Is myocardial ischemia really bad for you?

Abstract

The assessment of myocardial ischemia represents a cornerstone in our approach to coronary artery disease. Indeed many of the clinical decisions we make revolve around the results of stress testing, the assessment of coronary luminal stenoses and, more recently, fractional flow reserve measurements. Whilst the assessment of ischemia is often useful with respect to diagnosis and its treatment important in terms of symptom relief, whether ischemia directly leads to adverse cardiovascular outcomes, in particular myocardial infarction, is much more controversial. Indeed this is one of the key questions facing cardiology practice today and the focus of an ongoing multimillion-dollar study, the ISCHEMIA trial. In this editorial the authors examine some of the underlying evidence and ask the question: is ischemia itself really bad for you?

Keywords::

• atherosclerosis
• myocardial infarction
• myocardial ischemia
• percutaenous coronary intervention
• vulnerable plaque

The data that myocardial ischemia is associated with an adverse prognosis are unequivocal. Large-scale single-photon emission computed tomography (SPECT) studies in the 1990s established a stepwise increase in mortality with progressive ischemic burden [1], a finding that has been confirmed numerous times with multiple different techniques. Moreover, the Coronary Artery Surgery Study demonstrated that patients with a >50% lesion in the proximal left anterior descending artery were subsequently more likely to have an anterior myocardial infarction (MI) than those with less severe lesions [2]. On this basis, there seemed therefore a clear rationale to stenting tight obstructive lesions on coronary angiography in the belief that relieving the associated ischemia would improve not only the patient's angina but also his/her prognosis. Indeed, subsequent observational studies again using SPECT demonstrated that if a patient had an ischemic burden >10% then revascularization with either percutaneous coronary intervention (PCI) or coronary artery bypass surgery (CABG) conferred a significant survival benefit [3].

Unfortunately, when this approach has been tested in the context of large-scale randomized control trials, the results have been disappointing. The Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial compared optimal medical therapy (OMT) to PCI + OMT in 2287 stable patients with objective evidence of myocardial ischemia and significant (>70% luminal stenosis) coronary artery disease [4]. There was no benefit to PCI over and above OMT with respect to death or MI (hazard ratio for the PCI group: 1.05; 95% CI: 0.87–1.27; p = 0.62). Given that PCI is an excellent treatment for myocardial ischemia, its failure to impact on MI and death was somewhat of a surprise and questioned whether the link between ischemia and an adverse prognosis was a direct one. This was further investigated in the COURAGE nuclear substudy [5]. This examined a small subgroup of patients (n = 314) who also had SPECT imaging during the course of the COURAGE study. The primary focus of this study was to assess whether PCI + OMT or OMT was more effective in reducing ischemia. As expected, PCI emerged superior, but the authors also observed that those patients who achieved a greater than 5% reduction in ischemia using either strategy had a lower unadjusted rate of death or MI. This has been taken as evidence that patients with a large ischemic burden might still benefit from revascularization. However, this effect was no longer significant when risk-adjusted and, as clearly stated by the authors, the study was not sufficiently powered to assess such outcomes. Indeed, if we chose to examine the risk-adjusted results, as we probably should, then effectively the COURAGE nuclear substudy shows no benefit to ischemia reduction, consistent with the observations of the much larger and adequately powered parent trial.

There are several possible criticisms of COURAGE, including the randomization of patients after their diagnostic coronary angiogram, which may have led those with the most severe lesions to have been revascularized outwith the study. Moreover since its publication, stent technology has advanced, with drug-eluting stents now in widespread use, as has the functional assessment of coronary arterial stenoses with the introduction of the intra-coronary pressure wire. These provide measurements of the fractional flow reserve (FFR) that have allowed interventional cardiologists to identify with greater certainty those coronary lesions causing true hemodynamic obstruction. Indeed, FFR-guided PCI has now been evaluated in several outcome studies. The initial FAME trial compared FFR- with angiographic-guided PCI in patients with stable coronary artery disease [6]. It demonstrated that over a third of lesions thought to be obstructive on coronary angiography were in fact hemodynamically insignificant and that a PCI strategy driven by FFR resulted in a 5% reduction in death, non-fatal MI and repeat revascularization. On this basis, it was hoped that the subsequent FAME 2 study would succeed where COURAGE had failed and demonstrate a prognostic benefit to treating myocardial ischemia with FFR-guided PCI compared with OMT [7]. All patients received a drug-eluting stent and hopes were raised when the study of 1220 patients was halted early. However, whilst a significant benefit was observed in the primary outcome of death, MI and urgent revascularization (hazard ratio for PCI: 0.32; CI: 0.19–0.52; p < 0.0001), this was almost entirely driven by the need for urgent revascularization. Once again, there was no significant difference in the rates for MI or death.

One positive message that has emerged from both COURAGE and FAME 2 is the excellent outcomes associated with optimal medical therapy. However, it is interesting to note that of the available agents, aspirin and statins have demonstrated the most consistent reduction in MI rates, yet have no effect on ischemia. By contrast, specific anti-anginal therapies such as nitrates have no such impact on events in those with stable angina. In combination, the data to date would therefore suggest that while ischemia is associated with an adverse prognosis, reducing ischemia does not impact on hard clinical outcomes such as MI or death. In fact this is perhaps unsurprising. Patients with syndrome X are well known to have a good prognosis and since the 1980s, we have in fact known that the majority of MIs arise from lesions that appear non-obstructive on antecedent coronary angiography [8–10]. However, we also know that tightly stenosed lesions are more likely to result in subsequent luminal obstruction. There are two potential explanations for this apparent discrepancy. First, there is evidence that severe lesions are more likely to result in a chronic total occlusion not an acute MI, perhaps because of the collateral formation which progressive ischemia encourages [9]. Indeed, brief episodes of ischemia actually appear to reduce the volume of infarction compared with hearts not exposed to ischemic preconditioning [11]. The second explanation is that mild-to-moderate lesions more commonly result in MI simply due to sheer weight of numbers and the fact that they greatly outnumber the individually more risky severe lesions.

If not ischemia itself what factors might instead be responsible for precipitating MI? The majority of MIs are the consequence of atherosclerotic plaque rupture and the thrombotic response of the blood that ensues. Indeed, it is useful to divide the factors driving MI into those governing the tendency to rupture and those related to the thrombotic response of the coronary circulation. Plaques at risk of rupture, the so-called vulnerable plaque, have certain pathological characteristics that include a thin fibrous cap, a large necrotic core, inflammation, microcalcification, positive remodeling and angiogenesis. A recent post-mortem study of 295 victims of sudden cardiac death compared the features of ruptured atherosclerotic plaques with remote non-culprit lesions, identifying a thin fibrous cap <55 µm as the most closely associated characteristic, followed by increased macrophage density and the presence of a large necrotic core [12].

These features offer potential imaging targets for identification of vulnerable atherosclerotic plaque. In the recent PROSPECT study, virtual histology (VH)-intravascular ultrasound was used to characterize VH-thin capped fibroatheroma: lesions with both a thin fibrous cap and large necrotic core [13]. While the majority of events arose from these plaques, of the 595 VH-thin capped fibroatheromas identified only 6 ultimately resulted in MI, hinting at one of the potential limitations associated with current vulnerable plaque identification. Autopsy studies have demonstrated plaque rupture in the coronary arteries of approximately 10% of subjects who have died from non-cardiac causes [14], while a similar prevalence has been observed in angioscopy studies of patients with stable coronary artery disease [15]. Thus, coronary plaque rupture would appear to be a fairly frequent event that only rarely results in clinical sequelae and more commonly is believed to account for the staccato pattern of plaque growth that accompanies successive cycles of rupture and healing [16]. Vulnerable plaques are therefore likely to also be common, only rarely precipitating MI but more commonly contributing to the progression of coronary atheromatous lesions.

The factors governing thrombogenicity and the magnitude of the thrombotic response to plaque rupture must therefore also be carefully considered. Whilst in part this may come back to plaque composition, with larger more inflamed necrotic cores potentially stimulating a more extensive thrombotic response, other factors are also clearly at play. For example, the prothrombotic states induced by smoking and the peri-operative state are well known to be associated with increased rates of MI. However, in fact thrombogenicity is under the influence of a huge array of factors including stress, toxins, diet, pollution, inflammation, medication and genetic influences, which ensure that the there is almost minute-to-minute variation in its activity.

So ultimately MI results from a perfect storm scenario where multiple prorupture and prothrombotic events must align at precisely the same time in order for a clinical event to present [17]. Accurate prediction of MI on a lesion-by-lesion basis is therefore likely to be extremely difficult and lesion-specific therapies unsuccessful. An alternative approach is to estimate risk on an individual patient basis, and then target-intensive medical therapy to those at greatest need, the so-called vulnerable patient approach. Essentially this comes down to probabilities. The more high-risk plaques a patient has and the more sticky his/her blood then the increased chances of an event. Non-invasive imaging lends itself to this approach, even crude assessments of plaque burden such as the calcium score offer the most powerful contemporary prediction of myocardial events [18] and suggest that more refined techniques might ultimately provide accurate prognostic information. In particular, two approaches stand out: computed tomography (CT) coronary angiography and PET using 18F-NaF. The former provides increasingly detailed assessments of plaque morphology and burden [19], while the latter informs about plaque vulnerability and metabolic activity [20]. Moreover, with modern technology these two techniques can be performed simultaneously as part of the same PET/CT scan: an approach that holds real promise in identifying high-risk patients in whom aggressive medical therapy might help prevent MI [21].

So what about myocardial ischemia, could it too just be a surrogate marker of plaque burden? We know from VH-intravascular ultrasound and CT that when you see an isolated severe lesion on angiography in reality there is often eccentric positively remodeled plaque affecting large portions of that artery and indeed the coronary vasculature as a whole. This hypothesis is therefore certainly plausible. It would also explain why ischemia is associated with an adverse outcome, while ischemia-based therapies have failed to improve prognosis. Moreover, it would offer explanation for the 5-year MI rates in the recent SYNTAX and FREEDOM trials that were double in patients with multi-vessel disease undergoing PCI versus CABG [22,23]. While modern PCI offers equivalent reductions in ischemia burden to CABG, CABG has the added advantage of protecting against plaque rupture in all the adjacent non-obstructive lesions within that vessel. In effect, CABG also provides a treatment for plaque burden: an effect that appears to translate into improved clinical outcomes.

Conclusion

In conclusion, while myocardial ischemia is associated with an adverse prognosis, strategies aimed at its reduction have consistently failed to impact on the subsequent risk of death or MI. This would indicate that myocardial ischemia is not directly responsible for subsequent MI. Instead, ischemia might simply act as a surrogate for plaque burden. This hypothesis would explain the results of COURAGE, FAME 2 and the SYNTAX trials and suggests that we should move away from an ischemia-based paradigm for the management of coronary atherosclerosis.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending or royalties.

No writing assistance was utilized in the production of this manuscript.