Where does the Occluded Artery Trial leave the late open artery hypothesis?

Abstract

As of April 2007 the early open artery hypothesis is alive and well, but the late open artery hypothesis is adrift. For the foreseeable future, stable patients with persistent occlusion of the infarct artery late after myocardial infarction, and without severe ischaemia or uncontrollable angina, should be managed initially with optimal medical treatment alone, and not with percutaneous coronary intervention. Efforts should focus on establishing reperfusion earlier, including reducing the time to patient presentation.

Basic and clinical research elucidated the concept that an epicardial coronary artery occlusion leads to an advancing wavefront of necrosis, which progresses from endocardium to epicardium.¹,² This wavefront also advances radially and leads to a border zone of mixed infarcted, viable and dysfunctional, and viable but ischaemic myocytes. Coronary reperfusion before the wavefront of necrosis is complete diminishes infarct size, improves regional and global left ventricular (LV) function, and reduces mortality.

It was not until the development of pharmacotherapy to open infarct arteries, however, that studies like the Western Washington intracoronary streptokinase trial could demonstrate that interruption of the wavefront of myocardial necrosis could also take place in humans.³ At present, stenting of patients with ST elevation myocardial infarction leads to a patent artery in over 90% of patients.⁴,⁵ From the outset of these efforts at reperfusion it had been clear that the earlier after occlusion reperfusion occurred, the more complete the salvage of myocardium. Other lines of evidence, however, hinted that late reperfusion, too late for myocardial salvage, might also be advantageous.⁶,⁷,⁸

Late reperfusion

Experiments in late reperfusion of the experimental infarct suggested that preservation of LV geometry⁶,⁷ was possible, even in the absence of myocardial salvage. The improved survival of patients in thrombolytic trials appeared out of proportion to the improvement in ejection fraction (EF), suggesting that late infarct artery patency carried its own incremental benefit, above and beyond that of myocardial salvage. The total benefit of reperfusion was thought to be the sum of the benefit of myocardial salvage plus the benefit of having an open artery after myocardial infarction (MI). Retrospective analyses of clinical trials like SAVE suggested that even when a severe infarct occurred and differences in EF were controlled for, leaving the hospital with an open artery predicted a good prognosis.⁸ The concept that late opening of an occluded artery might be beneficial remains appealing, because recent data suggest that many patients with acute MI fail to receive timely reperfusion treatment either because of late or atypical presentation.

A number of mechanisms were proposed for this finding, and the most plausible of these centred on stabilisation or prevention of LV remodelling. Other lines of discussion focused on electrical stabilisation of an arrhythmogenic substrate, recruitment of hypofunctional border‐zone myocardium, and provision of collaterals to other coronary beds that might themselves become ischaemic.⁹,¹⁰,¹¹ Thus, more than a decade ago, multiple investigators called for a clinical trial to test the veracity of the late open artery hypothesis.

Four small trials of percutaneous coronary intervention (PCI) in late post‐MI in patients with occluded infarct arteries were designed, executed, and published.¹²,¹³,¹⁴,¹⁵ These trials had variable results, ranging from an optimistic reduction in LV remodelling and events, to, paradoxically, an increase in LV volumes and clinical events with PCI. Thus, as so often happens in the scientific evaluation of complex medical subjects, small‐scale studies that focus on surrogate physiological end points may give incomplete answers, but do prepare the field for a definitive trial.

Occluded Artery Trial (OAT)

The recently published OAT is the only robust trial in the field.¹⁶,¹⁷ OAT was sponsored by the National Heart, Lung and Blood Institute, and brought together 217 sites from 24 countries. A total of 2166 patients enrolled between February 2000 and December 2005. Enrolled patients had sustained an MI within 28 days, with certain high‐risk features, notably a proximal occlusion of a major epicardial artery, or LV dysfunction with an EF <50. The time window permitted patient enrolment just over 24 hours after MI. Patients underwent coronary angiography and, if an eligible occluded artery was found and they were clinically stable without three‐vessel disease, they were randomly assigned to a PCI‐stent, or to leaving the artery occluded without further intervention. We required that patients be receiving optimal medical treatment regardless of the treatment group to which they were assigned. Patients were followed up for an average of 1059 days.

There was no benefit from performing PCI (17.2% 4‐year rate of death, MI or class IV heart failure for PCI vs 15.6% for optimal medical treatment only), and a trend towards excess early and late reinfarctions in the PCI‐treated group was noted (fig 1). Within this large trial, several smaller mechanistic ancillary studies confirmed that sustained epicardial coronary patency was present¹⁸ and there was often retained infarct‐zone viability in enrolled patients. A subset of 332 OAT patients underwent repeat angiography at 1 year after study entry. In the PCI group 87% had sustained patency, whereas in the medical group 25% had recanalised. Moreover, resting sestamibi scanning in another small subgroup confirmed retained infarct‐zone viability (>40% of peak tracer uptake) in 69% of 124 patients.

Figure 1 Kaplan–Meier curves for the Occluded Artery Trial primary end point: death from any cause, non‐fatal reinfarction, or NYHA class IV heart failure requiring admission to hospital or time in a short‐stay unit. Reproduced with permission from Hochman JS, Lamas GA, Buller CE, et al. Coronary intervention for persistent occlusion after myocardial infarction. N Engl J Med 2006;355:2395–407.¹⁶ Copyright © 2007 Massachusetts Medical Society. All rights reserved.

Is the open artery hypothesis still viable?

Therefore, although the benefit of opening the infarct artery early after infarct onset is not in doubt,¹⁹ it has now become necessary to ask the question posed: where does OAT leave the late open artery hypothesis? Let us review the critiques of OAT, and critically determine whether the open artery hypothesis has been laid to rest.

1. The OAT enrolled a low‐risk population; patients at high risk, particularly those with low EFs and proximal left anterior descending coronary artery occlusions, were not included.

The annualised event rates for OAT reflect the benefits of modern post‐MI care as well as the selection of patients who survived the first few days of the MI and were clinically stable. In fact, an annualised event rate of 4% is not low and the rate was below the expected range only for the heart failure end point. We did, however, prospectively identify high‐risk subgroups such as those with a low EF and left anterior descending infarct artery, where the risk of events was high, and there was still no suggestion of benefit.

2. The most likely benefits of late opening of the infarct artery are in attenuation of LV remodelling. It would take a long time for the clinical benefit to become apparent.

The angiographic ancillary study assessed LV remodelling 1 year after the index infarction.¹⁸ The EF increased to a similar degree in both groups. A multivariable analysis in a subgroup of patients with paired volumetric ventriculograms found that PCI patients tended to have less LV dilatation in diastole (p = 0.02) and systole (p = 0.04). Thus we are left with a hint that there may be opposing processes at work. Perhaps attenuation of LV remodelling is counterbalanced by adverse consequences of myocardial injury from procedure‐related embolisation and both early and late reinfarction.

3. The PCI techniques were not the most modern treatments, because few drug‐eluting stents were used.

Recent meta‐analyses comparing bare metal with drug‐eluting stents have been published.²⁰,²¹,²²,²³ There is no evidence that drug‐eluting stents reduce the combined end point of death or MI. Furthermore, the LV remodelling benefit in the observational studies was demonstrated in patients with residual post‐MI stenoses, so restenosis should not adversely affect the potential protective effect of patency. Therefore, the stent type is not relevant to the interpretation of the primary end point.

4. In OAT, arteries were opened too late after the infarct; there might have been a benefit if the arteries had been opened earlier.

In OAT, the median time from MI to randomisation was 8 days, and patients assigned to PCI had their intervention within 24 hours after. By protocol, however, patients could be randomised just over 24 hours after MI. Subgroup analyses did not suggest any benefit from PCI at any time within the OAT time window.⁵

5. Does this mean that an occluded infarct artery should never be opened late?

The results of the OAT should not be extrapolated to clinical situations and patient subsets that were not included. For example, patients with cardiogenic shock revascularised up to 36 hours after onset of MI and 12 hours after shock onset derived a survival advantage over initial medical stabilisation.²⁴ Additionally, rescue PCI after failed fibrinolytic treatment, or primary PCI up to 12–24 hours after the event, is recommended for unstable or high‐risk patients with ST elevation myocardial infarction.¹⁹ Those with severe ischaemia were excluded; there was no suggestion of benefit for the subset with mild to moderate ischaemia on a stress test. In stable patients the efficacy of an initial intensive medical approach, even in the presence of recorded ischaemia, is supported by the findings of the COURAGE study.²⁵

Conclusions

The critiques above point out important considerations in the interpretation of OAT, but none temper its results. We respectfully submit that today, in April 2007, the early open artery hypothesis is alive and well, but the late open artery hypothesis is adrift.

Therefore, today and for the foreseeable future, stable patients, such as those enrolled in the OAT, with persistent occlusion of the infarct artery late after MI, and without severe ischaemia or uncontrollable angina, should be managed with an initial strategy of optimal medical treatment alone, and not with PCI. Efforts should focus on establishing reperfusion earlier, including reducing the time to patient presentation.

G Lamas

J Hochman

Abbreviations

EF - ejection fraction

LV - left ventricular

MI - myocardial infarction

OAT - Occluded Artery Trial

PCI - percutaneous coronary intervention