Coronary CT angiography in clinical practice: Experiences from Denmark

Coronary multislice computed tomography angiography (CTA) is increasingly used for non-invasive visualization of the coronary arteries. More than 100 single-center studies, numerous meta-analyses, and 4 prospective multicenter trials have documented consistently high sensitivities (90–99%) and negative predictive values (93–99%) of coronary CTA in detecting significant coronary artery disease (CAD) (1,2). Accordingly, in clinical practice, coronary CTA is often used as a gatekeeper for invasive coronary angiography (ICA). However, the worldwide expansion of coronary CTA has been a subject of controversy (3). Because of a suboptimal specificity of coronary CTA in quantifying CAD and the lack of functional assessment of lesions, it has been suggested that this technology will add unnecessary invasive diagnostic testing and coronary interventions in the workup of patients suspected of stable CAD (3). Another concern following the introduction of coronary CTA has been associated with the inherent exposure to ionizing radiation and the potential association with cancer risk (4). In the event of a well-indicated investigation, cancer risk relative to the diagnostic information obtained by coronary CTA may be small. However, with the increasing use of coronary CTA in low-risk patients, especially in younger women, and when repeated procedures are being performed, this risk should be acknowledged (5). Several radiation dose–saving techniques together with an increased awareness of physicians in these matters have the potential of significantly reducing the radiation exposure associated with coronary CTA (6).

In this issue of the Scandinavian Cardiovascular Journal, two Danish reports on local experiences related to the use of coronary CTA in routine clinical practice are presented (7,8). Both reports show an impressive decline over a 4-year time span in the effective radiation dose associated with “real-world” coronary CTA, both by using conventional 64-slice CT technology (from > 17 to ∼5 mSv) (8) and by using a more recent CT scanner technology (from > 4 to ∼ 2 mSv) (7). Of note, this reduction in radiation dose has developed without sacrificing image quality and thus the diagnostic performance of coronary CTA. With the most recent CT technology, coronary CTA may be performed in large subsets of patients with a sub-mSv radiation dose. Accordingly, the radiation dose associated with contemporary coronary CTA is substantially lower than that associated with SPECT or ICA. It must be anticipated that these improvements in coronary CTA practice, together with faster acquisition times, and higher spatial as well as temporal resolution may influence indications for coronary and cardiac CT in the future.

From registry data, N rgaar et al described the consequences of frontline coronary CTA assessment of patients suspected of CAD on downstream utilization of ICA and revascularization procedures. The fact that data were consecutively provided over a 4-year observation period, and thus, the study cohort was large, comprising > 3000 patients, is noteworthy. One interesting finding was a reduction in referrals to ICA during the observation period irrespective of unchanged patient characteristics. However, the authors did not provide information on symptoms (temporal changes in proportions of individuals with or without non-anginal, atypical, or anginal symptoms). Probably, symptoms are the most important “patient characteristic” driver for ICA in patients suspected of stable CAD. One might interpret the reduction in referrals to ICA during the observation period as a result of the introduction of coronary CTA. However, this would be a flawed conclusion since the authors did not compare their findings with patient flows during the standard diagnostic practice before the introduction of coronary CTA. Of note is the relatively high ICA rate ranging between 21 and 24% following coronary CTA assessment during the first three years of the observation period. The drop-off in referrals to ICA in 2013 to around 11% may, as suggested by the authors, be explained by the addition of myocardial perfusion imaging (MPI) testing in patients with a positive coronary CTA result before referring patients to ICA. Accordingly, several studies have documented increasing diagnostic specificity of coronary CTA by adding functional information from either MPI or coronary CT-derived fractional flow reserve. Moreover, it was shown by Nørgaard et al. that the proportion of patients having revascularization performed during follow-up was reduced from 2010 until 2013. However, this is not surprising since group observation times were not truncated (e.g., observation time was substantially longer for the “2010” versus the “2013” cohort). Thus, from the study by Nørgaard et al., it cannot be concluded that coronary CTA leads to “less invasive angiography or less revascularization.” Rather, this study underscores two important issues regarding frontline coronary CTA testing in patients suspected of CAD: 1. an inherent risk of subsequent “unnecessary” ICA even in a low-intermediate pre-test risk cohort, that is, between 2010 and 2012, ICA following coronary CTA was performed in more than 20% of the patients (of whom less than 30% were revascularized); 2. in patients with a positive coronary CTA result, the diagnostic performance of the test may be improved by adding the information from MPI testing (although data on subsequent clinical events were not reported in the study by Nørgaard et al.), that is, in 2013 after the introduction of MPI, the ICA referral rate after coronary CTA was reduced to 10% (of whom approximately 40% were revascularized). Expanding the study by Nørgaard et al. with information from a matched standard diagnostic practice (historical or parallel) cohort for comparison of the real-world influence of frontline coronary CTA with MPI testing on subsequent resource utilization and costs relative to clinical outcomes would be highly relevant.

It is well known that contrast (e.g., in relation to ICA or CT) is well tolerated. Serious short-term complications such as “real” allergic reactions and nephropathy occur very rarely. In the report by Pedersen et al. comprising 416 consecutive patients suspected of stable CAD, the renal function observation time was extended to 2 months. As expected in this rather low-risk cohort, contrast was well tolerated using estimated glomerular filtration rate (eGFR) as the reference standard. Rather paradoxically, eGFR improved at follow-up in the small subcohorts with diabetes mellitus or impaired pre-test renal function. Although this change in eGFR was statistically significant, it seems to have limited clinical relevance.

Even with the highest temporal resolution CT technology, the best image quality in coronary CTA is achieved in heart rates of < 60/min (9). In the study by Nørgaard et al., pre-coronary CTA beta-blockade was necessary in 85% of the patients in order to achieve a target scan heart rate of 60/min or less. The mean dose of metoprolol was not reported; however, the heart-rate-lowering regime was rather extensive: it involved applying 50 + 100 mg metoprolol on the day before and on the day of the scan, with the addition of i.v. boluses of 5 mg (maximum: 50 mg) in the event that the target heart rate was not achieved. The occurrence of side effects to metoprolol was rigorously evaluated using a detailed questionnaire. The beta-blocking regimen was well tolerated with no immediate hemodynamic complications or significant side effects before or after the scan. Of note, in the few patients, relevant for coronary CTA, with contraindications to beta-blockers, ivabradine or verapamil may be used as an alternative. The report by Nørgaard et al. underscores the safety of coronary CTA in clinical practice. However, in more susceptible patient groups (e.g., pre-procedural CT in relation to planning of TAVI or LAA closure), contrast or beta-blockers may be less well tolerated. In such “high risk” patients, the clinical question typically is not related to the coronary status, hence requirements to image quality are less than in standard coronary CTA. Therefore, in general, fast acquisition protocols with low contrast volumes (30–40 ml) and prevention of heart-lowering medication in such patients are feasible without affecting the diagnostic value of cardiac CT.

Do observational real-world reports as presented by Nørgaard and Pedersen add to the body of knowledge on the diagnostic performance and safety of coronary CTA? Guidelines and meta-analyses on the use of diagnostic tests are typically based on studies from large-scale investigations involving observers with the highest competence levels. Accordingly, the specificity of a diagnostic test tends to decline with time as the test may be implemented in less experienced laboratories and applied to a wider spectrum of patients. Moreover, refusal to publish negative reports may falsely have inflated the perceived diagnostic performance and safety of tests. Thus, there may be large discrepancies in the diagnostic performance, including consequences on the utilization of downstream test procedures and the incidence of adverse effects of a diagnostic test between what is reported in the literature versus what we experience in clinical practice. Although these issues may seem trivial, we often disregard them in our enthusiasm with new technologies, for example, in relation to coronary CTA with high availability and fast establishment of fascinating colorful 3D images of our coronaries. These issues underscore for our understanding of the truth the importance of real-world reports as provided by Nørgaard and Jensen in this issue of the Scandinavian Cardiovascular Journal.

Declaration of interest: The author has received unrestricted research grants from Siemens and Edwards Lifesciences. The author alone is responsible for the content and writing of the paper.