Article Text


Compounding uncertainties with first-line computed tomography coronary angiography
  1. Paul C Cremer
  1. Correspondence to Dr Paul C Cremer, Department of Cardiovascular Imaging, Heart and Vascular Institute, Cleveland Clinic, Cleveland, OH 44195, USA; cremerp{at}

Statistics from

Every year, hundreds of thousands of stable patients with chest pain are evaluated at rapid access clinics in England and in various other settings throughout the UK. The majority of these patients are at low risk for short-term adverse cardiovascular events, though to decrease lifetime risk, attention to proven preventative therapies is of paramount importance. Oftentimes, however, the primary focus is on near-term diagnostic testing, and the low yield of this testing has contributed to increasing healthcare costs and misallocation of resources.1 More recently, the National Institute for Health and Care Excellence (NICE) recommended CT coronary angiography (CTCA) as a first-line test, a dramatic departure from previous guidelines, centred on controversial interpretations of CTCA clinical trials.2

In the Heart paper by, Dreisbach et al, further the controversy by providing valuable insights into the challenges in implementing CTCA as a first-line test, specifically emphasising deficiencies in trained personnel and capable equipment.3 By their estimates, approximately 42 000 CTCAs are performed annually in the UK with an expected increase to 350 000. Even though the magnitude of this increase is alarming, the geographic variation in available resources is also striking. Throughout the UK, there are only 29 level 3 CTCA practitioners, including 3 in Scotland, 1 in Northern Ireland and 0 in Wales. The necessary expertise in these nations is undoubtedly a concern. Moreover, 31% of health regions do not have a single accredited CTCA practitioner. Importantly, and not addressed in this manuscript, how many nurses and technologists are trained in CTCA? Presumably, the deficit parallels the extent and geographic heterogeneity observed with physicians.

Regarding equipment, 304 CTCA-capable scanners are operated by the National Health Service, though as the authors emphasise, the distinction between ‘CTCA-capable’ and ‘currently performing CTCA’ is important. Nonetheless, several CTCAs per day would need to be performed on each of these scanners, which may not appear onerous, except that these scanners are likely already operating at or near full capacity. In fact, based on the estimated increase with implementation of the NICE guidelines, 7.6% of all CT examinations in the UK would be for coronary artery evaluation.

Dreisbach et al, as well as the British Societies of Cardiovascular Imaging and Cardiac CT, should be commended for highlighting these deficiencies, and for striving to address the inevitable initial differences in diagnostic quality and radiation dose these low-risk patients will receive.4 However, regarding the model developed to estimate annual CTCA volume, limitations related to external validity and expression of uncertainty warrant discussion. In general, the model employed a ‘back calculation’, starting with the number of percutaneous interventions, next assuming the number of coronary artery bypass graft surgeries, then accounting for the increase in revascularisation with CTCA, and finally extrapolating the ratio of CTCAs to revascularisation. The data for these assumptions were derived from SCOT-HEART, and the external validity of this trial for the entire UK is dubious.5 For one, SCOT-HEART screened nearly 10 000 patients of which nearly 9000 met eligibility criteria, but less than half were actually enrolled. Furthermore, given the resources afforded to patients in SCOT-HEART, contrasted with the current provision of CTCA, downstream management could differ substantially across the UK.

Even without these issues related to generalisability, statistical estimates of the uncertainty in the provided calculations would also help appreciate the possible range in number of CTCA scans. In making broader inferences from SCOT-HEART, the point estimate provides the single best guess, but it is also necessary to convey the error of this estimate as it applies to the UK population. Given the layers of assumptions in this analysis, this concept is especially relevant. For example, if the expected increase in revascularisation with CTCA based on the intention-to-treat data from SCOT-HEART is towards the lower bound of a 95% CI and the ratio of CTCA to revascularisation based on the as-treated data is also at the lower end, then the total number of CTCAs would be considerably less than the point estimate provided (figure 1).

Figure 1

Crude calculations to express uncertainty in estimates for total CTCA increase in UK population based on SCOT-HEART sample. Bootstrap resampling (n=1000) was performed to estimate 95% CIs for the number of CABGs as a percentage of total revascularisations, the increase in revascularisation with CTCA from the intention-to-treat analysis, and the ratio of CTCA to revascularisation from the as-treated analysis. The upper and lower bounds of these estimates were used in subsequent calculations to provide a general sense of the range of possible increases in CTCA volume as a result of implementation of the NICE guidelines. Analysis performed with R Core Team (2017). R: A language for environment and statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL CABG, coronary artery bypass grafting; CTCA, CT coronary angiography; NICE, National Institute for Health and Care Excellence; PCI, percutaneous coronary intervention.

Discussion of these limitations suggests substantial uncertainty regarding the actual increase in CTCAs. Still, with implementation of the NICE guidelines, Dreisbach et al have demonstrated that CTCAs will increase precipitously, and the UK is ill prepared both in terms of trained personnel and capable equipment. Given this predicament, despite increasing revascularisation, the failure of CTCA to demonstrate either superiority or non-inferiority for cardiac events should be emphasised.6 With imaging, the value of a test is contingent upon technical considerations and diagnostic accuracy, and downstream treatment considerations, costs and outcomes.7 An appreciation for this hierarchy is especially important for common problems such as stable chest pain, given the far-reaching and costly implications. With the inability of first-line CTCA to satisfy this standard for value, do the controversial NICE guidelines mandate substantial investment in developing infrastructure and expertise, especially when established alternative pathways exist for these low-risk patients? Undoubtedly, with the broad application of CTCA, many patients, perhaps hundreds of thousands annually in the UK, will receive unnecessary advanced imaging.


View Abstract


  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles