Limitation of Infarct Size and the Open Artery Hypothesis

Dr Rutherford asks: How did your interest in MI begin?

Dr Braunwald replies: I became aware of the importance of AMI in 1951 during an elective in cardiology during my senior year at the New York University School of Medicine. At the time, AMI was by far the most common cause of death in adults in the United States. Of the patients who survived to be admitted to hospitals (no one knew how many AMI patients failed to reach hospital emergency departments), about one third died before discharge, either suddenly, presumably from an arrhythmia, or secondary to pump failure. Of the patients who were discharged alive, about one quarter died within a year, most frequently as a consequence of heart failure or recurrent MI. Pathologists had taught that AMI was usually associated with coronary thrombosis. Indeed, during my cardiology elective, Herrick’s classic article describing AMI was required reading. Its title, “Clinical Features of Sudden Obstruction of the Coronary Arteries,” showed how these 2 conditions (coronary occlusion and AMI) were intertwined.¹ My direct involvement in the care of several patients who did not survive the acute event left a lasting impression on me and stimulated my interest in a career in cardiovascular research, with a special emphasis on AMI.

In 1955, I began my third fellowship in cardiovascular research in the Laboratory of Cardiovascular Physiology headed by the noted physiologist Stanley J. Sarnoff at the (then) National Heart Institute (now the National Heart, Lung and Blood Institute). I participated in the design and conduct of experiments in isolated dog hearts to identify the determinants of myocardial oxygen consumption (MVo₂) and of coronary blood flow.^2,3 We viewed these as interesting, perhaps important, physiological studies, but we did not grasp their potential clinical implications.

In the early 1960s, advances in continuous electrocardiographic monitoring, automatic alarms for arrhythmias, and closed chest resuscitation led to the development of coronary care units, which proved to be very effective in the prevention of primary arrhythmic deaths. This notable advance reduced the hospital mortality of AMI by about half, leaving pump failure as the major cause of in-hospital mortality. Pump failure, the thinking went, was a function of the extent and severity of the imbalance between myocardial O₂ supply and demand. As a consequence, the determinants of MVo₂ and of coronary blood flow rapidly assumed great clinical relevance. Therefore, when I was assigned my own research laboratories in the institute in 1963, together with John Ross, Edmund Sonnenblick, and James Covell, we redoubled our efforts in studying MVo₂. By 1968, we had concluded that 3 features of cardiac activity—myocardial tension development, frequency of contraction, and contractility—account for >90% of the MVo₂.^4,5

How did your discovery of the salutary influence of carotid sinus nerve stimulation on MVo₂ translate to patient care?

In 1947, Samuel Levine and Proctor Harvey at the (then) Peter Bent Brigham Hospital observed that massaging the carotid sinus could relieve episodes of angina pectoris. In the late 1960s, the management of angina pectoris consisted primarily of sublingual nitroglycerin and oral β-blockers. On the basis of Levine and Harvey’s observations, we hypothesized that an additional, perhaps more potent approach to restoration of the balance between myocardial O₂ supply and demand might be achieved by implanting a transcutaneous carotid sinus nerve stimulator, which could be activated externally by the patient at times of need. In 1967, my first wife, Nina Braunwald,* Steven Epstein, and I accomplished this, and we observed that carotid sinus nerve stimulation was indeed successful both in relieving severe angina and, when used prophylactically, in preventing exertion-induced ischemia.⁶ While we were conducting additional clinical studies with the device to obtain regulatory approval, Rene Favaloro at The Cleveland Clinic successfully performed cardiac revascularization using a coronary vein graft to bypass an occluded right coronary artery. Enhancing myocardial O₂ supply by this direct approach proved much more effective in preventing myocardial ischemia than carotid sinus nerve stimulation. Therefore, we discontinued our efforts with the latter, but not before I experienced a serendipitous event that would have a profound effect on my subsequent professional life.

Describe this serendipitous event and its consequences

In early 1968, one of our patients with the carotid sinus nerve stimulation system in place was admitted to the National Heart, Lung and Blood Institute during the early hours of a STEMI. His ECG was recorded when the stimulator was in both the on and off modes on several occasions. Subsequent review showed that the extensive ST-segment elevations in the absence of stimulation were markedly reduced during activation. I was surprised by this observation because I (and others at the time) had thought of thrombotic coronary occlusion and subsequent MI as sudden events, as had been reflected in the title of Herrick’s article.¹ However, the observations in this patient suggested that in the early hours of an AMI, a reduction of myocardial O₂ demands, as occurred with carotid sinus nerve stimulation, might actually reduce what had, until then, been considered to be sudden, irreversible ischemic myocardial damage. If so, it might be possible in patients with AMI to protect some of the ischemic but still viable myocardium by quickly restoring the balance between myocardial O₂ supply and demand.

What were the next steps your laboratory took to investigate salvage of ischemic myocardium?

After moving to the University of California, San Diego in 1968 with several colleagues from the National Institutes of Health, and joined by a talented postdoctoral fellow, the late Peter Maroko, we returned to the dog TIMI laboratory to test the hypothesis that the severity of ischemia and hence of ischemic damage produced by coronary occlusion could be reduced by altering the imbalance between myocardial O₂ supply and demand.^7,8 We related acute ischemia by correlating epicardial ST-segment alterations with myocardial necrosis by measuring myocardial creatine phosphokinaseconcentrations, a marker of myocardial viability, and by histology in the heart excised 24 hours after the intervention. We observed that increased myocardial O₂ supply by reperfusion of the ischemic zone as late as 3 hours after coronary occlusion could salvage jeopardized myocardium⁹ (Figure 1). In the presence of coronary occlusion, tachycardia induced by electric stimulation and β-adrenergic stimulation with isoproterenol, both of which augment MVo₂, increased ischemic damage. In contrast, intravenous propranolol, a β-blocker that reduced MVo₂, also reduced the extent of myocardial necrosis. From these studies, we proposed that severely ischemic myocardium destined to become necrotic could be salvaged by opening the IRA (hence the open artery hypothesis) but also by reducing MVo₂. We concluded the article as follows:

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Comparison between the relationships between ST-segment elevation 15 minutes after occlusion and creatine phosphokinase (CPK) activity 24 hours later at the same myocardial sites. Bottom (solid) line indicates the control group; middle (dotted) line, after administration of a combination of glucose-insulin-potassium and propranolol starting 3 hours after occlusion; and top (dashed) line, reperfusion of coronary arteries 3 hours after occlusion. The slope of the lower line is statistically different from the slope of the other lines, showing lesser CPK depression resulting from glucose-insulin-potassium and propranolol administration or reperfusion at 3 hours after occlusion than in the control group. Experiments in 198 sites on 26 dogs. Reprinted with permission from Maroko PR, Braunwald E. Modification of myocardial infarction size after coronary occlusion. Ann Intern Med. 1973;79:720–733.⁹ http://annals.org/article.aspx?articleid=688077. Copyright © 1973, American College of Physicians.

Of critical interest, from the clinical point of view, is the finding that the severity and extent of myocardial ischemic injury resulting from coronary occlusion could be radically altered not only by pretreatment of the animal but also by an appropriate intervention as late as 3 hours after the coronary occlusion. This suggests that measures designed for reduction of myocardial oxygen demands and improvement of coronary perfusion, when effected promptly after a patient has been brought to a hospital, might potentially reduce the ultimate size of the infarction.⁸

How did your investigations of clinical myocardial reperfusion evolve?

We were excited by these observations in open chest, anesthetized dogs and were anxious to extend them to patients undergoing AMI in the hope of limiting infarct size. Although the most effective intervention in our experiments had been myocardial reperfusion, in 1971, there was no practical way to reperfuse ischemic myocardium safely and rapidly during an evolving infarction. However, the picture changed dramatically in 1975 with the demonstration by Chazov and colleagues¹⁰ in a patient experiencing an STEMI that reperfusion could be accomplished by infusing streptokinase directly into the occluded IRA. After we moved to Boston and began treating patients with STEMI with this approach, it became critically important to test the hypothesis that opening the occluded IRA actually limited infarct size. Therefore, in 1981, John Markis, William Grossman, and I studied patients with STEMI who had just undergone thrombolytic reperfusion. We injected thallium-201 into the IRA that had been opened with intracoronary streptokinase. The uptake of thallium, as observed on scintigraphy, required myocardial viability. We found that the jeopardized myocardium that initially displayed no uptake of thallium-201 had in fact been salvaged by the reperfusion.¹¹ We considered this to be the evidence in patients that we had been after since 1968, when we first considered that infarct size might be limited in patients by opening an occluded IRA, and believed that it strengthened the open artery hypothesis. However, in the early 1980s, this hypothesis was not yet widely accepted, nor was the critical importance of infarct size to ventricular function and patient survival widely appreciated. Perhaps this was due to the finding that contractile dysfunction in severely ischemic myocardium was not perceived immediately to be reversed on opening of the IRA. In 1982, Robert Kloner and I called this “prolonged, postischemic ventricular dysfunction” and “myocardial stunning.”¹² We proposed that “when the myocardium is frequently stunned it may exhibit chronic postischemic left ventricular dysfunction,”¹² which could be responsible for some cases of ischemic cardiomyopathy that could be reversed by reperfusion.¹³

In a rat model of STEMI, developed by Janice and Marc Pfeffer and Bob Kloner, we demonstrated the inverse relation between infarct size and left ventricular ejection fraction^14,15 (Figure 2). The Pfeffers extended this observation by showing that the mortality of rats 1 year after infarction varied directly with infarct size.¹⁶ We then turned our full attention to clinical efforts to reduce infarct size.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Scatterplot showing relationship between ejection fraction index and histologically determined myocardial infarct size in rats with healed myocardial infarction (n=104; P<0.001). Reprinted from Pfeffer and Braunwald.¹⁵

How did the TIMI Study Group start, and where did it lead?

In 1984, I was invited by the National Heart, Lung and Blood Institute to chair a clinical trials group, the TIMI Study Group, to optimize thrombolytic therapy. This was an exciting time in the field. Schröder et al¹⁷ had just shown that thrombolysis could be achieved by the rapid intravenous injection of large doses of streptokinase, enhancing the likelihood that myocardial reperfusion could be accomplished relatively easily in almost any hospital emergency room. As TIMI began its first study, the first 2 mega-trials of streptokinase in AMI, the GISSI¹⁸ and ISIS 2¹⁹ trials, were underway but had not been completed. In the first TIMI trial, we compared the efficacy and safety of the intravenous administration of 2 thrombolytic drugs (streptokinase and tissue-type plasminogen activator, one of the first agents produced by the applications of recombinant biology to drug development) in patients with STEMI. The trial was halted early by the Data Safety Monitoring Board because of the clear superiority of tissue-type plasminogen activator in achieving full reperfusion, which we referred to as TIMI 3 flow.²⁰ In patients in whom the IRA had been opened, regardless of which thrombolytic agent was used, left ventricular function measured a week later was improved.²¹ Moreover, mortality at 1 year was ≈40% lower in patients who had an open compared with an occluded IRA 90 minutes after randomization,²² providing further support for the open artery hypothesis.

From the TIMI 2 trial, we learned that early intravenous injection of a β-blocker reduced the risk of reinfarction and recurrent ischemia,²³ a finding consistent with the cardioprotective action of β-blockade observed in our dog experiments 2 decades earlier.^7,8 However, the trial was underpowered to examine the effects on mortality. The results of the TIMI 3 trial on patients with non–ST-segment–elevation acute coronary syndromes came as something of a surprise. On arteriography, we observed coronary obstruction consistently but clear evidence of thrombus in only about one third of these patients.²⁴ Tissue-type plasminogen activator had no clinical benefit; it actually was found to be associated with numeric increases in intracranial hemorrhage, death, or AMI. However, we did observe that an invasive approach with balloon angioplasty to open the IRA (it was before the introduction of stents) resulted in a reduction in rehospitalization for recurrent severe myocardial ischemia.²⁵ It was not until we carried out TACTICS-TIMI 18 in 2001 that we observed clearly that an invasive strategy with early coronary arteriography followed by reperfusion resulted in improved clinical outcomes in patients with non–ST-segment–elevation acute coronary syndromes.²⁶ So, although achieving reperfusion by thrombolysis was not helpful in such patients, the clinical benefits derived from mechanical myocardial reperfusion were certainly consistent with the open artery hypothesis.

The most authoritative validation of the hypothesis came from the GUSTO investigators, who studied 2431 patients with STEMI by coronary arteriography and stated the following in 1993:

...the hypothesis that more rapid restoration of flow through the IRA after the initiation of thrombolytic therapy may better preserve left ventricular function and improve survival among patients with AMI has been controversial...This study [the GUSTO trial and its angiographic substudy] suggests...that more rapid and complete restoration of coronary flow through the IRA results in improved ventricular performance and lower mortality among patients with AMI.²⁷

As techniques for achieving reperfusion of jeopardized myocardium improved progressively, from streptokinase to tissue-type plasminogen activator, to tenecteplase,²⁸ to balloon angioplasty, to bare metal and several generations of drug-eluting stents, combined with an anticoagulant and dual antiplatelet therapy,²⁹ and as operator experience accumulated, there has been steady improvement in clinical outcomes in patients with STEMI. In our dog experiments in 1971, we suggested that the timing of myocardial reperfusion is important.^8,9 This concept was subsequently quantified by the important experiments by Reimer et al.³⁰ Enormous efforts worldwide were made to shorten the interval between the onset of symptoms of AMI and reperfusion. Hence, the credo for the treatment of STEMI has become “Open the IRA, and open it just as quickly as possible.” Although PCI is certainly superior to thrombolysis, when rapid PCI is not available or feasible, the vessel may be opened with a thrombolytic, begun as soon as possible, followed by PCI. This approach has been referred to as pharmacoinvasive therapy.³¹

What challenges remain?

Despite the salutary results of contemporary therapy for STEMI, several problems still require attention. The first is myocardial reperfusion injury. In 1985, Kloner and I called attention to the fact that reperfusion of jeopardized myocardium is a double-edged sword resulting simultaneously in myocardial salvage but also accelerated injury.³² Three decades later, prevention of the latter effect remains a challenge.³³

The second is the late open artery hypothesis. The benefit of late opening of the IRA, that is, when it can no longer be expected to salvage acutely ischemic, jeopardized myocardium, is sometimes referred to as the late infarct artery hypothesis. Late reperfusion of infarcts has been shown in rats, pigs, and dogs to produce thicker infarct scars and to limit, often even prevent, post-AMI infarct expansion.^34,35 We and others have suggested that late patency of the IRA permits additional collateral blood flow to the peri-infarct zone, may provide a “scaffold” that supports the surrounding myocardium, and may increase electric stability.^36–38 It has been observed that post-AMI patients with occluded IRAs may exhibit progressive ventricular dilatation, whereas patients with similar ventricular volumes but patent IRAs do not.³⁹ Although there are strong suggestive findings for these and other potential benefits of late reperfusion, rigorous evidence to support the clinical benefit of a late open artery is not available.

To address this question, the OAT, an National Institutes of Health–supported trial led by Hochman, enrolled patients with persistent total occlusion of the IRAs 3 to 28 days after AMI. All received optimal medical therapy; one half were randomized to PCI with stenting. Randomization to PCI did not reduce the primary end point (death, reinfarction, or heart failure.)⁴⁰ In an angiographic substudy, long-term (1 year) patency was present in 83% of the patients randomized to PCI compared with 25% of those initially assigned to medical therapy. The authors concluded that, on the basis of their angiographic findings “and the lack of clinical benefit over an average of 3 years follow-up, routine PCI is not recommended for persistent IRA occlusion in stable coronary artery disease.”⁴¹

I agree with this conclusion of this well-conducted trial and its angiographic substudy. However, it must be appreciated that OAT tested a therapy (PCI) but did not test the late open artery hypothesis directly. Indeed, in the angiographic substudy, the investigators reported that “regardless of initial treatment assignments, patients with a patent IRA at follow-uphad a greater increase in left ventricular ejection fraction than those with an occluded artery (absolute difference of 3.0%; P=0.003).”⁴¹ This observation actually supports the hypothesis. Subsequently, the authors of a meta-analysis of 5 trials of late PCI for totally occluded IRAs (including the above-mentioned substudy⁴¹) concluded that “a statistically significant improvement in cardiac function and a reduction in adverse remodeling...supports [sic] the pathophysiologic value of the open-artery hypothesis.”⁴²

Operator experience, equipment, and concomitant antithrombotic treatment to support percutaneous opening of totally occluded coronary arteries have all improved substantially in the 13 years since patients entered the OAT trial.⁴³ I believe that we have not heard the last word about the late open artery hypothesis.

After decades of preclinical research, after hundreds of thousands of patients with AMI have participated in clinical trials of various therapies, and after many hundreds of articles on the subject of the treatment of AMI have been published, the question “Does infarct size matter?” was recently posed again.⁴⁴ With all that we have learned about AMI and the many factors that influence outcome, do measurements of infarct size still provide information of clinical importance? Yes, infarct size still matters. In a recent analysis by Stone et al⁴⁵ of 2632 patients with AMI in 10 trials, infarct size measured within 30 days after early PCI remained a significant determinant of clinical outcome after hospital discharge. There is obvious room for improvement of post-AMI outcomes, and one approach could be early intravenous β-blockade, which I consider the third challenge.

As already indicated, our results with this intervention were encouraging both in the dog model^7,8 and in patients in the TIMI 2 trial.²³ Our work and that of others, including the excellent work of Reimer et al,⁴⁶ have indicated that the limitation of infarct size by β-blockade is time dependent (just as it is for myocardial reperfusion). Early intravenous administration should not be confused with long-term oral β-blockade begun after the infarction is complete. The latter has been shown repeatedly to reduce subsequent mortality and recurrent infarction, and oral β-blockade has become a cornerstone of post-AMI therapy. Until recently, despite a large number of trials, the jury has still been out on the benefit of early intravenous β-blockade,⁴⁷ perhaps because of the varying definitions of early used by different investigators.

This important issue has recently been readdressed in a series of rigorous studies by Ibanez and associates,⁴⁸ first in a pig model of STEMI and then in a double-blind trial of patients with anterior wall STEMI. Intravenous metoprolol or placebo was administered before reperfusion. Magnetic resonance imaging performed at 1 week and repeated at 6 months showed reduced IS, higher left ventricular ejection fraction, and reduced admissions for heart failure.⁴⁹ These and related studies using esmolol, the ultra– short-acting β-blocker,⁵⁰ have revived our suggestion that early intravenous β-blockade in suitable patients with AMI can limit infarct size and provide clinical benefit.⁵¹ This rejuvenated hypothesis should now be tested in a large trial on clinical outcomes in patients with STEMI treated with early PCI, examining patients with large, anterior wall infarcts with adequate doses of a β-blocker administered at least 15 minutes before reperfusion.

Do you have any final reflections on your career-long pursuit of this research question?

Looking back, “Limitation of Infarct Size and the Open Artery Hypothesis” has been the most gratifying of my adventures in cardiovascular research.⁵² This short history suggests that we simply moved from one victory to another. It has, in fact, not always been a smooth ride. There have been some bumps in the road, with a few false starts, blind alleys, errors (some on my part), missed opportunities, and plenty of sniping from competitors, but all of this is fairly typical of clinical science in a rapidly changing field with high stakes.

I have had 3 strokes of good fortune in my professional life. The first is that as a trainee I was stimulated to work in cardiovascular science at a time when this field was about to enter its golden period, during which the opportunities in translational research were immense. The second is that I landed in the right place at the right time. The National Heart, Lung and Blood Institute, University of California, San Diego, and Harvard/Brigham each provided me with the intellectual and clinical environments required for the type of research that our group conducted. The leaders of these institutions gave me the time and opportunity to conduct research even though my “day jobs” were largely administrative. The third and perhaps most important stroke of fortune has been the opportunity to work with dozens of brilliant, enthusiastic, and dedicated colleagues and trainees on this research, who deserve great credit for what has been accomplished. They are too numerous to mention, but in particular, this work would not have been possible without the stimulation of my mentor, Stanley J. Sarnoff, and the contributions of Peter R. Maroko^*, Robert A. Kloner, James E. Muller, Marc A. Pfeffer, John Ross, Jr, Burton E. Sobel^*, Edmund H. Sonnenblick^*, and in the past 3 decades, of course, my colleagues in the TIMI Group.