Author + information
- Received January 9, 2018
- Revision received May 4, 2018
- Accepted May 11, 2018
- Published online July 2, 2018.
- Dominik Buckert, MDa,∗ (, )
- Simon Witzel, MDa,
- Jürgen M. Steinacker, MDa,
- Wolfgang Rottbauer, MDa and
- Peter Bernhardt, MDb
- ↵∗Address for correspondence:
Dr. Dominik Buckert, Department of Internal Medicine II, University Hospital Ulm, Albert-Einstein-Allee 23, 89081 Ulm, Germany.
Objectives The purpose of this study was the prospective and randomized evaluation of cardiovascular endpoints and quality of life in patients with stable coronary artery disease comparing a cardiac magnetic resonance (CMR)–based management strategy with a coronary angiography–based approach.
Background Evidence from trials prospectively evaluating the role of CMR in clinical pathways and decision processes is limited.
Methods Patients with symptomatic CAD were randomized to diagnostic coronary angiography (group 1) or adenosine stress CMR (group 2). The primary endpoint was the composite of cardiac death and nonfatal myocardial infarction. Quality of life was assessed using the Seattle Angina Questionnaire at baseline and during follow-up.
Results Two hundred patients were enrolled. In group 1, 45 revascularizations (45.9%) were performed. In group 2, 27 patients (28.1%) were referred to revascularization because of ischemia on CMR. At 12-month follow-up, 7 primary events occurred: 3 in group 1 (event rate 3.1%) and 4 in group 2 (event rate 4.2%), with no statistically significant difference (p = 0.72). Within the next 2 years, 6 additional events could be observed, giving 4 events in group 1 and 9 events in group 2 (event rate 4.1% vs. 9.4%; p = 0.25). Group 2 showed significant quality-of-life improvement after 1 year in comparison to group 1.
Conclusions A CMR-based management strategy for patients with stable coronary artery disease was safe, reduced revascularization procedures, and resulted in better quality of life at 12-month follow-up, though noninferiority could not be proved. Optimal timing for reassessment remains to be investigated. (Magnetic Resonance Adenosine Perfusion Imaging as Gatekeeper of Invasive Coronary Intervention [MAGnet]; NCT02580851)
- cardiovascular magnetic resonance imaging
- inducible ischemia
- patient management
- prognosis and outcomes
- stable coronary artery disease
Current guidelines recommend proper risk stratification prior to invasive therapy for patients with stable coronary artery disease (CAD) (1–6). Detection or exclusion of moderate to severe myocardial ischemia is a crucial part of the workup process that designates patients with ischemia to the high-risk group (7).
Adenosine perfusion cardiac magnetic resonance (CMR) is a noninvasive imaging modality that can reliably detect reversible myocardial ischemia (8–12). It plays an increasing role in the diagnosis and risk stratification of patients with suspected or known CAD (13). Despite the detection of ischemia, it also offers the advantage of providing comprehensive information such as the assessment of myocardial fibrosis and scar (late gadolinium enhancement), which have been proved to further increase its prognostic significance (14–16). Although CMR therefore holds a Class Ia recommendation in the diagnostic workup of patients with symptomatic stable CAD, only a few studies have evaluated a randomized CMR-driven patient management approach.
In this study, a prospective randomized CMR-based management approach for patients with stable CAD was evaluated, with special emphasis on quality of life and occurrence of major cardiac endpoints, in comparison with a conventional approach.
Patients presenting to the outpatient clinic of our institution for the evaluation of symptoms indicating stable symptomatic CAD (e.g., exercise-related angina pectoris or dyspnea) were considered eligible and consecutively screened for enrollment. Patients had to be at intermediate to high CAD risk. Exclusion criteria were unstable angina pectoris, cardiac or respiratory instability, contraindication to CMR (17), age <18 years, and inability to give written informed consent. The study was approved by the institutional ethics committee.
Study protocol and design details have been previously published (18). All patients received thorough history, physical examination, and basic cardiac workup, including rest electrocardiography, treadmill testing, and echocardiography. Risk stratification scores were calculated according to general recommendations (19,20). To assess symptom burden and quality of life, the standardized Seattle Angina Questionnaire (SAQ) was carried out in each subject (21). All patients received guideline-directed medical therapy (1).
After enrollment, patients were randomized to 2 groups in a 1:1 fashion. Randomization was realized by blocked computer-generated random numbers. The allocation sequence was available only to a designated study nurse, who was phoned immediately after the enrollment of an individual patient. In group 1, patients underwent diagnostic coronary angiography after initial assessment. Percutaneous coronary intervention (PCI) was performed according to current guidelines in case of ≥70% stenosis in a coronary vessel with ≥2 mm diameter (1,2) or hemodynamic relevance on fractional flow reserve (FFR) testing. Hence, FFR testing was allowed but not mandated by the study protocol.
In group 2, functional stress testing by adenosine perfusion CMR was performed initially. Functional images were acquired by steady-state free precession images in contiguous short-axis orientation. Adenosine was infused intravenously at a constant rate of 140 μg/kg body weight for 3 min. First-pass perfusion images in 3 short-axis views were acquired using a bolus of gadoterate meglumine 0.05 mmol/kg body weight. The examinations were conducted using a 3.0-T whole-body clinical magnetic resonance scanner using a cardiac 32-channel phased-array receiver coil. Functional evaluation, perfusion imaging, and late gadolinium enhancement assessment were performed according to well-established standard protocols (22–24).
All CMR images were analyzed by 2 readers in consensus. To avoid bias, readers were blinded to initial clinical assessment and the results of other examinations (e.g., treadmill testing). In case myocardial ischemia affecting ≥10% of the myocardium was detected, patients were sent to coronary angiography and subsequent PCI afterward. Figure 1 depicts an example case of a 58-year-old male patient presenting with typical exercise-related angina pectoris. The study protocol is provided in Figure 2.
Follow-up information was gathered annually after enrollment by outpatient clinic visits and by telephone interviews of patients and their general practitioners. Any reported adverse event was documented and its significance evaluated by the endpoint adjudication committee formed by the investigators. Interim analyses were conducted after each year of follow-up to enable early recognition of safety issues. The primary endpoint was defined as the composite of cardiac death and nonfatal myocardial infarction (25,26). Cardiac death was defined as myocardial infarction leading to death, sudden cardiac death, death related to cardiovascular procedures, and cardiovascular hemorrhage. Any diagnostic or interventional coronary procedure not scheduled at time of initial diagnostic workup was recorded as an unplanned procedure. Decisions for these procedures were the responsibility of the treating physicians and were not influenced by the investigators. Symptom burden and quality of life were assessed using the SAQ each year.
Statistical analysis and sample size calculation
To test the relationship between categorical classification factors, the Fisher exact test was applied. Continuous variables were tested for normal distribution using the D’Agostino-Pearson test. In case of normal distribution, variables are reported as mean ± SD, and a 2-tailed Student’s t-test was applied. Cumulative event curves were compared using the Kaplan-Meier method using a log-rank test.
For sample size prediction and noninferiority testing, the Fisher exact test was applied (power 80%). The study was designed to prove noninferiority with a margin of 1% within 3-year follow-up (18).
Statistical analysis was performed using commercially available software (Stata version 13, StataCorp, College Station, Texas; MedCalc, MedCalc Software, Mariakerke, Belgium). Overall, p values ≤0.05 were considered to indicate statistical significance.
Patient enrollment was conducted from January 2012 to October 2014. During that time, 394 patients were prospectively screened for enrollment. The study population consisted of 200 subjects; randomization assigned 101 patients (50.5%) to group 1 and 99 patients (49.5%) to group 2. The mean age of the total cohort was 64.2 ± 9.6 years (group 1, 63.3 ± 9.4 years; group 2, 65.0 ± 9.7 years; p = 0.24). Patients were predominantly men (total cohort, n = 127 [64.8%]; group 1, n = 58 [59.2%]; group 2, n = 69 [70.4%]; p = 0.10). Ten-year risk according to the Framingham risk score (19) was 11.8 ± 7.0% (group 1, 11.4 ± 6.9%; group 2, 12.1 ± 7.2%; p = 0.53). Calculation of the PROCAM (Prospective Cardiovascular Münster Study) score (20) yielded 8.0 ± 6.2% for the total cohort, 8.1 ± 6.9% for group 1, and 8.0 ± 5.7% for group 2, respectively (p = 0.90). There were no significant differences concerning the distribution of cardiovascular risk factors between group 1 and group 2, except for the presence of diabetes mellitus (group 1, n = 13 [13.4%]; group 2, n = 25 [25.5%]; p = 0.04) and known CAD (group 1, n = 48 [49.0%]; group 2, n = 71 [72.4%]; p < 0.01). Symptom burden at baseline was comparable between groups 1 and 2. Table 1 depicts baseline characteristics and frequency of exercise-related angina, dyspnea, and pathological findings on treadmill testing.
Follow-up, treatment, outcomes, and event-rates
Among the 101 patients in group 1, 98 (97%) were treated according to protocol (2 withdrawals, 1 patient did not appear). Of the 99 patients assigned to group 2, treatment according to protocol was conducted in 96 patients (97%) (1 withdrawal, 2 patients did not appear). Complete follow-up information was available in 94 patients in group 1 and 96 patients in group 2. The study flowchart is depicted in detail in Figure 3.
PCI was performed in 45 patients (45.9%) in group 1 at initial examination. Fifty-three patients (54.1%) received medical treatment alone after coronary angiography. In group 2, 69 patients (71.9%) had no inducible ischemia on adenosine stress CMR and thus did not undergo coronary angiography. Of the 27 patients (28.1%) with inducible ischemia, PCI was performed in 23 cases (85.2%). The reasons for nonperformance of PCR were small target vessel (n = 1), no CAD at coronary angiography (n = 1), and patient did not appear for PCI (n = 2).
At 3-year follow-up, a total of 13 primary events could be observed (cardiac death, 2 cases; nonfatal myocardial infarction, 11 cases). Four events occurred in the cohort of patients with primary coronary angiography (group 1), leading to an event rate of 4.1%. For the CMR group (group 2), 9 events could be observed. The overall event rate for this group was 9.4%. The difference between group 1 and group 2 was not statistically significant (p for difference = 0.25), though noninferiority could not be proved (p for noninferiority = 0.88).
At 12-month follow-up, 7 primary events occurred (nonfatal myocardial infarctions in all cases). Three events occurred in group 1. The event rate was 1.9% for patients without PCI, 4.4% for patients with PCI, and 3.1% overall. For group 2, 4 events could be observed. The event rate was 1.4% for patients without ischemia on CMR, 25% for patients with ischemia and without subsequent PCI, 8.7% for patients with ischemia and subsequent PCI, and 4.2% overall. The difference between group 1 and group 2 was not statistically significant (p = 0.72, p for noninferiority = 0.35).
Figure 4 depicts event-free survival at 12-month and 3-year follow-up. The observed annual event rate for group 1 was lower than initially expected (observed annual event rate 1.4%, expected annual event rate ∼6%), while observed and expected annual event rates for the CMR group were fitting (observed annual event rate 3.1%, expected annual event rate ∼2.6%). Overall, there was an overestimation of events leading to an underestimation of sample size. Sex-based differences concerning event occurrence could not be observed.
Evaluation of SAQ at baseline yielded a moderate physical limitation (scale I: overall, 72.9 ± 22.8%; group 1, 71.2 ± 23.1%; group 2, 74.5 ± 22.4%; p = 0.32; scale III: overall, 73.2 ± 21.6%; group 1, 71.6 ± 22.8%; group 2, 74.8 ± 20.2%; p = 0.30), slightly worse angina stability (scale II: overall, 44.6 ± 24.2%; group 1, 41.2 ± 24.2%; group 2, 47.9 ± 23.8%; p = 0.06), excellent treatment satisfaction (scale IV: overall, 81.2 ± 17.1%; group 1, 80.4 ± 17.5%; group 2, 82.0 ± 16.9%; p = 0.53), and good quality of life (scale V: overall, 56.3 ± 26.0%; group 1, 53.1 ± 26.7%; group 2, 59.6 ± 25.0%; p = 0.08).
At 12-month of follow-up, improvements in scale II (angina stability), scale III (angina frequency), and scale V (quality of life) could be observed for both study arms. Scale I (physical limitation) and scale IV (treatment satisfaction) showed stable results in comparison with baseline. Patients in group 2 reported significantly higher values on scales I, IV, and V in comparison with group 1 (scale I, 80.1 ± 19.2 vs. 70.0 ± 25.2 [p < 0.01]; scale IV, 82.5 ± 16.1 vs. 77.0 ± 18.8 [p < 0.05]; scale V, 74.9 ± 21.4 vs. 64.5 ± 28.3 [p < 0.01]). At 3-year follow-up, these differences could not be reproduced. Development of SAQ scales is depicted in detail in Table 2 and Figure 5.
Number of revascularization procedures
Table 2 shows the number of revascularization procedures for both study arms. The initial assessment by diagnostic coronary angiography scheduled 45 (group 1) and 27 (group 2) revascularization procedures (p = 0.01). Within the first 12 months of follow-up, 31 additional revascularizations were performed (group 1, n = 15; group 2, n = 16; p = 0.85). At 3-year follow-up, a total of 154 revascularization procedures had been conducted. The significant difference at baseline between groups 1 and 2 did not persist (group 1, n = 83; group 2, n = 71; p = 0.08).
In this prospective and randomized clinical trial, a CMR-based management approach for patients with stable CAD was evaluated in comparison with a coronary angiography–based strategy. The initial performance of adenosine stress CMR correctly identified high-risk patients, proved to be safe with regard to endpoint occurrence, reduced the rate of revascularization procedures, and was accompanied by good patient satisfaction as measured by the standardized SAQ within the first 12 months of follow-up.
The groups did not differ significantly for primary endpoint occurrence. However, patients in the CMR group had a slightly higher rate of cardiac events than patients in the standard treatment arm. Consequently, noninferiority of the CMR-based strategy could not be proved after 3 years of follow-up.
Several studies have shown that patients with stable CAD have prognostic benefit from revascularization only if myocardial ischemia was detected beforehand (5). Consequently, current guidelines recommend the documentation of ischemia using functional testing before elective revascularization procedures (27). Stress perfusion CMR has proved its ability to reliably detect myocardial ischemia and therefore is highly recommended in the evaluation of patients presenting with stable CAD symptoms (1,2,28). Nevertheless, studies so far mainly have concentrated on CMR’s diagnostic accuracy in comparison with coronary angiography. Evidence from trials prospectively evaluating its role in clinical pathways and decision processes is still limited. The recently published CE-MARC 2 (Clinical Evaluation of Magnetic Resonance Imaging in Coronary heart disease) trial prospectively randomized 1,202 patients with stable CAD symptoms to 3 diagnostic strategies, including functional imaging modalities (multiparametric CMR and myocardial perfusion scintigraphy) as gatekeeper prior to invasive treatment (29). The primary endpoint was defined as unnecessary coronary angiography as determined by FFR measurement. This endpoint occurred in 36 patients (7.5%) in the CMR group with comparison with 69 patients (28.8%) receiving standard care (p < 0.001). Moreover, patients’ rates of major adverse cardiac events were documented within the first 12 months after initial assessment. Within this period, a significant difference concerning major adverse cardiac events between standard treatment and the CMR group could not be observed (1.7% vs. 2.5%). It is of note that patients in this study on average exhibited 2 cardiovascular risk factors and therefore were at lower risk than the subjects in the study at hand. Another study evaluating a CMR-based management of stable CAD is the MR-INFORM trial (30). This randomized multicenter study compared stress perfusion CMR with coronary angiography with FFR measurement. One-year results were presented at the 2017 Annual Scientific Sessions of the American College of Cardiology (31). Again, initial assessment by CMR significantly reduced the number of revascularization procedures (FFR 44.2%; CMR 36.0%; p = 0.005) within the first year of follow-up. Rates of major adverse cardiac events were not significantly different (FFR 3.9%; CMR 3.3%; p = 0.62). Patients in our study showed comparable event rates at 12-month follow-up in comparison with CE-MARC 2 and MR-INFORM (group 1 [primary diagnostic coronary angiography] 3.1%; group 2 [primary CMR] 4.2%). There was no significant difference in event rates between the study arms after the first year of follow-up. As in CE-MARC 2 and MR-INFORM, the number of revascularization procedures was significantly lower in the CMR arm (group 1, 45; group 2, 27; p = 0.01). It is of note that application of FFR was allowed but not mandatory in our study protocol. Because exclusive visual assessment of coronary stenosis often results in under- and overestimation of their hemodynamic significance, it can be assumed that strict use of FFR might have reduced the number of revascularization procedures in our study (32). Nevertheless, requiring FFR by study protocol as in MR-INFORM led to comparable results.
Growing evidence supports CMR-based guidance of revascularization. Nevertheless, it remains unclear for which time period a negative stress perfusion result will yield high negative predictive value for major cardiac endpoints and for improvement in quality of life. In our trial, a remarkable number of primary endpoints occurred after the first year of follow-up, eventually resulting in failure to prove noninferiority 3 years after initial assessment. Although it did not result in differences in Framingham or PROCAM score, in the CMR group significantly more patients had diabetes (25.5% vs. 13.4%; p = 0.04) and a history of CAD (72.4% vs. 49%; p < 0.01), resulting in higher risk for higher severity of CAD. These factors might have contributed to the fact that CMR did not meet noninferiority criteria in our trial. In a larger and preferably multicenter trial, these between-group imbalances most likely would not be apparent and might alter the results during late follow-up in favor of the CMR group, especially because it has been shown that CAD progression is faster in patients with diabetes. The latter aspect led to the recommendation of earlier reassessment in these patients, preferably at an interval of 2 years.
Several studies focusing on diagnostic accuracy and prognostic significance of CMR reported excellent negative and positive prognostic values for up to 3 years after the initial test. In a consecutive cohort of 1,152 patients with stable CAD undergoing adenosine perfusion CMR, a negative predictive value of 95.6% in case of a negative stress result was demonstrated during a mean follow-up period of 4.2 years (10). The hazard ratio for patients with positive stress results was 3.94 (p < 0.0001). Another study evaluating stress CMR in a comparable cohort of 3,138 patients produced similar results: for patients without wall motion abnormalities, annual event rates were about 1% within the first 3 years after initial assessment (33). However, these data have been derived mostly from retrospective and observational studies. To the best of our knowledge, no prospective controlled trial evaluating a CMR-based management strategy has reported event rates beyond 12 months of follow-up. Therefore, it remains unclear whether repetitive CMR examinations after 12 months could be beneficial in terms of cardiac endpoint avoidance, which patients would have to be repeatedly examined, and which examination intervals would be adequate.
Symptom burden and patient satisfaction play crucial roles in disease management, besides the number of major cardiac endpoints and invasive procedures. Remarkably, only a few studies have reported this important information. In our trial, patients experienced moderate physical limitations and worse angina stability at the time of initial assessment. Within 12 months, improvement of angina stability, angina frequency, and quality of life could be observed for both study arms.
There was a small but significant difference concerning physical limitation, treatment satisfaction, and quality of life in favor of the CMR group after 12 months of follow-up. This finding supports the appropriateness of stress perfusion CMR in patient management. Nevertheless, the differences in quality of life were not sustained during longer term follow-up. This finding might be consistent with the observation that more endpoints occurred and revascularization procedures were performed in this period. Further studies focusing on long-term management of patients with stable CAD on the basis of symptoms and already performed diagnostic and therapeutic interventions thus are warranted.
Within 12 months of follow-up, a CMR-based strategy for the management of patients with intermediate risk and stable CAD was safe, reduced revascularization procedures, and resulted in better symptom control and patient satisfaction in comparison with a conventional coronary angiography–based approach. The appropriate time for reassessment of patients at high risk for adverse events remains unclear and must be further evaluated.
COMPETENCY IN MEDICAL KNOWLEDGE: With its ability to noninvasively detect or exclude myocardial ischemia, CMR plays an important role in the diagnostic evaluation of patients with symptomatic stable CAD. Although its diagnostic accuracy has extensively been proved, there is only little information on how to include this emerging modality in clinical pathways and decision processes. In this trial, it was demonstrated that a CMR-based management approach was safe with regard to the occurrence of major cardiac events, resulted in reduction of revascularization procedures, and led to good symptom control and patient satisfaction within 12 months after initial assessment.
TRANSLATIONAL OUTLOOK: Although the CMR-based strategy exhibited good results within the first year of follow-up, the appropriate time for reassessment of patients at high risk for adverse events remains unclear. Further prospective management studies must be conducted to correctly identify those patients and to establish improved clinical pathways.
This trial has been partially funded by Guerbet. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery disease
- cardiac magnetic resonance
- fractional flow reserve
- percutaneous coronary intervention
- Seattle Angina Questionnaire
- Received January 9, 2018.
- Revision received May 4, 2018.
- Accepted May 11, 2018.
- 2018 American College of Cardiology Foundation
- Fihn S.D.,
- Gardin J.M.,
- Abrams J.,
- et al.
- Fihn S.D.,
- Blankenship J.C.,
- Alexander K.P.,
- et al.
- Buckert D.,
- Cieslik M.,
- Tibi R.,
- Radermacher M.,
- Rottbauer W.,
- Bernhardt P.
- Shaw L.J.,
- Berman D.S.,
- Picard M.H.,
- et al.,
- for the National Institutes of Health/National Heart, Lung, and Blood Institute-Sponsored ISCHEMIA Trial Investigators
- Buckert D.,
- Dewes P.,
- Walcher T.,
- Rottbauer W.,
- Bernhardt P.
- Bingham S.E.,
- Hachamovitch R.
- Bernhardt P.,
- Spiess J.,
- Levenson B.,
- et al.
- Buckert D.,
- Kelle S.,
- Buss S.,
- et al.
- Steel K.,
- Broderick R.,
- Gandla V.,
- et al.
- Prasad S.K.,
- Pennell D.J.
- Buckert D.,
- Witzel S.,
- Cieslik M.,
- Tibi R.,
- Rottbauer W.,
- Bernhardt P.
- D’Agostino R.B. Sr..,
- Vasan R.S.,
- Pencina M.J.,
- et al.
- Assmann G.,
- Cullen P.,
- Schulte H.
- Spertus J.A.,
- Winder J.A.,
- Dewhurst T.A.,
- et al.
- Kramer C.M.,
- Barkhausen J.,
- Flamm S.D.,
- Kim R.J.,
- Nagel E.,
- for the Society for Cardiovascular Magnetic Resonance Board of Trustees Task Force on Standardized Protocols
- Schulz-Menger J.,
- Bluemke D.A.,
- Bremerich J.,
- et al.
- Hicks K.A.,
- Tcheng J.E.,
- Bozkurt B.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Foley J.R.J.,
- Plein S.,
- Greenwood J.P.
- ↵Nagel E. MR-INFORM: stress perfusion imaging to guide the management of patients with stable coronary artery disease. Presented at: American College of Cardiology Annual Scientific Session (ACC 2017); Washington, DC; March 17, 2017.
- Gewirtz H.
- Kelle S.,
- Nagel E.,
- Voss A.,
- et al.