Computed tomography coronary angiography: Is it such a useful test for emergency department chest pain patients?

Assessment and investigation of patients presenting to EDs with chest pain is an increasing challenge world-wide. The dual priorities are to identify patients who have an acute coronary syndrome (ACS) so that emergent treatment can be given, as well as those patients who have significant coronary artery disease (CAD) without infarction, so that interventions to minimise morbidity and mortality can be undertaken. Identifying the first group relies on clinical assessment, biomarkers and ECG analysis. Identifying the second group has, until recently, relied largely on functional tests and nuclear medicine studies. Multi-slice computed tomography (CT) coronary angiography (CTCA) is an emerging technology for the assessment of coronary arteries that is proposed as an alternative testing strategy. Three recent studies investigating the use of CTCA in ED chest pain patients have reported several benefits including a shorter ED stay, fewer admissions, good prognostic performance of a negative scan and possible improvement in the accurate identification of CAD compared with alternative strategies.1-3 There is good evidence that CTCA is sensitive for the detection of CAD. A recent systematic review and meta-analysis concluded that CTCA had 94% sensitivity (95% confidence interval [CI], 61–99%) and 87% specificity (95% CI, 16–100%) for CAD.4 This review reported sensitivity as high as 99% with ≥64-slice CT technology, but that about 9% of tests were non-diagnostic or inconclusive. It also reported very good negative predictive value (NPV) for future cardiac events identifying just one subsequent ACS death among 1334 patients with a negative CTCA scan, equating to a rate of 0.07% (95% CI, 0.01–0.4%). However, not all patients are suitable to undergo this test because of known pre-existing CAD, or other contraindications such as irregular heart rhythm, high pulse rate (with rate control contraindicated), renal or thyroid disease, contrast allergy, metformin use, and pregnancy.5 The proportion of eligible patients has been estimated that about 50%.5 In the American College of Radiology Imaging Network – Pennsylvania Department of Health study,1 CTCA halved the admission rate (23% vs 50%) and reduced the ED length of stay (LOS) by about 25% (from ∼25 to ∼18 h). No patient with a negative CTCA died or had an myocardial infarction (MI) within the next 30 days. However, the rate of subsequent MI was very low: only 2 of 1357 patients not diagnosed with an MI at the index visit had an MI within 30 days (0.15%). In addition, the participants were at low risk of ACS: more than 85% were Thrombolysis In Myocardial Infarction (TIMI) 0 or 1. In the Australasian context, approximately 50% of ED chest pain patients are TIMI 0 or 1.6 Regarding LOS, data from Victoria, Australia report the average ED LOS for discharged chest pain patients is of the order of 7–8 h7 and that it is uncommon for patients to have further tests for CAD during the index ED visit. These data highlight differences in practice between the US and Australasian contexts, and the very low risk of this group for adverse events – serious challenges to the role of CTCA in this context. In the Coronary Computed Tomographic Angiography for Systematic Triage of Acute Chest Pain Patients to Treatment (CT-STAT) study,2 CTCA reduced the time to diagnosis by 54% (∼3 h vs ∼6 h) compared with myocardial perfusion imaging (MPI). Time to diagnosis was defined as time for randomisation for inclusion in the study to when test results were available; LOS was not reported. ED costs were reduced by 38%, but total costs of care (ED and non-ED components combined) were not reported. There were no significant differences in either early or late clinical outcomes between the CTCA and MPI groups. Once again, the trial population was at low risk of ACS. Although TIMI score <4 could be included, the median TIMI score was 1. The time to diagnosis from randomisation is not a clinically meaningful end-point, and cost data are difficult to interpret when limited to ED costs. Of particular importance, patients having CTCA were six times more likely to have additional non-invasive tests after ED discharge. This adds risks of complications and additional radiation exposure, depending on the test chosen. The Rule Out Myocardial Infarction using Computer Assisted Tomography (ROMICAT) II study3 also identified a reduced LOS in patients randomised to CTCA compared with standard care (∼9 h vs ∼27 h), and a substantial reduction in ED costs (19%). However, total costs were similar between the groups. There were no undetected ACS and no significant differences in major adverse cardiovascular events (MACE) at 28 days. CTCA patients had a higher rate of additional testing (27% vs 12%) with a trend towards a higher rate of coronary angiography (12% vs 8%). Importantly, this study excluded patients with known CAD, who make up approximately 28% of Australasian chest pain cohorts.8 Again, there are stark contrasts with Australasian practice and LOS data. In these studies, access to CTCA was expedited or patient enrolment was limited to ‘office’ hours rather than real world CTCA access in Australasia. The CT-STAT study2 compared CTCA with MPI and found no difference in early or late clinical outcomes. The ROMICAT II study3 compared CTCA with standard care (various) with the same finding. Conti et al.9 compared exercise ECG and exercise MPI in ED chest pain patients without known CAD. They reported that MPI had a sensitivity of 87% with NPV for MACE of 97.2% as compared with exercise ECG sensitivity of 19.5% and NPV 86%. The Randomised Assessment of Treatment using Panel Assay of Cardiac Markers study10 reported a sensitivity of 40% for exercise ECG for MACE at 3 months. In a meta-analysis of the cost-effectiveness of various prognostic strategies for the investigation of troponin-negative patients, Goodacre et al. reported that exercise ECG was dominated (resulted in a lower incremental cost-effectiveness ratio despite higher cost) by other strategies, including no testing.4 The same meta-analysis reported that cost-effectiveness of CTCA depended on the rate of future adverse events, with CTCA likely to be cost-effective if the combined risk of death and non-fatal MI within the time period assumed to be influenced by testing exceeded 2% (using a £30 000/quality-adjusted life year (QALY) threshold) or 2.9% (using a £20 000/QALY threshold). This is unlikely to be the case for TIMI 0 or 1 patients, who have normal ECG and biomarkers in ED. One study of TIMI 0 patients reported a 3-month event rate (death, subsequent MI or revascularisation) of 0.47%.11 In another study, Goodacre et al. report a 3-month event rate (death or non-fatal MI) of 0.4% (95% CI, 0.02–0.08).12 This highlights the importance of patient selection for optimising the use of CTCA. At present, there are insufficient data to define a suitable ‘at-risk’ group. That said, in an MPI study, Conti et al.12 found that patients with diabetes or metabolic syndrome had double the rate of MACE as other patients. Perhaps, this can inform future research into case selection for CTCA. The risks and costs of any flow-on effects of CTCA are also important in determining its utility. Shreibati et al.13 analysed an administrative dataset of non-emergent tests for CAD in patients aged over 65 years and found that CTCA resulted in approximately twice as many angiograms, percutaneous interventions and coronary artery bypass grafts but without a significant difference in 180-day mortality compared with MPI. Taken together, these data challenge the place of routine exercise ECG in prognostic testing for CAD in troponin-negative patients without known CAD. It also raises the possibility that no further testing may be appropriate in very low risk patients, with MPI and CTCA being reserved for higher risk groups – although exactly how these would be defined is unclear. Cancer risk from radiation exposure cannot be ignored. The risk of cancer arising from CTCA radiation exposure using retrospectively gated protocols has been estimated at 0.11–0.13% for men and 0.27–0.37% for women.14 The same study estimated that using prospectively gated protocols, the risk was 0.014–0.017% for men and 0.035–0.06% for women. The risk is also higher in younger patients and in people of Asian ethnicity. Given the high number of chest pain patients and low prevalence of clinically relevant disease, widespread use of CTCA would pose a significant risk for radiation-related cancers for apparently limited gain. Feasibility of CTCA is also a significant barrier to its utility. To be clinically useful, it necessitates access to a scanner, the availability of experienced readers and an extended hour service to match the ED workflow. Current access to CTCA in Australasia is usually limited to ‘office hours’, which would require longer ED stays or hospital admission, and inconvenience to patients awaiting its availability. In theory, CTCA has the capacity to rule-out CAD for ED patients with chest pain, but its impact on clinical outcome is unproven. The rationale for any test is that it improves clinical outcome(s) compared with no test or an alternative test. The event rates in the key clinical trials of CTCA to date have been so low in all arms (<1%) that it has not been possible to determine if CTCA has provided net clinical benefit. The assumption of CTCA use in this context is that the CAD lesions identified are clinically relevant to the patient's presentation. This assumption is questionable given that about 5% of asymptomatic individuals in the relevant age group in the general population have asymptomatic occlusive coronary lesions. As previously discussed, there is some evidence that CTCA increases the rate of angiography and revascularisation without impacting on mortality. If the cost of the test and its consequences do not result in a clinically meaningful benefit to patients, it is hard to justify. Given the lack of evidence of demonstrated clinical benefit, low cost-effectiveness and the potential for increased risk for patients, I believe that CTCA is not suitable for broad application in ED chest pain patients with a negative ACS work-up. It may prove to be of benefit in selected patient groups, but these are yet to be defined. A-MK is a co-author of the NHF Guidelines for the management of Acute Coronary Syndrome and its addenda, and is also a member of the editorial board of Emergency Medicine Australasia.