Early computed tomography coronary angiography in adults presenting with suspected acute coronary syndrome: the RAPID-CTCA RCT

Diagnostic accuracy

Without a doubt, there is a need for novel investigations to support the evaluation of suspected or provisionally diagnosed acute coronary syndrome that enable improved diagnosis of acute coronary syndrome and risk stratification for subsequent clinical events. Moreover, better case selection for invasive coronary angiography is important given the increasing recognition of the occurrence of MI with non-obstructive coronary artery (MINOCA) disease 23–25 and the lack of specificity (false-positive rates) with the use of high-sensitivity cardiac troponin level assays. 13,14 Ultimately, any investigation should provide information to tailor subsequent management and improve clinically important outcomes. CTCA may fulfil all of these requirements. CTCA is quicker, simpler, substantially cheaper and more readily delivered than invasive coronary angiography, and should translate into a highly effective and safe imaging strategy.

A systematic review of 21 diagnostic accuracy studies of CTCA reported a pooled sensitivity of 99% and specificity of 89% for the detection of coronary artery disease. 25 This Health Technology Assessment (HTA) comprehensive systematic review assessed the role of 64-slice multidetector computed tomography as an alternative to invasive coronary angiography. In keeping with previous analyses,26 it confirmed the excellent accuracy of multidetector computed tomography in the identification of coronary artery disease. However, this systemic review highlighted several areas that need further research and highlighted, among other things, the need to evaluate the usefulness of multidetector CTCA in patients with suspected coronary artery disease.

A more recent evidence synthesis27 evaluated the diagnostic and prognostic accuracy and cost-effectiveness of CTCA in suspected acute coronary syndrome. This evidence review of eight trials found that CTCA had good diagnostic accuracy for detecting coronary artery obstruction: sensitivity of 94% (95% predictive intervals 61% to 99%) and specificity of 87% (95% predictive intervals 16% to 100%). Economic analysis was subject to substantial uncertainty, but CTCA was likely to be cost-effective if the incidence of subsequent major adverse cardiac events was > 2% [£30,000 per quality-adjusted life-year (QALY) threshold] or > 2.9% (£20,000 per QALY threshold). Decision-analysis modelling was unable to draw reliable conclusions about the clinical effectiveness and cost-effectiveness of CTCA in suspected acute coronary syndrome. The review suggested that further research regarding the effect of testing and treatment on major adverse cardiac events is needed.

Trial evidence: acute chest pain

Four US trials28–31 investigating CTCA in patients with low-risk acute chest pain presenting to the ED promoted the use of CTCA and its widespread adoption. Meta-analyses of these trials32,33 concluded that CTCA is safe, is cost-effective and reduces the length of stay in the US health-care system. However, the event rates in these studies were extremely low, with no difference between trial arms for clinically important outcomes. Moreover, the participants had relatively long hospital stays and many additional tests compared with contemporary practice in the UK.

Since the publication of these initial trials, there have been a number of other trials from North America, Europe and Australia, with none powered to evaluate the impact of CTCA use on longer-term clinically important outcomes. 34–39 However, in a secondary analysis of the CATCH trial40 of 600 patients with suspected acute coronary syndrome and normal troponin levels and ECG, CTCA was associated with a reduction in major adverse cardiac events at 18–20 months, although the absolute number of events was small: five patients (MI, n = 2; unstable angina, n = 3) in the CTCA group compared with 14 patients (cardiac death, n = 1; MI, n = 7; unstable angina, n = 5; coronary revascularisation, n = 1) in the standard-care group [hazard ratio (HR) 0.36, 95% confidence interval (CI) 0.16 to 0.95; p = 0.04]. 40

These trials have been assimilated by a number of recent meta-analyses. 41–44 Ten acute chest pain trials of 6285 patients were synthesised to examine the benefits and risks of CTCA compared with other standard-care approaches. 42 There were no significant differences in all-cause mortality [relative risk (RR) 0.48, 95% CI 0.17 to 1.36; p = 0.17], MI (RR 0.82, 95% CI 0.49 to 1.39; p = 0.47) or major adverse cardiac events (RR 0.98, 95% CI 0.67 to 1.43; p = 0.92) between the groups. However, there were higher rates of invasive coronary angiography (RR 1.32, 95% CI 1.07 to 1.63; p = 0.01) and revascularisation (RR 1.77, 95% CI 1.35 to 2.31; p < 0.0001) with CTCA than with standard-of-care approaches. A further meta-analysis by Foy et al.,43 comparing the clinical effectiveness of CTCA with that of functional stress testing for patients with suspected coronary artery disease, included 13 trials (acute chest pain, n = 9; stable chest pain, n = 4). There were 10,315 patients in the CTCA arm and 9777 patients in the functional stress testing arm. CTCA was associated with a reduction in rates of MI (0.7% vs. 1.1%; RR 0.71, 95% CI 0.53 to 0.96). Patients undergoing CTCA were more likely than those who were not to require invasive coronary angiography (11.7% vs. 9.1%; RR 1.33, 95% CI 1.12 to 1.59) and revascularisation (7.2% vs. 4.5%; RR 1.86, 95% CI 1.43 to 2.43). The patients receiving CTCA were also more likely to receive a new diagnosis of coronary artery disease and to start new preventative therapies, such as aspirin or statin therapy, than those who were not.

The latest meta-analysis44 included data from 16 trials of both acute and stable chest pain, enrolling 21,210 participants. There was no difference in all-cause mortality (103 vs. 110; RR 0.93, 95% CI 0.71 to 1.21; p = 0.58) between the CTCA arm and the standard-care arm. There was a reduction in subsequent MI in the CTCA arm (115 vs. 156; RR 0.71, 95% CI 0.56 to 0.91; p < 0.006). This was largely because of a reduction in MI in the subgroup of patients with stable chest pain (80 vs. 120; RR 0.66, 95% CI 0.50 to 0.88; p = 0.004). There was no difference found between arms in the acute chest pain subgroup (35 vs. 36; RR 0.88, 95% CI 0.54 to 1.44; p = 0.61). The CTCA arm had higher invasive angiography rates than the standard-care arm (1044 vs. 701; RR 1.41, 95% CI 1.28 to 1.55; p < 0.00001). This was found in patients with either acute (311 vs. 205; RR 1.35, 95% CI 1.13 to 1.62; p = 0.001) or stable (733 vs. 496; RR 1.44, 95% CI 1.30 to 1.61; p < 0.00001) chest pain. This led to subsequent comparable changes in revascularisation (percutaneous coronary intervention or coronary artery bypass graft surgery) rates, again for the groups together [789 vs. 472; odds ratio (OR) 1.84, 95% CI 1.44 to 2.35; p < 0.00001] and for both acute (175 vs. 82; OR 1.95, 95% CI 1.42 to 2.69; p < 0.0001) and stable (614 vs. 390; OR 1.70, 95% CI 1.16 to 2.51; p = 0.007) chest pain. There was also a demonstrable reduction in both subsequent ED visits and hospital admissions (570 vs. 616; RR 0.75, 95% CI 0.60 to 0.94; p = 0.01) and downstream investigations in the CTCA arm (242 vs. 342; OR 0.45, 95% CI 0.22 to 0.90; p = 0.02). These findings were not significant when looking at the acute chest pain group alone.

Trial evidence: stable chest pain

The use of CTCA in the evaluation of patients with stable chest pain has been explored in several randomised controlled trials (RCTs). The two largest trials were the Scottish computed tomography of the heart (SCOT-HEART) and prospective multicentre imaging study for evaluation of chest pain (PROMISE) trials. 45–47 These trials demonstrated that CTCA was associated with improved diagnostic certainty,45 reduced rates of normal invasive coronary angiography,47,48 increases in preventative therapy and early coronary revascularisation,47,49 and reduced rates of subsequent coronary heart disease (CHD) death or non-fatal MI. 45,46 NICE guidelines recommend the use of CTCA as the first-line diagnostic test in patients with suspected CHD. 50

In a recent evaluation of a Danish country-wide registry of 86,705 stable patients being evaluated for suspected coronary artery disease, CTCA was associated with greater use of statin therapy, aspirin and invasive procedures and higher costs than functional testing. CTCA was associated with a lower risk of MI, but a similar risk of all-cause mortality. 42

Computerised tomography coronary angiography in patients with intermediate- or high-risk acute chest pain

A small, single-centre, three-arm trial investigated an imaging-first strategy of either CTCA or cardiac magnetic resonance imaging, compared with standard care, in patients with acute chest pain who had a non-diagnostic ECG and an elevated high-sensitivity cardiac troponin level. 51 An initial non-invasive imaging strategy reduced the proportion of patients referred for invasive coronary angiography during initial hospitalisation (cardiac magnetic resonance imaging 87%, CTCA 66% and routine clinical care 100%; p < 0.001 for imaging vs. routine care), with the fewest invasive coronary angiograms performed in those undergoing CTCA (p < 0.004 vs. cardiac magnetic resonance). The reduction in invasive coronary angiography persisted for at least 1 year (88%, 70% and 100%. respectively; p < 0.003 for imaging vs. routine care). Unlike previous studies, this trial showed that in higher-risk groups of patients with acute chest pain, CTCA reduced the rates of invasive coronary angiography rather than increasing it and increased the proportion of patients receiving revascularisation who had invasive coronary angiography.

Trial aims

This trial aimed to investigate the effect of receiving an early CTCA in patients with suspected or provisionally diagnosed acute coronary syndrome presenting to the ED, acute medicine or cardiology service on health-care interventions, clinical event rates and health-care costs in a clinical trial and economic evaluation up to 1 year after the trial intervention.

The primary objective was to investigate the effect of early CTCA on all-cause death or subsequent non-fatal type 1 or type 4b MI at 1 year.

Inclusion criteria

Patients aged ≥ 18 years with symptoms mandating investigation for suspected or provisionally diagnosed acute coronary syndrome with at least one of:

• History of ischaemic heart disease (for which the clinician assessing the patient confirms history based on patient history or available health-care records).

• Troponin level elevation above the 99th centile of the normal reference range or increase in high-sensitivity troponin levels meeting European Society of Cardiology criteria for ‘rule-in’ of MI. Troponin level assays varied from site to site (see Appendix 7, Table 32); local laboratory reference standards were used.

• ECG abnormalities, such as ST segment depression of > 0.5 mm.

Exclusion criteria

• Signs of, symptoms of or investigations supporting high-risk acute coronary syndrome:

  • ST elevation myocardiaI infarction (STEMI).

  • Acute coronary syndrome with signs or symptoms of arrhythmia, acute heart failure or circulatory shock.

  • Crescendo episodes of typical anginal pain.

  • Marked or dynamic ECG changes, such as ST depression of ≥ 3 mm.

  • Clinical team had scheduled an urgent invasive coronary angiography on the day of the trial eligibility assessment. This was added as an exclusion criterion on 15 January 2016.

• Patient inability to undergo computerised tomography (CT):

  • severe renal failure (serum creatinine of > 250 µmol/l or estimated glomerular filtration rate of < 30 ml/minute/1.73 m²)

  • known contrast allergy

  • beta-blocker intolerance (if no alternative heart rate-limiting agent available or suitable) or allergy

  • inability to hold breath

  • atrial fibrillation for which the mean heart rate was anticipated to be > 75 beats per minute (b.p.m. ) after beta-blockade.

• Patient has had invasive coronary angiography or CTCA within the last 2 years and the previous investigation revealed obstructive coronary artery disease, or patient had either investigation within the last 5 years and the result was normal.

• Previous recruitment to the trial.

• Known pregnancy or currently breastfeeding.

• Inability to consent.

• Further investigation for acute coronary syndrome would not be in the patient’s interest owing to limited life expectancy, quality of life or functional status.

• Prisoners.

Consent

All eligible participants provided written informed consent after approach and discussion with appropriately trained and delegated members of the research or clinical team.

Randomisation

After assessment for eligibility and consent, the clinical research nurse or a delegated member of the clinical team collected the baseline data necessary to complete the pre-randomisation information. Randomisation was carried out using a web-based randomisation service (managed by the Edinburgh Clinical Trials Unit) to ensure allocation concealment. Randomisation was carried out within 18 hours of arrival at the hospital, extending to 24 hours after 25 November 2016. Consented patients were randomised on a 1 : 1 basis to CTCA in addition to standard care or standard care alone and were stratified by study site.

Blinding

This was an open trial. The patient, recruiting and treating clinicians, and radiologist were not blinded to the intervention, including the results. Members of the adjudication committee completing the primary outcome assessment were blinded to the intervention.

Computerised tomography coronary angiography

The minimum technology requirement for CT was a 64-slice or greater multidetector CT scanner that was enabled to perform ECG-gated cardiac studies. The examination included a non-contrast ECG-triggered acquisition for calcium scoring (if part of local protocol) and a post-contrast ECG-gated acquisition covering the whole of the heart and the root of the aorta. Patients had to be able to hold their breath for > 20 seconds. The intervention lasted for no longer than 30 minutes and patients were observed for a period of 30 minutes afterwards. Radiation reduction techniques were employed and, when appropriate, intravenous or oral beta-blockades (or alternative heart-rate-limiting agent) were used to reduce heart rate (target of < 70 b.p.m.), enabling radiation dose-saving protocols. The use of glyceryl trinitrate for coronary artery dilatation was at the discretion of individual centres.

Given the range of conversion factors used across sites to convert dose–length product (DLP) to effective dose in mSv, radiation dose was reported as a DLP. A DLP to mSv conversion factor of 0.014 mSv/mGy/cm was calculated, similar to other recent publications. 54 A typical participant with a heart rate of < 70 b.p.m. in sinus rhythm and a body mass index of < 25 kg/m² would be anticipated to have a DLP of ≤ 686 mGy/cm. Values exceeding this were considered to be protocol deviations. Deviations were reviewed by a radiologist or cardiologist in the central trials team, and any that were deemed to have an impact on patient safety or study outcomes were reported as protocol violations. For other participants, without the typical characteristics stated above, DLP values of > 686 mGy/cm were anticipated as part of routine clinical practice (e.g. owing to the need for continuous retrospectively gated scanning in some participants with arrhythmia). For such participants, any DLP that exceeded 1500 mGy/cm was considered to be a protocol deviation. Deviations were reviewed by a radiologist or cardiologist in the central trials team, and any that had an impact on patient safety or study outcomes were reported as protocol violations.

All participating centres were required to verify that their CTCA imaging protocol complied with the doses outlined in the research protocol prior to recruitment. Patient doses were recorded and monitored as part of the study. Iodine-based contrast media were administered intravenously using the standard local procedure at each site.

Computerised tomography coronary angiography result reporting

Computerised tomography coronary angiography results were reported by a trained radiologist or cardiologist at recruiting centres as soon as possible, ideally within 2 hours, and were immediately communicated to the treating clinician.

The clinical report detailing the results was reported in accordance with the Society of Cardiovascular CT guidelines using the American Heart Association coronary artery segment model, and included both the calcium score, if calculated, and the presence of cardiac and non-cardiac findings. 55,56 Stenoses were quantified as no significant coronary artery disease (estimated stenosis of < 10%), mild non-obstructive disease (estimated stenosis of 10–49%), moderate non-obstructive disease (estimated stenosis of 50–70%) or obstructive coronary artery disease (estimated stenosis of > 70% or > 50% for left main stem disease).

The research scan report was completed by the radiologist or cardiologist, and recorded scanner technology, acquisition protocol, DLP and patient characteristics (see Appendices 1 and 2).

Quality assurance reporting of computerised tomography coronary angiography scans

The first 10 CT scans and reports from each site were reviewed and checked by experts who were independent of the trial site and blinded to the initial report to measure interobserver reliability. This process and the reporting form are detailed in Appendices 3 and 4.

Computerised tomography coronary angiography images for future research

The scans sent for quality assurance (QA) reporting have been retained for future research along with all of the CT scans from the lead recruiting sites in a research repository at the University of Edinburgh (Edinburgh, UK).

Impact of computerised tomography coronary angiography on participant care

A trial guideline on the management of trial participants depending on CTCA result was developed (see Appendix 5). Its use was not mandated because the trial had a pragmatic approach and was investigating the impact of the diagnostic intervention on clinical practice and outcomes.

Compliance and withdrawal of study participants

Study participants were free to withdraw from the trial at any time. Reasons, if given, were recorded and data collected up to that time point were used in the final analyses, unless the patient specifically requested that their data were not used. If the patient withdrew consent to have their data stored, this was documented on the trial Consolidated Standards of Reporting Trials (CONSORT) flow diagram as ‘withdrawn’ and their data were not used in the final analyses. Patients were able to withdraw from participation in active follow-up, but data continued to be collected unless the patient requested otherwise. The patient may have been willing to give a reason for withdrawal, but this was not obligatory.

Crossover

Any patient in the standard-care arm who received a CTCA as part of routine care within 30 days of randomisation was defined as a crossover and was not recorded as a deviation.

Investigation guidelines and strategies for each centre were collected and the use of CTCA was monitored during the trial. Each centre was requested not to use CTCA in the routine assessment of suspected or confirmed acute coronary syndrome during trial recruitment, and was asked to inform the trials team about any changes to local practice.

Non-adherence

This was defined as any participant not receiving a reported CTCA if they had been randomised to receive it within 72 hours of the randomisation. This was recorded as a deviation. This allowed ambulatory CTCA to be delivered when appropriate. Individual site retention, crossover and non-adherence were monitored and reviewed at the Project Management Group, Trial Steering Committee and Data Monitoring Committee meetings.

Other interventions in the computerised tomography coronary angiography arm and the standard-care arm

All other management and admission or discharge decisions were at the discretion of the treating clinicians.

Primary end point

The primary end point was all-cause death or subsequent non-fatal type 1 or 4b MI at 1 year, measured as the time to first such event. MI was defined in accordance with the third universal definition,11 and events were adjudicated by two independent cardiologists blinded to the trial intervention.

Secondary end points

Key secondary end points included:

• CHD death or subsequent non-fatal MI

• cardiovascular disease death or subsequent non-fatal MI

• subsequent non-fatal MI

• CHD death

• cardiovascular death

• all-cause death.

Other end points

• Coronary heart disease death or subsequent non-fatal MI (type 1 or 4b).

• Subsequent non-fatal MI (type 1 or 4b).

• Non-cardiovascular death.

• Invasive coronary angiography.

• Coronary revascularisation.

• Percutaneous coronary intervention.

• Coronary artery bypass graft surgery.

• Proportion of patients prescribed acute coronary syndrome therapies during index hospitalisation.

• Proportion of patients discharged on preventative treatment or had alteration in dosage of preventative treatment during index hospitalisation.

• Length of stay for index hospitalisation.

• Re-presentation or rehospitalisation with suspected acute coronary syndrome or recurrent chest pain within 12 months after index hospitalisation.

• Chest pain symptoms up to 12 months.

• Patient satisfaction at 1 month.

• Clinician certainty of presenting diagnosis after CTCA.

• Quality of life [measured by EuroQol-5 Dimensions, five-level version (EQ-5D-5L) up to 12 months].

Trial management and oversight

The Edinburgh Clinical Trials Unit (ECTU) was responsible for trial management, including the organisation of Trial Management Group meetings, the organisation of the Trial Steering Committee (TSC) and the Data Monitoring Committee (DMC), contracting with other organisations, the preparation of Research Ethics Committee and research and development office applications, the standard operating procedures, the provision of the randomisation system, database development, data management, data analysis, writing the report and the dissemination of findings.

Project Management Group

The trial was led by Alasdair Gray, and was co-ordinated by a trial manager from the ECTU and an emergency medicine research nurse co-ordinator, with support from the ECTU. A Project Management Group comprising the applicants and relevant members of the ECTU team oversaw trial delivery. The Academic and Clinical Central Office for Research & Development in Edinburgh provided sponsorship and monitoring oversight for the trial, which was conducted in line with relevant sponsor standard operating procedures, which are available at www.accord.ed.ac.uk/standardopprocs/CRSOPs.html (accessed January 2021). A delegation log at each site detailed the roles and responsibilities of each member of staff working on the trial.

Trial Steering Committee

A TSC was established to oversee the conduct and progress of the study. The terms of reference of the TSC, the draft template for reporting, and the names and contact details were detailed in the TSC charter.

Data Monitoring Committee

The DMC was composed of independent members, including a statistician, a cardiologist, a radiologist and an emergency or acute medicine physician. The Peto–Haybittle rule was used by the DMC as a guideline on the primary end point to trigger discussions on stopping the trial. Importantly, a decision to stop the trial did not rely on p-values alone and considered whether or not the results were convincing to the clinical community and patients.

Pre-funding preparation

Professor Steve Goodacre (co-applicant) met with members of the Sheffield Emergency Care Forum to consult them on the development of the study design for the grant application. The forum is a patient and public representative group that provides independent advice on emergency care-related research. They reviewed the proposal and provided advice on study design, patient procedures and ethics issues, which helped inform the final submission and subsequent study design.

Post-funding preparation

The Sheffield Emergency Care Forum was consulted again during the development of the trial information (patient information letters, consents, general practitioner letter) and the documents were amended to incorporate their feedback, which helped improve the usability of the documents.

During the trial

Patient and public involvement representatives were invited to participate in the TSC and were involved in the oversight of the trial throughout its duration. They provided valuable feedback about the patient perspective throughout the trial, which helped guide the decision-making of the trial team.

Report writing, academic paper preparation and dissemination

Patient and public involvement representatives helped to develop the Plain English summary for the final report and for dissemination of the results, which helped us ensure that this was clear and easy to interpret. The research findings will be presented at one of the Sheffield Emergency Care Forum’s regular meetings and members of the forum will help to develop material to allow us to disseminate the trial findings to the public.

Outcomes and conclusions

The inclusion of PPI representatives at each stage of the trial was beneficial because it provided continuous input throughout the project and advice when we needed it. It was helpful to have several PPI members involved because it provided a varied perspective and each member brought different strengths to the project. We received very positive feedback from the PPI members in the TSC. We received positive feedback about our level of engagement with the members and about how we had created an inclusive environment. An area for improvement is to ensure that lay language is consistently used during discussions in meetings to ensure that PPI representatives can follow complex discussions and can engage fully whenever possible. There were no negative impacts from PPI involvement in this case.

Participant primary symptoms, assessment and management pathways

Overall, 1549 (89%) participants reported chest pain as the primary symptom (see Table 1). The characteristics and details of the chest pain can be found in Appendix 7, Table 34. The time from presenting symptom onset to randomisation is reported in Table 3. Patients had a variety of referral pathways before attending hospital. The majority of participants (n = 957; 54.7%) telephoned emergency ambulance services. However, 521 (29.8%) patients self-presented to hospital and 188 (10.8%) were referred by a general practitioner. Most participants (n = 1557; 89.1%) were initially assessed in the ED (see Appendix 7, Table 35).

The majority of participants (n = 1611; 92%) were in sinus rhythm at the time of their initial hospital 12-lead ECG. The characteristics of the participants’ ECG findings and troponin level results during index hospitalisation are detailed in Appendix 7, Table 36.

After initial assessment, the attending clinical team were asked how likely they thought acute coronary syndrome was as the presenting clinical diagnosis. In 642 (36.7%) patients, the clinician was highly suspicious of an acute coronary syndrome. The median level of certainty was 7 on a scale from 0 to 10, where 0 is least certain and 10 is most certain (Q1 6 to Q3 8; mean 7.1, SD 1.8) (see Appendix 7, Table 37). After initial assessment, most participants (n = 1372; 78%) were admitted to hospital. The details of the participants’ admission specialty are given in Appendix 7, Table 38.

Adherence to trial allocation: computerised tomography coronary angiography plus standard care arm

The majority of participants randomised to receive CTCA (n = 767, 87.5%) underwent CTCA, with 757 (86.3%) participants receiving the CTCA within the protocol-defined 72-hour window. The other 10 (1.1%) participants received the intervention in the first 10 days. Of the 110 patients who did not receive CTCA as the allocated trial intervention, five subsequently had a CTCA scan outside the trial protocol. Six of the 767 patients who underwent CTCA as the trial intervention had a second CTCA scan during the first year of the trial. The reasons for not receiving a CTCA are detailed in Table 4.

Adherence to trial allocation: standard-care arm

In the standard-care arm, 48 (5.5%) participants received a CTCA within 30 days of randomisation and, therefore, met the predetermined criteria for cross over. Thirty-three of these patients received CTCA within 3 days of randomisation, eight between 4 and 10 days after randomisation, three between 11 and 20 days after randomisation, and four between 21 and 30 days after randomisation. A further 25 (2.9%) patients in the standard-care arm received a CTCA, but more than 30 days after randomisation.

Computerised tomography coronary angiography delivery and quality

The median time from randomisation to CTCA was 4.2 [interquartile range (IQR) 1.6–21.6] hours, with CTCA being delivered on the day of randomisation in the majority of cases (Figure 6). Table 5 details the CTCA delivery and quality. CT scanners varied across the sites, from 64-slice to 320-slice CT scanners. The majority of scans were delivered during the index hospitalisation using a 64-slice (n = 358; 47.7%) or 128-slice (n = 256; 34.1%) scanner. Beta-blockade was used in 539 (70.3%) participants and sublingual glyceryl trinitrate was used in 521 (67.9%) participants as a pre treatment to optimise CT scan acquisition. The CTCA was of diagnostic quality for 700 (91.3%) participants. The first 10 CTCA scans from each site were reviewed centrally for QA purposes. The details of this process are summarised in Appendix 3.

FIGURE 6.

Time from randomisation to receiving CTCA as trial intervention for patients allocated to the CTCA arm. There were three patients allocated to the CTCA arm for whom the time from randomisation to receiving CTCA as the trial intervention was > 120 hours, and their times to undergoing CTCA of 140, 163 and 218 hours are not shown in the histogram. Reproduced with permission from Gray et al. 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Reproduced with permission from Gray 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Computerised tomography coronary angiography findings

Computerised tomography interpretation was available for 759 out of the 767 participants who underwent CTCA. The intervention identified normal coronary arteries in 178 (23%), non-obstructive disease in 222 (29%) and obstructive disease in 359 (47%) participants. The severity of coronary artery disease was associated with increasing age, male sex, prior CHD, troponin level elevation and GRACE score, as well as the use of invasive coronary angiography and subsequent revascularisation (Table 6). Other cardiac and non-cardiac diagnoses identified on CTCA are detailed in the safety outcomes.

Other clinical outcomes

Other clinical outcomes reported were CHD death or non-fatal MI (type 1 or 4b), non-fatal MI (type 1 or 4b) and non-cardiovascular death. Once again, there was no evidence of a difference in outcome between trial arms (Table 8) (see Appendix 8, Figure 24).

Treatment during index hospitalisation and secondary preventative treatment

Fewer participants in the CTCA arm than in the standard-care arm received invasive coronary angiography: 474 (54.0%) compared with 530 (60.8%), respectively (adjusted HR 0.81, 95% CI 0.72 to 0.92; p = 0.001) (Table 9 and Figure 9). Despite there being relatively less invasive coronary angiography in the CTCA arm, there was no evidence of a difference in the rates of coronary revascularisation by trial allocation (adjusted HR 1.03, 95% CI 0.87 to 1.21; p = 0.76) (Figure 10). There was also no evidence of a difference in percutaneous intervention or coronary artery bypass surgery between arms (see Table 9) (see Appendix 8, Figures 27 and 28). The proportion of participants receiving revascularisation following invasive angiography was 63.3% in the CTCA arm (300/474 who received invasive angiography) compared with 54.3% in the standard-care arm (288/530 who received invasive angiography).

FIGURE 9.

Cumulative probability of invasive coronary angiography. The date of invasive coronary angiography was not known for one patient in the CTCA arm and for one patient in the standard care alone arm, and these participants are not included in the estimates of cumulative probability. Reproduced with permission from Gray et al. 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Reproduced with permission from Gray 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

FIGURE 10.

Cumulative probability of coronary revascularisation. The date of coronary revascularisation was not known for one patient in the CTCA arm, and this participant is not included in estimates of cumulative probability. Reproduced with permission from Gray et al. 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Reproduced with permission from Gray 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Overall, there was no evidence of a difference in the in-hospital prescription of medications for acute coronary syndrome treatment (adjusted OR 1.06, 95% CI 0.85 to 1.32; p = 0.63). At hospital discharge, the change in prescription (increased or decreased dose, treatment started or stopped) of preventative therapies (adjusted OR 1.07, 95% CI 0.87 to 1.32; p = 0.52) was similar between trial arms (Table 10). Each component of prescription change (start, stop, increase or decrease dose) was also similar between trial arms (Table 11).

Length of hospital stay

The median length of hospital stay was longer in the CTCA arm than in the standard-care arm: 2.2 (IQR 1.1–4.1) days compared with 2.0 (IQR 1.0–3.8) days, respectively (Hodges-Lehmann estimator of location shift 0.21, 95% CI 0.05 to 0.40 days; p = 0.009).

Patient satisfaction

Overall, 1322 (75.6%) participants responded to the trial patient satisfaction questionnaire at 1 month (see Appendix 7, Table 40). Participant satisfaction (rated excellent or very good on a five-point Likert scale) was higher in the CTCA arm than in the standard-care arm: 83.3% compared with 79.7%, respectively (Figure 11).

FIGURE 11.

Patient satisfaction with the care that they received when they attended hospital. Q, question; SC, standard care.

Clinician diagnostic certainty in the computerised tomography coronary angiography group

The attending clinician reported increased diagnostic certainty following CTCA. The mean increase was 1.4 (2.2) on a 10-point scale, from 7.1 (diagnostic certainty before CTCA scan) to 8.5 (diagnostic certainty after CTCA scan). The scale was from 0 to 10, with 10 being the most certain (Table 12).

Symptoms and hospital presentations during follow-up

During 1 year of follow-up, 268 (15.3%) participants presented to hospital with suspected acute coronary syndrome. The rate of re-presentation was similar in both trial arms (adjusted HR 1.06, 95% CI 0.83 to 1.34; p = 0.66) (Figure 12). In addition, there was no evidence of a difference in chest pain symptoms between trial arms at 1, 6 and 12 months (Table 13).

FIGURE 12.

Cumulative probability of re-presentation or rehospitalisation with suspected acute coronary syndrome or recurrent chest pain. Reproduced with permission from Gray et al. 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Reproduced with permission from Gray 61 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure.

Discharge diagnosis from index hospitalisation

At discharge from the index hospitalisation, 857 (49%) participants had a diagnosis of acute coronary syndrome [non-ST elevation myocardial infarction (NSTEMI), STEMI, unstable angina]. There was no difference in the discharge diagnosis of acute coronary syndrome (MI or unstable angina) between trial arms: 440 (50.2%) in the CTCA arm compared with 417 (47.9%) in the standard-care arm. Other discharge diagnoses are detailed in Table 14.

Investigations during index hospitalisation and up to 1 year

There was no evidence of a difference between trial arms and the completion of other cardiac investigations. The most common additional cardiac investigation was an echocardiogram, which was performed in 932 (53.3%) participants (Table 15) (see Appendix 8, Figure 29). The details of additional non-cardiac investigations are reported in Appendix 7, Table 41.

Patient safety

There were 32 adverse events reported in 29 (1.7%) participants. There were three non-serious adverse events definitely related to the CTCA: three problems associated with the intravenous cannula. There was one serious adverse event possibly related to the CTCA: an admission to hospital with a non-cardiac condition (Table 16).

Alternative diagnoses on computerised tomography coronary angiography

Alternative cardiovascular and non-cardiovascular diagnoses identified on CTCA are delineated in Tables 17 and 18. Other cardiac findings were found in 225 (29.3%) participants, but very few were directly related to the participant’s presenting complaint. An alternative non-cardiac finding was found in 237 (30.9%) participants. Again, very few had alternative findings directly related to the participant presentation, such as pulmonary emboli (n = 5) or thoracic aortic dissection (n = 1).

Radiation exposure from computerised tomography coronary angiography in the CTCA arm during index hospitalisation

The median effective radiation dose was 3.1 (IQR 1.9–5.5) mSv (0.014 mSv/mGy/cm conversion factor). Calcium scores, when calculated and available, are categorised in Table 19.

Radiation doses were monitored during the trial, and doses that were higher than those specified in the protocol were reviewed by a radiologist and cardiologist in the central trial team to determine whether these should be reported as a protocol deviation or, in the case that the dose may have had an impact on patient safety, a protocol violation. This process was introduced in protocol version 4 (implemented on 29 June 2016) after it was identified that high radiation doses were frequently occurring at two of the trial sites and universal definitions were needed in the protocol. There were 48 (6.3%) participants who were allocated to the CTCA arm and received CTCA as the trial intervention had their radiation dose reported as either a protocol deviation (n = 31) or a protocol violation (n = 17).

Methods for the within-trial analysis

The within-trial cost-effectiveness analysis was performed by comparing the observed costs and QALYs of the CTCA arm with those of the control arm during the trial period. The QALYs and costs for each arm of the trial (i.e. for each strategy) during the follow-up were used to estimate the incremental cost-effectiveness ratio (ICER) of the CTCA arm compared with the standard-care arm. Confidence intervals for the within-trial ICER were estimated to capture the sampling uncertainty.

Of the 1748 participants, who had a mean age of 62 years (64% male), 877 were randomised to receive early CTCA and 871 were randomised to receive standard care. Non-fatal MI within 12 months occurred in 39 out of the 877 participants in the early CTCA arm and in 40 out of the 871 participants in the standard-care arm. In the 12 months, there were 19 deaths among the 877 participants in the early CTCA arm and 17 deaths among the 871 participants in the standard-care arm. Along with the clinical end points, a range of data from questionnaires and case report forms were also collected. These data were used to estimate the within-trial costs and QALYs. These data sources are briefly described in the following sections.

EuroQol-5 Dimensions, five-level version

The EQ-5D-5L is the most widely used instrument used to estimate patients’ quality of life (i.e. utility). These EQ-5D-5L questionnaires were administered using paper format to the patients at baseline and 1, 6 and 12 months after index admission. The responses from the patients were then collated into electronic format for health economic analysis.

Health-care resource use

Hospital records and patient self-reported questionnaires were used to estimate the health-care resource use. Hospital resource use was primarily determined from items recorded from the hospital records by the research nurse in the case report form, whereas the patient questionnaires were used to estimate other resource use (such as general practitioner surgery visits, general practitioner home visits, nurse home visits and social worker visits) that was not captured within the case report form. In these questionnaires, the patients were asked for items of resource use in the last month (at month 1) or last 3 months (at months 6 and 12).

Methods for estimating quality-adjusted life-years from EuroQol-5 Dimensions, five-level version

The EQ-5D-5L questionnaire responses at baseline and 1, 6 and 12 months after index admission were used to estimate the patients’ quality of life (i.e. utility) at each time point. In line with the NICE recommendations, utility scores were estimated using the mapping algorithm by van Hout et al. 63 rather than the EQ-5D-5L value set for England. Given that the baseline mean utility values were similar between treatment arms, there was no need to adjust utilities using regression techniques. The curve of utility over different time points was constructed for each patient (accounting for missing data using the approach mentioned in Dealing with missing data in the EuroQol-5 Dimensions, five-level version, questionnaires) and the QALYs for each arm were estimated by calculating the area under the curve for health utility over the 1-year period. The patients who died during the trial were included with zero utility from the time of death and the average of all of the patients was used to estimate the overall within-trial QALYs.

Dealing with missing data in the EuroQol-5 Dimensions, five-level version, questionnaires

The primary analysis included all patients who had any follow-up data, and we used multiple imputation techniques for estimating the missed utilities for patients with missing data in the follow-up points. For cases with only interim time point(s) missing (e.g. 3 months), we used the average value from the previous time point (i.e. utility at baseline) and the next time point (i.e. 6 months). This allowed us to include all cases with at least some follow-up data, that is only those with no utility values at any follow-up point were excluded from the primary analysis. Multiple imputation was performed using the R package MICE (multiple imputation using chain equations) (The R Foundation for Statistical Computing, Vienna, Austria).

We estimated the baseline utilities for patients without missing follow-up data and patients with missing follow-up data (i.e. those who did not respond to any questionnaires) to identify if there were any systematic differences in utility values between the two arms (i.e. responders and non-responders to questionnaires). Given that the non-responder cases had slightly different baseline values to the responders, we used only the data for responders in the base-case analysis. We also undertook a secondary analysis in which we included all patients and used multiple imputation to estimate the utilities for non-responders. The results of this scenario analysis are presented in Appendix 11, Tables 49 and 50.

Methods for estimating health-care resource use

All health-care consumption and costs within the trial period were estimated from a health-care perspective using hospital records and from patient self-reported questionnaires. Resource use data were primarily determined from items recorded in the case report form from the case notes; data from the health service resource use questionnaire was used only for additional items not recorded in the case report form or notes (which included telephone consultations, general practitioner surgery visits, general practitioner home visits, nurse home visits and social worker visits).

All cardiac-related resource use, including the need for continued hospitalisation, additional invasive or non-invasive imaging, drug therapy and rehospitalisation for myocardial ischaemia, was captured. The difference in use of CTCA between the two arms of the trial was translated into the costs of CTCA included in the within-trial analysis. The differences in the rates of related procedures (invasive coronary angiography, other tests for coronary artery disease, coronary interventions and coronary artery bypass grafting) were captured from the case report form and were included in the estimation of costs. Other items of resource use (ED visits, outpatient visits and inpatient stays) were also captured from the case report form, as it is considered to be more accurate than a patient questionnaire. The medication costs were estimated according to their dosage, which was either just for the hospital stay (assumed as an average duration of 5 days) or for the whole year based on expert clinical input (see Appendix 10).

The general practitioner surgery visits, general practitioner home visits, nurse home visits and social worker visits were captured from the patient questionnaire. In the questionnaires, the patients were asked for items of resource use in the last month (at month 1) or last 3 months (at months 6 and 12). For this reason, we estimated costs incurred in the ‘missing’ months (e.g. months 2 and 3) using the average monthly resource use carried back (e.g. assuming resource use in months 2 and 3 were the same as the average of months 4, 5 and 6).

Dealing with missing data in resource use questionnaires

The case report form data were complete, so there were no missing data, whereas the patient questionnaire data had the same missing pattern as the EQ-5D-5L data. We, therefore, included the same cases in the primary analysis for costs and outcomes (i.e. those with at least one follow-up point) and used multiple imputation techniques for missing time points. Multiple imputation was performed using the R package MICE.

Estimating the within-trial costs

The overall resource use for each patient over the 12 months was multiplied with the unit costs (i.e. national average costs) to provide the estimated cost for each patient in the trial. Full details of the unit costs used are presented in Appendix 10, Tables 47 and 48. In brief, NHS reference costs64 were used to estimate the health-care resource use (see Table 47) and the British National Formulary and electronic market information tool65 were used to estimate the drug costs (see Table 48). The patients who died during the trial were included with zero costs from the time of death, and the costs for all of the patients in each arm were averaged to estimate the overall within-trial costs.

Estimating within-trial cost-effectiveness

The QALYs and costs for each arm of the trial (i.e. for each strategy) during follow-up were used to estimate the ICER of the CTCA arm compared with the standard-care arm. Confidence intervals for the within-trial ICER were estimated to capture the sampling uncertainty.

Within-trial quality-adjusted life-years

All patients (877 randomised to early CTCA and 871 randomised to standard care) completed the EQ-5D-5L questionnaire at baseline. The mean utility at baseline was 0.752 in the CTCA arm and 0.760 in the standard-care arm. However, there were 113 patients in the CTCA arm and 158 patients in the standard-care arm without any follow-up data (i.e. did not respond to any EQ-5D-5L questionnaires). For this reason, these patients were excluded from the analysis estimating the QALYs. The scenario analysis using all patients (i.e. including those with no follow-up data) is presented in Appendix 11, Tables 49 and 50.

For patients with follow-up EQ-5D-5L data, the average baseline utility was 0.765 in the CTCA arm (n = 764) and 0.768 in the standard-care arm (n = 713), suggesting similar baseline values. In the CTCA arm, there were 185, 215 and 243 patients who did not respond to the EQ-5D-5L questionnaires at 1, 6 and 12 months, respectively. Similarly, in the standard-care arm, there were 250, 275 and 301 patients who did not respond to the EQ-5D-5L questionnaires at 1, 6 and 12 months, respectively. The responses on the individual items of the EQ-5D-5L questionnaire were used to estimate utility scores using the mapping algorithm by van Hout et al. ,63 as recommended by NICE. The value set for the UK was used to estimate the utilities using the Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA) crosswalk calculator. The patients who died during the trial were included with zero utility from the time of death and the average of all of the patients was used to estimate the mean values of the utilities at different time points in both arms. These are presented in Table 20.

Multiple imputation techniques were used to impute the utility values for patients with missing data at 1, 6 and 12 months. This multiple imputation was performed using the package MICE in R software. The missing utility values were estimated as the average of values of data sets estimated using multiple imputations.

The QALYs were estimated using the trapezoidal rule for calculating the area under the curve. The utility values at baseline and 1, 6 and 12 months were multiplied with the corresponding time that patients spent in these utilities. It was assumed that the patients stay in the same utility until the midpoint of the time difference to the next follow-up point. For example, utility in month 1 is assumed to last until 3.5 months (i.e. the midpoint of the follow-ups at months 1 and 6) and the utility at month 6 is used for 3.5 to 9 months (i.e. the midpoint of the follow ups at months 6 and 12). Mean QALYs and 95% CIs estimated using bootstrapping are presented in Table 21.

Costs in the index hospital stay

The mean resource use during the index hospital stay for each arm was estimated from the case report form and included costs of hospital stay, costs associated with MI, costs of CTCA, costs of invasive coronary angiography, cost of percutaneous coronary intervention, costs of coronary artery bypass graft surgery, costs of diagnostic tests (including echocardiogram, radionuclide scan, 24-hour tape, exercise test, magnetic resonance imaging angiography, stress echocardiogram, other ECG monitoring, cardiac magnetic resonance imaging and stress magnetic resonance imaging) and the cost of drugs in the index hospital stay. Detailed resource use and unit costs for diagnostic tests and drug costs can be found in Appendix 10, Tables 47 and 48.

The average of the resource use of all patients in each corresponding arm was multiplied with the unit costs (i.e. national average costs) to provide the mean costs for the index hospital stay. The mean resource use in the CTCA arm and standard-care arm during the index hospital stay is presented in Table 22. Unit costs derived from the NHS reference costs64 are also presented here, and more detail about the Healthcare Resource Group codes used to estimate these costs is presented in Appendix 10, Table 47.

The mean costs in the CTCA arm and the standard-care arm for the index hospital stay are estimated as £4646 and £4394, respectively. A detailed breakdown of these costs is presented in Figure 13 and Table 22, which are estimated by multiplying the mean resource use during the index hospital stay for each arm (Table 23) with the unit costs derived from the NHS reference costs. 64 Detailed resource use and unit costs for diagnostic tests and drug costs can be found in Appendix 10, Tables 47 and 48. As observed in Figure 13 and Table 23, most of the costs are similar between the two arms, but the higher costs in the CTCA arm are mainly a result of the increased hospital length of stay (3.93 days in the CTCA arm vs. 3.50 days in the standard-care arm) and the additional CTCA use (average of 0.888 scans in the CTCA arm vs. 0.084 scans in the standard-care arm).

FIGURE 13.

Breakdown of the costs for index hospital stay for the CTCA and standard-care arms. CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention.

Costs in the 12 months post index hospitalisation

The mean resource use during the 12 months post index hospital admission was estimated for each arm from both the case report form and the patient questionnaire data.

The case report form reported data on the number of ED attendances, number of outpatient visits, days spent on a coronary care unit (CCU), days spent on an intensive care unit (ICU) and total days spent in hospital (any location, excluding CCU/ICU). Patient questionnaires reported the number of telephone consultations, general practitioner surgery visits, general practitioner home visits, nurse home visits and social worker visits in the last month (at month 1) or last 3 months (at months 6 and 12). Multiple imputation techniques were used to estimate the missing data in the patient questionnaires. Resource use was also incurred in the ‘missing’ months (e.g. months 2 and 3) using the average monthly resource use carried back (e.g. assuming that resource use in months 2 and 3 was the same as the average of that in months 4, 5 and 6).

The mean resource use in the CTCA arm and standard-care arm in the 12 months post index hospitalisation, from the case report form and patient questionnaires, respectively, is presented in Tables 24 and 25. Unit costs derived from the NHS reference costs64 are also presented here, and more detail about the Healthcare Resource Group codes used to estimate these costs is presented in Appendix 10, Table 47.

The mean costs in the CTCA arm and the standard-care arm for the 12 months post index hospitalisation are estimated as £2768 and £2451, respectively. A detailed breakdown of these costs, which are estimated by multiplying the mean resource use with the unit costs derived from the NHS reference costs, is presented in Figure 14 and Table 25. 64

FIGURE 14.

Breakdown of costs 12 months post index hospital stay in the CTCA and standard-care arms. GP, general practitioner.

The total within-trial costs are estimated as £7414.13 and £6845.11 for the CTCA arm and the standard-care arm, respectively. The mean costs and bootstrapped CIs are presented in Table 26.

Estimating life expectancy

Based on the data from the recent NICE guidelines,18 the life expectancy was estimated by applying relevant standardised mortality ratios (SMRs) to the general population mortality rates. Age- and sex-adjusted general population mortality were based on national life tables for the UK from 2017 to 2019. The average age of the population entering the model was 61.6 years and the proportion of males was 64%, which was estimated as the average of the mean values in the CTCA and standard-care arms. SMRs were sourced from the economic analysis in the recent acute coronary syndrome NICE guidelines,18 which suggested a SMR of 2.00 (95% CI 1.99 to 2.01) for patients without reinfarction and a SMR of 3.00 (95% CI 2.95 to 3.05) for patients with reinfarction. These SMRs were combined with age- and sex-adjusted general population mortality to estimate the proportions of patients alive over time. This resulted in mean (undiscounted) life-years of 17.69 for patients without reinfarction and 14.83 for patients with reinfarction.

Estimating long-term quality-adjusted life-years

Long-term QALYs were estimated by multiplying the age-adjusted utility values of the different health states with the proportion of patients alive over time in the respective health states.

The utility values at the start of the model for patients without and with reinfarction were extracted as 0.842 and 0.821, respectively, from the NICE acute coronary syndrome guidelines. 18

These utility values were adjusted for age to account for the decreasing quality of life over time; this adjustment was performed by using the ratio of age- and sex-specific general population EQ-5D-5L utilities at different ages compared with the general population utility at the start of the mode. The age- and sex-specific general population utility values were derived using the formula from Ara and Brazier,66 which states:

Multiplicative method was used to estimate the utility values over time, that is the utility values at the start of the model were multiplied with utility decrements.

This resulted in mean undiscounted QALYs of 13.97 for patients without reinfarction and 11.52 for patients with reinfarction. Using the NICE recommended discount rate of 3.5% per annum67 resulted in discounted QALYs of 10.10 and 8.67 for patients without with reinfarction, respectively.

Estimating long-term costs

Long-term costs were estimated by multiplying the annual costs of the different health states with the proportion of patients alive over time in the respective health states. The annual costs for patients without and with reinfarction were extracted as £943 and £1415, respectively, from the NICE acute coronary syndrome guidelines. 18 Using these annual costs, along with the life expectancy estimated earlier, resulted in mean undiscounted costs of £16,685 for patients without reinfarction and £20,979 for patients with reinfarction. Using the NICE recommended discount rate of 3.5% per annum67 resulted in discounted long-term costs of £11,929 and £15,650 for patients without and with reinfarction, respectively.

Summary of modelling input parameters

The decision-analytic model assigned patients in the CTCA arm and standard-care arm with probability of death and MI. Each patient who was alive then accumulated costs and QALYs based on the cost parameters, life expectancy and utility values based on whether or not they suffered MI. A summary of the model parameters along with the distributions and sources is provided in Table 28.

Estimating long-term cost-effectiveness

The cost-effectiveness of the different interventions was estimated and uncertainty was incorporated in the modelling by performing a probabilistic sensitivity analysis.

The lifetime QALYs and costs for each arm of the trial (i.e. for each strategy) were used to estimate the ICER of the CTCA arm compared with the usual care arm. An ICER measures the relative value of two strategies and is calculated as the mean incremental cost divided by the mean incremental benefits. A strategy is dominated when another strategy accrues more QALYs for less cost. The willingness-to-pay threshold is the amount of money that the decision-maker is willing to pay to gain 1 additional QALY. The usual threshold for decision-making at NICE is considered to be around £20,000–30,000 per QALY.

Parameter uncertainty was included in a probabilistic sensitivity analysis (PSA) based on Monte Carlo simulation. PSA model results include the effects of accounting for uncertainty in the model parameters (the costs, utilities, risk of mortality and MI), characterised as probability distributions. PSA is undertaken whereby the model is rerun (1000 times), each time with a different value for the risks, costs and utilities, which are independently sampled from the probability distributions.

The cost-effectiveness plane shows the incremental costs (y-axis) and incremental QALYs (x-axis) compared with usual care. In this chart, if a model run for a strategy had exactly the same costs and QALYs as usual care, the ‘sample’ for that model run would appear at the origin. Samples plotted to the right of the y-axis have more QALYs than usual care, and samples plotted above the x-axis have more costs. Samples plotted to the right of a straight line with slope lambda passing through the origin are cost-effective whereas those plotted to the left are not.

Cost-effectiveness acceptability curves (CEACs) were plotted to identify the probability of the CTCA arm being cost-effective compared with the standard-care arm for a range of threshold values for an additional QALY. A CEAC shows the proportion of model runs for which each strategy is cost-effective over a range of potential willingness-to-pay thresholds (i.e. lambda).

Statement of principal findings

In the within-trial analysis, the mean costs for the CTCA arm (£7414.13) were higher than for the standard-care arm (£6845.11). The greater costs in the CTCA arm were mainly a result of the longer hospital length of stay and the additional CTCA use, despite slightly lower costs of invasive coronary angiography. The QALYs are lower in the CTCA arm (0.7488) than in the standard-care arm (0.7577). Given that standard care is more effective and less expensive than CTCA, standard care is dominant (i.e. CTCA is dominated by the standard-care arm).

The long-term economic analysis suggested that CTCA was slightly less effective than standard care, with 0.025 QALYs lost per patient treated, and was more expensive, with additional costs of £481 per patient treated. Given that standard care is more effective and less expensive than CTCA, standard care is dominant, with 76% probability of being cost-effective at the £20,000 per QALY willingness-to-pay threshold. The greater costs in the CTCA arm are mainly a result of the higher within-trial costs, and the lower QALYs are a result of the larger number of deaths in the CTCA arm than in the standard-care arm.

Strengths and limitations of the assessment

Our within-trial analysis was based on resource use and EQ-5D-5L data from a large pragmatic trial with a representative population, and included multiple imputation for missing data and bootstrapping analysis for estimating the mean costs and QALYs. Other strengths in the within-trial analysis included detailed costing based on resource use data in the trial to estimate per-patient costs, and valuing outcomes as QALYs based on EQ-5D-5L questionnaires completed by the patients.

Our long-term analysis took the economic perspective of the NHS in England and Wales, was based on an established model, used robust sources for input parameters (effectiveness estimates were from the large pragmatic trial, with a representative population, and the long-term costs and QALYs were estimated from recent NICE acute coronary syndrome guidelines), valued outcomes as QALYs, used a lifetime horizon and included a PSA.

Despite these strengths, our analysis had some limitations. Within the trial, standard-care patients were less inclined to return their questionnaires than CTCA patients (158 patients in the standard-care arm did not respond to any EQ-5D-5L questionnaires compared with 113 patients in the CTCA arm). Given that these patients without any follow-up data were excluded from the analysis, this differential response to EQ-5D-5L questionnaires could result in bias in the estimation of within-trial QALYs. In the scenario analysis using all patients (i.e. including those with no follow-up data), the QALY difference between the CTCA arm and the standard-care arm is approximately half of the difference using the base-case analysis for responders only (see Appendix 11, Table 50). However, this is unlikely to affect overall conclusions around cost-effectiveness. In addition, our analysis was not able to incorporate short-term benefits related to patient satisfaction and diagnostic certainty, as these may not be captured in the EQ-5D-5L questionnaires.

The long-term modelling assumed benefits related only to reducing deaths and non-fatal MI by 1 year, and other benefits, such as reduced rates of life-threatening arrhythmia or new-onset heart failure outcomes, were not included. The event rate for heart failure or life-threatening arrhythmia in the standard-care arm did not exceed 5% and a meaningful effect from CTCA was not observed; therefore, including the long-term costs and effects of these events in the model would not have significantly changed the results. We assumed that all patients without non-fatal MI have CHD in the model and, although this is not true, the proportion without CHD would be the same across the standard-care and CTCA arms, suggesting that it is reasonable to assume that this will not markedly affect overall cost-effectiveness. Incorporating those with and without CHD separately would increase model complexity without having any effect on the incremental results.

Impact on clinical outcomes

There was no evidence that early CTCA had an effect on the 1-year rate of all-cause death or subsequent non-fatal type 1 (spontaneous) or 4b (stent thrombosis) MI. The trial design deliberately included a broad and representative group of patients focusing on those at intermediate risk of acute coronary syndrome. To provide increased precision around the impact of CTCA in specific risk categories, the prespecified analyses included several subgroup analyses. These included the individual inclusion criteria of troponin level elevation, previous diagnosis of coronary artery disease and ECG abnormality, as well as important risk categories such as age and sex. The GRACE score,59 an internationally accepted prognostic risk score in acute coronary syndrome and recommended in contemporary guidelines,9,10 was categorised into low, intermediate and high risk to mirror the groupings in these guidelines. Furthermore, an analysis comparing the primary outcome between sites with and without on-site invasive angiography facilities was included, as this may affect thresholds for invasive management. However, there was no evidence of benefit of CTCA for any of the prespecified subgroup analyses.

A number of key secondary end points were also investigated to ensure that all potentially important cardiovascular and CHD outcomes were examined. Once again, there was no evidence of benefit of CTCA for any of these key outcomes. In line with previous trials, we explored the impact of CTCA on all non-fatal MIs only, irrespective of subtype. Further exploration of all of the subtypes of MI, as defined in the universal definition,11 may have some value, but we did not assess this here.

It could be argued that we recruited too many patients who were at high risk of disease and, therefore, candidates for inpatient invasive coronary angiography. However, over three-quarters of participants had a low or an intermediate GRACE risk score, most did not have obstructive coronary artery disease, and, in the prespecified subgroup analyses, there was no evidence of heterogeneity of effect for the primary outcome according to the level of risk. As previously stated, these subgroups are the focus of contemporary guidelines9,10 that attempt to define intermediate-risk groups suitable for observation and additional testing, including CTCA. We were unable to demonstrate an effect of CTCA on 1-year clinical outcomes in any of these subgroups or risk categories.

Impact on treatment

The use of early CTCA had no impact on the overall rates of treatments for acute coronary syndrome, with similar rates of coronary revascularisation, prescription of acute coronary syndrome medications and discharge preventative therapies. This was similar between trial allocation whether the treatment was started, the treatment was stopped or the dose was altered. This is likely to reflect the high sensitivity of contemporary clinical assessment combining cardiac troponin level testing with a 12-lead ECG. Contemporary practice, therefore, provided limited opportunity for CTCA to identify unrecognised cases of acute coronary syndrome owing to CHD. Moreover, for CTCA to improve adverse coronary events, it would need to alter patient management. Such changes have been seen in prior studies of patients with stable chest pain in which CTCA improved the detection of unrecognised coronary artery disease, increasing the use of both preventative therapies and coronary revascularisation. These changes in management were associated with reduced rates of subsequent CHD death or non-fatal MI. 45,46 In our population of patients with acute chest pain and relatively high rates of invasive coronary angiography and coronary revascularisation, we observed no overall changes in rates of treatment or outcome. The most likely explanation for the difference between RAPID-CTCA and previous studies showing benefit from CTCA in patients with stable chest pain is that troponin level testing and the 12-lead ECG are powerful predictors of the need for intervention in acute chest pain, leaving less potential for CTCA to detect clinically important unrecognised pathology. They have much less value in stable chest pain.

Impact on processes of care

Computerised tomography coronary angiography was associated with a reduction in invasive coronary angiography, contrasting with previous acute chest pain trials in which CTCA was associated with increased rates of invasive angiography. 28–31 This disparity is likely to be a result of the differences in trial populations, especially the baseline risk and the prevalence of CHD. We recruited patients with a high prevalence of disease (50–75% with coronary artery disease) in contrast to the low prevalence (< 10%) in prior studies. 28–31

While recent adoption of high-sensitivity troponin level testing has supported earlier clinical decision-making,9,10 it has led to the misidentification of many patients who do not have MI. 13,14 The strength of CTCA is its very high negative predictive value for coronary artery disease, which is applicable to all risk groups, including those patients with non-ST segment elevation acute coronary syndrome. 68 Indeed, we found that 40–50% of patients with normal or non-obstructive coronary artery disease had an elevated cardiac troponin level. Our finding of a 19% relative reduction in the hazard for invasive coronary angiography is likely to reflect the exclusion of obstructive coronary artery disease, thereby avoiding unnecessary invasive coronary angiography, especially in patients with elevations in cardiac troponin levels not attributable to MI or obstructive coronary artery disease. Importantly, this reduction in invasive coronary angiography was not associated with differences in overall rates of coronary revascularisation or clinical events, suggesting excellent diagnostic accuracy and continued appropriate coronary revascularisation for those with obstructive coronary artery disease.

Previous meta-analysis of low-risk acute chest pain trials has shown that CTCA reduces hospital length of stay and subsequent rehospitalisation with suspected acute coronary syndrome. 44 In addition, CTCA was associated with a downstream reduction in cardiovascular investigation. 44 Once again, there was no evidence of this effect in the RAPID-CTCA trial, with no difference between arms in rehospitalisation with chest pain or subsequent cardiovascular or non-cardiovascular investigation rates. Unlike previous data, CTCA in the RAPID-CTCA trial increased rather than decreased the length of hospital stay. These are all likely to reflect the much higher prevalence of underlying coronary artery disease and consequentially high invasive angiography rates in our population than with previous trials. In addition, previous trials in low-risk populations were undertaken in other settings in which standard care for this group involves high rates of admission and intervention. Our finding that CTCA increased length of hospital stay in the intermediate-risk group suggests that the relatively low rate of admission and short time in the NHS leaves little scope for CTCA to achieve the reductions in resource use that have been shown elsewhere and used to support its use in low-risk patients.

Early CTCA was associated with some benefits. Patients valued the use of CTCA, possibly reflecting a more rapid evaluation of their clinical condition. In addition, there was enhanced diagnostic certainty for the attending clinician.

Impact on cost-effectiveness

Economic analysis showed that CTCA was not cost-effective compared with standard care. The lifetime analysis suggested that CTCA was associated with 0.025 fewer QALYs and £481 additional costs per patient treated. Standard care, therefore, dominated CTCA, with 76% probability of being cost-effective at the £20,000 per QALY threshold. The within-trial analysis showed a similar difference in mean costs per patient (£7414 vs. £6845) to the lifetime analysis, suggesting that the difference in costs in the lifetime analysis was driven by the difference in within-trial costs, which were mainly owing to longer hospital length of stay and the additional CTCA use. The cost-effectiveness plane (see Figure 28) showed that most of the estimates of the incremental cost per QALY were in the upper-left quadrant (CTCA less effective and more expensive) or upper-right quadrant (CTCA more effective and more expensive). This suggests that, although there is uncertainty around the effectiveness estimates in the probabilistic lifetime analysis, the cost estimates are more certain. A reasonable overall conclusion is, therefore, that CTCA increases costs without improving effectiveness.

Chief investigator

Professor Alasdair Gray.

Grant co-applicants/Project Management Group

Dr Alan Fletcher (Sheffield Teaching Hospitals NHS Foundation Trust), Professor Steve Goodacre, Professor Alasdair Gray, Mr Robert Lee, Professor Steff Lewis, Dr Graham McKillop (NHS Lothian), Professor David Newby, Professor Carl Roobottom, Surgeon Captain Jason Smith, Professor Robert Storey and Dr Praveen Thokala.

Trial manager

Ms Katherine Oatey.

Research nurse co-ordinator

Ms Rachel O’Brien.

Principal investigators

Professor Alasdair Gray, Professor Steve Goodacre, Professor Carl Roobottom, Dr Jason Smith, Dr Dirk Feldman, Dr Andrew Kinnon, Dr Ajay Yerramasu, Dr Robert Huggett, Dr Liza Keating, Dr Sudantha Bulugahapitiya, Dr Jehangir Din, Dr Andrew Mitchell, Dr Anne Scott, Dr Anna Beattie, Dr Khalid Alfakih, Dr Adrian Brady, Dr Atilla Kardos, Dr Hefin Jones, Dr Derek Connolly, Dr Ronak Rajani, Dr Rangasamy Muthusamy, Dr Simon Smith, Dr Abdel-Rahman Saif-El-Dean, Dr Ben Holloway, Dr Ansuman Saha, Professor Nick Curzen, Dr Matthias Schmitt, Dr Christopher Travill, Dr Ceri Davies, Professor Tim Harris, Dr Will Roberts, Dr Patrick Donnelly, Dr Justin Carter, Dr John Irving, Dr Chris Vorwerk, Dr Ash Basu, Dr Jason Dungu, Dr Elisa McAlindon, Dr Sandeep Hothi, Dr David Rosewarne, Dr Arivalagan Bapusamy, Dr Jonathan Watt and Dr Claire McGroarty.

Trial research team

Royal Infirmary of Edinburgh: Polly Black, Caroline Blackstock, Julia Grahamslaw, Mark Jones, Collette Keanie, Margaret MacLeod, Siobhan McLaughlin, Rachel O’Brien, Fergus Perkes, Alyson Phillips, Ewan Pirie, Janet Summerside, Gordon Truong, Kirsty Weston and Jennifer Wooton.

Sheffield Northern General Hospital: Sarah Bird, Peter Brown, Hridesh Chatha, Alan Fletcher, Catherine Hill, Shery Mofidi, Hasan Qayyum, Rob Storey, Judith Sugden and Anna Wilson.

Derriford Hospital, Plymouth: Alison Jeffrey, Memory Mwadeyi, Rosalyn Squire and Peter Wafer.

Torbay Hospital: Dr Lesley Archer, Lisa Felmeden, Dr Guy Gribbin, Dr Sarah Harrison, Debbie Hughes, Dr Philip Keeling, Dr Ian Mahy, Allison Summerhayes, Justine Sutton and Dr Abdullah Yonis.

Victoria Hospital: Susan Fowler, Amanda McGregor, Karen Gray, Dr Tom Hartley, Dr David Szapiro, Lorraine Dinnel and Dennis Sandeman.

Russells Hall Hospital: Julie Dean and Amy Pugh.

Royal Berkshire NHS Foundation Trust: Parminder Bhuie, Dr James Briggs, Claire Burnett, Abby Gandy, Nicola Jacques, Sarah MacGill, Dr Archie Speirs and Niamh Tolan.

Bradford Teaching Hospitals: Craig Atkinson, Dr Mark Kon, Carita Krannila and Manitha Thomas.

Royal Bournemouth Hospital: Dr Russell Bull, Stephanie Horler, Nicki Lakeman, Jane McLeod, Sara Nix and Dr Sue Thomas.

Jersey General Hospital: Dr Daniel Ahlert, Dr Christopher Edmond, Dr Christopher Hare, Kelly Anne Kinsella, Dr Jessica Langtree, Dr James Speakman and Dr Ranjit Thomas.

Borders General Hospital: Gillian Donaldson, Fiona Hall, Terry Fairbairn and Christopher Rofe.

Royal Victoria Infirmary: Jennifer Adams-Hall, Ange Bailey, Dr Kris Bailey, Leslie Bremner, Dr Ifti Haq and Angela Phillipson.

Lewisham University Hospital: Saroj David, Osman Najam and Samia Pilgrim.

Glasgow Royal Infirmary: Claire Adams, Ammani Brown, Andrew Dougherty, Ailsa Geddes, Karen Lang, David Lowe, Ross MacDuff, Lorraine McGregor, Giles Roditi, Susan Thornton and Joyce Triscott.

Milton Keynes Hospital NHS Foundation Trust: Felicia Adjei, Antoanela Colda, Caitlin Chapman, Veronica Edgell, Michael Fell, Laszlo Halmai, Aarzoo Khan, John Northfield, Cheryl Padilla-Harris, Mike Pashler, Gill Richie, Diane Scaletta, Sarah-Beth Sunderland, Joanne Turner, Lois Vickery, Sonya Walia, Felicity Williams, Lynn Wren and Nicola Wright.

University Hospitals of the North Midlands: Holly Maguire and Resti Varquez.

Sandwell Hospitals: Julie Colley, Anthony D’Sa, Vinoda Sharma, Ashley Turner and Sylvia Willets.

Guy’s and St Thomas’ NHS Foundation Trust: Megan Bell, Dr Giulia Benedetti, Kirsty Gibson, Dr Sze Mun Mak, Dr Rebecca Preston, Amy Raynsford and Ruth Sanchez-Vidal.

Rotherham General Hospital: Susan Biggins, Kathryn Dixon, Peter Kraut, Mwada Lawan, Victoria Murray, Dr Tom Mwambingu, Rachel Walker and Carol Weston.

Leeds General Infirmary: Roo Byrom-Goulthorp, Dr Michael Darby, Dr Annette Eunice Ikongo, Johnstone, Alan Lin and Melanie McGinlay.

Queen Elizabeth Hospital Birmingham: Tania Albutt, Vicky Dawson, Claire Dowling, Karen Isaacs, Cheyanne Kaila, Dr Gareth Lewis, Nicky Mortimer, Sunitha San, Kelly Tabor and Kealy Wright (nee Tombs).

Surrey and Sussex Hospitals: Dr Riaz Ahmed, Sally Collins, Sarah Davies and Nokukhanya Ndlouu.

University Hospital Southampton NHS Foundation Trust: Ausami Abbas, Karen Banks, Julie Bigg, Alison Calver, Simon Corbett, Peter Cowburn, Zoe Duke, Andrew Flett, Sam Gough, Huon Gray, Stephen Harden, Paul Haydock, Michael Mahmoudi, Zoe Nicholas, John Paisey, Charles Peebles, Drew Rakhit, John Rawlins, Paul Roberts, Benoy Shah, James Shambrook, Iain Simpson, Rohit Sirohi, Wagas Ullah, Katharine Vedwan, James Wilkinson and Arthur Yue.

Manchester University NHS Foundation Trust: Sarra Giannopoulou, Melanie Greaves, Stephen McGlynn, Chris Miller, Akhila Muthuswamy, Lindsay Murray, Anie Nicholas and Matthias Schmitt.

Luton and Dunstable University Hospital: Susan Gent and Nafisa Hussain.

Barts Health NHS Trust Royal London Hospital: Raine Astin-Chamberlain, Ben Bloom, Olivia Bolton, Dan Martin, Lyrics Noba, Georgia Norman, Shelley Page, Helen Power, Imogen Skene, David Smith and Jon Walters.

Whipps Cross University Hospital: see Barts Health NHS Trust Royal London Hospital.

Worcestershire Acute Hospitals NHS Trust: Angela Doughty, Elaine Byng-Hollander and Dr Helen Routledge.

Ulster Hospital Belfast: Leah Hammond, Jayne Hutchinson, Stephanie Kelly, Susan Regan and Aileen Smith.

North Tees and Hartlepool NHS Trust: Julie Gray, Sarah Purvis and Pam Race.

Ninewells Hospital: Christine Almaden-Boyle, Kim Bissett, Carol Blues, Jackie Duff, Scot Dundas, Shirley Fawcett, Dr Graeme Houston, Emma Hutchison, Debbie Letham, Ann Mackintosh, Laura Meach, Laura Jayne Queripcz and Alan Webster.

Queen Alexandra Hospital: Julian Atchley, Zoe Daly and Kat Ellinor.

Betsi Cadwaladr University Health Board: Richard Cowell, Helen Craddock, Rachel Hughes, Lynda Sackett, Victoria Saul, Fiona Smith, Jane Stockport and Clare Watkins.

Basildon and Thurrock University Hospitals NHS Foundation Trust: Edward Barden, Jackie Colnet, Swamy Gedeza, Laura Hoskin, Lauren Kittridge, Gracie Maloney, Claire McCormick, Anne Nicholson, Stacey Pepper, Joanne Riches and Annaliza Sevillano.

The Royal Wolverhampton NHS Trust: Vincent Amoah, Stacey Aulton, Dr Arivalagan Bapusamy, Victoria Cottam, Stella Metherell, Sarah Milgate, Elizabeth Radford, Dr David Rosewarne and Andy Smallwood.

Raigmore Hospital: Charlotte Barr, Jonathan Broadie, David Eason, Ing-Marie Logie, Debbie McDonald, Laura O’Keeffe, Donna Patience and Lesley Patience.

Queen Elizabeth University Hospital: Dr Faheem Ahmad, Nicola Baxter, Ammani Brown, Dr John Byrne, Dr Damien Collison, Tracey Hopkins, Hayley King, Dr David Lowe, Evonne McLennan, Dr Giles Roditi, Dr David Stobo, Mark Wilson and Rosie Woodward.

Heart rate control

Heart rate reduction to < 65 b.p.m. should be achieved in all patients in whom it is safe and practicable to do so to ensure absence of motion artefacts on derived images. The development of high-specification CT scanners (e.g. dual source, > 64-slice scanners) means that the likelihood of diagnostic image quality is higher than for 64-slice CT when heart rate control is not achieved, but image quality is improved even on these high-specification CT scanners when the heart rate is < 65 b.p.m.

Beta-blockers

Beta-blockers should be used as the first-line drugs for lowering heart rate prior to CTCA. They can also have the beneficial effect of reducing ectopic activity and heart rate variability. Beta-blockers should not be administered to patients already taking verapamil owing to the risk of ventricular standstill and cardiac arrest. Metoprolol is the most commonly used beta-blocker in this context. It can be administered intravenously or orally.

General considerations

Minimum requirements:

• A maintenance, servicing and repair programme should be in place for all scanners within a department.

• A QA programme should be in place, supervised by appropriately trained radiographers and medical physicists.

Scanner capabilities

Minimum requirements:

• ECG gating with prospective and retrospective gating capability is required.

• Dose modulation should be available for retrospectively gated scanning to reduce the radiation dose to the patient.

• Pre-scan contrast timing assessment, either by automated or by visual bolus tracking or with assessment of a test bolus, must be possible.

Hardware specification

Minimum requirements:

• A 64-slice detector (or above) multidetector CT is required. The non-diagnostic rate of scanners with fewer detector rows is 10% greater than those with 64-slice detectors.

• Detector width should be ≤ 0.625 mm.

• Gantry rotation time should be < 350 milliseconds.

• z-axis coverage of at least 20 mm and preferably 32 mm is required to ensure a realistic scan time (and therefore breath hold) and minimise misalignment artefact.

• Temporal resolution should be < 175 milliseconds for a single sector.

Prospective gating

Minimum technical requirements:

• Prospective gating, which permits a considerable reduction in radiation dose without reduction in image quality, should be used routinely unless there is a clear indication otherwise.

• Where heart rate exceeds the recommended threshold for standard prospective gating, additional tube-on time may be utilised, a technique known as ‘padding’. This allows a greater proportion of the cardiac cycle to be imaged and, although this means a higher radiation dose than with standard prospective gating, this should be used in preference to retrospective gating where possible.

Retrospective gating

• Retrospective gating should be reserved for use as a problem-solving tool, for example in patients with very high heart rates, or rarely where full-cycle acquisition is required, such as functional or heart valve assessment. This is less sensitive to alterations in rhythm than prospective gating. 6

• Except where there is significant heart rate variability (such as atrial fibrillation), dose modulation techniques should be adopted, using the narrowest selectable window of diagnostic tube current. 7

• The pitch should not be less than 0.2 (defined as table travel per rotation divided by the collimation of the X-ray beam).

• Adaptive technologies should be utilised to permit the deletion of data from premature ventricular beats, the insertion of undetected R peaks, and the shifting of R peaks to adjust for arrhythmia, during retrospective gating. 7

Novel gating options

• Prospective helical scanning may be utilised with dual-source scanners where patient heart rate is ≤ 60 b.p.m. and patients meet the limitations of the field of view.

Image quality and radiation dose optimisation

The objective is to obtain diagnostic-quality images while delivering the lowest reasonably achievable radiation dose to the patient. Image quality and radiation dose are intrinsically linked.

Setting the image quality

Temporal resolution is optimised by using a low tube rotation time. Dual-source scanners have intrinsically better temporal resolution than single-source scanners without any impact on patient dose. Multiphase reconstructions incur a dose penalty over single cardiac phase reconstructions because they require projection data to be acquired over a longer part of the cardiac cycle. Multisegment reconstructions incur a dose penalty over single-segment reconstructions because they require projection data to be acquired using retrospective ECG gating, whereas single-segment reconstructions can be generated from data using prospectively gated acquisitions, which are generally more dose efficient.

Spatial resolution is determined by the reconstruction kernel and the slice thickness. Both in turn affect image noise and, therefore, radiation dose.

Contrast resolution is determined by the iodine concentration in the coronary arteries and the tube voltage setting. The sensitivity to iodine is higher at lower tube voltages, so 80 kVp or 100 kVp is advantageous for coronary CTCA. If the tube voltage is reduced without changing the mAs value, the dose to the patient will drop but image noise will increase. Lower-energy X-ray beams may fail to penetrate through the anatomy of medium and large patients, leading to streak artefacts. The optimum choice of tube voltage is, therefore, a compromise dependent on the size of the patient.

Setting scan protocol parameters to optimise the radiation dose

• Where possible, prospective electrocardiographic gating should be used.

• In exceptional cases in which retrospective electrocardiographic gating is required, dose modulation or mA pulsing techniques should be used to keep the radiation dose as low as possible.

• The scan range and field of view should be tailored for each patient. If possible, the scan range must be set inferior to the shoulders to avoid the prescribed mAs being set for the width of the shoulders rather than the thorax. The greater the detector coverage, the harder it is to tailor the scan range to the patient’s anatomy. In retrospective ECG-gated CTCA, the irradiated length may be up to 6 cm longer than the imaged length as a result of helical over-ranging.

• Scan protocols should be size specific. This is achievable by activating tube current modulation with patient size or setting mAs in accordance with the patient’s weight or BMI.

• The tube voltage should be reduced to 100 kVp for small and medium patients: iodine contrast agent is more conspicuous at lower tube voltages. The mAs value can be increased modestly to compensate for the increase in image noise. (The prescribed CTDIvol should still be lower at 100 kVp than at 120 kVp.)

• Iterative reconstruction should be used if built into the CTCA protocol.

• The CTDIvol and DLP predicted for the scan should be used to determine the effect of dose-reduction measures before the scan takes place.

Auditing patient doses

The DLP is the standard radiation dose measure that is used in CTCA. The DLPs recorded should be included in the report.

Contrast protocols

Providing precise recommendations on contrast regimens that cover all clinical scenarios and CT scanner types is very difficult. However, a typical example of injection flow rates for CTCA using a 64-slice CT system for coronary artery imaging is given in Table 31 (assuming a contrast concentration of 350 mgI/ml). Contrast volumes are typically lower in CT scanners that are capable of single heart beat acquisition (either wide-area detector or high-pitch dual-tube scanners) (see Table 31).

Patient history/clinical question

Patient characteristics:

• heart rhythm

• initial heart rate, amount of beta-blockade administered and acquisition heart rate and rhythm during scan

• patient’s BMI and scan compliance (e.g. achieved breath hold).

Scan acquisition parameters

• type of scanner (e.g. 64-slice company name and model)

• scanning mode (e.g. retrospective), including degree of dose modulation (as percentage of cardiac cycle) and padding (in milliseconds) if relevant

• kVP and mA used

• total DLP.

Imaging findings: coronary

• The results from calcium scoring – if performed – describing the distribution (localised or diffuse) and total Agatson score.

• The quality of the CTCA data (excellent, diagnostic, moderate, non-diagnostic).

• Describe the coronary anatomy, including dominance and presence of coronary artery anomalies.

• Each coronary artery is separately commented on with regard to the presence and type of plaques (i.e. non-calcified, mixed or calcified).

• Semiquantitative reporting of the degree of stenosis using the terms mild (< 50% luminal narrowing), moderate (50–75%), severe (≥ 75%) and occlusion (100%) for each coronary artery.

• When bypass grafts are present, the type, origin and site of anastomosis, as well as the degree of stenosis, should be included.

Imaging findings: cardiac

This section is dedicated to information about other non-coronary cardiac findings, such as the morphology of the valves, aortic root, size of the pulmonary trunk and arteries, the ventricular and atrial septum, as well as the presence of myocardial disease (e.g. areas of infarction, scars or aneurysm formation).

Imaging findings: non-cardiac

Extra cardiac findings in the mediastinum, lung and chest wall, including the analysis of the images from the upper abdomen.

Conclusions

Specify conclusions and recommendations.

Further guidance

For further advice please contact:

Professor Carl Roobottom

E-mail address:

Telephone number:

Dr Graham McKillop

E-mail address:

Telephone number:

This reporting guideline closely follows guidance from The Royal College of Physicians. 72