In ISCHEMIA, Stable CAD Treatment Effect Hinged on MI Definition

A new analysis shows that spontaneous MIs were reduced with invasive strategy and these events were linked with CV death.

In ISCHEMIA, Stable CAD Treatment Effect Hinged on MI Definition

A new analysis of the ISCHEMIA trial reveals its results are highly contingent on how myocardial infarction is defined, with investigators reporting a significant treatment difference between the invasive and conservative strategies when counting MIs using the more-sensitive biomarker: cardiac troponin (cTn).

When investigators analyzed events based on the secondary MI definition, which assessed PCI- and CABG-related infarctions using cTn levels as opposed to CK-MB, the conservative strategy of guideline-directed medical therapy was associated with a significantly lower risk of the study’s primary endpoint, a composite of cardiovascular death, MI, hospitalization for unstable angina, heart failure, or resuscitated cardiac arrest, and a lower risk of cardiovascular death or MI, which was a secondary endpoint.

However, the difference between the treatment approaches was driven by a higher number of procedural MIs with the invasive strategy. Type 1 MIs, or spontaneous infarctions, were significantly less frequent with the invasive revascularization strategy of PCI or CABG surgery regardless of the MI definition used in the trial, and these nonprocedural MIs were associated with worse prognosis.

“In contrast to procedural MIs, spontaneous type 1 MIs were more strongly associated with an increased risk of death and were significantly reduced in patients randomized to the invasive strategy,” write lead author Bernard Chaitman, MD (Saint Louis School of Medicine, St. Louis, MO), and colleagues in their paper published online December 3, 2020, in Circulation.

The researchers point out that COURAGE, BARI 2D, and FAME 2 did not show reductions in spontaneous MIs with PCI compared with optimal medical therapy, noting that “the reasons why PCI conferred protection against type 1 MI is not clear since spontaneous MI events occur in non-stented vessels or non-flow-limiting lesions.”

It’s possible that greater use of dual antiplatelet therapy (DAPT) contributed to the reduction in events, although in most patients DAPT is stopped after 12 or 18 months, or that the decrease stems from ascertainment bias (ie, doctors more likely to admit a patient with chest pain/diagnose MI if they were treated conservatively), according to the investigators.

Sanjay Kaul, MD (Cedars-Sinai Medical Center, Los Angeles, CA), who wasn’t involved in the study, also zeroed in on this reduction in spontaneous MIs with the invasive strategy. While such a reduction has been seen previously with CABG surgery, which can bypass vulnerable plaques, it has never been reported with PCI. Without prior supportive evidence or a biologically plausible mechanism, it’s a “shaky finding,” he said.

Additionally, Kaul pointed out that 20% of patients in the invasive-strategy arm didn’t undergo coronary revascularization because they had no obstructive CAD, and yet these patients also had a reduction in spontaneous MIs. As Kaul wryly observed: “Simply doing a coronary angiogram is unlikely to impact outcomes.”

Speaking with TCTMD, Ajay Kirtane, MD (NewYork-Presbyterian/Columbia University Irving Medical Center, New York, NY), took a different tack. When ISCHEMIA was presented in 2019 at the American Heart Association Scientific Sessions, there was uncertainty as to why the late MI rates favored the invasive strategy, but this new analysis provides some clarity, he said.

“For me, what this new study shows is that the invasive strategy—whatever that invasive strategy might be—is associated with a reduction in type 1 MI,” said Kirtane. “That’s an important finding. It’s not something conventional wisdom would say. Conventional wisdom says that we will never prevent MIs, any type of MI, by doing a procedure.”

Different Definitions, Different Outcomes

The landmark ISCHEMIA trial showed that an invasive strategy on top of optimal medical therapy did not reduce the risk of its primary composite endpoint—the aforementioned cardiovascular death, MI, hospitalization for unstable angina, heart failure hospitalization, or resuscitation—in patients with stable ischemic heart disease and moderate to severe myocardial ischemia when compared with a conservative strategy of medical therapy alone. The invasive strategy also failed to lower the risk of death or MI, the study’s original “harder” primary endpoint.

When the trial was designed, there were already a couple of MI definitions floating around, including the Third Universal Definition of myocardial infarction (UDMI) and one from the Society for Cardiovascular Angiography and Interventions (SCAI). The Third UDMI, which is now in its fourth iteration, recommended use of cTn plus clinical evidence, such as ECG changes, the development of new pathological Q waves, or angiographic or imaging evidence supporting MI. The SCAI definition, on the other hand, prefers CK-MB but also incorporates supporting criteria.  

In ISCHEMIA, researchers chose two MI definitions for adjudication and analysis. The primary MI definition used cTn as the preferred biomarker for nonprocedural events based on the UDMI and used CK-MB for type 4a and 5 procedural-related infarctions. For procedural MIs, a CK-MB > 5-fold the upper limit of normal (ULN) within 48 hours plus clinical evidence defined a PCI-related MI while the postprocedure CK-MB threshold was set higher for surgery (> 10-fold the ULN). In contrast, the secondary MI definition used cTn as the preferred biomarker for both procedural and nonprocedural events, using the same thresholds as in the primary MI definition for type 4a and 5 MIs (cTn > 5- and > 10-fold the ULN, respectively).

Of note, both the primary and secondary definitions set significantly higher biomarker thresholds for capturing events in the absence of other clinical criteria.

Among the 5,179 patients randomized in the trial, there were 443 MIs (8.6%) according to the primary definition and 593 Mis (11.5%) based on the secondary definition. Procedural MIs accounted for 20.1% of all MI events with the primary definition but amounted to 40.6% of all infarctions using the secondary definition.

Overall, first MI events based on the primary definition occurred more frequently in patients treated with medical therapy, but when the secondary definition was used there were more MI events with the invasive strategy. Regardless of definition, there were more procedural MIs with the invasive strategy. As noted, the cumulative incidence of spontaneous type 1 events—no matter the definition—was lower among patients randomized to the invasive strategy and this lower incidence was observed in those treated with PCI, CABG, and those who had no obstructive CAD on cardiac catheterization.

In terms of clinical outcomes, the mean time free from the composite primary endpoint over 5 years, as well as the secondary endpoints of CV death/MI and all-cause death/MI, did not differ between treatment strategies when researchers used the primary MI definition. In contrast, clinical event rates—both the primary and secondary endpoints—were significantly higher in the invasive arm when cTn was used to capture MI events with the secondary MI definition. The higher event rates were driven by the higher risk of procedural MIs with the invasive strategy. 

Prognostically, the risk of CV death after a spontaneous MI according to the primary definition was significantly higher when compared with patients who didn’t have an infarction according to this definition (HR 3.38; 95% CI 2.03-5.61). The same association held true using the secondary spontaneous MI definition (HR 3.52; 95% CI 2.11-5.88). Procedural MIs based on the primary and secondary definitions, however, were not associated with an increased risk of CV mortality, nor were they linked to all-cause mortality.

Everyone assumes ISCHEMIA was a PCI trial and it wasn’t. Ajay Kirtane

Kirtane noted that many patients randomized to conservative treatment wouldn’t require any form of revascularization, just as there were roughly 20% of patients in the invasive strategy arm who had no obstructive CAD and didn’t receive PCI or CABG. Those low-risk, conservatively managed patients without obstructive disease would have a very low rate of MI. The treatment effect in terms of reducing MI with PCI/CABG would be even larger when comparing only patients in the conservative arm who had obstructive CAD, said Kirtane.

The reduction in spontaneous MI is an actionable finding that physicians should take into account when risk-stratifying patients, said Kirtane. The cath lab, he also stressed, does not automatically mean a stent.

“Everyone assumes ISCHEMIA was a PCI trial and it wasn’t,” he said. “It was a trial of cath to triage the patient appropriately. In that trial, 20% of patients didn’t even get revascularization, and of those who did, a substantial number were treated surgically. For me, I think in appropriately selected patients, there is a benefit to risk-stratifying the patient by anatomy. Maybe CT angiography is the way to do it. You don’t have to take the patient to the cath lab and you’re not obligated to do anything, but at least looking at the anatomy allows physicians to triage patients appropriately.”

Procedural MIs and Clinical Outcomes

Kaul, in contrast, said he didn’t find this analysis “particularly illuminating,” stating that “it appears to be an overenthusiastic post hoc exercise in making the invasive strategy look more favorable than justified by the ISCHEMIA trial evidence.”

He took issue with the strong conclusions of the researchers, particularly their assertion that “there were more frequent type 1 MIs with an initial conservative strategy and [these events] were associated with subsequent CV death.” However, in ISCHEMIA, the invasive approach had no discernible impact on CV mortality. “Once a trial fails to meet the primary or key secondary outcome, statistical rules of engagement dictate that all subsequent analyses, including analyses of individual component endpoints, should only be conducted to generate hypotheses and not to draw definitive conclusions or inferences,” said Kaul.   

There were 50% fewer spontaneous MIs among patients treated with the invasive approach compared with medical therapy (75 versus 147 type 1 MIs based on the primary definition), he said. Given that these were prognostically relevant in this latest analysis, the invasive strategy would be expected to reduce CV death in the overall trial, but it did not. 

The ISCHEMIA investigators have previously argued that follow-up is possibly insufficient for a survival benefit to emerge with the invasive strategy, and they have also pointed to the STICH trial, where the mortality signal favoring revascularization with CABG surgery took 10 years to be seen. However, Kaul noted there already was a trend toward a mortality benefit with surgery at 5 years in STICH and no material mortality signal between 5 and 10 years.

The newest analysis also throws a wrench into some of the arguments recently made by the EXCEL investigators as part of their justification for counting procedural MIs as part of the study’s primary endpoint, said Kaul. Surgeons have long argued that doing so biases the study against CABG, but the EXCEL investigators have claimed these procedural MIs are prognostically relevant.

“Yet, the ISCHEMIA investigators argue the opposite,” said Kaul. “[It’s] difficult to reconcile these findings.”

While Kirtane said that not all periprocedural MIs should be discounted, he pointed out that biomarkers aren’t routinely collected after PCI or CABG. He also noted that when using the primary MI definition, procedural MIs were associated with a trend toward subsequent CV death, but the results weren’t significant and confidence intervals wide (HR 1.99; 95% CI 0.73-5.43). Nonetheless, Kirtane believes that events that happen outside the hospital are more important for clinical practice.  

He also highlighted published this week in Circulation showing that elevations in cTn were detected by three different high-sensitivity assays in 34 healthy patients who underwent balloon inflation of the LAD for just 90 seconds. Inducing ischemia transiently with balloon inflation didn’t cause myonecrosis, just the elevation in troponin I and T, he said.     

Michael O’Riordan is the Managing Editor for TCTMD. He completed his undergraduate degrees at Queen’s University in Kingston, ON, and…

Read Full Bio
  • The National Institutes of Health funded ISCHEMIA.
  • Chaitman reports grants from the National Heart, Lung, and Blood Institute and personal fees from Merck, Novo Nordisk, Sanofi, Lilly, Johnson & Johnson, Daiichi Sankyo, Tricida, Relypsa, Imbria, and Xylocor.
  • Kirtane reports grant support/research contracts to his institution from Medtronic, Abbott Vascular, Boston Scientific, CSI, Siemens, Philips, and ReCor Medical.
  • Kaul reports no relevant conflicts of interest.