Is CT Angiography the Best First Test for Chest Pain?

Patrice Wendling

October 11, 2017

HERSHEY, PA — A new meta-analysis shows that use of coronary computed tomography angiography (CTA) is associated with significantly fewer MIs than standard functional stress testing in patients with suspected CAD and acute or stable chest pain[1].

CTA, however, is also associated with significantly more downstream invasive coronary procedures, CAD diagnoses, and aspirin and statin prescriptions—all without an overall reduction in mortality or cardiac hospitalizations.

"If you look at the strength and robustness of the individual findings, what my coauthors and I can say with high certainty is that cardiac CT compared with functional testing will lead to a significant increase in downstream procedures, whether that's just catheterization or whether it's also revascularization, and we believe that a lot, if not all, of that excess is unnecessary," corresponding author Dr Andrew J Foy (Penn State College of Medicine, Hershey, PA), told | Medscape Cardiology.

The study, with senior author Dr Rita Redberg (University of California, San Francisco), was published October 3, 2017 in JAMA Internal Medicine.

Dr Matthew Budoff (Los Angeles Biomedical Research Institute, Torrance, CA), who was not involved in the study, said it's not surprising the MI finding is being downplayed by the researchers, given Redberg's well-known negative view of the added value of coronary CTA that she has discussed in several studies, editorials, and even a previous debate with Budoff.

Furthermore, he said, patient-level data were not used in the analysis, nor did any of the studies look at the appropriateness of the revascularizations.

"This represents a large number of trials, more than 10,000 patients, and shows very unequivocally that if you undergo CT angio you have fewer heart attacks by 29%, and that's a huge advantage for a diagnostic test," he said.

"We talk about using statins to reduce myocardial infarction and they reduce MIs about 30%, so we're talking about an effect that's considered among the largest that we've ever seen, and now we're just talking about using a different diagnostic test that's cheaper and faster than functional testing in most cases."

Mining the MI Finding

The meta-analysis included 13 randomized clinical trials that assigned 10,315 patients to coronary CTA and 9777 to a functional stress-testing strategy that included no testing, myocardial perfusion imaging, exercise-treadmill or bike electrocardiography testing, and stress echocardiography.

The overall quality of evidence was moderate, with 45 of 98 domains (46%) judged to be at high or questionable risk for bias.

After a mean follow-up of only 18 months, the results showed CTA and functional stress testing had similar rates of death (1.0% vs 1.1%) and cardiac hospitalizations (both 2.7%).

Use of coronary CTA, however, significantly lowered the risk of MIs overall (0.7% vs 1.1%, risk ratio [RR] 0.71; 95% CI 0.53–0.96) and for patients with stable chest pain (RR 0.68, 95% CI 0.49–0.95), but not for those with acute chest pain (RR 0.84, 95% CI 0.44–1.61).

Foy said the overall MI reduction is very small on an absolute scale and driven by the SCOT-HEART trial, in which 85% of patients in the CTA arm received functional stress testing. After the trial was removed from the sensitivity analysis, the risk estimate increased 17% and was no longer significant (RR 0.88, 95% CI 0.70–1.21).

Also undercutting the robustness of the MI reduction is the interesting finding that in the stable chest pain subgroup that SCOT-HEART falls into, there was actually a signal that just missed statistical significance for increased cardiac hospitalizations in the CTA arm, despite the reduction in MIs in this subgroup.

Further, "Even if we're to take that MI reduction at face value, it's a number needed to test of 250 to reduce one MI, whereas it's only a number needed to test in the 30 to 40 range to lead to an excess invasive procedure. So we don't think that's a very good trade-off personally or from a public-health standpoint," Foy said.

Subsequent invasive coronary angiography (ICA) was significantly higher with the CTA strategy overall (RR 1.33, 95% CI 1.12–1.59) as were revascularizations (RR 1.86, 95% CI 1.43–2.43).

"We hypothesize that at least some of these additional procedures are associated with the finding of incidental CAD that is not causing symptomatic ischemia and would not have been detected with functional stress testing alone. Results from SCOT-HEART support this contention," the investigators write.

Budoff countered that the revascularization rate was quite low, at only 7.2% among patients undergoing coronary CTA (vs 4.5% with functional testing) and that meta-analyses prior to SCOT-HEART have shown similar MI reductions, as did the PROMISE trial, which was included in the present analysis.

Furthermore, Budoff said the new meta-analysis downplays the "most important finding," that patients who underwent coronary CTA were significantly more likely to receive prescriptions for aspirin (21.6% vs 8.2%; RR 2.21; 95% CI 1.20–4.04) and statins (20% vs 7.3%; RR 2.03; 95% CI 1.09–3.76).

"We know the long-term benefits of statins and aspirin, so if you went out another 5 years, you'd expect a compounded benefit of getting the CT angio," he said.

Did NICE Get It Right?

In an accompanying editorial[2], Dr Todd Villines (Uniformed Services University of the Health Sciences, Bethesda, MD) and Dr Leslee Shaw (Emory University School of Medicine, Atlanta, GA) write that "the benefits of CTA are likely largely driven by changes in preventive therapies—and perhaps better selectivity for ICA and revascularization."

They note that CTA studies consistently show that more than 70% of patients referred for ICA have obstructive CAD, whereas more than half referred for ICA after functional testing have no or nonobstructive CAD based on cath-lab testing.

Still, single-photon emission computed tomography (SPECT) remains the dominant noninvasive test for CAD in the US, with more than 50 SPECT studies currently performed for every CTA.

In a dramatic move, however, the UK's National Institute for Health Care Excellence (NICE) recently updated its guidelines, making coronary CTA the initial test for all patients with chest pain and suspected CAD.

Villines and Shaw say the current findings support the NICE approach and that coronary CTA should be more broadly available to patients with acute and stable chest pain.

They call for further exploration of the findings but conclude, "The evidence to date supports that CCTA may improve important clinical outcomes, a fact that should be shared with patients during the decision-making process regarding test choice."

A novel finding from the meta-analysis is that CTA was associated with an increase in new CAD diagnoses compared with functional testing alone (18.3% vs 8.3%; RR 2.80; 95% CI 2.03–3.87).

"This increase in CAD diagnoses likely drove the associated increase in aspirin and statin use and might explain the reduction in MIs," the investigators suggest.

The biggest limitation of the study was that they did not use individual patient-level data, which would have allowed them to look at the impact of the patient's sex, age, and individual CV risk on the different modalities, Foy said. Also, they were unable to make an assessment of differences in radiation exposure or the impact of incidental cardiac and pulmonary findings identified with cardiac CTA.

"There's more to this question than just does it reduce MI by a very tiny fraction," Foy said. "There are other implications that are really important—the radiation exposure, incidental findings, and of course the cost for all this downstream testing.

"It's extremely unlikely that the reduction in MI, even if it's real, can be offset or cost-effective when taking into account all these downstream tests and follow-up for incidental findings that would come from cardiac CT."

In a recent large observational study of patients with stable CAD from Denmark[3], which the editorialists noted had "strikingly similar" results to the meta-analysis, the mean costs of subsequent testing, invasive procedures, and medications were higher after coronary CTA than functional testing ($995 vs $718; P<0.001).

Commenting to | Medscape Cardiology in an email, lead author Dr Mads E Jørgensen (University of Copenhagen, Denmark) said, "The initial choice of noninvasive testing in patients with suspected CAD affects downstream patient management, but the exact mechanisms or causal relations between changes in patient management and lower risks of MI remain unknown. If future observational studies are to provide answers, further data on blood work, ie, lipid levels, and results from invasive tests and revascularizations must be incorporated.

"Meanwhile, it is important that clinicians are aware of the downstream effects of initial testing, which have now been confirmed over and over again," he added.

Foy, his coauthors, Villines, and Shaw reported no relevant financial relationships. Redberg is the editor of JAMA Internal Medicine but was not involved in decisions regarding review or acceptance of the manuscript. Coauthor Dr John Mandrola is a columnist for on Medscape. Budoff reported conducting research for the National Institutes of Health and General Electric. Jørgensen reported no relevant financial relationships.

Follow Patrice Wendling on Twitter: @pwendl. For more from | Medscape Cardiology, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.