Off Script: Will ISCHEMIA Answer Our Most Pressing Questions?

I had a funny dream the other night. I was somehow supposed to be the lead discussant at the upcoming American Heart Association (AHA) Scientific Sessions on the ISCHEMIA trial results, except there was a problem. Not only did I not have the data, but I had also missed Dr. Judith Hochman’s late-breaking clinical trial presentation (which may indeed end up happening, since my daughter has her last high school dance recital the same day). In the dream, as time stood still, I started furiously searching the web and Twitter for something—ANYTHING—that would tell me the results. I woke up saying to myself, “This probably isn’t going to go so well; Judy eats unprepared people for lunch . . . ”

Aside from what it says about me that I’m having this ridiculous dream, given my genuine and patient-oriented interest in the data both for and against coronary angiography and revascularization, I suppose my dream actually does describe the current situation many of us face while waiting another few days for the top-line results of this landmark clinical trial. It really seems like we’ve been waiting forever. Despite years of discussion and, at times, vitriolic speculation, our actual knowledge is limited to what has been presented and published about the study population thus far. We really don’t know a whole lot more. And since we’ve already learned that there will not be a simultaneous publication filled with data supplements, we’ll have to make the best of whatever data get presented at AHA later this week.

The rest we don’t know and can only speculate about—for now. The fundamental premise of the trial is to assess whether there is a risk to initial medical management of stable patients with moderate-severe ischemia on stress testing. This potential risk is weighed against the risk of proceeding down an invasive pathway with coronary revascularization (CABG or PCI) as well as the risk of recurrent events that can occur after initial intervention. We can only hope that the population enrolled was “high risk enough” to ascertain a prognostic benefit of revascularization in complex coronary artery disease, if one truly exists.
We DO know that the study group is top-notch, with deep clinical trial experience as well as a penchant for attention to detail (which partially explains my nightmare). We DO know that 5,179 stable patients with moderate-to-severe ischemia by either stress imaging or exercise tolerance test (ETT) and preserved ejection fraction were randomized to a strategy of invasive coronary angiography followed by revascularization on top of optimal medical therapy (OMT) versus an initial conservative strategy of OMT alone. We DO know that ISCHEMIA had a sound trial design, with randomization done prior to angiography so as to avoid the obvious treatment bias that would have ensued from exclusion of anatomy too severe to randomize if physicians had seen the angiograms before randomization. We DO know that blinded CT angiography (CTA) was done in approximately two-thirds of the enrolled patient cohort to either exclude life-threatening left main disease or to ensure inclusion of patients with some (although not necessarily severe or critical) CAD. We DO know that despite external criticism over changing endpoints, if the rates of crossover from one treatment arm to the other are low, we should still be able to look at the study endpoints individually—as well as the timing of occurrence of these endpoints—to guide our clinical interpretation of the results. In order to keep the crossover rates low, we DO know that most patients were not that symptomatic, with only monthly anginal frequency, which actually helps test the hypothesis of whether anatomic delineation/treatment of ischemia without severe symptoms helps patients or not.

A high-risk population is the correct population in which to test this hypothesis. Based upon the published baseline data from ISCHEMIA, it certainly appears that the investigators tried their best to enroll this population. But while many randomized patients had multivessel disease, it is unclear whether the criteria for eligibility for randomization which included CTA-based stenosis severity of only 50% or greater (70% or greater for those qualifying with ETT alone) and severe ischemia in only 45% of patients, allowed for enough randomized patients with several truly critical, severe, or at least flow-limiting epicardial stenoses. Recall that most of the early revascularization trials have shown a benefit of coronary revascularization in anatomically high-risk subsets of disease, and stress testing’s initial aim was to reliably identify patients with anatomically severe disease. It is notable that in prior trials, including COURAGE, anatomic burden of disease (and not ischemic burden) was correlated with adverse events. That’s why the angiographic findings will have so much relevance to the interpretation of treatment effects, if they exist.

One of the best things about our field is that the exercise of trial interpretation isn’t just academically fulfilling, but it also has real clinical application for our patients with a highly prevalent disease. Ajay Kirtane

I do wonder how much we will be able to generalize the ISCHEMIA results to a greater number of patients with moderate and especially severe ischemia by stress testing, without forgetting how much of a role both the limited baseline symptoms of the population and the CTA played prior to randomization. Significant left main disease following a higher-risk stress test was likely seen (by my estimate) in almost 10% of enrolled patients (8% of the CTA cohort who were not randomized plus a likely higher proportion in patients who couldn’t get the CTA due to chronic kidney disease). This doesn’t even account for screen failures for other reasons. Given the overall prevalence of CAD, that number is not small, and suggests that even if the ISCHEMIA trial is negative that we need to be certain left main disease isn’t present in a patient with a high-risk stress test before simply concluding that patients don’t need cath or revascularization.

Beyond the primary ISCHEMIA results, quality-of-life data will be presented. These results may be challenging to interpret, given that so many patients had minimal symptoms at baseline. It’s hard to make patients feel much better if they themselves say they are feeling well to begin with, but perhaps the prespecified stratified analyses of patients with more severe symptoms at baseline will be informative.

Many physicians have expressed concerns that patients were randomized at centers that may not practice state-of-the-art stress testing or coronary revascularization. Fewer than 20% of ISCHEMIA patients were enrolled in the US. While this concern is legitimate, I tend to fully agree with the study group’s anticipated assertion that this is likely representative of how most testing and coronary revascularization is currently practiced in the “real world.” It will be fascinating to see how well or poorly stress testing actually performs even when the results are higher risk, and does this differ by modality of the test? Here again the CTA data will be helpful. Any interventional cardiologist with reasonable volume can point to a high-risk stress test they have seen in the last week or two that led to a cath with mild-to-moderate, nonobstructive coronary lesions by angiography. While some of this may represent microvascular disease, “false-positive” stress tests (particularly those done by less-sophisticated stress labs) are also a rampant finding in this clinical scenario.

Another question many physicians will have is: how well or how poorly was coronary revascularization actually done in this global trial, and did study investigators appropriately refer complex disease to CABG? Yes, CABG is part of the trial—a fact that is frequently overlooked. In fact, the highest-quality data in favor of a prognostic benefit of coronary revascularization in more complex disease comes from the CABG literature. If on the other hand PCI was performed, was it “generic PCI” or PCI with an aim for functionally complete revascularization using optimized interventional techniques, both of which have been linked to improved outcomes? Speaking to this concern, Dr. Hochman once said to me something along the lines of: “If you’re concerned about that, then you and centers like you should randomize more patients,” and she was absolutely right. While I personally tried to (and actually did) refer patients at our center for randomization, it was not easy at all to convince many physicians or potentially randomizable patients to actually enroll in the trial. This is not solely because physicians had incentives to take patients to the cath lab. It’s also because it’s not so easy to randomize a nervous patient sitting in front of you with his/her family, all of whom know the patient’s stress test was “high risk.” It’s even harder when the patient and family have been told by their referring physician that angiography is the next step. This is another reason that the study investigators should be commended for their efforts, and maybe in this way ISCHEMIA also sets the stage for future trials that were previously thought to be impossible.

If the stars align, we should get some answers to many of the important questions posed by the ISCHEMIA trial in the coming week. I hope that the discussion of the results is fair and balanced with an aim to understand this important area dispassionately and without accusations. The trial leadership, investigators, and patients who consented to be enrolled deserve a ton of credit. Even in anticipation of the trial results, I personally have adopted the ISCHEMIA investigators’ approach of CTA to rule out left main or critical proximal disease for screening asymptomatic patients with higher-risk stress tests in lieu of a diagnostic coronary angiogram. I am certain that final answers will take additional analyses and more months to sort out, because experience from many other trials tells us that in order to inform patient care, it is usually prudent to wait for more details beyond the top-line results. For example, I would be surprised if we had in-depth analyses on anatomic lesion complexity and completeness of revascularization at the time of the initial presentations.

Rest assured that there will be many opportunities to analyze and dissect the trial going forward. Noninvasive clinicians, imagers, interventionalists, cardiac surgeons—in fact any physicians who take care of patients with coronary artery disease—should all have interest in these important trial results. One of the best things about our field is that the exercise of trial interpretation isn’t just academically fulfilling, but it also has real clinical application for our patients with a highly prevalent disease. Am I a nerd for having dreams about clinical trials? Maybe. But this nerd cares.