Objectives: To assess the acceptability and feasibility of functional tests as a gateway to angiography for management of coronary artery disease (CAD), the ability of diagnostic strategies to identify patients who should undergo revascularisation, patient outcomes in each diagnostic strategy, and the most cost-effective diagnostic strategy for patients with suspected or known CAD.
Design: A rapid systematic review of economic evaluations of alternative diagnostic strategies for CAD was carried out. A pragmatic and generalisable randomised controlled trial was undertaken to assess the use of the functional cardiac tests: angiography (controls); single photon emission computed tomography (SPECT); magnetic resonance imaging (MRI); stress echocardiography.
Setting: The setting was Papworth Hospital NHS Foundation Trust, a tertiary cardiothoracic referral centre.
Participants: Patients with suspected or known CAD and an exercise test result that required non-urgent angiography.
Interventions: Patients were randomised to one of the four initial diagnostic tests.
Main outcome measures: Eighteen months post-randomisation: exercise time (modified Bruce protocol); cost-effectiveness compared with angiography (diagnosis, treatment and follow-up costs). The aim was to demonstrate equivalence in exercise time between those randomised to functional tests and those randomised to angiography [defined as the confidence interval (CI) for mean difference from angiography within 1 minute].
Results: The 898 patients were randomised to angiography (n = 222), SPECT (n = 224), MRI (n = 226) or stress echo (n = 226). Initial diagnostic tests were completed successfully with unequivocal results for 98% of angiography, 94% of SPECT (p = 0.05), 78% of MRI (p < 0.001) and 90% of stress echocardiography patients (p < 0.001). Some 22% of SPECT patients, 20% of MRI patients and 25% of stress echo patients were not subsequently referred for an angiogram. Positive functional tests were confirmed by positive angiography in 83% of SPECT patients, 89% of MRI patients and 84% of stress echo patients. Negative functional tests were followed by positive angiograms in 31% of SPECT patients, 52% of MRI patients and 48% of stress echo patients tested. The proportions that had coronary artery bypass graft surgery were 10% (angiography), 11% (MRI) and 13% (SPECT and stress echo) and percutaneous coronary intervention 25% (angiography), 18% (SPECT) and 23% (MRI and stress echo). At 18 months, comparing SPECT and stress echo with angiography, a clinically significant difference in total exercise time can be ruled out. The MRI group had significantly shorter mean total exercise time of 35 seconds and the upper limit of the CI was 1.14 minutes less than in the angiography group, so a difference of at least 1 minute cannot be ruled out. At 6 months post-treatment, SPECT and angiography had equivalent mean exercise time. Compared with angiography, the MRI and stress echo groups had significantly shorter mean total exercise time of 37 and 38 seconds, respectively, and the upper limit of both CIs was 1.16 minutes, so a difference of at least 1 minute cannot be ruled out. The differences were mainly attributable to revascularised patients. There were significantly more non-fatal adverse events in the stress echo group, mostly admissions for chest pain, but no significant difference in the number of patients reporting events. Mean (95% CI) total additional costs over 18 months, compared with angiography, were 415 pounds (-310 pounds to 1084 pounds) for SPECT, 426 pounds (-247 pounds to 1088 pounds) for MRI and 821 pounds (10 pounds to 1715 pounds) for stress echocardiography, with very little difference in quality-adjusted life-years (QALYs) amongst the groups (less than 0.04 QALYs over 18 months). Cost-effectiveness was mainly influenced by test costs, clinicians' willingness to trust negative functional tests and by a small number of patients who had a particularly difficult clinical course.
Conclusions: Between 20 and 25% of patients can avoid invasive testing using functional testing as a gateway to angiography, without substantial effects on outcomes. The SPECT strategy was as useful as angiography in identifying patients who should undergo revascularisation and the additional cost was not significant, in fact it would be reduced further by restricting the rest test to patients who have a positive stress test. MRI had the largest number of test failures and, in this study, had the least practical use in screening patients with suspected CAD, although it had similar outcomes to stress echo and is still an evolving technology. Stress echo patients had a 10% test failure rate, significantly shorter total exercise time and time to angina at 6 months post-treatment, and a greater number of adverse events, leading to significantly higher costs. Given the level of skill required for stress echo, it may be best to reserve this test for those who have a contraindication to SPECT and are unable or unwilling to have MRI. Further research, using blinded reassessment of functional test results and angiograms, is required to formally assess diagnostic accuracy. Longer-term cost-effectiveness analysis, and further studies of MRI and new generation computed tomography are also required.