Reflections on the Appropriate Use Criteria for PCI

As a clinician interested in helping my patients live better, longer or more comfortably, I have always tried to see where the tools I have at my disposal may be best utilized. That goes for everything I can do for a patient, from educating them and their families, prescribing just the right amount of medication (no more and no less), all the way to implanting stents or sending them off to coronary artery bypass grafting. Although there is much literature to guide us in these decisions, all told they probably offer guidance for maybe 10-20% of clinical decision-making. We have been comfortable in knowing that, although insufficient to cover all patients all times, we in cardiology still have some of the most robust data available in the form of randomized controlled trials, observational data and even case series.

But, how then do we try to provide guidance for the other 80-90% of clinical issues that arise, for which little high-level data exists? Having served on official ACCF/AHA guidelines, I know that we sometimes offer expert consensus, extrapolate from what data there is or just say simply “we don’t know.” And, these are then placed into neat little categories from Class 1 for those therapies we feel positively sure we should do to Class 3 for others we feel positively sure we should not do, either because of overwhelming high-level evidence (Level of Evidence A), strong beliefs (Level of Evidence C) or some combination of the two (Level of Evidence B). In this process, importantly, we leave the bulk of room for Class II, where clinical judgment and the unique patient circumstances allow the clinician to tailor care.

Not so with the Appropriate Use Criteria. Magically, there is no longer that large gray zone where clinical judgment might come into play. Apparently, indications for PCI are black and white. The appropriate use criteria for PCI look at (a) symptoms of angina, (b) number of anti-ischemic medications, (c) extent of ischemia or risk on stress testing and (c) coronary anatomy, and assign a score of Appropriate, Uncertain or Inappropriate based on the various combinations. Those with high-risk stress test findings, lots of angina (CCS Class III or greater), or three-vessel disease, left main or two-vessel disease involving the proximal LAD all would receive Appropriate revascularization, whereas those with lesser degrees of angina or low risk stress test findings, especially if they are not on at least 2 medications, would often fall as Inappropriate.

There are two problems with this that I’d like to highlight. First, there is no evidence that performing even these “Inappropriate” PCI’s would be providing more harm than good, and thus most patients falling into “Inappropriate” would not be Class III were they in an official guideline. In real-world practice, really only those patients with no ischemia and no symptoms are Class III, where you have no business implanting a stent; the rest falls in the gray zone, where some would argue reducing ischemia as much as possible, or eliminating all angina, is a noble goal, especially if you might be able to improve exercise tolerance or quality of life, or even reduce the number (or dose) of medications a patient may need to take. So, the appropriate use criteria in effect imply that we have harmful data for a large portion of PCIs when we do not really have that data.


Second, the criteria are written by physicians who are making value judgments on behalf of patients. That’s fine when there are clear mortality benefits, such as advocating for primary PCI over thrombolytics, but not when the benefits include reduced cost or avoidance of invasive procedures. As a physician we may advise a patient to add a second mediation instead of performing a PCI, but that same patient may turn around and ask for the PCI so they don’t need medications and can more quickly and more thoroughly reduce angina.


As a whole, I applaud and welcome the appropriate use criteria taskforce for taking complicated clinical scenarios and boiling them down to help decision-making at the point of care. It is important for us to utilize the criteria so we can see where our patients fall on the spectrum, and thereby include our patients more in the decision-making process. But, they should not pretend to be what they are not. Like everything else in medicine where the data is not clear-cut, tough decisions about how to effectively use interventional techniques to improve quality of life or survival for our patients should be guided but not dictated.


Dr. Srihari S. Naidu is Director of the Cardiac Catheterization Laboratory, Interventional Cardiology Fellowship Program and Hypertrophic Cardiomyopathy Treatment Center at Winthrop University Hospital, and Associate Professor of Medicine at SUNY – Stony Brook School of Medicine.