Should NICE Guidelines Be Universally Accepted for the Evaluation of Stable Coronary Disease?

A Debate

Harvey S. Hecht; Leslee Shaw; Y.S. Chandrashekhar; Jeroen J. Bax; Jagat Narula


Eur Heart J. 2019;40(18):1440-1453. 

In This Article

The Balanced View

Cardiology is one of the most evidence-based specialties, with a surfeit of large randomized clinical trials to answer pressing questions. It might, therefore, seem perplexing to the outside observer that cardiologists are still arguing on the best test to detect and diagnose CAD. As summarized above, the cardiology community is split between these two viewpoints. Part of the problem—akin to the three blind men describing an elephant, stems from the fact that we may often unknowingly confuse and conflate the goals of diagnostic testing in CAD. This becomes important since each of the diagnostic tests and pathways is good at identifying a certain endpoint, while being not as good at identifying another related one.

The first issue of confusion is the central question when one is looking at a patient with possible CAD—what is the clinical cardiologist, armed with the pertinent patient information and presentation, most interested in evaluating? (i) the presence of CAD (of any degree), or (ii) high-risk plaque that portends hard events, or (iii) stenosis of a certain anatomic/physiologic severity, or (iv) stenosis that compromises myocardial blood flow significantly (ischaemia), or (v) stenosis that needs intervention with varying urgency, or (vi) stenosis that can and should be fixed to reduce adverse outcomes?

Many of these entities might need a different diagnostic strategy—some tests can answer more of these questions than others but it is not realistic to think that one test will adequately answer all of them. It becomes even more complicated when we realize that anginal symptoms are often, but not always, a consequence of lumen compromise (which is what the majority of our non-invasive testing tries to detect). Risk, on the other hand, is predicated on the amount of plaque and the nature of the arterial wall (plaque morphology and local dynamics, something most such tests don't address) and can be high in asymptomatic patients who are not typically seen or offered non-invasive testing at all by the cardiology community. Even if they were tested, the usual non-invasive tests looking for severe stenosis would often be normal and not able to predict future hard events, which will unfairly make these tests seem less optimal.Table 5

Even within the simple task of identifying stenosis, some tests are good at 'ruling out' flow limiting disease (e.g. CTA), while others are better at 'ruling in' flow limiting disease (MRI or PET). The data supporting these statements often come from different populations with different Bayesian probability (largely low to intermediate risk for CTA and all comers, with probably a larger proportion of intermediate to higher risk populations in nuclear imaging studies). There are very few head to head comparisons in the same patients, and many consensus recommendations use systematic reviews or meta-analyses to support their conclusions and these are often fraught with their own limitations which explains the profusion of such studies. It is, therefore, difficult to say CTA is the best test—this might be acceptable if we are 'ruling out' significant CAD and stenosis (given its high NPV) in a low to intermediate risk patient but may be inappropriate if the goal is to 'rule in' significant stenosis in patients with higher risk (given its low PPV). CT-FFR and CTP will certainly improve the positive predictive ability of CTA but many of these techniques are still awaiting robust outcome data about whether this diagnostic premium improves hard event prediction, tailored medical therapy or surgical planning. Thus one cannot firmly make a recommendation for their routine use at this time. Furthermore, there is some variability in the accuracy among different studies and populations, and few studies have been done without vendor involvement.

One needs to have consensus on what is the best gold standard for comparative studies since this strongly influences the results. An anatomic test such as CTA will correlate well with another anatomic test such as ICA but may not correlate well with a gold standard looking at physiology, myocardial blood flow or objective evidence for ischaemia. Even a purportedly physiologic test such as invasive FFR (being a surrogate of myocardial blood flow during maximal vasodilatation) may not compare well to another test that more directly (PET) or indirectly (SPECT) evaluates the same reduction in myocardial blood flow. Furthermore, especially in meta-analyses of studies, a significant proportion of patients may not have had the correct gold standard. In the Takx et al. meta-analysis[10] invasive FFR was not actually measured but was assigned based on angiographic appearance in a significant proportion of patients; three vessel FFR was available in only 14% of the studies, thereby rendering the data less robust.

Finally, from the practical standpoint, the availability of up to date CT scanners and reader expertise will vary from country to country in Europe, with local variations in the United States as well. The choice of imaging modality should be strongly influenced by the best available technology. Concomitantly, cardiology training should familiarize clinicians with the pros and cons of all the imaging modalities and the cardiologist/radiologist collaboration necessary to improve patient care should be encouraged.