Computed tomography versus invasive coronary angiography: design and methods of the pragmatic randomised multicentre DISCHARGE trial
- First Online:
- Cite this article as:
- Napp, A.E., Haase, R., Laule, M. et al. Eur Radiol (2016). doi:10.1007/s00330-016-4620-z
More than 3.5 million invasive coronary angiographies (ICA) are performed in Europe annually. Approximately 2 million of these invasive procedures might be reduced by noninvasive tests because no coronary intervention is performed. Computed tomography (CT) is the most accurate noninvasive test for detection and exclusion of coronary artery disease (CAD). To investigate the comparative effectiveness of CT and ICA, we designed the European pragmatic multicentre DISCHARGE trial funded by the 7th Framework Programme of the European Union (EC-GA 603266).
In this trial, patients with a low-to-intermediate pretest probability (10–60 %) of suspected CAD and a clinical indication for ICA because of stable chest pain will be randomised in a 1-to-1 ratio to CT or ICA. CT and ICA findings guide subsequent management decisions by the local heart teams according to current evidence and European guidelines.
Major adverse cardiovascular events (MACE) defined as cardiovascular death, myocardial infarction and stroke as a composite endpoint will be the primary outcome measure. Secondary and other outcomes include cost-effectiveness, radiation exposure, health-related quality of life (HRQoL), socioeconomic status, lifestyle, adverse events related to CT/ICA, and gender differences.
The DISCHARGE trial will assess the comparative effectiveness of CT and ICA.
• Coronary artery disease (CAD) is a major cause of morbidity and mortality.
• Invasive coronary angiography (ICA) is the reference standard for detection of CAD.
• Noninvasive computed tomography angiography excludes CAD with high sensitivity.
• CT may effectively reduce the approximately 2 million negative ICAs in Europe.
• DISCHARGE addresses this hypothesis in patients with low-to-intermediate pretest probability for CAD.
KeywordsComputed tomography Angiography Invasive coronary angiography Adverse events Comparative effectiveness
Coronary artery disease (CAD) is the leading cause of death in high-income countries, and the World Health Organisation predicts that cardiovascular diseases will become the main cause of death in low- and middle-income countries by 2030 . Invasive coronary angiography (ICA) is the reference standard for the diagnosis of CAD and allows immediate intervention if indicated in the case of stenosis. However, only 38–40 % of the patients undergoing ICA in Europe  and the USA  actually have obstructive CAD (defined as at least 50 % coronary diameter stenosis). ICA involves relatively rare but considerable risks for patients such as death, myocardial infarction and stroke [4, 5]. An effective noninvasive test to rule out CAD would be pivotal to reduce the approximately 2 million annual ICAs in Europe that yield negative results . Coronary computed tomography (CT), including coronary calcium score and CT angiography, is the most accurate noninvasive diagnostic imaging strategy for CAD [6, 7] and recommended by current guidelines at low to intermediate pretest probability . Yet, in clinical routine, ICA is still often performed in such patients leading to overdiagnosis. It should also be noted that CT is highly cost-effective, thus promising the greatest societal benefit [9, 10]. With its high sensitivity of ca. 95 % [6, 7], it is the best noninvasive option to exclude CAD in patients with an intermediate pretest probability of CAD , e.g. patients with equivocal stress test results . However, CT is not reimbursed by all European national health systems. CT applied as the first-line imaging modality to determine further workup may result in early and safe discharge of the majority of patients with intermediate risk of CAD and stable chest pain.
Only two published randomised trials in patients with stable chest pain and suspected CAD were published so far: PROMISE  and SCOT-HEART . However, these two trials did not compare CT and ICA but CT with functional testing and standard care, respectively. Thus, we designed the multicentre DISCHARGE trial (Diagnostic Imaging Strategies for Patients with Stable Chest Pain and Intermediate Risk of Coronary Artery Disease: Comparative Effectiveness Research of Existing Technologies). This pragmatic randomised controlled trial (PRCT) will ultimately compare the effectiveness of CT and ICA by randomly evaluating CT versus ICA in patients with stable chest pain and low-to-intermediate pretest probability of CAD who have a clinical indication for ICA.
Registration and study website
Study protocol and key study features
The study protocol (Appendix 1) was developed in accordance with the WHO recommendation (http://www.who.int/rpc/research_ethics/format_rp/en/) and with SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) [15, 16]. The SPIRIT checklist and the WHO checklist are available in Appendices 2 and 3, respectively. Over a period of 2 years, patients with stable chest pain and a pretest probability of CAD of 10–60 % who are indicated to undergo ICA will be randomised to CT or ICA. Chest pain will be assessed using the Rose angina questionnaire . In both groups, patients with negative test results will be discharged and patients with positive test results (at least one obstructive coronary stenosis of at least 50 %) will undergo management by the local heart teams according to guidelines and guided by CT or ICA and noninvasive imaging ischaemia tests. Plaque features and coronary artery calcium scores will be evaluated by CT at each clinical site to guide local decisions about further patient management including optimal medical therapy and intensified risk factor modification . Feedback for each examination is given by the core lab at Charité where image quality, stenosis, plaque characteristics and noncardiac findings will be assessed. Follow-up for up to 4 years is planned.
Pragmatic versus explanatory trials
Explanatory trials are concerned with the question of how and why an intervention works under ideal circumstances [19, 20]. In contrast, for generating high-quality evidence for decision-makers, the pragmatic approach addresses practical questions about the risks, benefits and costs of an intervention as they occur in everyday clinical practice . Therefore, they help facilitate decision-making about whether certain diagnostic procedures or therapies should be used more widely by analysing their effectiveness.
Pragmatic trial design
Primary outcome, secondary outcomes and other outcomes
Primary outcome, secondary and other outcomes
Composite endpoint: major cardiovascular event (MACE) consisting of cardiovascular death, myocardial infarction and stroke
MACE in different composites
MACE in subgroups
Composite endpoint: minor cardiovascular events (MICE) include coronary revascularisation (at least 1 month after initial ICA in order to remove test-driven outcomes), peripheral artery revascularisation, hospitalisation for chest pain/discomfort, emergency department visit for chest pain/discomfort, transient ischaemic attack and congestive heart failure
Influence of computed tomography angiography and invasive coronary angiography on angina pectoris
Procedural complications in the computed tomography angiography and invasive coronary angiography group
Procedural complications of invasive coronary angiography in the computed tomography angiography and invasive coronary angiography group
Comparison of incidental findings in computed tomography angiography and invasive coronary angiography group and potential benefits and harms of findings
Radiation exposure in the computed tomography angiography and invasive coronary angiography group
Socioeconomic status, quality of life and lifestyle
Analysis of differences in Europe
Computed tomography angiography and invasive coronary angiography image-based outcomes
Computed tomography image-based outcomes: image quality
Computed tomography image-based outcomes: heart rate and dose
Computed tomography image-based outcomes: plaques
Invasive coronary angiography outcomes
Planned cross-over in accordance with management recommendations
Imaging ischaemia tests
Comparison of pretest probability calculators
Predictive value of DISCHARGE calculator
Development of novel pretest probability calculator
The primary outcome is the occurrence of major adverse cardiovascular events (MACE). It is a composite endpoint consisting of cardiovascular death, myocardial infarction and stroke. This primary outcome in different composites, e.g. only myocardial infarction and cardiovascular death, will also be analysed as a secondary outcome.
Minor cardiovascular events
Minor cardiovascular events (MICE) are also a composite endpoint. They include coronary revascularisation (at least 1 month after initial ICA in order to remove test-driven outcomes), peripheral artery revascularisation, hospitalisation for chest pain/discomfort, emergency department visit for chest pain/discomfort, transient ischaemic attack and congestive heart failure.
Major procedural complications are a composite endpoint and include death, stroke, myocardial infarction and other complications if they lead to a hospital stay of at least 24 h. Complications not leading to this prolonged stay, e.g. slight bleeding or mild allergic reactions, are classified as minor.
Health-related quality of life (HRQoL)
HRQoL encompasses individual perceptions of physical and mental health as well as functional capacities in everyday life. HRQoL will be assessed using established generic questionnaires, which have been validated previously in all languages of the participating sites and will be completed by participants at baseline (prior to randomisation) and at each follow-up: the Short Form (SF)-12v2,  the EuroQoL (EQ-5D-3 L)  and the Hospital Anxiety and Depression Scale (HADS) .
Expected costs concerning diagnostic tests, treatments and handling of adverse events are calculated by using an analytic decision tree model . The main analysis is the calculation of the incremental cost-effectiveness ratio (ICER) which focuses on differences between the two study groups in baseline costs plus treatment costs of MACE according to differences in the occurrence of MACE. Additional ICERs address serious adverse events (SAEs), adverse events (AEs) as well as mortality due to incidental and non-cardiovascular findings, applying adequate definitions of costs . Downstream treatments after CT/ICA such as medication, further coronary diagnoses and interventions will be recorded. A cost-utility analysis, addressing health-related quality of life, will also be conducted.
Radiation and EU CT quality criteria
Image quality and radiation exposure in all patients will be assessed and monitored. The experience gained will be used to develop guidance on acquisition protocols and radiation protection that will be published as EU CT quality criteria. For this purpose, a guideline on how to generally perform cardiac CT (a 10-step guide) was developed at Charité with input from the DISCHARGE consortium and complemented by scanner-specific protocols.
Other outcomes consist of analyses within the CT and ICA group, European differences and the development of a novel pretest probability calculator.
Eligibility criteria for the DISCHARGE trial
• Patients with suspected* coronary artery disease with stable** chest pain and intermediate pretest probability (10–60 %)*** of CAD clinically referred for ICA
• Patient is or was on haemodialysis
• Patients at least 30 years of age
• No sinus rhythm or pregnancy
• Written informed consent
• Any medical condition that raises concern that participation is not in the best interest of health (e.g. extensive comorbidities)
Methods against bias
Bias will be reduced mainly because the patient population under investigation is eligible for randomisation to either CT or ICA at all sites. Blinding patients to the groups (CT or ICA) is not possible. Allocation concealment and equal allocation to the two trial arms will be ensured by block randomisation with central assignment. In addition, patients at each clinical site will be stratified according to gender to minimise covariate imbalance .
Clinical monitoring and clinical data management
On-site clinical monitoring will be carried out by European Clinical Research Infrastructure Network (ECRIN) in collaboration with KKS Charité and the coordinator team. Clinical monitoring will also be done centrally by checking the electronic case report forms in the study-specific database. (Serious) adverse events will also be entered in the database, and reporting procedures to meet legal requirements have been defined. This database is compliant with Good Clinical Practice and was specially designed for remote data entry to store and manage all study data. Automated checks for plausibility, ranges, consistency and data completeness ensure high quality of the data. These data will be prepared and exported for statistical and other analyses.
Statistical plan for interim analysis
The sample size estimation for the DISCHARGE PRCT is outlined in Appendix 1.
The objectives, methods and design outlined above are unique in that the DISCHARGE trial will assess the comparative effectiveness of CT versus that of ICA in patients with stable chest pain and a clinical indication for ICA based on a low-to-intermediate pretest probability of disease. The primary outcome is MACE and further secondary and other outcomes have been defined prospectively (Table 1).
Two large multicentre randomised studies in patients with suspected CAD based on stable chest pain or chest pain of recent onset have been published so far. The SCOT-HEART study included 4146 patients and found a (nonsignificant) reduction in fatal and nonfatal myocardial infarction after 1.7 years in the CT group, indicating that CT may have benefits in patients with recent onset angina. This was supported by a common reclassification of disease by CT in comparison to standard of care and relevant changes in clinical management, e.g. withholding or prescribing medication . The SCOT-HEART trial also showed that CT leads to more appropriate use of ICA but is more expensive than standard care . For calculating costs, SCOT-HEART used reimbursement in Scotland. After a mean follow-up of 25 months, the PROMISE trial in 10,003 patients found no difference between CT and functional testing in the combined endpoint, which included death, myocardial infarction, hospitalisation for unstable angina, and major procedural complications . This may relate to the composite nature of the endpoint which, contrary to recommendations , reflected both safety (e.g. complications) and effectiveness (e.g. deaths). The events combined in this endpoint clearly differ in clinical relevance, e.g. hospitalisation for unstable angina versus death. Thus a secondary analysis of PROMISE, at best with longer follow-up if funded by the National Institutes of Health (NIH), with regard to the composite of cardiovascular death, myocardial infarction and stroke might better reflect the capabilities of CT . Additionally, statin therapy initiated after plaque detection on CT may reduce mortality but only after longer follow-up than currently reported for SCOT-HEART and PROMISE . A randomised study mainly recruiting in Korea, the CONSERVE study, is also investigating CT versus ICA in a randomised design and is currently recruiting patients (NCT01810198). CONSERVE is potentially biased by the financial support provided by a medical technology company. Also, MACE is defined to incorporate additional events such as rehospitalisation for angina, which may dilute the results for the harder outcomes such as cardiovascular death, for which the planned patient number may not be sufficient. In conclusion, it is still open as to whether CT might be a valuable noninvasive alternative in certain situations where coronary angiography is already indicated . This highlights the importance of the planned DISCHARGE study.
The radiation exposure of CT is lower than that of ICA using most recent technology . But because of additional ICA examinations following positive CTs, the overall radiation exposure might be increased in a subgroup of the patients randomised to CT. Thus, the study was approved by the German Federal Office for Radiation Protection. Since plaque characterisation is included and scanners using 64 slices or up to 320 rows are part of the centres’ technology, average effective doses of 5–10 mSv can be expected. The large sample size of DISCHARGE is an important advantage that will allow us to draw representative conclusions about the radiation exposure.
The DISCHARGE trial has some limitations. It is a multicentre study conducted in Europe based on funding provided by the European Union (grant agreement no. 603266). Therefore, other regions of the world cannot be represented in the study. Second, the CT arm of the study is more explanatory than the ICA arm, which is almost entirely pragmatic. The DISCHARGE trial will include patients with a clinical indication for ICA because of suspected CAD with low-to-intermediate probability. Low-to-intermediate probability is defined as 10–60 % and will be estimated in all individuals eligible for the study using a pretest probability calculator derived from data obtained in the individual-patient data meta-analysis COME-CCT . Patients with lower or higher probability of CAD are less likely to benefit from CT since positive and negative predictive values are reduced in these patient groups, respectively . Still, this should be explored in the future since some studies have shown potential in such groups when comparing CT with exercise ECG  or standard care . Patients are excluded from the DISCHARGE trial if they are or were on haemodialysis, are pregnant or not in sinus rhythm. These very limited exclusion criteria reflect the limitations of current CT technology, which are heavy calcification posing issues in patients with haemodialysis and sinus rhythm providing the best basis for diagnostic image quality. This is important because nondiagnostic results are rather common in such situations and would be considered to reflect ‘positive’ findings. Nevertheless, we will not exclude patients with contraindications to beta-blockade or high heart rates, thereby making the study more representative and generalizable. The exclusion criteria reflect good clinical practice because we will avoid radiation exposure in pregnant women. An important advantage of the DISCHARGE trial is that we will not exclude patients above certain body mass index levels or body weights in order to be as pragmatic as possible. It is because of this background that the radiation exposure of DISCHARGE will not be very low but has to be tailored to each individual patient with the aim to have as few nondiagnostic CT examinations as possible  since these will lead to the need to perform additional ICA. Another practical advantage of the trial is that CT will not be withheld in patients with increased Agatston scores estimated from non-contrast CT performed before CT. Therefore, and because of the limited exclusion criteria, the results will be representative for most patients with low-to-intermediate probability of suspected CAD.
In summary, the multicentre European DISCHARGE trial is designed to pragmatically assess the comparative effectiveness of CT and ICA in patients with suspected CAD and a clinical indication for ICA based on low-to-intermediate probability of disease.
The scientific guarantor of this publication is Prof. Marc Dewey, M.D. Ph.D. The authors of this manuscript declare relationships with the following companies, foundations and government or university grants: Abbott, Academy of Finland, Centre of Excellence programs, Actelion, American Thoracic Society, Amgen, AP Møller og hustru Chastine McKinney Møllers Fond, Astra-Zeneca, Athera Biotechnologies AB, B.Braun Medical, Bayer, Bayer-Schering, Berlin Chemie, Beuth Hochschule für Technik Berlin, Biosensors, Biotie Therapies, Biotronik, Boehringer-Ingelheim, Boston Scientific, Bracco, Braun, Cardiac MR Academy Berlin, CardiRad Ltd, Comac-Medical, Cordis, Covance, Directorate-General for Health and Food Safety, Eckert & Ziegler Radiopharma GmbH, Edwards Lifesciences, Endocyte Inc, European Regional Development Fund, European Society of Cardiology, European Union, F.Hoffmann-La Roche, Finnish Foundation for Cardiovascular Research, Fundação para a Ciência e Tecnologia, Portugal, GE Healthcare, Gedeon Richter, German Federal Ministry of Education and Research (BMBF), German Research Foundation (DFG), Guerbet, Guidotti,Intermeetings Padova, Irish Lung Foundation, Jansen-Cilag, Johnson & Johnson, KRKA, Lantheus Inc, Medtronic, Medtronik, Menarini, Merck, Merck Sharp & Dohme, MSD, National Heart, Lung, and Blood Institute (NHLBI), Novartis, Orion Pharma, Pfizer, Pharma Swiss, Philips, Piramal, Portuguese Society of Cardiology, Quintiles, Research Council of Rigshopitalet, Roche, Sandoz, Sanofi Aventis, Schering, Servier, Siemens, Sigma Tau, Southeastern Health and Social Care Trust Innovation Research Development Group Fund, Spanish Society of Cardiology, St. Jude Medical, Stiftung Charité (Berlin Institute of Health), Takeda,TEKES Finland, The Danish Agency for Science Technology and Innovation by The Danish Council for Strategic Research, The Danish Heart Foundation, The John and Birthe Meyer Foundation, The Lundbeck Foundation, Toshiba Medical Systems, Turku University Hospital research funds, VITAL Images Inc., Zealand A/S
This study has received funding by the Seventh Framework Programme of the European Union (EC-GA 603266). Prof. Peter Schlattmann kindly provided statistical advice for this manuscript. Institutional review board approval was obtained. Written informed consent will be obtained from all subjects (patients) for the study. Study subjects or cohorts have not been previously reported in another journal/study. Methodology: prospective/retrospective, randomised controlled trial, multicenter study.