Share this article on:

00005768-201412000-0000400005768_2014_46_2216_martin_prognostication_12article< 106_0_29_5 >Medicine & Science in Sports & Exercise© 2014 American College of Sports MedicineVolume 46(12)December 2014p 2216–2223Arm Exercise Myocardial Perfusion Imaging for Prognostication of Long-Term Outcome[CLINICAL SCIENCES]MARTIN, WADE H. III1; XIAN, HONG2; WAGNER, DANIEL L.3; CHANDIRAMANI, POOJA4; BAINTER, EMILY5; ILIAS-KHAN, NASREEN51Division of Cardiology, Department of Internal Medicine, St. Louis Veterans Administration Medical Center and Washington University School of Medicine, St. Louis, MO; 2Department of Biostatistics, College for Public Health and Social Justice, St. Louis University, St. Louis, MO; 3Division of Cardiology, Department of Internal Medicine, University of Maryland School of Medicine, Baltimore, MD; 4Department of Epidemiology, College for Public Health and Social Justice, St. Louis University, St. Louis, MO; and 5Department of Internal Medicine, Washington University School of Medicine, St. Louis, MOAddress for correspondence: Wade H. Martin III, M.D., F.A.C.S.M., Division of Cardiology 111A/JC, St. Louis Veterans Administration Medical Center, 915 North Grand, St. Louis, MO 63106; E-mail: Wade.Martin@va.gov .Submitted for publication February 2014.Accepted for publication March 2014.ABSTRACTIntroduction: Pharmacologic evaluations constitute ≥50% of imaging stress tests, but exercise reduces adverse effects, improves myocardial perfusion imaging (MPI) quality and diagnostic results, and provides powerful prognostic and clinically important information on exercise capacity and cardiovascular responses to the relevant physiologic stress of exercise. Thus, our purpose was to determine whether arm exercise and MPI variables predict long-term outcome in patients who cannot perform leg exercise.Methods: We performed arm exercise MPI stress tests in 253 consecutive patients age 64.5 (10.7) yr (mean (SD)) from 1997 to 2002 and investigated associations of arm exercise and abnormal MPI variables with all-cause mortality, myocardial infarction (MI), coronary artery bypass grafting (CABG), and percutaneous coronary intervention (PCI) during follow-up of 12.0 (1.3) yr.Results: There were 156 deaths (61.7%), 47 patients suffered MI (18.6%), 24 underwent CABG (9.5%), and 50 had PCI (19.8%). Arm exercise capacity and delta HR (peak − resting) were strongly associated with survival after adjustment for significant demographic and clinical variables (Cox multivariate P < 0.0001 and 0.001, respectively). MPI was abnormal in 157 patients (62.1%). An abnormal arm exercise MPI was borderline predictive of mortality by Cox analysis (71.8% vs 46.4% for normal study; univariate P < 0.0001; multivariate P = 0.07) but resulted in 58% relative incremental integrated discrimination improvement over clinical variables for predicting death. Perfusion defect size also strongly predicted mortality (Cox multivariate P = 0.003). An abnormal arm exercise MPI study, perfusion defect type, and size all prognosticated PCI (all P ≤ 0.03) but not MI or CABG.Conclusions: Arm exercise MPI is a valuable approach for outcome prediction in patients unable to perform leg exercise.Although treadmill exercise capacity and cardiovascular responses are powerful predictors of all-cause and cardiovascular mortality (4,6,11,12,15), the number of patients who cannot perform lower extremity exercise because of claudication, amputations, orthopedic, or other disabilities has been increasing steadily for many years and now exceeds 50% of all stress test referrals (17). Pharmacologic testing has become the method of choice for evaluation of cardiac ischemia in these individuals but may elicit adverse reactions, including myocardial infarction (MI) and death, and result in lesser image quality when performed without exercise (19,22). Pharmacologic testing also does not provide powerful prognostic or clinically useful information on exercise capacity, symptomatic, cardiovascular, or ECG responses to the relevant physiologic stress of exercise. The number and extent of ischemic perfusion defects and frequency of transient ischemic dilatation are reported to be higher with arm exercise plus pharmacologic stress than those with pharmacologic evaluation alone (19). Arm exercise myocardial perfusion imaging (MPI) results are also predictive of anatomic findings at cardiac catheterization (3,8). Although arm exercise elicits approximately 40% lower peak oxygen uptake than treadmill exercise, blood pressure responses at similar oxygen uptake are greater (2,5,13), possibly resulting in more severe myocardial ischemia relative to pharmacologic MPI. We have recently shown that arm exercise capacity and HR responses during and after exercise are robust independent predictors of all-cause and cardiovascular mortality over 12 yr in a population of veterans unable to perform leg exercise (14). Despite these results and the large number of patients with disabilities precluding lower extremity stress testing, arm exercise is currently not widely used in clinical practice, the number of patients in previous arm exercise MPI studies was small, and there are no long-term data for prognostication of clinical outcome after arm exercise MPI stress testing. Thus, the purpose of this investigation was to determine whether arm exercise MPI findings are predictive of all-cause mortality, MI, or coronary revascularization over a follow-up of ≥10 yr.METHODSPatient characteristicsA total of 253 consecutive patients (249 men, four women) age 64.4 (10.7) yr (mean (SD)) completed an individualized arm ergometer exercise testing protocol with same-day rest and subsequent stress single-photon emission computed tomographic imaging (20) at the St. Louis Veterans Administration Medical Center (VAMC) between 1997 and 2002. Participants were referred for MPI stress testing by their primary VA health care provider, were unable to perform treadmill or leg cycle ergometer exercise, and were willing to complete the arm exercise protocol. There were no exclusions, but those having upper extremity disabilities or medical conditions for which stress testing is contraindicated were not evaluated. Patients reported to the laboratory after an overnight fast. Those with diabetes were instructed to take one-half of their usual insulin dose, and all patients were instructed to withhold beta adrenergic blocking agents on the morning of the test. Otherwise, all medications were consumed as prescribed. Preexisting clinical diagnoses and procedures were identified and documented by a brief history and physical examination and review of VA electronic medical records. The study was approved by the institutional review board of the St. Louis VAMC. All participants provided a voluntary written informed consent.MPI protocolBaseline single-photon emission computed tomographic imaging was performed in the supine posture 30 min after intravenous administration of 8–10 mCi of Tc-99m sestamibi. Once baseline images had been acquired, arm ergometer stress testing was conducted and 22–30 mCi of Tc-99m sestamibi was injected 30–60 s before attainment of peak exercise. Baseline and stress images were acquired, processed, displayed, and interpreted as described previously (20). Studies were interpreted independently by VA board-certified nuclear medicine physicians who were aware of arm exercise stress testing results but were not involved in subsequent patient management. Images were determined to be normal or abnormal, and perfusion defects were further characterized as large, moderate sized, or small and fixed, reversible, or mixed on the basis of semi-quantitative visual interpretation. Because defect severity was not determined in most patients between 1997 and 2002, we did not calculate summed rest, stress, or difference scores. For the 20-segment left ventricular model, small defects were defined as having 1–3 abnormal myocardial segments and moderate-sized and large defects were defined as having 4–6 and ≥7 abnormal segments, respectively. Individual defects were defined as reversible if ≥75% of abnormal stress segments were normal at baseline and fixed if ≥75% of abnormal stress and baseline myocardial segments were concordant. Intermediate defects were considered mixed. Defect number was determined on the basis of the presence of perfusion abnormalities in distinct anatomic coronary territories. Total myocardial defect size was based on summation of separate perfusion defects. Transient ischemic dilatation was defined as a 23%–24% stress-induced increase in left ventricular cavity volume by AutoQuant processing (1).Arm exercise protocolPatients exercised in the seated posture on a wall-mounted, electronically braked cycle ergometer (Angio 2000; Lode BV, Groningen, The Netherlands) at a target cadence of 60 rpm to an end point of fatigue or symptom limitation. A progressive multistage protocol, designed to elicit exhaustion within 5–12 min, was used, with constant work increments of 50–200 kpm·min−1 every 2 min, depending on pretest-estimated exercise capacity, and developed from earlier investigations (13), as described previously (9). Additional pilot testing at the St. Louis VAMC was performed with respiratory gas analysis during arm ergometer and treadmill exercise in five young healthy subjects (four men, one woman) age 28 (7) yr before initiation of the study protocol. These results demonstrated that a lower (vs treadmill exercise) but still adequate peak HR was obtainable with arm exercise (168 (16) bpm (88% age-predicted maximum) vs 184 (10) bpm for treadmill testing (96% age-predicted maximum)). Peak oxygen uptake and RER for the two exercise modalities were 25.8 (6.1) versus 46.1 (1.8) mL·kg−1·min−1 and 1.26 (0.05) versus 1.28 (0.08), respectively, consistent with both the extremely limited literature data (2) and attainment of peak effort with arm exercise. HR was determined from the ECG, and blood pressure was determined from manual sphygmomanometry, as described (9). Exercise ECG interpretations were performed using standard criteria (5,9).Outcome and participant dataOccurrence and date of death, MI, and coronary revascularization by either coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI), demographic, medication, and resting ECG data, and clinical diagnoses of study participants were determined by review of all VA electronic medical records that preceded and followed stress testing until documented mortality or November 5, 2012. This information was supplemented by scanned reports of non-VA episodes of care and data from the Missouri Department of Health and Senior Services and Social Security Death Index. Mortality and ascertainment reliability of these records are comparable with those of the National Death Index (7). Research staff had no knowledge of stress test findings at the time of record review. The criteria for MI were elevated creatine kinase-MB or troponin above the 99th percentile and either ischemic ECG changes or symptoms. Patients were not censored after MI or coronary revascularization.Data analysisUnivariate analyses were performed with Student’s t-tests to evaluate differences among group means. Chi-square statistics were used to examine associations among categorical variables. Multivariate analyses were performed with logistic and Cox regression models to identify significant relations for associations of demographic, clinical, exercise, MPI, and other variables with mortality, MI, and coronary revascularization. Significant variables by univariate analysis were entered into multivariate models. Kaplan–Meier curves were generated to compare survival and other outcome events among patient groups. The incremental value of MPI results over demographic and clinical variables was also calculated by relative integrated discrimination improvement analysis. Differences were considered statistically significant at P < 0.05. All analyses were performed with SAS 9.3 (18).RESULTSParticipant characteristicsAs illustrated in Table 1, our population included a substantial proportion of older individuals having multiple coronary risk factors and major comorbidities. By multivariate analysis, age, dyslipidemia, history of congestive heart failure, atrial fibrillation, peripheral arterial disease, chronic obstructive pulmonary disease, and cancer were significantly associated with mortality (all P ≤ 0.01).TABLE 1 Demographic, resting hemodynamic and ECG findings, medications, and clinical characteristics (mean (SD)).OutcomeThe average follow-up interval for survivors in our population was 12.0 (1.3) and 7.9 (4.3) yr for the entire cohort. There were 156 deaths (61.7%), 47 patients suffered MI (18.6%), 24 underwent CABG (9.5%), and PCI was performed in 50 patients (19.8%). Combined events occurred in 194 individuals (76.7%). The average annual event rate was 7.8% for death, 2.4% for MI, 1.2% for CABG, 2.6% for PCI, and 9.8% for combined events.Arm exercise test resultsArm exercise MPI stress tests were requested to evaluate chest pain (46%), perioperative risk (29%), CAD (10%), functional capacity (7%), dyspnea (6%), and for various other reasons (2%). Treadmill exercise could not be performed because of orthopedic conditions (34%), spinal cord and lower extremity neurologic or myopathic abnormalities (24%), severe claudication or foot ulcers (20%), one or both lower extremity amputations (5%), and various other conditions (mostly severe chronic obstructive pulmonary disease or morbid obesity) in 17%. The end point of arm exercise was fatigue (75%), severe dyspnea (15%), angina (2%), and other signs or symptoms (8%).Arm exercise responses are shown in Table 2. Arm exercise capacity and delta HR were greater in survivors than those in decedents (multivariate P < 0.0001 and 0.001, respectively). Peak HR was also higher in survivors by univariate analysis (P < 0.0001) but did not retain multivariate significance (P = 0.34), whereas peak systolic and diastolic blood pressures were not different in the two groups. There were no significant differences between groups in occurrence of nonlimiting angina (6.2% in survivors vs 7.7% in nonsurvivors) or limiting angina (2.1% vs 2.6%). Arm exercise-induced ST segment depression of at least 1 mm was observed in 16.0% of survivors and 21.8% of decedents, a difference that did not achieve statistical significance by chi-square or Kaplan–Meier survival evaluations.TABLE 2 Arm exercise and MPI data.Prediction of MI and coronary revascularization.Variables predictive of MI by univariate analysis were history of hypertension (P = 0.04), an arm exercise end point of angina (P < 0.05), and peak systolic blood pressure (P = 0.02), but none of these retained multivariate significance. Coronary revascularization by either CABG or PCI was prognosticated by dyslipidemia (P = 0.03), previous PCI (P < 0.01), nonlimiting angina (P = 0.02), and arm exercise-induced ST segment depression (P = 0.004) by univariate evaluation, but none of these remained significant with multivariate analysis.MPI resultsMPI data are shown in Table 2. Imaging studies were abnormal in 157 patients (62.1%). Kaplan–Meier survival plots for normal and abnormal MPI results are shown in Figure 1. An abnormal MPI result prognosticated mortality by univariate analysis (71.8% abnormal vs 46.4% for normal study, P < 0.0001) but was only borderline predictive (P = 0.07) after inclusion of demographic (age) and clinical variables in a Cox multivariate model. However, in a net reclassification model, an abnormal MPI evaluation resulted in absolute and relative incremental integrated discrimination improvements of 0.14 and 58%, respectively, over age and clinical variables for prediction of death.FIGURE 1. Kaplan–Meier plots of normal and abnormal arm exercise MPI studies in relation to survival.Additional Kaplan–Meier survival curves were generated to illustrate effects on mortality of specific MPI abnormalities, categorized by defect number (none, 0, 1, 2, or 3), predominant defect type (none, reversible, mixed, or fixed), and total left ventricular defect size (none, small, moderate, or large), as shown in Figure 2A–C. A bar graph depicting the association of various MPI defect characteristics with death is illustrated in Figure 3. Mortality ranged from 46.4% in the 37.9% of participants with no perfusion defect to 88.9% for the 3.6% of patients with three defects (univariate P < 0.0001 for effect of defect number), 74.2% and 76.3% for the 15.2% and 35.5%, respectively, of individuals with predominately mixed and fixed defects (univariate P = 0.0002 for defect type), and 82.9% for the 27.8% of participants with a large defect (univariate P < 0.0001 for defect size). Absolute and relative incremental integrated discrimination improvements for prediction of death ranged from 0.15 to 0.16 and 61% to 63%, respectively, for defect number, type, and size after inclusion of age and clinical variables. Transient ischemic dilatation was observed in five patients (2.0%), all of whom died, in comparison with 60.7% mortality in those without this abnormality (P = 0.07). After adjustment for age and clinical variables, defect size remained predictive of mortality by Cox analysis (P = 0.003).FIGURE 2. Kaplan–Meier plots of arm exercise MPI defect number (A), type (B), and size (C) in relation to survival.FIGURE 3. The relation of arm exercise MPI defect number, type, and size with mortality. Statistically significant differences among various groups are denoted by horizontal brackets, with P values for intergroup differences shown above the brackets. The number of patients in various subgroups is delineated below the defect characteristic descriptions.An abnormal arm exercise MPI study did not predict MI or CABG but was associated with greater likelihood of PCI (24.2% vs 12.5% with normal MPI, P = 0.02). Defect type and size also prognosticated PCI (both P ≤ 0.03).Interactions of arm exercise variables with MPI resultsArm exercise-induced ST segment depression was associated with an abnormal MPI study (78.7% abnormal MPI with vs 55.7% without ST segment change, P = 0.004) as well as perfusion defect number (P < 0.01), type (P < 0.01), and size (P < 0.0001). A normal MPI evaluation was associated with stress-induced ST segment depression in 10.4% of participants versus 25.5% in those having an abnormal MPI study (P = 0.004). For prediction of mortality, an abnormal MPI study had a higher sensitivity (71.3% vs 21.8%, P < 0.0001) and lower specificity (53.6% vs 84.0%, P < 0.0001) than an abnormal arm exercise ECG. A normal MPI evaluation had greater negative predictive value (54.2% vs 40.7%, P < 0.01), whereas the positive predictive value of an abnormal study for death was similar for the two modalities (71.3% vs 68.1%, P = 0.38).There was a significant trend for a lower percentage of abnormal MPI studies with increasing arm exercise capacity (69.2% lowest vs 48.7% highest tertile of exercise capacity, P = 0.01). For patients with an abnormal arm MPI study, there was also a gradient of mortality from highest to lowest exercise capacity (48.7% highest vs 90.8% lowest exercise capacity tertile, P < 0.0001). Nevertheless, an abnormal arm exercise MPI evaluation was still associated with higher mortality for participants in the top (48.7% abnormal vs 20.5% normal MPI study, P = 0.01) but not the intermediate or lowest tertile of exercise capacity (64.8% vs 51.9% and 90.8% vs 75.9%, respectively).Decreasing tertiles of peak HR were associated with a higher percentage of participants with an abnormal MPI study (52.3% for highest, 58.3% for middle, and 75.9% for the lowest tertile; P = 0.005), whereas this trend was not significant for an abnormal arm exercise ECG (24.1%, 21.5%, and 12.7% for highest-, middle-, and lowest-peak HR tertiles, respectively; P = 0.16). However, a higher delta HR tertile was associated with greater frequency of an abnormal arm exercise ECG (7.6% for lowest, 21.8% for middle, and 28.6% for highest tertile; P = 0.003). For the top and middle but not the lowest tertile of peak HR, mortality was increased in association with an abnormal MPI evaluation (60.0% vs 31.7%, P = 0.009; 71.4% vs 48.6%, P = 0.03; and 79.4% vs 70.0%, P = 0.39, respectively) but peak HR tertile did not interact significantly with the effect of an abnormal arm exercise ECG on mortality.Effects of beta blockadeThere was a residual effect of beta adrenergic blocking agents on the day of stress testing that was associated with lower resting and peak exercise HR (71 (16) vs 75 (16) bpm and 115 (24) vs 127 (20) bpm; P < 0.05 and P < 0.001, respectively). However, there was no significant difference in frequency of arm exercise-induced ST segment depression or MPI abnormalities for patients who were and were not being treated with beta blocking agents.Outcome interactionsMortality of individuals undergoing PCI was lower than that of those who did not have this procedure (46.0% vs 65.5%, P = 0.01). Mortality was also reduced for those undergoing either CABG or PCI (47.8% vs 66.7%, P < 0.01) but not for CABG alone (45.8% vs 63.3%, P = 0.09). Survival of patients with an abnormal arm exercise MPI evaluation was 48.9% for those undergoing subsequent CABG or PCI versus 20.0% for those without coronary revascularization (P = 0.0002). In contrast, for patients with a normal MPI study, survival with and without CABG or PCI was similar (60.0% vs 52.6%, P = 0.56). Survival of patients with stress-induced ST segment depression who underwent CABG or PCI also exceeded that for those who did not undergo coronary revascularization (55.0% vs 14.8%, P = 0.004), whereas survival with and without coronary revascularization did not differ significantly in the absence of arm exercise-induced ST segment depression (51.2% vs 37.8%, P = 0.11)DISCUSSIONOur results are the first to demonstrate that arm exercise MPI stress test findings are predictive of long-term all-cause mortality and stratification for coronary revascularization in patients conventionally referred for pharmacologic evaluations because of disabilities precluding lower extremity exercise. More than 50% of patients now undergo pharmacologic stress tests in large civilian referral centers (17). Such patients were observed to have a much higher all-cause mortality and greater prevalence of perfusion defects than those undergoing treadmill MPI evaluations (17). The 7.8% annual all-cause mortality observed in our veteran population is even higher than the 5.9% annual all-cause mortality reported by Rozanski et al. (17) for patients undergoing pharmacologic MPI tests during a similar period and follow-up interval as our cohort. This is likely due to the elevated burden of conditions such as congestive heart failure, peripheral arterial disease, atrial fibrillation, and chronic obstructive pulmonary disease associated with increased mortality in our population. Despite these high-risk conditions, arm exercise MPI testing was found to be a feasible and valuable alternative to pharmacologic stress testing.Previous studiesThe high prevalence of perfusion defects in our cohort is consistent with the finding of Rozanski et al. (17) that imaging abnormalities are more than twice as frequent in individuals undergoing pharmacologic studies as in those performing treadmill exercise. However, perfusion defects were found in more than 60% of our population in comparison with 23% of patients undergoing pharmacologic MPI evaluations at Cedars-Sinai during the same 1998–2000 epoch (17). Besides the higher prevalence of comorbidities in our cohort, another explanation for this discrepancy is that the number and extent of perfusion defects may be greater with arm exercise than with pharmacologic MPI stress testing alone, as suggested by the findings of Stein et al. (19). However, we did not observe the very high frequency of transient ischemic dilatation reported in the latter study and the prevalence of transient ischemic dilatation in our cohort was similar to that in the population of Rozanski et al. (17).Importance of exerciseAn additional major limitation of pharmacologic stress is its failure to provide powerful prognostic and clinically important information conferred by exercise. In our cohort, arm exercise capacity and delta HR were more strongly related to survival than overall MPI results and recovery HR also independently prognosticates survival (6,14). In addition, arm exercise responses yielded valuable complementary and interactive information for more comprehensive interpretation of the clinical context of MPI data. These findings further emphasize the potential role of arm exercise as a premier and less expensive approach than pharmacologic stress for optimal evaluation of the cardiac status of patients with adequate upper extremity function who cannot perform leg exercise.SensitivitySensitivity of ECG and MPI findings for detection of myocardial ischemia is generally thought to be compromised for exercise tests with attainment of ≤85% age-predicted peak HR (5). Although this may be true for ST segment depression, the percentage of patients in our cohort having abnormal MPI studies was greater for the lowest compared with that for the highest tertile of peak HR. These findings provide evidence that arm exercise MPI sensitivity is not seriously impaired by a blunted peak HR. In addition and in contrast to the reported effect of beta blockade on pharmacologic MPI stress test sensitivity (21), there was no significant association of beta blocker treatment of our patients with reduced percentage of abnormal MPI evaluations or perfusion defect findings despite an association with lower resting and peak HR. Thus, sensitivity of arm exercise MPI evaluations may be less affected by beta blockade than pharmacologic studies. Increased mortality has been associated with blunted HR responses to both treadmill exercise and pharmacologic stress (10,11). Therefore, rather than being viewed primarily as a marker of decreased test sensitivity for detection of CAD, exercise chronotropic incompetence would be perceived more accurately as a marker of adverse survival outcome. Our data do not support reflexive referral for pharmacologic MPI testing in patients with chronotropic incompetence.In a somewhat analogous fashion, deconditioning and low exercise capacity are often regarded as justification for pharmacologic stress testing or as a contributor to decreased test sensitivity (17). Peak oxygen uptake for arm ergometer exercise approximates 60% of treadmill maximal oxygen uptake (2,9). Nevertheless, we observed a very high (69%) prevalence of abnormal MPI studies and perfusion defect findings and a greater percentage of MPI abnormalities in patients with the lowest versus those with the highest tertile of arm exercise capacity. However, abnormal MPI results still provided additional value for prognostication of death in the tertile with the highest exercise capacity, although average arm exercise capacity in this group (4.2 METs or multiples of resting metabolic rate) was still only equivalent to a treadmill exercise capacity (7 METs), about 10%–15% below average for healthy sedentary men of similar age (16). These findings indicate that deconditioning and low exercise capacity do not necessarily require pharmacologic stress testing.As reported by many other investigators (23), the prevalence of arm exercise-induced MPI abnormalities was considerably greater than abnormal ECG responses in the same patients, consistent with higher sensitivity of arm exercise MPI testing for detection of CAD. However, there was a strong correlation of abnormal arm exercise ECG results with MPI defect number, size, and type, suggesting that patients at greater risk, based on MPI findings, are more likely than individuals at lower risk to be identified with arm exercise ECG evaluations alone. Nevertheless, an arm exercise MPI evaluation had substantially greater sensitivity and negative predictive value for prognostication of mortality, albeit with considerably lower specificity than an abnormal arm exercise ECG.Overall, an abnormal MPI study was of only borderline significance for prediction of mortality by multivariate Cox analysis. This is likely explained by the significant number of patients with small or single perfusion defects with no or only minor effects on mortality as illustrated in Figure 3, the favorable effect of coronary revascularization on mortality in patients with large, reversible, or mixed perfusion defects, and our relatively small sample size.LimitationsThis study was restricted to a relatively small population of veterans at a single institution and included mostly middle-age to older men with a high prevalence of cardiovascular disease. Thus, our results could be less relevant to younger, healthier populations or to women. In addition, this investigation was not a randomized comparison of arm exercise with pharmacologic or treadmill exercise testing and the potential advantages of arm exercise versus these other stress testing modalities require confirmation with a randomized controlled clinical trial. Although MPI evaluations may be very useful for identification of cardiac ischemia in patients with an abnormal resting ECG, for quantification of the size of myocardial ischemic regions, and have higher sensitivity than that of an exercise ECG for detection of ischemia, their role in predicting subsequent outcome relative to arm exercise ECG findings alone has not been well defined and also merits investigation.CONCLUSIONSOur data provide evidence that arm exercise MPI stress testing is a valuable alternative to pharmacologic imaging for detection of myocardial ischemia and prediction of all-cause mortality. Arm exercise also provides clinically important information on functional capacity, symptomatic, cardiovascular, and ECG responses to the relevant physiologic stress of exercise in patients with disabilities precluding treadmill testing.W. H. M., H. X., P. C., and E. B. are supported by a Merit Review research award from the Department of Veterans Affairs, Washington, DC. N. I. K. was supported by a Seed Grant research award from the American Medical Association Foundation, Chicago, IL, and by a Mentors in Medicine research grant from Washington University School of Medicine, St. Louis, MO.The authors declare no conflict of interest and relationship with industry.The results of the present study do not constitute endorsement by the American College of Sports Medicine.REFERENCES1. Abidov A, Bax JJ, Hayes SW, et al. Transient ischemic dilatation ratio of the left ventricle is a significant predictor of future events in patients with otherwise normal myocardial perfusion SPECT. J Am Coll Cardiol. 2003; 42: 1818–25. [CrossRef] [Medline Link] [Context Link]2. Astrand PO, Saltin B. Maximal oxygen uptake and heart rate in various types of muscular activity. J Appl Physiol. 1961; 16: 977–81. [Context Link]3. Balady GJ, Weiner DA, Rothendler JA, Ryan TJ, et al. Arm exercise-thallium imaging testing for detection of coronary artery disease. J Am Coll Cardiol. 1987; 9: 84–8. [CrossRef] [Medline Link] [Context Link]4. Blair SN, Kohl HW III, Paffenbarger RS Jr, Clark DG, Cooper KH, Gibbons LW. Physical fitness and all-cause mortality: a prospective study of healthy men and women. JAMA. 1989; 262: 2395–401. [CrossRef] [Medline Link] [Context Link]5. Chaitman BR. Exercise stress testing. In: Zipes DP, Libby P, Bonow RO, Braunwald E, editors. Braunwald’s Heart Disease. A Textbook of Cardiovascular Medicine. 7th ed, vol. 1. Philadelphia (PA): Elsevier Saunders; 2005. pp. 153–85. [Context Link]6. Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-rate recovery immediately after exercise as a predictor of mortality. New Eng J Med. 1999; 341: 1351–7. [Context Link]7. Fisher SG, Weber L, Goldberg J, Davis F. Mortality ascertainment in the veteran population: alternatives to the National Death Index. Am J Epidemiol. 1995; 141: 242–50. [Medline Link] [Context Link]8. Grover-McKay M, Milne N, Atwood JE, Lyons KP. Comparison of thallium-201 single-photon computed tomographic scintigraphy with intravenous dipyridamole and arm exercise. Am Heart J. 1994; 127: 1516–20. [CrossRef] [Medline Link] [Context Link]9. Ilias NA, Xian H, Inman C, Martin WH III. Arm exercise testing predicts clinical outcome. Am Heart J. 2009; 157: 69–76. [CrossRef] [Full Text] [Medline Link] [Context Link]10. Iqbal FM, Al Jaroudi W, Sanam K, et al. Reclassification of cardiovascular risk in patients with normal myocardial perfusion imaging using heart rate response to vasodilator stress. Am J Cardiol. 2013; 111: 190–5. [CrossRef] [Medline Link] [Context Link]11. Leeper NJ, Dewey FE, Ashley EA, et al. Prognostic value of heart rate increase at onset of exercise testing. Circulation. 2007; 115: 468–74. [CrossRef] [Full Text] [Medline Link] [Context Link]12. Mark DB, Shaw L, Harrell FE Jr, et al. Prognostic value of a treadmill exercise score in outpatients with suspected coronary artery disease. N Eng J Med. 1991; 325: 849–53. [Context Link]13. Martin WH III, Berman WI, Buckey JC, Snell PG, Blomqvist CG. Effects of active muscle mass size on cardiopulmonary responses to exercise in congestive heart failure. J Am Coll Cardiol. 1989; 14: 683–94. [CrossRef] [Medline Link] [Context Link]14. Martin WH III, Xian H, Wagner D, Chandiramani P, Bainter E, Ilias-Khan N. Arm exercise as an alternative to pharmacologic stress testing: arm exercise testing and outcome. Am Heart J. 2014; 167: 169–77. [CrossRef] [Full Text] [Medline Link] [Context Link]15. Myers J, Prakash M, Froelicher V, Do D, Partington S, Atwood JE. Exercise capacity and mortality among men referred for exercise testing. New Eng J Med. 2002; 346: 793–801. [CrossRef] [Full Text] [Medline Link] [Context Link]16. Ogawa T, Spina R, Martin WH III, et al. Effects of aging, sex, and physical training on cardiovascular responses to exercise. Circulation. 1992; 86: 494–503. [CrossRef] [Full Text] [Medline Link] [Context Link]17. Rozanski A, Gransar H, Hayes SW, et al. Temporal trends in the frequency of inducible myocardial ischemia during cardiac stress testing: 1991 to 2009. J Am Coll Cardiol. 2013; 61: 1054–65. [CrossRef] [Medline Link] [Context Link]18. SAS software. Version 9.3 of the SAS System for Windows; [accessed 2013 December 19]; Copyright 2011 SAS Institute, Inc. Cary, NC, USA. [Context Link]19. Stein L, Burt R, Oppenheim B, Schauwecker D, Fineberg N. Symptom-limited arm exercise increases detection of ischemia during dipyridamole tomographic thallium stress testing in patients with coronary artery disease. Am J Cardiol. 1995; 75: 568–72. [CrossRef] [Full Text] [Medline Link] [Context Link]20. Stratmann HG, Williams GA, Wittry MD, Chaitman BR, Miller DD. Exercise technetium-99m sestamibi tomography for cardiac risk stratification of patients with stable chest pain. Circulation. 1994; 89: 615–22. [Context Link]21. Taillefer R, Ahlberg AW, Masood Y, et al. Acute beta-blockade reduces the extent and severity of myocardial perfusion defects with dipyridamole Tc-99m sestamibi SPECT imaging. J Am Coll Cardiol. 2003; 42: 1475–83. [CrossRef] [Medline Link] [Context Link]22. Thomas GS, Prill NV, Majmundar H, et al. Treadmill exercise during adenosine infusion is safe, results in fewer adverse reactions, and improves myocardial perfusion image quality. J Nucl Cardiol. 2000; 7: 439–46. [CrossRef] [Medline Link] [Context Link]23. Udelson JE, Dilsizian V, Bonow RO. Nuclear cardiology. In: Zipes DP, Libby P, Bonow RO, Braunwald E. Braunwald’s Heart Disease. A Textbook of Cardiovascular Medicine. 7th ed. Philadelphia (PA): Elsevier Saunders; 2005, vol. 1, chapter 13, pp. 311–8. [Context Link] CORONARY REVASCULARIZATION; MORTALITY; NUCLEAR CARDIOLOGY; STRESS TESTINGovid.com:/bib/ovftdb/00005768-201412000-0000400004481_2003_42_1818_abidov_dilatation_|00005768-201412000-00004#xpointer(id(R1-4))|11065213||ovftdb|SL00004481200342181811065213P76[CrossRef]10.1016%2Fj.jacc.2003.07.010ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_2003_42_1818_abidov_dilatation_|00005768-201412000-00004#xpointer(id(R1-4))|11065405||ovftdb|SL00004481200342181811065405P76[Medline Link]14642694ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_1987_9_84_balady_detection_|00005768-201412000-00004#xpointer(id(R3-4))|11065213||ovftdb|SL00004481198798411065213P78[CrossRef]10.1016%2FS0735-1097%2887%2980086-4ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_1987_9_84_balady_detection_|00005768-201412000-00004#xpointer(id(R3-4))|11065405||ovftdb|SL00004481198798411065405P78[Medline Link]3794114ovid.com:/bib/ovftdb/00005768-201412000-0000400005407_1989_262_2395_blair_prospective_|00005768-201412000-00004#xpointer(id(R4-4))|11065213||ovftdb|SL000054071989262239511065213P79[CrossRef]10.1001%2Fjama.262.17.2395ovid.com:/bib/ovftdb/00005768-201412000-0000400005407_1989_262_2395_blair_prospective_|00005768-201412000-00004#xpointer(id(R4-4))|11065405||ovftdb|SL000054071989262239511065405P79[Medline Link]2795824ovid.com:/bib/ovftdb/00005768-201412000-0000400000429_1995_141_242_fisher_ascertainment_|00005768-201412000-00004#xpointer(id(R7-4))|11065405||ovftdb|SL00000429199514124211065405P82[Medline Link]12251096ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_1994_127_1516_grover_scintigraphy_|00005768-201412000-00004#xpointer(id(R8-4))|11065213||ovftdb|SL000004061994127151611065213P83[CrossRef]10.1016%2F0002-8703%2894%2990379-4ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_1994_127_1516_grover_scintigraphy_|00005768-201412000-00004#xpointer(id(R8-4))|11065405||ovftdb|SL000004061994127151611065405P83[Medline Link]8197977ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_2009_157_69_ilias_exercise_|00005768-201412000-00004#xpointer(id(R9-4))|11065213||ovftdb|00000406-200901000-00010SL0000040620091576911065213P84[CrossRef]10.1016%2Fj.ahj.2008.09.007ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_2009_157_69_ilias_exercise_|00005768-201412000-00004#xpointer(id(R9-4))|11065404||ovftdb|00000406-200901000-00010SL0000040620091576911065404P84[Full Text]00000406-200901000-00010ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_2009_157_69_ilias_exercise_|00005768-201412000-00004#xpointer(id(R9-4))|11065405||ovftdb|00000406-200901000-00010SL0000040620091576911065405P84[Medline Link]19081399ovid.com:/bib/ovftdb/00005768-201412000-0000400000416_2013_111_190_iqbal_reclassification_|00005768-201412000-00004#xpointer(id(R10-4))|11065213||ovftdb|SL00000416201311119011065213P85[CrossRef]10.1016%2Fj.amjcard.2012.09.013ovid.com:/bib/ovftdb/00005768-201412000-0000400000416_2013_111_190_iqbal_reclassification_|00005768-201412000-00004#xpointer(id(R10-4))|11065405||ovftdb|SL00000416201311119011065405P85[Medline Link]23111139ovid.com:/bib/ovftdb/00005768-201412000-0000400003017_2007_115_468_leeper_prognostic_|00005768-201412000-00004#xpointer(id(R11-4))|11065213||ovftdb|00003017-200701300-00009SL00003017200711546811065213P86[CrossRef]10.1161%2FCIRCULATIONAHA.106.666388ovid.com:/bib/ovftdb/00005768-201412000-0000400003017_2007_115_468_leeper_prognostic_|00005768-201412000-00004#xpointer(id(R11-4))|11065404||ovftdb|00003017-200701300-00009SL00003017200711546811065404P86[Full Text]00003017-200701300-00009ovid.com:/bib/ovftdb/00005768-201412000-0000400003017_2007_115_468_leeper_prognostic_|00005768-201412000-00004#xpointer(id(R11-4))|11065405||ovftdb|00003017-200701300-00009SL00003017200711546811065405P86[Medline Link]17242274ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_1989_14_683_martin_cardiopulmonary_|00005768-201412000-00004#xpointer(id(R13-4))|11065213||ovftdb|SL0000448119891468311065213P88[CrossRef]10.1016%2F0735-1097%2889%2990111-3ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_1989_14_683_martin_cardiopulmonary_|00005768-201412000-00004#xpointer(id(R13-4))|11065405||ovftdb|SL0000448119891468311065405P88[Medline Link]2768718ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_2014_167_169_martin_pharmacologic_|00005768-201412000-00004#xpointer(id(R14-4))|11065213||ovftdb|00000406-201402000-00006SL00000406201416716911065213P89[CrossRef]10.1016%2Fj.ahj.2013.10.022ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_2014_167_169_martin_pharmacologic_|00005768-201412000-00004#xpointer(id(R14-4))|11065404||ovftdb|00000406-201402000-00006SL00000406201416716911065404P89[Full Text]00000406-201402000-00006ovid.com:/bib/ovftdb/00005768-201412000-0000400000406_2014_167_169_martin_pharmacologic_|00005768-201412000-00004#xpointer(id(R14-4))|11065405||ovftdb|00000406-201402000-00006SL00000406201416716911065405P89[Medline Link]24439977ovid.com:/bib/ovftdb/00005768-201412000-0000400006024_2002_346_793_myers_mortality_|00005768-201412000-00004#xpointer(id(R15-4))|11065213||ovftdb|00006024-200203140-00002SL00006024200234679311065213P90[CrossRef]10.1056%2FNEJMoa011858ovid.com:/bib/ovftdb/00005768-201412000-0000400006024_2002_346_793_myers_mortality_|00005768-201412000-00004#xpointer(id(R15-4))|11065404||ovftdb|00006024-200203140-00002SL00006024200234679311065404P90[Full Text]00006024-200203140-00002ovid.com:/bib/ovftdb/00005768-201412000-0000400006024_2002_346_793_myers_mortality_|00005768-201412000-00004#xpointer(id(R15-4))|11065405||ovftdb|00006024-200203140-00002SL00006024200234679311065405P90[Medline Link]11893790ovid.com:/bib/ovftdb/00005768-201412000-0000400003017_1992_86_494_ogawa_cardiovascular_|00005768-201412000-00004#xpointer(id(R16-4))|11065213||ovftdb|00003017-199208000-00019SL0000301719928649411065213P91[CrossRef]10.1161%2F01.CIR.86.2.494ovid.com:/bib/ovftdb/00005768-201412000-0000400003017_1992_86_494_ogawa_cardiovascular_|00005768-201412000-00004#xpointer(id(R16-4))|11065404||ovftdb|00003017-199208000-00019SL0000301719928649411065404P91[Full Text]00003017-199208000-00019ovid.com:/bib/ovftdb/00005768-201412000-0000400003017_1992_86_494_ogawa_cardiovascular_|00005768-201412000-00004#xpointer(id(R16-4))|11065405||ovftdb|00003017-199208000-00019SL0000301719928649411065405P91[Medline Link]1638717ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_2013_61_1054_rozanski_myocardial_|00005768-201412000-00004#xpointer(id(R17-4))|11065213||ovftdb|SL00004481201361105411065213P92[CrossRef]10.1016%2Fj.jacc.2012.11.056ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_2013_61_1054_rozanski_myocardial_|00005768-201412000-00004#xpointer(id(R17-4))|11065405||ovftdb|SL00004481201361105411065405P92[Medline Link]23473411ovid.com:/bib/ovftdb/00005768-201412000-0000400000416_1995_75_568_stein_dipyridamole_|00005768-201412000-00004#xpointer(id(R19-4))|11065213||ovftdb|00000416-199503150-00004SL0000041619957556811065213P94[CrossRef]10.1016%2FS0002-9149%2899%2980618-3ovid.com:/bib/ovftdb/00005768-201412000-0000400000416_1995_75_568_stein_dipyridamole_|00005768-201412000-00004#xpointer(id(R19-4))|11065404||ovftdb|00000416-199503150-00004SL0000041619957556811065404P94[Full Text]00000416-199503150-00004ovid.com:/bib/ovftdb/00005768-201412000-0000400000416_1995_75_568_stein_dipyridamole_|00005768-201412000-00004#xpointer(id(R19-4))|11065405||ovftdb|00000416-199503150-00004SL0000041619957556811065405P94[Medline Link]7887379ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_2003_42_1475_taillefer_dipyridamole_|00005768-201412000-00004#xpointer(id(R21-4))|11065213||ovftdb|SL00004481200342147511065213P96[CrossRef]10.1016%2FS0735-1097%2803%2901046-5ovid.com:/bib/ovftdb/00005768-201412000-0000400004481_2003_42_1475_taillefer_dipyridamole_|00005768-201412000-00004#xpointer(id(R21-4))|11065405||ovftdb|SL00004481200342147511065405P96[Medline Link]14563595ovid.com:/bib/ovftdb/00005768-201412000-0000400041879_2000_7_439_thomas_myocardial_|00005768-201412000-00004#xpointer(id(R22-4))|11065213||ovftdb|SL000418792000743911065213P97[CrossRef]10.1067%2Fmnc.2000.108030ovid.com:/bib/ovftdb/00005768-201412000-0000400041879_2000_7_439_thomas_myocardial_|00005768-201412000-00004#xpointer(id(R22-4))|11065405||ovftdb|SL000418792000743911065405P97[Medline Link]11083192Arm Exercise Myocardial Perfusion Imaging for Prognostication of Long-Term OutcomeMARTIN, WADE H. III; XIAN, HONG; WAGNER, DANIEL L.; CHANDIRAMANI, POOJA; BAINTER, EMILY; ILIAS-KHAN, NASREENClinical Sciences1246