We evaluated the prognostic value of cardiac magnetic resonance (CMR) stress testing with direct comparison of adenosine stress first-pass perfusion and dobutamine stress wall motion imaging among 513 subjects with known or suspected coronary heart disease over a median follow-up period of 2.3 years. Positive results on CMR stress testing identified subjects at high risk for subsequent cardiac events (nonfatal myocardial infarction or cardiac death), whereas normal CMR stress test results were associated with a very low annual cardiac event rate.
The superior diagnostic capabilities of cardiac magnetic resonance (CMR) stress testing have been extensively studied and are well recognized. Cardiac magnetic resonance stress testing is currently used successfully in numerous centers for the detection of inducible wall motion abnormalities during dobutamine stress magnetic resonance (DSMR) imaging1-3 or inducible perfusion deficits during adenosine stress magnetic resonance perfusion (MRP) imaging.4,5 This superiority has mainly been attributed to the consistently high image quality of CMR scans, with limiting acquisition windows/near-field artifacts (echocardiography) or diaphragmatic/tissue attenuation artifacts (nuclear techniques) no longer representing a diagnostic pitfall. Consequently, CMR stress testing is considered a class II indication for the detection of coronary artery disease (CAD) under current guidelines.6
An important prerequisite of any stress test is the ability not only to correctly diagnose the presence of obstructive CAD but also to discriminate those patients at high risk for future cardiac events from those with a low cardiac event rate. However, our knowledge about the prognostic value of CMR stress testing has been limited. Thus, we evaluated the role of adenosine stress MRP and DSMR wall motion imaging for the prediction of cardiac events (myocardial infarction [MI] or cardiac death) in patients with known or suspected CAD.
Subjects and Methods
The sample included 513 subjects with known or suspected CAD using a combined single-session CMR examination consisting of adenosine stress and rest first-pass perfusion imaging as well as dobutamine stress wall motion analysis (standard high-dose dobutamine/atropine protocol). A total of 473 subjects successfully completed the combined CMR stress examination, with 12 subjects lost to follow-up. Thus, the final analysis included 461 subjects. Subjects who underwent a revascularization procedure within 3 months of CMR stress testing were excluded from the analysis, and those who underwent a revascularization procedure more than 3 months after CMR stress testing were censored at the time of revascularization.
Clinical variables were assessed at the time of the CMR examination based on subjects' Framingham Risk Score; the total number of concomitant cardiac risk factors (range, 0-7) was also identified, including a history of CAD, a family history of CAD, hypertension, current or prior cigarette smoking, age (>45 years for men, >55 years for women), hyperlipoproteinemia, and diabetes mellitus. Survival curves were estimated by the Kaplan-Meier method and compared by the log-rank test. Univariate and stepwise multivariate Cox proportional hazards models were used to calculate hazard ratios and 95% confidence intervals (CIs). In addition, a 3-step modeling procedure was carried out to test for an incremental prognostic value of CMR stress imaging over clinical parameters and resting wall motion abnormalities.
Table 1. Univariate predictors of cardiac death and nonfatal myocardial infarction in subjects
undergoing combined adenosine stress MRP and DSMR wall motion imaging.
MRP indicates magnetic resonance perfusion; DSMR, dobutamine stress magnetic resonance;
HR, hazard ratio; CI, confidence interval; LVEF, left ventricular ejection fraction; MRP,
magnetic resonance perfusion.
†Per 10% LVEF points.
‡Per 10-mL change.
(Reprinted with permission from Jahnke C, Nagel E, Gebker R, et al. Prognostic value of cardiac
magnetic resonance stress tests: adenosine stress perfusion and dobutamine stress wall
motion imaging. . 2007;115:1769-1776.)
During a mean follow-up period of 2.26 ± 1.03 years (median, 2.30 years; range, 0.06 to 4.55 years), 19 hard cardiac events were observed (4.1%), consisting of 10 nonfatal MIs and 9 cardiac deaths. Clinical variables that were univariate predictors of hard cardiac events were more than 4 cardiac risk factors, known CAD, and diabetes mellitus. Magnetic resonance imaging results that predicted cardiac events were resting wall motion abnormalities, left ventricular ejection fraction, and positive results on CMR stress testing (Table 1). However, on multivariate analysis, an abnormal result on either adenosine stress MRP (hazard ratio = 10.57; 95% CI, 2.86-39.07; < .001) or DSMR wall motion analysis (hazard ratio = 4.72; 95% CI, 1.76-12.64; = .002) remained the only independently associated predictors of hard cardiac events. Representative imaging examples are shown in Figure 1.
Figure 1. (A) Dobutamine stress magnetic resonance (DSMR) wall motion and adenosine
stress magnetic resonance perfusion (MRP) images of a 74-year-old man with known
coronary heart disease (CAD). At rest, the anterior and anteroseptal segments are
hypokinetic, with improvement in contractile function under low-dose dobutamine stress
but deterioration of wall motion at maximum stress (akinetic anterior and anteroseptal
segments; white arrows). During stress MRP, an inducible subendocardial perfusion deficit
in the identical segments was found (white arrows). Sudden cardiac death occurred 1.24
years after the magnetic resonance examination. (B) DSMR wall motion and adenosine
stress MRP of a 66-year-old woman with known CAD. Regional wall motion at rest and
under low-dose dobutamine infusion was shown to be normal, whereas at maximum stress,
an akinetic inferolateral segment was identified. Stress MRP showed an extensive inducible
circumferential perfusion deficit, most prominent in the anterior and anteroseptal segments.
Nonfatal myocardial infarction occurred 3 months later.
Sequential Cox regression models were used to further examine whether CMR imaging data added to the predictive capabilities of clinical data. During a 3-step modeling procedure, variables were included in the same order as in clinical practice, that is: (1) clinical data, (2) wall motion abnormalities and left ventricular systolic function at rest, and (3) results of CMR stress testing. With more than 4 cardiac risk factors, the presence of wall motion abnormalities at rest further increased the likelihood of hard cardiac events (χ2 test, 10.6 vs 16.0; <.001). In addition, a positive result on MRP or DSMR further strengthened the model and demonstrated an incremental value over clinical data and wall motion imaging at rest (χ2 test, 34.3 and 29.5, respectively; <.001).
Table 2. Cumulative event rate during the 3-year follow-up period based on results of adenosine
stress MRP and DSMR testing.
MRP indicates magnetic resonance perfusion; DSMR, dobutamine stress magnetic resonance.
(Reprinted with permission from Jahnke C, Nagel E, Gebker R, et al. Prognostic value of cardiac
magnetic resonance stress tests: adenosine stress perfusion and dobutamine stress wall motion
Figure 2 shows the Kaplan-Meier curves for event-free survival based on the results of CMR stress testing. Notably, a low annual cardiac event rate was demonstrated in the case of a negative CMR stress study (Table 2). A combined analysis of the CMR stress test results (ie, the combined test is regarded normal if both tests are normal, and it is regarded abnormal if either test shows an abnormality), showed that combined testing was not better for prognosis of event-free survival than each CMR stress test independently (log-rank, = .18).
Figure 2. Click to enlarge.
We assessed the prognostic value of CMR stress testing in subjects with suspected or known CAD. The utility of adenosine stress MRP and DSMR wall motion imaging for identifying those subjects at high risk for future cardiac events (cardiac death or MI) could be assessed directly and comparatively. Results of our study showed that MRP and DSMR were equally valuable in identifying subjects at risk for future cardiac events. In addition, a negative CMR stress test result was associated with a very low cardiac event rate, and CMR stress testing provided incremental prognostic information over clinical risk factors and wall motion assessment at rest.
Although the diagnostic value of adenosine MRP and DSMR wall motion imaging is well known, data regarding the prognostic value are still limited. A clinically important prerequisite for any stress modality, however, is the ability not only to distinguish the presence from the absence of obstructive CAD, but also to segregate subjects at high risk for future cardiac events from those with a low cardiac event rate. Our data showed that CMR stress testing provides incremental prognostic information compared with clinical risk factors and wall motion assessment at rest. Moreover, during multivariate analysis, a positive CMR stress test result proved to be the only independent predictor of future cardiac events. Thus, CMR stress testing allows for the identification of high-risk subjects with a positive test result during MRP or DSMR, resulting in a 12- or 5-fold increased risk, respectively, of experiencing a major cardiac event (nonfatal MI or cardiac death).
The prognostic impact of noninvasive stress testing using well-established nuclear imaging techniques has already been investigated. With nuclear myocardial perfusion imaging, a normal test result indicates a major cardiac event rate of approximately 1% per year during the following 3-year period.7,8 Similarly, a normal test result during dobutamine stress echocardiography identifies subjects with a low cardiac event rate in the range of 1% to 3% per year.9 Cardiac magnetic resonance imaging offers the opportunity to combine stress perfusion and stress wall motion imaging during a single session examination, thereby allowing for a direct comparison of both stress tests. Our data showed that a normal test result during MRP or DSMR resulted in similarly low cardiac event rates during the following 2 years (0.7% and 2.6%, respectively). This result is in accordance with previously reported data on nuclear imaging or echocardiography.
It is noteworthy that the predictive power of a negative test result of MRP or DSMR may be considered limited to a 2-year "warranty period": event-free survival seemed to be prolonged only in subjects with a normal MRP a normal DSMR result, as indicated by a consistently low event rate of 0.8%. Further long-term and large-scale multicenter trials are needed to verify this observation.
In our study, adenosine stress MRP and DSMR wall motion imaging were shown to be equally effective in identifying subjects at high risk for future nonfatal MI or cardiac death, whereas a negative result on CMR stress testing indicated a very low cardiac event rate during a 2-year period. In addition, CMR stress testing proved to be superior to clinical risk factors and wall motion assessment at rest for the prediction of cardiac outcome.