lipolysis (and thus decreasing lipid peroxidation) and/or by decreasing auto-oxidation of catecholamines as a result of the inhibition of norepinephrine release.^{17 39} Adenosine may also protect against the effects of oxygen metabolites.⁴¹ However, our results clearly demonstrate that the primary mechanism of action of adenosine in this model of stunning is a decrease in the severity of ischemic injury, not reperfusion injury. Mitigation of ischemic injury by adenosine may take place via various mechanisms. For example, adenosine may attenuate calcium overload during ischemia by inhibiting calcium influx through the L-type calcium channel.^{17 40 42} Adenosine may also improve energy balance during ischemia by stimulating glycolysis,⁴³ by inhibiting the release of norepinephrine from nerve endings,⁴⁴ and/or by antagonizing the effect of catecholamines on contractility.^{39 45 46} In this regard, in the present study adenosine exerted a negative inotropic effect, as evidenced by a 15% to 20% decrease in wall thickening before coronary occlusion in both groups II and V (Figs 1⇑ and 4⇑), which is consistent with previous observations.^{47 48 49 50 51} Therefore, it is possible that the protection afforded by adenosine in group II was due in part to its antiadrenergic effects reflected by the decrease in ventricular contractility. Interestingly, adenosine failed to enhance recovery in group V, in which ischemia was initiated 10 minutes after discontinuation of the adenosine infusion, when the negative inotropic effect had largely dissipated (Fig 4⇑).

Our finding that adenosine acts primarily by mitigating ischemic injury is not in conflict with the notion that a free radical–mediated reperfusion injury plays a prominent role in the pathogenesis of myocardial stunning. First, it has been pointed out^{32 36 52} that free radicals are not the sole factor contributing to the pathogenesis of myocardial stunning. Second, and most important, since the magnitude of free radical generation after reperfusion in the stunned myocardium is directly related to the severity of the antecedent ischemic insult,^{26 27 28 29 31 36} any intervention that alleviates the injury incurred during ischemia would be expected to indirectly attenuate the radical-mediated injury inflicted after reperfusion.

Conclusions

In summary, the present study demonstrates that infusion of adenosine before, during, and after a 15-minute coronary occlusion results in a significant and sustained improvement in the recovery of myocardial function, which reflects a direct cardioprotective action of adenosine. This beneficial effect cannot be reproduced when adenosine is infused solely at reperfusion, indicating that in the 15-minute occlusion model, the purine acts primarily by decreasing the severity of ischemic injury. Intracoronary adenosine fails to precondition against myocardial stunning, a finding that emphasizes the differences between the pathophysiology of stunning and that of infarction. Although adenosine itself would not be a practical clinical therapy, adenosine receptor agonists or adenosine modulators may be useful in the management of myocardial stunning in patients.