Is CT Angiography Overkill for Low-Risk Chest Pain?

John M. Mandrola, MD; Andrew J. Foy, MD


January 09, 2017

This feature requires the newest version of Flash. You can download it here.

John M. Mandrola, MD: Hi, everyone. This is John Mandrola from on Medscape. I am here in New Orleans at the 2016 American Heart Association Scientific Sessions. I am very pleased to have with us Andrew Foy, a cardiologist at Penn State Hershey Medical Center. Andrew has a very provocative abstract[1] presented here at the meeting. Andrew, welcome.

Andrew J. Foy, MD: Thank you.

Dr Mandrola: Full disclosure: I was involved in this abstract. Andrew and I wrote it together, but Andrew did most of the work. Tell us about the study question.

Dr Foy: The main question for this analysis was to look at the efficacy of cardiac CT versus standard care in emergency department (ED) patients with low-risk chest pain.

Dr Mandrola: Compared with what?

Dr Foy: The standard care in most of these trials tended to be some form of stress testing for all or the majority of patients.

Dr Mandrola: There are a ton of patients presenting with chest pain. This is a huge problem.

Dr Foy: Yes, it's the second leading cause of ED visits behind abdominal pain, at 6-8 million per year.

Dr Mandrola: CT angiography (CTA) is a pretty sensitive test.

Dr Foy: It's a very sensitive test for coronary atherosclerosis. The question is, does the increased detection of that lead to improvement in patient outcomes? Or does it lead to more revascularization? Or both?

Dr Mandrola: How did you go about answering the questions? What were your methods?

Dr Foy: We did a systematic review specifically looking for randomized controlled trials of cardiac CT versus standard care for patients with low-risk chest pain as defined by a negative workup, negative troponin, nonischemic ECG, and time of admission to the ED.

Dr Mandrola: There have been studies looking at low-risk patients with negative enzymes, negative ECGs, and low-risk clinical history. How many randomized controlled trials were included in this analysis and how many patients?

Dr Foy: Several thousand patients were included in this analysis. Six randomized controlled trials were included in the primary analysis and three were included in a sensitivity analysis. They didn't strictly meet the inclusion criteria, but the patients presented to the ED with chest pain, so they were included in the sensitivity analysis.

Dr Mandrola: What endpoints did you look at?

Dr Foy: The efficacy endpoints were myocardial infarction (MI), death, MI plus death as a composite endpoint, recidivism or return to the ED, cardiac hospitalization, and revascularization defined as percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG) and invasive coronary angiography.

Dr Mandrola: What did you find?

Dr Foy: There was no difference in the efficacy endpoints, so there was no statistically significant reduction in MI or death plus MI. There was no reduction in return visits to the ED and no reduction in cardiac hospitalization. There was a significant increase in coronary angiography, of about 50%, which was highly statistically significant, and there was a 100% increase in revascularization procedures. That was also highly significant. It came out to a number needed to test of 40-50 to lead to one excess coronary angiogram and also one excess revascularization procedure.

Dr Mandrola: The CTA compared with standard care led to more angiograms and more procedures, but no difference in outcomes.

Dr Foy: That is correct for the outcomes that we looked at in this study.

Dr Mandrola: Were there any limitations? This was a meta-analysis. Talk to us about that.

Dr Foy: The main limitation to this analysis is that it used trial-level data, not individual patient-level data. We didn't have access to those, but I don't think that represents a significant limitation to the conclusions that were drawn.

Dr Mandrola: Why not? I read that patient-level meta-analyses are stronger than trial-level meta-analyses.

Dr Foy: That is true depending on the question. In this case, I just don't see a reason why the individual patient-level data would lead to any difference in the conclusion. We accept the fact that the trials, in this case, enrolled patients who we were interested in looking at. Whether we had individual patient-level data would not necessarily have changed the conclusions. We may have been able to look at differences between age groups or gender if we had that type of information, but ultimately it would not have fundamentally changed the conclusion.

Dr Mandrola: There have been previous meta-analyses looking at this question of CTA and low-risk chest pain. How is this paper different?

Dr Foy: This is several years later. It includes two more trials in the primary analysis and a total of five more trials if you include the sensitivity analysis that was performed. It strengthens the conclusions of the earlier meta-analyses of this low-risk chest pain population, which is that cardiac CT increases utilization and does not improve cardiac events.

One of the earlier meta-analyses[2,3] looked at time spent in the ED, which seemed to have a benefit with cardiac CT. But, as Dr Redberg[4] pointed out in an editorial, those trials had a specific accelerated chest pain protocol. It may not be that that would manifest in real-world populations.

Dr Mandrola: The individual trials that you put in the meta-analysis weren't powered to detect differences in outcomes. Is that correct?

Dr Foy: That is correct.

Dr Mandrola: Did the individual trials show any change in utilization?

Dr Foy: All of the trials showed a signal that there was an increase in utilization, but they were not powered for that. There was one exception. One of the newer trials that was included, called the PERFECT trial,[5] showed a statistically significant increase in utilization with cardiac CT. The main standard-care modality in that trial was stress echo, which is interesting but for a different conversation.

Dr Mandrola: The big question here is, how do you explain the higher utilization of PCI and revascularizations but no difference in cardiac outcomes? This goes against the grain.

Dr Foy: I think the explanation is that the excess detection in the CTA group is not detection related to acute coronary syndrome. It may just be true—true and unrelated based on the complaint that they came to the ED with.

Dr Mandrola: True that they have chest pain and true that they have incidental coronary disease, but not related or with any connection.

Dr Foy: That is correct.

Dr Mandrola: It gets to the story of the difference between coronary atherosclerosis and coronary artery disease.

Dr Foy: It does. If a disease requires some symptoms, we can attribute ischemia on functional testing to be more representative of the symptoms that brought them to the ED than just anatomic disease.

Dr Mandrola: What are the clinical implications of this going forward? There are tons of patients presenting to EDs with chest pain.

Dr Foy: For me, the take-home is that if we are going to test these patients, it should be functional stress testing. There is also a debate about whether most of these patients need any testing at all. The observational data would suggest that maybe they don't, but there has not been a good randomized controlled trial done in that population. Many people would say that it would require too many patients to show any difference, and that it would be futile to perform. In my opinion, that makes the point that maybe we don't need to be doing this.

Dr Mandrola: There is no change in outcomes and it's a pretty powerful meta-analysis, yet there are more procedures. These procedures have a fixed complications rate; it's a low rate of complications but it's not zero.

Dr Foy: It is very possible that CTA could increase harm. We looked at our data and we looked at the number of patients needed to test to lead to excess revascularization. If you multiply that number with the number of patients needed to harm with the revascularization procedure—periprocedural MI, bleeding, and stroke—you get a total number needed to harm with CT of about 400-500 compared with stress testing.

Dr Mandrola: Thank you for being here.

Dr Foy: Thank you for having me.

Dr Mandrola: Thank you for watching. This is John Mandrola from on Medscape, at the American Heart Association Scientific Sessions in New Orleans.


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.