CCTA doesn’t help: The evidence

Since publishing my 5 part series illustrating that stress tests are useless (also available as an EM Cases Journal Jam Episode and on EM:RAP), I have received a number of requests to review coronary CT angiography (CCTA). After all, stress testing is old technology. CCTA is newer and therefore obviously better. We can get immediate images of the coronary arteries, and with the newer FFR technology, we can also get physiologic information about potentially flow limiting lesions. What’s not to like?

This post will review the current evidence for (or against) CCTA, but before we jump into the evidence, we should review some important background information. After a negative emergency department workup, our patients have an incredibly low risk of MI and death. Although this risk has traditionally been cited as 2%, a review of the literature clearly demonstrates that it is much lower, probably in the range of 2-3/1000. Thus, whether we are talking about using stress tests or CCTA, our pretest probability is incredibly low. Furthermore, in order for a test to help patients, it has to result in some change in management. However, the evidence is very clear that the low risk patients in whom we use CCTA or stress testing (those with negative troponins) do not benefit from invasive management. Because all these patients will benefit from risk factor modification, it isn’t clear how CCTA could possibly help these patients (perhaps with the one exception of deciding who needs antiplatelet therapy).

Therefore, I engaged in this literature review with a low pretest probability. It is very hard to see how CCTA could possibly help patients. It is not good enough to have an accurate test. In order to consider adopting this technology, we really need to see some evidence that it results in clinical benefit for our patients. With that being said, let’s look at the evidence.

Is CCTA accurate?

This question is a little more complicated than you might think, because there is a difference between accurately identifying coronary artery disease (CAD) and determining whether that anatomic lesion is resulting in symptoms or clinically important limitations in blood flow. In general, CCTA is pretty accurate at identifying anatomic lesions. When compared directly with invasive angiography, CCTA has an excellent sensitivity, but only a moderate specificity. In one study of 230 low risk patients who underwent both CT and invasive angiography, CCTA had a sensitivity of 95% (95% CI 85-99%) and a specificity of 83% (95% CI 76-88%). (Budoff 2008) This translates to a positive likelihood ratio of 5.6 and a negative likelihood ratio of 0.06 (with the corresponding wide confidence intervals). In another study of 291 patients with suspected CAD, CCTA had a sensitivity of 85% (95% CI, 79 to 90) and a specificity of 90% (95% CI, 83 to 94) compared to invasive angiography. (Miller 2008) Finally, another study looked at 360 patients with stable angina, all of whom underwent both CCTA and invasive angiogram, and the CCTA had a sensitivity of 99% (95% CI 98% to 100%), but a specificity of only 64% (95% CI 55-73%). (Meijboom 2008)

One major problem with this data, however, is that it assumes that larger stenoses are more important. However, a significant proportion of ACS originates from coronary lesions less than 50%, and CCTA doesn’t help us identify these lesions.

Overall, when looking for anatomical disease, CCTA is a reasonably accurate test, but it certainly isn’t perfect. It is a sensitive test, but the specificity appears to be moderate. When used in a low risk population it will almost certainly suffer from a low positive predictive value. Admittedly, it is almost certainly more accurate than stress testing for identifying CAD. However, identifying CAD isn’t helpful if we can’t do anything about it. What we really want to know is: does this test improve clinical outcomes?

Does CCTA change outcomes? RCTs of ED chest pain patients

There are 6 RCTs looking at CCTA in emergency department populations. I summarize them all here, but if you are looking for a quick summary, these trials show no clinical benefit over standard care.

Goldstein JA, Gallagher MJ, O’Neill WW, Ross MA, O’Neil BJ, Raff GL. A randomized controlled trial of multi-slice coronary computed tomography for evaluation of acute chest pain. Journal of the American College of Cardiology. 2007; 49(8):863-71. [pubmed] [free full text]

This is a single center RCT that looked at 203 low risk ED chest pain patients after 2 negative biomarkers and normal ECGs. They randomized the patients to either CCTA or standard care (nuclear stress test). In the CCTA group, 67% of patients had a negative test and were discharged, but 24% were indeterminate and had to have a nuclear stress test anyway. In the nuclear stress test group, 95% of patients had negative scans and were sent home. There was no statistical difference between the groups in terms of the number of patients who went on to have invasive angiograms (12% with CCTA vs 7% with standard care, p=0.24). Most importantly, there were no deaths, MIs, or patients diagnosed with unstable angina during the 6 month follow up. These patients were at 0% risk of ACS, and therefore really didn’t need any further testing at all. Unfortunately, even if CCTA was an excellent test, there is no way it could possibly help in this zero risk population. Length of stay was shorter with CCTA, but considering that none of these patients needed to be kept in the hospital for more testing, that is sort of a silly comparison. It’s like saying to a patient with a sprained ankle: you can either stay 1 hour for an x ray or 24 hours for an MRI. Neither test is needed, and the length of stay should have been close to 0.

Comments: It will be a recurrent theme in these papers, but it is going to be impossible to demonstrate a clinical benefit from CCTA if you apply the test to such a low risk population that not a single patient has a real clinical outcome. 

The CT-STAT TRIAL: Goldstein JA, Chinnaiyan KM, Abidov A, et al. The CT-STAT (Coronary Computed Tomographic Angiography for Systematic Triage of Acute Chest Pain Patients to Treatment) trial. Journal of the American College of Cardiology. 2011; 58(14):1414-22. [pubmed] [free full text]

This is a multi-center randomized trial comparing CCTA to nuclear stress test in 749 low risk ED chest pain patients after a negative work up. (They don’t actually say if it was after 1 or 2 troponins). CCTA was negative in 82% of patients, allowing for discharge, but was indeterminate in 14%, so these patients had a nuclear stress test. The nuclear stress test was negative in 90% of patients. There was no difference in the rate of invasive angiography (6.7% vs 6.2%) or early revascularization (3.6% vs 2.4%). There was no difference in repeat ED visits over the next 6 months. Most importantly, not a single patient died, had an MI, or was diagnosed with unstable angina in the 6 months after their negative workup, so once again, none of these patients needed any testing. There is no way to demonstrate clinical benefit from CCTA in such a low risk population. Once again, the CCTA group went home a few hours faster, but seeing as all the patients could have been sent home without either CCTA or stress test, that’s sort of irrelevant. 

Comments: There were a large number of exclusions (6640 screened to find 749 patients) including 9% just for “physicians discretion”, so there will be selection bias. 

THE ACRIN TRIAL: Litt HI, Gatsonis C, Snyder B, et al. CT angiography for safe discharge of patients with possible acute coronary syndromes. The New England journal of medicine. 2012; 366(15):1393-403. [pubmed] [free full text]

This mutli-center RCT randomized 1370 low to intermediate risk ED chest pain patients either CCTA or standard care after a normal ECG. Patients could be enrolled before their biomarkers were reported. 83% of the CCTA group had no disease or non-obstructive CAD, and 14% were indeterminate and needed a stress test. Only 64% of the standard care group underwent any kind of diagnostic testing. There was no difference in the rate of invasive testing or initial revascularization. 1% of both groups were diagnosed with MI before being sent home (presumably on the first or second troponin). In the 30 day follow up period, there was 1 MI in the CCTA group and 2 in the standard care group (that’s a risk of 1 or 4 in 1,000, if you are comparing to the stress test data). There were no deaths in follow up, and no difference in repeat ED visits, hospitalizations, or cardiologist office visits. Compared to usual care, more patients in the CCTA groups were discharged home (9.6% vs 22.7%: ARR 26.8%; 95% CI 21.4 to 32.2) and length of stay in the ED was also decreased.

Comments: Although CCTA increased the number of patients discharged home, it isn’t clear that any of these patients needed to be admitted. If you waited for the negative biomarkers (which would have presumably caught the 1% of patients diagnosed with MI during their admission), these patients had essentially 0% risk, and yet they admitted 80% of the standard care group! In Canada, we would have discharged almost 100% of these patients, so CCTA could not have possibly helped.

The fact that only ⅔ of the standard care group had any kind of provocative testing is important. It means that this is closer to the study we actually want: CCTA versus no further testing. It is also a good reminder than not everyone needs a stress test. Even during a study at big academic centers, ⅓ of patients are discharged without stress testing, so stress testing is clearly not a standard of care.

Another publication looking at 1 year follow-up data from this cohort reports that CCTA did not reduce subsequent ED visits, or change subsequent resource use at all. (Hollander 2016)

ROMICAT-II: Hoffmann U, Truong QA, Schoenfeld DA, et al. Coronary CT angiography versus standard evaluation in acute chest pain. The New England journal of medicine. 2012; 367(4):299-308. [pubmed] [free full text]

This is a multicenter RCT that randomized 1000 patients aged 40-74 with suspected ACS, but negative ECGs and 1 negative troponin, to CCTA or standard care. In the standard care group, about 25% had no provocative testing, while those that did were evenly split among nuclear stress, stress echo, and exercise stress testing. The primary outcome was length of hospital stay, and was lower in the CCTA group. The mean stay was 23 hours vs 30 hours (p<0.001). There were no cases of missed ACS (defined as a bounceback within 72 hours). In the 1 month follow-up, there were 2 major adverse cardiac events (MACE), including 1 MI, in the CCTA group and 6 in the standard care group, including 4 MIs (p=0.18). There were no differences in repeat ED visits or hospitalizations. Radiation was higher in the CCTA group.

Comments: The length of stay numbers are honestly incomprehensible to anyone working outside of America. My chest pain patients are essentially all discharged within 4-6 hours, so the comparison of 23 vs 30 hours has no external validity. 

CT COMPARE: Hamilton-Craig C, Fifoot A, Hansen M, et al. Diagnostic performance and cost of CT angiography versus stress ECG–a randomized prospective study of suspected acute coronary syndrome chest pain in the emergency department (CT-COMPARE). International journal of cardiology. 2014; 177(3):867-73. [pubmed] [free full text]

This is an unblinded, single-center RCT from Australia. They randomized 717 low to intermediate risk chest pain patients in the ED to either CCTA or exercise stress test. (They lost the consent documents for 84 patients, so only include 562 in the analysis.) Their primary outcome was the accuracy of CCTA (but it is an unblinded trial with no real gold standard.) Despite calling these patients low to intermediate risk, only 5.2% of the CCTA group and 2.9% of the stress test group had a final diagnosis of ACS. They report a sensitivity of 100% and a specificity of 94%, but they numbers seem wrong to me. In the text, they talk about 16 patients with moderate stenosis on CCTA, and 8 of these were false positive based on either angiogram or follow up nuclear test – so that doesn’t really fit with the 94% specificity. CCTA increased the rate of downstream testing (13.4% vs 7.5%), mostly by increasing invasive angiography. At 30 day follow up, there were no deaths or MIs in either group. By 1 year, there were 3 events in the CCTA group and 1 in the stress testing group, but most were deaths from other causes. The only real miss seems to have been an MI in the CCTA group. There was a decrease in hospital stay with CCTA, but these Australians were a bit like the Americans, with a 21 hour length of stay.

Comments: Once again, it is very hard to show any benefit from a test if you use it in a group of patients so low risk that none of them have the disease. With 1 MI in 717 patients over a year follow up period, it’s pretty clear these patients didn’t need any testing at all after their second troponin. ACS was diagnosed more often in the CCTA group, but in a properly randomized trial the outcomes should be even. This could have something to do with the lost patients, but I expect we are seeing the effects of incorporation bias. The trial is not blinded, so you can imagine that in borderline cases, a positive CCTA might sway clinicians to diagnose ACS, whereas if the same patient didn’t get a CCTA, they might receive a non-ACS diagnosis. For the same reason, the reported sensitivity and specificity numbers are probably overinflated. 

BEACON TRIAL: Dedic A, Lubbers MM, Schaap J, et al. Coronary CT Angiography for Suspected ACS in the Era of High-Sensitivity Troponins: Randomized Multicenter Study. Journal of the American College of Cardiology. 2016; 67(1):16-26. [pubmed] [free full text]

This is an open-label, multicenter RCT based in the Netherlands that enrolled 490 patients with acute chest pain in the ED. After a normal ECG and a single negative troponin, they randomized them to CCTA or standard care. There was no difference in the primary outcome, revascularization at 30 days (9% vs 7%, p=0.4). There were no important differences. There was 1 death in the CCTA and none with standard care. There was one missed ACS with CCTA and 3 with standard care (p=0.62). MACE was unchanged (10% vs 9%, p=0.54). There was also no difference in the number of patients discharged from the ED, nor the hospital length of stay.

Comments: The primary outcome, “patients requiring revascularization”, is problematic because of incorporation bias. It is really just a way of asking “did physicians trust CCTA?” It is completely unrelated to the accuracy or value of the test. 

Bottom line

These trials give a pretty good impression that CCTA does not provide any clinical benefits over standard care. There is a decrease in length of stay, but only if you have incredibly long (and unnecessary) lengths of stay to begin with. If you just sent your patients home after the second negative troponin, there would be no benefit. CCTA causes a small increase in radiation.

These studies are pretty definitively negative, but they do not exclude a role for CCTA. The patients enrolled in these studies were so low risk that they could not have possibly benefited from further testing. If you enroll a group of patients that has essentially 0 risk of MI, you can’t possibly demonstrate a decrease in MI from any management strategy. It is possible that we would see different outcomes in a higher risk group of patients.

CCTA RCTs looking at non-ED patients

There are 6 more RCTs that look at CCTA in non-ED populations. 3 of these trials looked at patients who were admitted from the ED with acute chest pain, and so probably could be combined with the above ED trials, because the populations are pretty similar. The other trials look at stable outpatients. Once again, if you want the quick summary, CCTA provides no benefit over standard care.

The PROSPECT TRIAL: Levsky JM, Spevack DM, Travin MI, et al. Coronary Computed Tomography Angiography Versus Radionuclide Myocardial Perfusion Imaging in Patients With Chest Pain Admitted to Telemetry: A Randomized Trial. Annals of internal medicine. 2015; 163(3):174-83. [pubmed] [free full text]

This is a single center RCT of 400 patients admitted to hospital because their doctor thought they required a noninvasive test. They were randomized to either CCTA or myocardial perfusion scan. The same number of patients underwent coronary angiography in the next year (15% vs 16%), half of which included PCI. There was no difference in MACE between the groups at 3 years (5% vs 5%). The overall radiation burden was statistically lower with CCTA, but both groups got a lot of radiation (24 vs 29 mSv). Hospital length of stay was the same.

Comments: After reviewing all the evidence, it is pretty clear that nuclear stress test is not a great comparison. Stress tests don’t help, and probably hurt. A better comparison group would have been no testing at all.

THE CATCH TRIAL: Linde JJ, Hove JD, Sørgaard M, et al. Long-Term Clinical Impact of Coronary CT Angiography in Patients With Recent Acute-Onset Chest Pain: The Randomized Controlled CATCH Trial. JACC. Cardiovascular imaging. 2015; 8(12):1404-1413. [pubmed] [free full text]

This is an RCT of 600 low and intermediate risk chest pain patients who were admitted to hospital for noninvasive testing after a negative ED workup including 2 negative troponins. Patients were randomized to CCTA or a control group that got either exercise stress testing or SPECT imaging. The study was designed to look at long term outcomes, with a median follow-up of 19 months. The CCTA group was treated differently during that year, with more aspirin use (47% vs 36%) and more other antiplatelet medications (14% vs 6%). Their primary outcome was a composite of cardiac death, MI, unstable angina, late symptom driven revascularization, and readmission for hospital. They report a statistical difference between the two groups (11% vs 16%, p=0.04), but as I discuss below, I am skeptical. None of the individual components of the composite were statistically significant, and the numbers are quite small. There were 2 deaths in the CCTA group as compared to 1 in with standard care. There was 1 MI in the CCTA group as compared to 7 with standard care (p=0.07). They also report that MACE was statistically better in the CCTA group (2% vs 5%; p=0.04), but the stats here are also questionable.

Comments: I have some significant concerns with this study.  If you look at their clinicaltrials.gov listing, they changed their primary outcome 6 months after they finished collecting data. The original composite outcome included all cause mortality, but they changed it to only include cardiac death. That is concerning, because there were 2 “non-cardiac” deaths in the CCTA group, but none with standard care. I plugged the numbers into a fragility index calculator to see if those 2 deaths would have changed the results, and the calculator says that the trial is ALREADY not statistically significant, despite the reported p value of 0.04. So I plugged the number for MACE in as well, and I get a p value of 0.06 instead of the 0.04 they report. (Just in case you think the calculator is broken, I plugged in the MI numbers, and get the same p value – 0.07 – that the authors report, so it is just the “statistically significant” results that are fishy.) Thus, I have a hard time trusting these results, and we haven’t even addressed the fact that this is a composite outcome. The majority of the difference between the two groups was in “readmission for chest pain”. That is hardly a patient important outcome, and it is probably driven entirely by a lack of blinding. (For the exact same patient, if you see a CCTA result that is normal you send the patient home, but if you see a stress test you admit him, but that doesn’t mean you made the correct decision.)

PROMISE TRIAL: Douglas PS, Hoffmann U, Patel MR, et al. Outcomes of anatomical versus functional testing for coronary artery disease. The New England journal of medicine. 2015; 372(14):1291-300. [pubmed] [free full text]

This is a pragmatic, open-label, multicenter RCT. They included 10,003 high risk, older, symptomatic outpatients requiring noninvasive testing according to their physician, and randomized them to either CCTA or functional testing. The stress testing was primarily nuclear (68%), followed by stress echo (22%), and exercise (10%). They followed patients for a median of 25 months. There was no difference in the primary outcome, a composite of death, MI, hospitalization for unstable angina, and complications of cardiovascular procedures or diagnostic imaging (3.3% with CCTA vs 3.0% with functional testing, HR 1.04, 95% CI 0.83-1.29, p=0.75). The trial was designed such that if CCTA was not shown to be superior, it was powered to be a noninferiority trial. The 95% confidence intervals exceeded the 10% noninferiority margin, so CCTA is NOT noninferior to functional testing. Looking at just the most important parts of the composite, there was no difference in “death and nonfatal MI” (2.1% vs 2.2%, p=0.35). Mortailty was 1.5% in both groups. More patients in the CCTA group had invasive testing (12.2% vs 8.1%). 

Comments: This trial was unblinded, so softer outcomes like unstable angina, and even MI could easily have been influenced by the results of the tests. That being said, it is clearly a negative trial. There was no difference in any of the clinical outcomes, but CTCA did result in more invasive testing, so it appears CCTA caused net harm.

SCOT-HEART TRIAL: The SCOT-HEART investigators. CT coronary angiography in patients with suspected angina due to coronary heart disease (SCOT-HEART): an open-label, parallel-group, multicentre trial. Lancet (London, England). 2015; 385(9985):2383-91. [pubmed] [free full text]

This is an open label, multicenter RCT than enrolled 4146 patients with suspected stable angina, but no active chest pain, from cardiologists offices. All patients underwent a full assessment, including stress test if the cardiologist thought it was warranted, and then they were randomized to CCTA or nothing. This was a pretty high risk group – the cardiologists thought that ½ had CAD and ⅓ had angina after the baseline assessment. The primary outcome was very weird – it was the proportion of patients diagnosed with angina at 6 weeks (in an open label trial.) In other words, they purposely and fundamentally designed their study to include incorporation bias. Not surprisingly, when presented with the CCTA results, the cardiologist changed some diagnoses and felt more confident in their diagnoses. However, how do we know that the new diagnoses were correct? We don’t. In fact, there is pretty good evidence that the cardiologists were wrong. Patients were followed for a median of 1.7 years, and there were no differences in any clinical outcomes. Death was 0.8% with CCTA and 1.0% without (p=0.64). MI was 1.1% with CCTA and 1.7% without (p=0.08). Revascularization was 11.2% with CCTA and 9.7% without (p=0.06). Hospitalization for chest pain was 11.9% with CCTA and 12.7% without (P=0.39).

Comments: This is pretty clearly a negative study. If anything, it seems to show that CCTA falsely reassures or confuses cardiologists. Although they felt more confident in their diagnosis with CCTA, there were absolutely no differences in any clinical outcomes, so they were clearly wrong. 

SCOT-HEART TRIAL long term follow up:The SCOT-HEART investigators, Newby DE, Adamson PD, et al. Coronary CT Angiography and 5-Year Risk of Myocardial Infarction. The New England journal of medicine. 2018; 379(10):924-933. [pubmed]

This is long term follow-up data from the SCOT-HEART trial. They followed the 4146 patients for a median of 4.8 years. The big news from this paper is that their primary long term outcome, a composite of death from coronary heart disease and non-fatal MI, was decreased (2.3% vs 3.9%, p=0.004). Although that sounds like we might have the first evidence of a clinical benefit from CCTA, there are some significant problems. They use a composite outcome that combines 2 very nebulous outcomes. (There are a lot of reasons not to like disease specific mortality.) All cause mortality was not changed (2.1% in both groups), and the disease specific mortality was also statistically unchanged (0.2% vs 0.6%). That means the positive outcome was driven by the 1.4% difference in non-fatal MI. However, as I have discussed before, nonfatal MI is a surrogate outcome without clear implications for our patients. This is especially true in an unblinded trial, without any formal adjudication of the end points. (You can imagine a doctor saying, “I know that troponin was positive, but this definitely wasn’t MI, because their CCTA was negative.”) Given the significant possible bias, the fact that this is a small change in a surrogate outcome, and the fact that other trials don’t show any clinical benefit from CCTA, I don’t think this study changes practice in any way.

Nabi F, Kassi M, Muhyieddeen K, et al. Optimizing Evaluation of Patients with Low-to-Intermediate-Risk Acute Chest Pain: A Randomized Study Comparing Stress Myocardial Perfusion Tomography Incorporating Stress-Only Imaging Versus Cardiac CT. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2016; 57(3):378-84. [pubmed] [free full text]

This was a single center RCT that randomized 598 inpatients to either CCTA or single-photon emission computerized tomography (SPECT). All patients were adults who were admitted to the hospital specifically for a SPECT. Length of stay was statistically shorter with CCTA, although the actual difference (20 vs 23 hours) is not very impressive. There were no deaths and only 1 MI over the 6 month follow up period. They don’t tell us which group the MI occurred in. There were 21 diagnoses of unstable angina, 18 of which occurred during the index admission. There was no difference in ACS between the two groups (3% vs 4.6%, p=0.39).

Comments: Once again, there was no clinical benefit, but the rate of cardiac events was so low, a benefit would have been impossible. More than 40% of eligible patients refused to participate. Length of stay is not a very important primary outcome, and seeing as the SPECT probably wasn’t needed, it is a false comparator anyway. 

PERFECT TRIAL: Uretsky S, Argulian E, Supariwala A, et al. Comparative effectiveness of coronary CT angiography vs stress cardiac imaging in patients following hospital admission for chest pain work-up: The Prospective First Evaluation in Chest Pain (PERFECT) Trial. Journal of nuclear cardiology. 2017; 24(4):1267-1278. [pubmed]

This is a randomized trial of 411 chest pain patients admitted to one of two hospitals. They were randomized to either CCTA or stress testing (with modality chosen by the treating physicians). There was no difference in length of hospital stay. There was no difference in medication use over the next year. CCTA did lead to more downstream testing and more revascularizations. 11% of the CCTA group underwent an invasive angiogram as compared to 2% of the stress test group. Once again, there were essentially no clinical outcomes over the 1 year follow up period. There were no deaths. There was 1 MI in the stress testing group and 2 in the CCTA group. There was also one unstable angina in the CCTA group.

Comments: Again, you can’t show a benefit from CCTA if no one has an important outcome. This trial had 4 primary outcomes, and they also changed their primary outcomes on clinicaltrials.gov after the trial was over. But either way, it’s a definitely negative trial for CCTA – no benefit, but an increase in invasive procedures.

Bottom Line

This group of trials is more heterogenous than the first, but I think they still clearly demonstrate no clinical benefits, but some likely harm from CCTA in the forms of more unnecessary invasive tests. Two trials reported benefits, but there were some significant problems with both of those trials, so I don’t think they can overrule the majority of the evidence, which is clearly negative. 

Systematic Review

Gongora CA et al. Acute Chest Pain Evaluation using Coronary computed tomography Angiography Compared with Standard of Care: A Meta-Analysis of Randomised Clinical Trials. HEART 2018. PMID: 28855273

This systematic review and meta-analysis included RCTs of ED patients, comparing CCTA to standard care. There were 10 trials, encompassing 6285 patients. CCTA did not result in a decrease in any of mortality (RR 0.48; 95% CI 0.17-1.36; p = 0.17), MI (RR 0.82; 95% CI 0.49 -1.39; p = 0.47), or MACE (RR 0.98; 95% CI 0.67-1.43; p = 0.92), although the mortality and MI numbers probably warrant further study. Despite the lack of benefit, CCTA did result in an increase in invasive angiography (RR 1.32; 95% CI 1.07 -1.63; p = 0.01) and revascularization (RR 1.77; 95% CI 1.35-2.31; p<0.0001), which means that CCTA caused net harm. None of the individual trials were perfect, as is discussed above.

Harms

There are a few direct harms of CCTA. Obviously it is a contrast CT. I doubt that means anything from a renal standpoint, and these are stable patients so hypotension shouldn’t be a big issue, but there will be a few anaphylactoid reactions and other adverse events using this strategy. The radiation is important to consider. Early studies reported exposures of 10-25 mSv. (Goldstein 2007; Meijboom 2008) The radiation dose of the single CT has come down with time, and is now probably in the range of 5 mSv. (SCOT-HEART 2015) However, CCTA also increases downstream testing, and so the total radiation dose for a CCTA strategy might be expected to be around 25 mSV. (Levsky 2015)

In my mind, the biggest harm is the increase in unnecessary invasive testing and procedures. Almost all of these trials showed an increase in invasive angiography and revascularization with CCTA. That would be fine if it was preventing downstream MIs or deaths, but it isn’t. There was no benefit seen in any of these trials, so all of those patients underwent invasive procedures and the significant risk of harm for no reason.

Summary

At this point, I think there is pretty strong evidence that we should not be using CCTA in clinical practice. These trials paint a picture of net harm, at least when CCTA is used in the low risk patient populations studied here. (These trials have mostly been done in the United States, where chest pain is managed unlike anywhere else in the world.)

Could there be a role for CCTA? The benefit of angiography and PCI in patients with NSTEMI is marginal. Our current practice of taking all NSTEMI patients to the cath lab is not saving lives, and the only benefit seen is a small decrease in the surrogate outcome of nonfatal MI. (Fanning 2016) However, NSTEMI is a broad category of patients. You could imagine that a subset of these patients truly might benefit from PCI, while others do not, resulting in the net neutral clinical trials. I could imagine that CCTA (especially with FFR) might be helpful in this high risk group of patients to determine who actually requires PCI, although obviously I would like to see some evidence supporting that hypothesis. (There is, in fact, an ongoing RCT looking at this question, so I will update when it is published: NCT02284191.)

To think about this from a different perspective, we can go back to the original data looking at the accuracy of CCTA. The sensitivity of CCTA is excellent, but the specificity is only moderate. What other tests do we have that are highly sensitive but not very specific? The DDimer is an excellent example. We don’t use DDimer in low risk patients, because we know there will be too many false positives, but for some reason we have focused on low risk patients with CCTA. However, in a patient you are considering sending for a CT pulmonary angiogram, the DDimer can be a valuable test. Similarly, if you have a patient you have already decided to send for an invasive angiogram, it is possible that CCTA could be used to eliminate the need for the more invasive traditional test. There may be a role in differentiating between type 1 and type 2 MIs. There may be a role, but after reviewing this evidence, I think it is very clear that CCTA has no role in the workup of low risk emergency department patients. 

Unlike stress testing, which has been widely adopted without being adequately studied, we have multiple RCTs looking at CCTA, and they are clearly negative. No benefit has been demonstrated, but there is clear harm from radiation and increased invasive testing. So, unless you are participating in some kind of research, there appears to be no role for CCTA.

Other FOAMed

REBEL EM: Coronary Computed Tomography Angiography (CCTA): The Holy Grail of “Low Risk” Chest Pain Evaluation?

Multiple posts from EM Literature of Note: 

• Pooled CCTA Outcomes
• Excitement For/Failure of CCTA
• And Here’s Why CTCA Is Useless
• CCTA Is A Bad Shortcut Around Bad Care

EM Nerd-The Case of the Anatomic Heart Revisited

Five Reasons I Don’t Believe an Imaging Test Improves Outcomes

References

Budoff MJ, Dowe D, Jollis JG, et al. Diagnostic performance of 64-multidetector row coronary computed tomographic angiography for evaluation of coronary artery stenosis in individuals without known coronary artery disease: results from the prospective multicenter ACCURACY (Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography) trial. Journal of the American College of Cardiology. 2008; 52(21):1724-32. [pubmed] [free full text]

Dedic A, Lubbers MM, Schaap J, et al. Coronary CT Angiography for Suspected ACS in the Era of High-Sensitivity Troponins: Randomized Multicenter Study. Journal of the American College of Cardiology. 2016; 67(1):16-26. [pubmed] [free full text]

Douglas PS, Hoffmann U, Patel MR, et al. Outcomes of anatomical versus functional testing for coronary artery disease. The New England journal of medicine. 2015; 372(14):1291-300. [pubmed] [free full text]

Fanning JP, Nyong J, Scott IA, Aroney CN, Walters DL. Routine invasive strategies versus selective invasive strategies for unstable angina and non-ST elevation myocardial infarction in the stent era. The Cochrane database of systematic reviews. 2016.

Goldstein JA, Gallagher MJ, O’Neill WW, Ross MA, O’Neil BJ, Raff GL. A randomized controlled trial of multi-slice coronary computed tomography for evaluation of acute chest pain. Journal of the American College of Cardiology. 2007; 49(8):863-71. [pubmed] [free full text]

Goldstein JA, Chinnaiyan KM, Abidov A, et al. The CT-STAT (Coronary Computed Tomographic Angiography for Systematic Triage of Acute Chest Pain Patients to Treatment) trial. Journal of the American College of Cardiology. 2011; 58(14):1414-22. [pubmed] [free full text]

Gongora CA et al. Acute Chest Pain Evaluation using Coronary computed tomography Angiography Compared with Standard of Care: A Meta-Analysis of Randomised Clinical Trials. HEART 2018. PMID: 28855273

Hamilton-Craig C, Fifoot A, Hansen M, et al. Diagnostic performance and cost of CT angiography versus stress ECG–a randomized prospective study of suspected acute coronary syndrome chest pain in the emergency department (CT-COMPARE). International journal of cardiology. 2014; 177(3):867-73. [pubmed] [free full text]

Hoffmann U, Truong QA, Schoenfeld DA, et al. Coronary CT angiography versus standard evaluation in acute chest pain. The New England journal of medicine. 2012; 367(4):299-308. [pubmed] [free full text]

Hollander JE, Gatsonis C, Greco EM, et al. Coronary Computed Tomography Angiography Versus Traditional Care: Comparison of One-Year Outcomes and Resource Use. Annals of emergency medicine. 2016; 67(4):460-468.e1. [pubmed]

Levsky JM, Spevack DM, Travin MI, et al. Coronary Computed Tomography Angiography Versus Radionuclide Myocardial Perfusion Imaging in Patients With Chest Pain Admitted to Telemetry: A Randomized Trial. Annals of internal medicine. 2015; 163(3):174-83. [pubmed] [free full text]

Linde JJ, Hove JD, Sørgaard M, et al. Long-Term Clinical Impact of Coronary CT Angiography in Patients With Recent Acute-Onset Chest Pain: The Randomized Controlled CATCH Trial. JACC. Cardiovascular imaging. 2015; 8(12):1404-1413. [pubmed] [free full text]

Litt HI, Gatsonis C, Snyder B, et al. CT angiography for safe discharge of patients with possible acute coronary syndromes. The New England journal of medicine. 2012; 366(15):1393-403. [pubmed] [free full text]

Meijboom WB, Meijs MF, Schuijf JD, et al. Diagnostic accuracy of 64-slice computed tomography coronary angiography: a prospective, multicenter, multivendor study. Journal of the American College of Cardiology. 2008; 52(25):2135-44. [pubmed] [free full text]

Miller JM, Rochitte CE, Dewey M, et al. Diagnostic performance of coronary angiography by 64-row CT. The New England journal of medicine. 2008; 359(22):2324-36. [pubmed] [free full text]

Nabi F, Kassi M, Muhyieddeen K, et al. Optimizing Evaluation of Patients with Low-to-Intermediate-Risk Acute Chest Pain: A Randomized Study Comparing Stress Myocardial Perfusion Tomography Incorporating Stress-Only Imaging Versus Cardiac CT. Journal of nuclear medicine : official publication, Society of Nuclear Medicine. 2016; 57(3):378-84. [pubmed] [free full text]

SCOT-HEART investigators. CT coronary angiography in patients with suspected angina due to coronary heart disease (SCOT-HEART): an open-label, parallel-group, multicentre trial. Lancet (London, England). 2015; 385(9985):2383-91. [pubmed] [free full text]

SCOT-HEART investigators, Newby DE, Adamson PD, et al. Coronary CT Angiography and 5-Year Risk of Myocardial Infarction. The New England journal of medicine. 2018; 379(10):924-933. [pubmed]

Uretsky S, Argulian E, Supariwala A, et al. Comparative effectiveness of coronary CT angiography vs stress cardiac imaging in patients following hospital admission for chest pain work-up: The Prospective First Evaluation in Chest Pain (PERFECT) Trial. Journal of nuclear cardiology. 2017; 24(4):1267-1278. [pubmed]