Cardiovascular outcomes are predicted by exercise-stress myocardial perfusion imaging: Impact on death, myocardial infarction, and coronary revascularization procedures

Imaging and Diagnostic Testing

Cardiovascular outcomes are predicted by exercise-stress myocardial perfusion imaging: Impact on death, myocardial infarction, and coronary revascularization procedures Douglas S. Lee, MD, PhD, a,b,c Flavia Verocai, MD, a,b Mansoor Husain, MD, a,b Darar Al Khdair, MD, b Xuesong Wang, MSc, c Michael Freeman, MD, d and Robert M. Iwanochko, MD a,b Toronto, Canada

Background The aim of this study was to determine the impact of myocardial perfusion imaging (MPI) on the outcomes of death, myocardial infarction (MI), and late coronary revascularization procedures. Methods

In patients undergoing exercise-stress MPI (January 1, 2003–March 31, 2007), we determined the impact of summed stress score (SSS) and percent left ventricular (LV) ischemia on (a) death or MI and (b) composite of death, MI, or late coronary revascularization occurring more than 90 days post-MPI.

Results

During 35,007 person-years of follow-up among 9,605 patients (mean ± SD age 54.4 ± 13.2 years, 60.3% men), there were 290 deaths, 175 MIs, and 525 coronary revascularization procedures. Of those who attained ≥10 metabolic equivalents (METS) workload, major stress perfusion defects (SSS ≥7) were present in 4.2% overall and in 3.7% without ST-segment shifts, whereas large ischemic defects (≥10% LV ischemia) were present in 1% overall and 0.7% without STsegment shifts. For those with 1% to 4%, 5% to 9%, and ≥10% LV ischemia, adjusted hazard ratios were 1.40 (95% CI 1.131.73, P = .002), 2.07 (95% CI 1.56-2.74, P b .001), and 3.03 (95% CI 2.21-4.16, P b .001) for the outcome of late revascularization, MI, or death versus no ischemia. Summed stress scores ≥7 were associated with increased risk of death or MI, with an adjusted hazard ratio of 1.57 (95% CI 1.16-2.13, P = .004) compared with those with no stress perfusion defects.

Conclusion

Although workload ≥10 METS conferred lower frequency of major ischemia (≥10%), %LV ischemia predicted the occurrence of cardiovascular events and death (eg, MI, late coronary revascularization, or death). Presence of a large stress perfusion defect (SSS ≥7) predicted increased risk of MI or death. (Am Heart J 2011;161:900-7.)

Approximately 1 in 3 individuals have some form of cardiovascular disease, and many of these may need diagnostic testing with exercise treadmill stress during their lifetime.1 Exercise treadmill testing with or without myocardial perfusion imaging (MPI) is a widely used diagnostic test modality for detection of coronary artery disease. The advantages of exercise treadmill tests are its wide availability and the relatively short time duration needed to perform the test. In contrast, MPI requires a longer duration for completion of the entire test, may have more limited availability, and is more expensive than basic exercise treadmill testing. From the aRobert J. Burns Nuclear Cardiology Laboratory, Toronto, Canada, bUniversity Health Network, Toronto, Canada, cInstitute for Clinical Evaluative Sciences, University of Toronto, Toronto, Canada, and dSt. Michael’s Hospital, University of Toronto, Toronto, Canada. Submitted July 12, 2010; accepted January 31, 2011. Reprint requests: Douglas S. Lee, MD, PhD, Institute for Clinical Evaluative Sciences, University of Toronto, Rm G-106, 2075 Bayview Ave, Toronto, ON, M4N 3M5. E-mail: [email protected] 0002-8703/$ - see front matter © 2011, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2011.01.019

There is heightening interest in the evaluation of noninvasive test modalities for diagnosis and prognosis in those with suspected coronary heart disease. The ability to exercise is a prognostic marker, and recent data have suggested that patients who are able to exercise to a workload of 10 metabolic equivalents (METS) with a negative stress electrocardiogram (ECG) have normal perfusion.2 However, this study was relatively small in size and did not examine outcomes including revascularization procedures.3 Prognostication may be enhanced by quantitation of outcome events when the degree of ischemia is cross-indexed with workload attained. Recent studies have highlighted the need for evaluations of noninvasive imaging that can predict those who are more likely to undergo coronary revascularization in addition to the traditionally evaluated outcomes of myocardial infarction (MI) and death.4 In this context, the impact of MPI on process of coronary revascularization has not been fully evaluated. We examined the association of MPI with clinical outcomes and coronary revascularization care accounting for exercise capacity. We hypothesized that MPI results

American Heart Journal Volume 161, Number 5

would be associated with increased event rates and increased late coronary revascularization throughout the range of exercise workloads attained.

Methods Patients We examined consecutive patients (≥18 years of age) who underwent exercise-stress MPI from January 1, 2003, to March 31, 2007, at the Robert J. Burns Nuclear Cardiology Laboratory. Patients with a valid health card number who were residents of Ontario, Canada, were included. Ethical approval for this study was obtained from the research ethics board of the University Health Network.

Exercise protocol Before exercise testing, the patient's resting heart rate (HR) and blood pressure (BP) were measured, resting 12-lead ECG was performed, and a baseline survey was administered for symptom history and presence of coronary heart disease risk factors. Patients underwent exercise testing using the Bruce protocol until age-predicted maximal exercise HR ≥85% was achieved, or a positive stress ECG response (eg, ≥1-mm horizontal/downsloping ST-depression), inability to continue treadmill because of limiting symptoms, or life-threatening arrhythmias occurred during exercise stress. Workload attained was categorized as follows: b7, 7-9, and ≥10 METS. Those who had limitations to stress ECG interpretation (eg, bundle-branch blocks, left ventricular [LV] hypertrophy, resting ST-abnormalities, or paced rhythm) were excluded.

MPI protocol Imaging was performed using 99mtechnetium-sestamibi–gated single-photon emission computed tomography, with rest images performed 30 to 60 minutes after tracer injection (370-444 MBq [10-12 mCi]). Images were acquired supine with both arms raised above the head using a dual-head gamma-camera in the 90°-setting (Siemens, SMV, or ADAC) with LEAP or VXGP/VXGP collimator. Patients were connected to 3 nonradiopaque ECG leads so that the study could be gated by R-R interval. Two to 3 hours after rest images, patients were exercise stressed and injected with 925 to 1110 MBq (25-30 mCi) of 99mTc-sestamibi. Images were taken 15 to 60 minutes following stress. Sixty-four projections were acquired. All images were stored in a 64 × 64 matrix, processed, and reconstructed according to the American College of Cardiology/American Society of Nuclear Cardiology algorithm. Three orthogonal slices, short, horizontal, and vertical long axis, were obtained for display and interpretation. Attenuation correction was applied as required when the body mass index was ≥30 kg/m2. Images were reported using the 17-segment reporting model of the American Heart Association, with each segment scored from 0 (normal uptake) to 4 (absent tracer uptake), yielding summed stress scores (SSS) and summed rest scores ranging from a normal score of 0 to maximum of 68.5 The extent of reversible ischemia was defined by the summed difference score (SDS), which is calculated as: SDS = SSS − summed rest score. The percent myocardium with reversible ischemia was defined as (SDS ÷ 68) × 100%, and thus, SDS ≥7 corresponded to

Lee et al 901

10% LV ischemia. Normal and abnormal stress perfusion were defined as SSS = 0 and SSS N 0, respectively.

Data linkages Clinical, exercise, and MPI data were linked to administrative databases using the patients' unique, encrypted health card number. We used the Canadian Institute for Health Information (CIHI) Discharge Abstract Database and CIHI Same-Day Surgery Database to determine the presence of prior cardiac disease and procedures using the International Classification of Diseases 9th or 10th (ICD-9/10) revision. Prior MI (ICD-9 code 410; ICD10 codes I21, I22), heart failure (HF; ICD-9 428, ICD-10 I50), and percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) surgery (ICD-9 CCP 4802, 4803, 481, 482, 483; ICD-10 CCI 1IJ76, 1IJ50, 1IJ57GQ, and 1IJ54GQAZ) were identified. We examined the Registered Persons Database for vital status. These databases have been extensively validated previously and found to have high accuracy.6-9

Outcomes evaluation We examined the following composite outcomes: (a) death or MI and (b) death, MI, or revascularization procedure. For the latter composite outcome, we included late revascularization procedures that were performed more than 90 days post-MPI, as a study endpoint. We did not include early revascularization (within 90 days post-MPI) as an outcome because of the potential that it would be driven by MPI results.

Statistical analysis Descriptive results are reported as mean ± SD or median (25th, 75th percentile) for continuous variables and frequencies for categorical variables. Comparisons were performed using analysis of variance or Student t test for continuous variables and χ2 test for categorical variables. We performed multiple Cox regression analysis to determine the effects of workload, STsegment shift, and extent of reversible ischemia (eg, %LV ischemic or SDS) adjusted for age and sex. Subsequent models were adjusted further for additional clinical variables. Patients were followed for ≥2 years from MPI date and censored at last follow-up date of March 31, 2009. For models examining late revascularization outcomes, we censored procedures that occurred within the first 90 days after MPI. The proportional hazards assumption was tested for Cox regression models. Adjusted survival curves were constructed using the corrected group prognosis method.10 A P value b.05 was considered statistically significant. All analyses were performed using SAS, version 9.2 (SAS Institute, Cary, NC).

Results Patient characteristics Among 9,732 patients who underwent exercise-stress MPI, 127 were excluded because of invalid health card numbers. The total study sample was 9,605 patients (54.4 ± 13.2 years), with 5,789 men and 3,816 women. Of the cohort, 1,676, 2,893, and 5,036 patients completed b7, 7-9, or ≥10 METS, and 4,215 attained ≥10 METS without exercise-induced ST-segment shifts. Cohort characteristics are shown in Table I.

American Heart Journal May 2011

902 Lee et al

Table I. Patient characteristics

Age (y), mean (SD) Male, n (%) Symptoms Angina, n (%) Dyspnea, n (%) Chest pain NYD, n (%) Prior cardiovascular history MI, n (%) HF, n (%) PCI, n (%) CABG, n (%) Cardiac risk factors Hypertension, n (%) Hyperlipidemia, n (%) Diabetes, n (%) Cigarette smoking, n (%) Family history, n (%) 1 or 2 risk factors, n (%) ≥3 risk factors, n (%) Physical examination Systolic BP, mean (SD) Diastolic BP, mean (SD) Resting HR, mean (SD) Rest ECG Normal rhythm, n (%) Atrial fibrillation, n (%) Any Q-waves, n (%)

<7 METS, n = 1676

7-9 METS, n = 2893

≥10 METS, n = 5036

P

62.4 (12.5) 783 (46.7)

58.5 (11.5) 1422 (49.2)

49.4 (12.1) 3584 (71.2)

b.001 b.001

564 (33.7) 578 (34.5) 518 (30.9)

971 (33.6) 917 (31.7) 1101 (38.1)

1380 (27.4) 1156 (23.0) 2048 (40.7)

b.001 b.001 b.001

184 86 178 119

(11.0) (5.1) (10.6) (7.1)

213 44 260 153

(7.4) (1.5) (9.0) (5.3)

312 35 393 206

(6.2) (0.7) (7.8) (4.1)

b.001 b.001 .001 b.001

772 624 313 234 650 876 547

(46.1) (37.2) (18.7) (14.0) (38.8) (52.3) (32.6)

1308 1128 483 475 1340 1660 869

(45.2) (39.0) (16.7) (16.4) (46.3) (57.4) (30.0)

1394 1537 385 801 2320 2907 902

(27.7) (30.5) (7.6) (15.9) (46.1) (57.7) (17.9)

b.001 b.001 b.001 .078 b.001 b.001 b.001

132.1 (30.9) 80.3 (19.0) 77.1 (27.9)

131.6 (79.6) 80.7 (16.0) 73.8 (13.2)

123.9 (16.7) 79.9 (14.1) 71.9 (12.6)

b.001 .073 b.001

856 (51.1) 77 (4.6) 165 (9.8)

1429 (49.4) 60 (2.1) 213 (7.4)

2488 (49.4) 35 (0.7) 280 (5.6)

.461 b.001 b.001

NYD, not yet diagnosed.

Table II. MPI findings at different exercise treadmill workloads 7-9 METS, n = 2893

≥10 METS, n = 5036

(21.3) (78.7) (9.4) (5.2) (6.7)

750 2143 354 179 217

(25.9) (74.1) (12.2) (6.2) (7.5)

943 (18.7) 4093 (81.3) 550 (10.9) 184 (3.7) 209 (4.2)

708 3507 420 132 156

(16.8) (83.2) (10.0) (3.1) (3.7)

b.001 b.001 .013 b.001 b.001

(16.1) (83.9) (9.5) (3.7) (2.9)

518 2375 308 125 85

(17.9) (82.1) (10.6) (4.3) (2.9)

619 (12.3) 4417 (87.7) 447 (8.9) 124 (2.5) 48 (1.0)

447 (10.6) 3768 (89.4) 330 (7.8) 86 (2.0) 31 (0.7)

b.001 b.001 .035 b.001 b.001

<7 METS, n = 1676 Stress defect: SSS, n (%) N0 357 0 1319 1-3 158 4-6 87 ≥7 112 LV ischemia, n (%) N0% 270 0% 1406 1%-4% 160 5%-9% 62 ≥10% 48

Exercise MPI findings Exercise workload and corresponding MPI results are shown in Table II. Ischemia was present on perfusion imaging in 12.3% and 10.6% of those who attained ≥10 METS overall and without ST-segment changes, respectively. Large ischemic defects (≥10% LV ischemia) were present in 1% and 0.7% of those who attained ≥10 METS overall and without ST-segment changes, respectively. Of those who attained ≥10 METS overall and without ST-

≥10 METS + no ST change, n = 4215

P

segment changes, stress perfusion defects occurred in 18.7% and 16.8%, and large stress defects (SSS ≥7) were present in 4.2% and 3.7%, respectively.

Outcomes after MPI Mean follow-up was 3.64 ± 1.11 years, with a total of 35,007 person-years of follow-up examined. During follow-up, 290 (3.02%) patients died, 175 (1.82%) experienced MI, and 525 (5.47%) underwent coronary

American Heart Journal Volume 161, Number 5

Lee et al 903

Table III. Event rates per 100 person-years according to exercise workload and MPI findings b7 METS Death or MI SSS 0 2.67 1-3 2.01 ≥4 4.82 LV ischemia 0% 2.68 1%-4% 3.09 ≥5% 4.86 Death, MI, or late coronary revascularization (N90 d) SSS 0 3.94 1-3 4.16 ≥4 9.70 LV ischemia 0% 4.03 1-4% 6.15 ≥5% 9.78

7-9 METS

≥10 METS

1.14 1.14 2.40

0.67 0.89 1.59

1.13 1.87 2.56

0.69 1.33 1.30

1.87 3.29 8.46

1.10 1.59 3.66

2.06 5.09 9.69

1.20 2.14 2.97

Figure 1

Event-free survival probability

100%

0%

90%

1-4% 80%

5-9%

70% 60% 50%

≥10%

40% 30% 20% 10% 0% 0

500

1000

1500

2000

Time (days) Adjusted survival curves for the composite outcome by %LV ischemia.

artery revascularization. Table III shows the rates per person-time of the following endpoints: (a) death or MI and (b) late coronary revascularization (N90 days), death or MI stratified by exercise capacity and MPI results. Among those who exercised ≥10 METS with no stress perfusion defects (SSS = 0), the annualized mortality rate was 0.5%/year, and the rate of death or MI was 0.8% /year. The rates of death or MI increased with decreasing workload and higher stress defect score or greater %LV ischemic (Table III). The composite of late revascularization, death, or MI also increased with increasing SSS %LV ischemia, and lower workload attained. Adjusted curves

showing survival free of MI or coronary revascularization are shown (Figure 1).

Cox regression analysis Univariate predictors of the composite outcomes of (a) death, MI, or revascularization procedure and (b) death or MI are shown in Table IV. Multivariable-adjusted hazard ratios for increasing SSS are shown in Figure 2 for (a) death or MI and (b) death, MI, or revascularization, after adjustment for age, sex, workload attained, and presence of exercise-induced ST-segment shifts. Those with SSS ≥7 had a significantly increased risk of death or

American Heart Journal May 2011

904 Lee et al

Table IV. Univariate predictors of cardiovascular outcomes Death or MI

1.53

Referent 1.57 (1.19-2.07)† 2.06 (1.38-3.08)⁎ 2.58 (1.60-4.14)⁎ Referent 0.98 (0.71-1.36) 1.78 (1.23-2.55)† 2.57 (1.92-3.42)⁎

‡ P < .001

1.5

HR (95%CI)

HR (95%CI)

1.30

2.0 0.83

1.0 0.5

5.0

3.44 ‡

4.0 3.0

1.92 ‡

2.0

1.06

1.0

0.0 0

1-3 4-6

≥7

Summed Stress Score

0.0 0

1.23

1.5 1.0

5.0 4.0

2.41 ‡

3.0 1.52 ‡

2.0

0.5

1.0

0.0 0

1-4 5-9

≥10

% LV Ischemia

0.0 0

1-4 5-9 ≥10 % LV Ischemia

Death, MI, or late revascularization

(0.66-1.01) (1.31-2.08)⁎

6.0

1.57 †

2.0

Table V. Multivariable MPI predictors of death, MI, or late revascularization (N90 days)

7.0

2.5

1.46

(0.95-1.49) (0.37-0.73)⁎ (1.30-3.53)† (1.13-2.08)† (2.28-3.68)⁎ (2.42-5.35)⁎

Death, MI or Revascularization

† P < .01

4.95 ‡

‡ P < .001

Effect of %LV ischemia on risk of death or MI (left) and death, MI, or coronary revascularization (right).

Figure 2

3.0

6.0

Referent (0.36-0.57)⁎ (0.21-0.33)⁎

⁎ P ≤ .001. † P ≤ .01. ‡ Hazard ratio for rest LVEF missing vs ≤50% was 0.37 (95% CI 0.29-0.48, P b .001). § P ≤ .05.

Death or MI

2.5

HR (95%CI)

1.04 (1.04-1.05)⁎ 1.85 (1.50-2.29)⁎

0.46 0.27 1.19 0.52 2.14 1.53 2.90 3.60 0.81 1.65

7.0

3.0

Hazard ratio (95%CI) Age, per 10 y 1.05 (1.04-1.05)⁎ Male 2.49 (2.11-2.93)⁎ LV ischemia 0% Referent 1%-4% 2.15 (1.77-2.61)⁎ 5%-9% 3.81 (2.97-4.88)⁎ ≥10% 10.81 (8.60-13.58)⁎ SSS 0 Referent 1-3 1.34 (1.11-1.74)† 4-6 2.87 (2.26-3.65)⁎ ≥7 6.55 (5.52-7.77)⁎ Workload attained b7 METS Referent 7-9 METS 0.64 (0.54-0.75)⁎ ≥10 METS 0.31 (0.26-0.36)⁎ Exercise ST shift vs none 2.33 (2.02-2.69)⁎ Rest LVEF N50% vs ≤50%‡ 0.53 (0.42-0.67)⁎ Atrial fibrillation on rest ECG 1.52 (1.00-2.32)§ Q-waves on rest ECG 2.10 (1.71-2.56)⁎ Prior MI 2.69 (2.25-3.22)⁎ Prior HF 2.84 (2.06-3.92)⁎ Angina 1.26 (1.10-1.46)⁎ Prior revascularization 2.19 (1.87-2.56)⁎ procedure (PCI or CABG)

Death, MI or Revascularization

Death or MI

HR (95%CI)

Death, MI, or revascularization

Figure 3

1-3 4-6

≥7

Summed Stress Score

Effect of SSS on risk of death or MI (left) and death, MI, or coronary revascularization (right).

MI, whereas those with SSS ≥4 had increased risk of the composite outcome including coronary revascularization. The corresponding multivariable-adjusted hazard ratios for %LV ischemia are shown in Figure 3. In multivariable-adjusted analysis, there was a significant

Hazard ratio (95%CI) Stress defect Age Male SSS 0 1-3 4-6 ≥7 Workload attained b7 METS 7-9 METS ≥10 METS Exercise ST shift vs none %LV ischemia Age Male LV ischemia 0% 1%-4% 5%-9% ≥10% Workload attained b7 METS 7-9 METS ≥10 METS Exercise ST shift vs none

1.02 (1.02-1.03)⁎ 2.41 (2.01-2.89)⁎ Referent 0.98 (0.77-1.25) 1.82 (1.40-2.37)⁎ 2.65 (2.14-3.27)⁎ Referent 0.60 (0.50-0.71)⁎ 0.35 (0.28-0.43)⁎ 1.65 (1.40-1.95)⁎ 1.03 (1.02-1.03)⁎ 2.52 (2.10-3.01)⁎ Referent 1.40 (1.13-1.73)† 2.07 (1.56-2.74)⁎ 3.03 (2.21-4.16)⁎ Referent 0.61 (0.51-0.73)⁎ 0.35 (0.29-0.43)⁎ 1.62 (1.37-1.91)⁎

⁎ P ≤ .001. † P ≤ .01.

relationship between the composite outcome of death, MI, or revascularization with %LV ischemia (Table V). Higher SSS was also a significant predictor of late coronary revascularization events, death, or MI, after

American Heart Journal Volume 161, Number 5

multivariable adjustment, although the magnitude of the association was reduced compared with %LV ischemia (Table V). The model χ2 for the outcome of death, MI, or revascularization, accounting for age, sex, workload attained, and ST-segment shift, was 607.88. The addition of %LV ischemia to this model increased the model χ2 to 756.35 (P b .001), and the addition of SSS increased the model χ2 to 765.91 (P b .001). The corresponding model χ2 for the outcome of death or MI was 240.23. The addition of %LV ischemia to this model tended to increase the χ2 to 246.57 (P = .096), and the addition of SSS significantly increased the model χ2 to 251.87 (P = .009).

Sensitivity analysis After further adjustment for age, sex, workload attained, exercise ST shift, cardiac risk factors, prior revascularization, systolic BP, HR, angina, presence of Qwaves on resting ECG, atrial fibrillation, and LV ejection fraction, the adjusted hazard ratios for the composite outcome of death, MI, or revascularization remained more pronounced for LV ischemia than the SSS. The adjusted hazard ratios for 1% to 4%, 5% to 9%, and ≥10% LV ischemia were 1.55 (95% CI 1.25-1.93), 2.42 (95% CI 1.84-3.17), and 4.85 (95% CI 3.73-6.32) compared with no ischemia, respectively (all P b .001). The adjusted hazard ratios for SSS 1 to 3, 4 to 6, and ≥7 were 1.19 (95% CI 0.92-1.53, P = .190), 2.09 (95% CI 1.59-2.74, P b .001), and 3.60 (95% CI 2.85-4.56, P b .001) compared with an SSS of 0, respectively. The association with death or MI continued to be more significant with SSS ≥7 after multivariable adjustment for the above-mentioned factors. Hazard ratios for SSS 1 to 3, 4 to 6, and ≥7 were 0.85 (95% CI 0.60-1.21, P = .366), 1.31 (95% CI 0.88-1.94, P = .186), and 1.47 (95% CI 1.022.12, P = .040), respectively. :Left ventricular ischemia was associated with a trend toward increased death or MI with hazard ratios of 1.25 (95% CI 0.92-1.70, P = .162), 1.48 (95% CI 0.97-2.26, P = .072), and 1.60 (95% CI 0.972.66, P = .068) for 1% to 4%, 5% to 9%, and ≥10% ischemia, respectively.

Discussion An important role of noninvasive testing for coronary heart disease is the evaluation of patient prognosis and outcomes. A noninvasive test that is useful for both diagnosis and prognosis would impact subsequent processes of care (eg, coronary revascularization) and demonstrate an association with clinical events (eg, death or MI). In this study, we examined an array of downstream outcomes among patients who underwent exercise-stress MPI. The extent of myocardial ischemia was a major predictor of the composite outcomes of death, MI, and late coronary revascularization, and it was predictive at all exercise treadmill workloads attained.

Lee et al 905

When the extent of LV ischemia was 1 to 4%, 5 to 9%, or ≥10%, the risk of death, MI, or revascularization was 1.5-, 2.4-, and 4.9-fold higher than those with no ischemia. The presence of stress perfusion defects was associated with greater risk of death or MI, with a 57% increase in risk when the SSS was ≥7. Prior studies have found that exercise capacity is a more powerful predictor of mortality than cardiac risk factor profile,11,12 and each 1-MET increase conferred a 12% reduction in mortality risk.13 Our study findings confirm the prognostic impact of exercise workload, because it was a substantial predictor of all outcomes including death, MI, and coronary revascularization. However, our findings extend the literature by the observation that a substantial proportion of patients (4.2%) can attain ≥10 METS despite the presence of large stress perfusion defects (SSS ≥7), which were also independently associated with a statistically significant increase in risk of death or MI. Prior studies have demonstrated that MPI has incremental value beyond exercise treadmill parameters.14-18 One prior study of 388 patients reaching stage IV of the Bruce protocol found that an abnormal MPI was more predictive of cardiac events and death than the stress ECG.19 Hachamovitch et al20 reported that quantitation of defect size is important for prognostication and, in particular, suggested that presence of ≥10% LV ischemia defined the crossover between the benefit of revascularization versus medical therapy.21 However, few studies have comprehensively quantified the implications of MPI findings over the range of potential outcomes and exercise capacity. Our study was consistent with a recently published study which demonstrated that achievement of ≥10 METS without ST-segment shift conferred a low frequency of major LV ischemia ≥10%2 and excellent prognosis over an intermediate follow-up duration.22 We found that only 0.7% of patients who achieved ≥10 METS without ST-segment shift had a large ischemic defect that was ≥10%. Our study extended upon these prior reports by examining a broad range of fatal and nonfatal outcomes in a patient sample that was nearly 10-fold that of the aforementioned study,2 with complete follow-up for these events in the full study cohort using linked population-based databases. In a large patient sample, we considered death, MI, and late coronary revascularization procedures that occurred more than 90 days after MPI because early revascularization may be driven by MPI results. We found that the composite of death, MI, and late revascularization events were significantly increased in the presence of greater %LV ischemia, attesting to the sustained prognostic implications of MPI findings over time. It is important to note that the SDS reflects ischemic burden,23,24 and thus, we did not include cardiac catheterization without revascularization as an outcome, because it is not directly associated with the presence of angiographically significant coronary disease.

American Heart Journal May 2011

906 Lee et al

There are several implications of our study. Our study confirms the utility of exercise ECG when patients are able to achieve a high workload ≥10 METS without STsegment shifts, because in such cases, the presence of a large reversible defect size of ≥10% is infrequent. Thus, in many cases, a simple exercise treadmill test may be sufficient to exclude large ischemic defects, which may be amenable to revascularization. However, attainment of high workload in itself is not sufficient to rule out stress perfusion defects that may impact upon prognosis, or lesser degrees of reversible ischemia that may correlate with symptoms. Exercise myocardial perfusion abnormalities portend increased risk of death or MI and are upstream findings that alter patient care by identifying candidates for coronary revascularization. Finally, although exercise-stress MPI has value beyond that of exercise electrocardiography alone and provides an array of prognostic information, it entails radiation exposure and should be performed after considering the absolute benefit of study findings and the role of information provided by MPI in changing patient management. There were some notable limitations to our study. This was a single-center evaluation of a teaching hospital– based nuclear cardiology laboratory that conducts singlephoton emission computed tomography–MPI studies on a wide range of patients of varying complexity. However, we anticipate that our findings would be generalizable given the high reproducibility and low user-dependent variability of MPI. Furthermore, although MPI was performed at a single center to ensure study homogeneity, outcome events and revascularization procedures could have occurred in distant geographical locations, because we tracked outcomes that occurred throughout the Ontario provincial health care system. In conclusion, MPI findings had significant implications for morbid cardiac and fatal outcomes at all levels of treadmill exercise even after accounting for clinical and exercise parameters. At all levels of exercise workload attained, MPI identified stress perfusion defects, which were associated with an increased risk of death or MI when large stress defects were present. Although large ischemic defects were infrequent at workloads greater than 10 METS in the absence of ST-segment shifts, reversible perfusion defects were strongly associated with death, MI, and downstream coronary revascularization procedures.

Disclosures The Institute for Clinical Evaluative Sciences is supported in part by a grant from the Ontario Ministry of Health and Long Term Care. The opinions, results, and conclusions are those of the authors, and no endorsement by the Ministry of Health and Long-Term Care or by the Institute for Clinical Evaluative Sciences is intended or should be inferred. This research was supported by a

Canadian Institutes of Health Research clinician-scientist award (D.S.L.). The authors are solely responsible for the design and conduct of this study, all study analyses, drafting, and editing of the paper and its final contents. Dr Iwanochko has received speaker's fees from Lantheus. All other authors have no conflicts of interest to declare.

References 1. Lloyd-Jones D, Adams R, Carnethon M, et al. Heart disease and stroke statistics–2009 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2009;119:e21-e181. 2. Bourque JM, Holland BH, Watson DD, et al. Achieving an exercise workload of N or =10 metabolic equivalents predicts a very low risk of inducible ischemia: does myocardial perfusion imaging have a role? J Am Coll Cardiol 2009;54:538-45. 3. Bouzas-Mosquera A, Peteiro J, Alvarez-Garcia N. Value of a high exercise workload to rule out myocardial ischemia. J Am Coll Cardiol 2010;55:265-6. 4. Patel MR, Peterson ED, Dai D, et al. Low diagnostic yield of elective coronary angiography. N Engl J Med 2010;362:886-95. 5. Klocke FJ, Baird MG, Lorell BH, et al. ACC/AHA/ASNC guidelines for the clinical use of cardiac radionuclide imaging–executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (ACC/AHA/ ASNC Committee to Revise the 1995 Guidelines for the Clinical Use of Cardiac Radionuclide Imaging). Circulation 2003;108:1404-18. 6. Ghali WA, Rothwell DM, Quan H, et al. A Canadian comparison of data sources for coronary artery bypass surgery outcome “report cards.” Am Heart J 2000;140:402-8. 7. Tu JV, Austin PC, Walld R, et al. Development and validation of the Ontario acute myocardial infarction mortality prediction rules. J Am Coll Cardiol 2001;37:992-7. 8. Austin PC, Daly PA, Tu JV. A multicenter study of the coding accuracy of hospital discharge administrative data for patients admitted to cardiac care units in Ontario. Am Heart J 2002;144:290-6. 9. Lee DS, Donovan L, Austin PC, et al. Comparison of coding of heart failure and comorbidities in administrative and clinical data for use in outcomes research. Med Care 2005;43:182-8. 10. Ghali WA, Quan H, Brant R, et al. Comparison of 2 methods for calculating adjusted survival curves from proportional hazards models. JAMA 2001;286:1494-7. 11. Peterson PN, Magid DJ, Ross C, et al. Association of exercise capacity on treadmill with future cardiac events in patients referred for exercise testing. Arch Intern Med 2008;168:174-9. 12. Snader CE, Marwick TH, Pashkow FJ, et al. Importance of estimated functional capacity as a predictor of all-cause mortality among patients referred for exercise thallium single-photon emission computed tomography: report of 3,400 patients from a single center. J Am Coll Cardiol 1997;30:641-8. 13. Myers J, Prakash M, Froelicher V, et al. Exercise capacity and mortality among men referred for exercise testing. N Engl J Med 2002;346:793-801. 14. Shaw LJ, Hachamovitch R, Peterson ED, et al. Using an outcomesbased approach to identify candidates for risk stratification after exercise treadmill testing. J Gen Intern Med 1999;14:1-9. 15. Berman DS, Hachamovitch R, Kiat H, et al. Incremental value of prognostic testing in patients with known or suspected ischemic heart disease: a basis for optimal utilization of exercise technetium-99m

American Heart Journal Volume 161, Number 5

16.

17.

18.

19.

20.

sestamibi myocardial perfusion single-photon emission computed tomography. J Am Coll Cardiol 1995;26:639-47. Hachamovitch R, Berman DS, Kiat H, et al. Value of stress myocardial perfusion single photon emission computed tomography in patients with normal resting electrocardiograms: an evaluation of incremental prognostic value and cost-effectiveness. Circulation 2002;105:823-9. Iskandrian AS, Chae SC, Heo J, et al. Independent and incremental prognostic value of exercise single-photon emission computed tomographic (SPECT) thallium imaging in coronary artery disease. J Am Coll Cardiol 1993;22:665-70. Elhendy A, Schinkel A, Bax JJ, et al. Long-term prognosis after a normal exercise stress Tc-99 m sestamibi SPECT study. J Nucl Cardiol 2003;10:261-6. Chatziioannou SN, Moore WH, Ford PV, et al. Prognostic value of myocardial perfusion imaging in patients with high exercise tolerance. Circulation 1999;99:867-72. Hachamovitch R, Hayes SW, Friedman JD, et al. A prognostic score for prediction of cardiac mortality risk after adenosine stress

Lee et al 907

21.

22.

23.

24.

myocardial perfusion scintigraphy. J Am Coll Cardiol 2005;45: 722-9. Hachamovitch R, Hayes SW, Friedman JD, et al. Comparison of the short-term survival benefit associated with revascularization compared with medical therapy in patients with no prior coronary artery disease undergoing stress myocardial perfusion single photon emission computed tomography. Circulation 2003;107:2900-7. Bourque JM, Charlton GT, Holland BH, et al. Prognosis in patients achieving ≥10 METS on exercise stress testing: was SPECT imaging useful? J Nucl Cardiol 2010. Gimelli A, Rossi G, Landi P, et al. Stress/rest myocardial perfusion abnormalities by gated spect: still the best predictor of cardiac events in stable ischemic heart disease. J Nucl Med 2009;50:546-53. Hachamovitch R, Berman DS, Shaw LJ, et al. Incremental prognostic value of myocardial perfusion single photon emission computed tomography for the prediction of cardiac death: differential stratification for risk of cardiac death and myocardial infarction. Circulation 1998;97:535-43.