Additive value of thallium single-photon emission computed tomography myocardial imaging for prediction of perioperative events in clinically selected high cardiac risk patients having abdominal aortic surgery

Additive Value of Thallium Single-Photon Emission Computed Tomo raphy Myocardial Imaging for Pre %iction of Perioperative Events In Clinically Selected High Cardiac Risk Patients Having Abdominal Aortic Surgery Gerald Vanzetto, MD, Jacques Machecourt, MD, Dan Blendea, Daniel Fagret, MD, PhD, Elisabeth Borrel, MD, Jean Luc Magne, Francoise Gattaz, MD, and Henri Guidicelli, MD evalThe present study was desi ned to prospectively uate whether reinjection tf allium-201 single-photon emission computed tomography (SPECT) has a significant additive redictive value for occurrence of perioperative car criac events in clinically selected patients at high cardiac risk undergoing abdominal aortic surgery. Of a group of 5 17 consecutive patients referred, 134 had 22 of the following clinical or electrocardiographic cardiac risk variables: age 270 years; history of myocardial infarction, angina, or congestive heart failure; diabetes mellitus; hy ertension with severe left ventricular hypertrophy; an s Q waves or ischemic STsegment abnormalities on electrocardiogram at rest. Operation was performed after thallium SPECT study. Twelve patients (9%) had major perioperative events (cardiac death or nonfatal myocardial infarction) and 18 patients had other cardiac events (unstable angina, con estive heart failure, or severe ventricular tachyarrhyt 3, mia). Variables correlated with the occurrence of

he high prcvalcnce of coronary artcry disease in paT tients with abdominal aortic disease (aneurysm or athcroma) makes cardiac ischcmic events the most fre-

MD, MD,

major events were history of myocardial infarction (p ~0.05) and the presence (p
dergoing many types of vascular” ‘4.‘x and cvcn nonvascular procedures.” I7 Planar lhalliu’n imaging was pcrformed in most of these studies. Thallium single-photon qucnt complication 01’ abdominal aortic surgery.‘-” Pre- emission computed tomography (SPECT), which has an operative evaluation of cardiac risk is therefore very improved accuracy for detection of ‘nyocardial ischemia important for reducing o\;eralI morbidity and mortalily and assessing cardiac prognosis, was the method used both intra- and postoperatively. Bcsidcs clinical and elec- in only a few of these studies.‘7.‘X Redistribution studies trocardiographic factors:.‘-” thallium imaging has shown were performed in previous studies. but the predictive itself to bc a powerl’ul predictor for the occurrence 01 value of thallium imaging with reinjection has not been perioperative coronary cv~nls”--~“: a significant prcdic- extensively evaluated. Kcccntly, 2 important points have live value was found to be associated with the presence been raised which refocus interest on this problem. First, of a reversible dcfcct on the thallium scan.” ‘f).“).20The the predictive value of thallium i’naging has been quesnumber of reversible thallium dcfccts, indicating the tioncd in 2 rccenlly published studics.2’.22 Second, beextent of viable ischcmic m]Iocardiuml was also found C~LIW thalliu’n SPECT scintigaphy is relatively expento bc r&ted to the probability of cardiac events. I3--I6 sive and also time-consuming, it cannot be pcrformcd in However. most of thcsc studics”-‘s included patients un- cvcry patient before abdominal aortic surgery. Thus. a subset of patients with maximal benelit from thallium examination needs to be defined. Most of the previous studies were performed retrospcctivcly. WC have prospcclivcly examined a group of high-risk paticnls sclccted bcl’orc abdominal aortic surgery (presence of >_2clinical risk factors), with the aim of assessing whether abnormal thallium SPECT results have a sieniticant additivc prcdictivc value for the prediction (;f perioperative events.

METHODS Study patients: Between January 1990 and December 1993, 517 patients were referred to our institution for elective abdominal aortic surgery. They were examined preoperatively by a cardiologist and an anesthetist. Surgery was canceled in 60 patients: in 23 for noncardisc reasons (advanced age, n = 7), patient refusal (n = 5), surgical reasons (n = 7), and other reasons (n = 4). Surgery was canceled in 37 patients for cardiac reasons (nonstabilized heart failure and/or class fII to IV angina [ 15 underwent myocardial revascularization]). The remaining 457 patients were prospectively considered according to the following clinical and electrocardiographic criteria of high cardiac risk: (1) age >70 years, (2) history of myocardial infarction, (3) history of angina, (4) history of congestive heart failure, (5) diabetes mellitus, (6) hypertension with severe left ventricular hypertrophy, (7) presence of Q waves, or (8) ST-segment ischemic abnormalities during rest electrocardiogram.* Three hundred seventeen patients had <2 of these clinical prognostic risk factors, and underwent operation on clinical and electrocardiographic preoperative assessment only. One hundred forty patients satisfied our high cardiac risk selection criteria and underwent SPECT evaluation. Six patients in whom thallium-SPECT images were unsuitable for evaluation were excluded. One hundred thirty-four patients underwent elective abdominal aortic surgery after clinical, electrocardiographic, and thallium SPECT evaluation and represented the study Short

Short

Axis - Apex

Axis - Basal

group. None of these 134 patients underwent myocardial revasculatization before abdominal aortic surgery. Thallium 201 singlephoton emission computed tomogmphy acquisition and analysis: Thallium examination

was performed 120 days before surgery. The protocol of image recording and analysis has been previously described.23 Dipyridamolc infusion was the stress method utilized. After dipyridamole infusion (0.56 mg/kg body weight intravenously for 4 minutes), 2.5 mCi (92.5 MBq) of thallium-201 was injected intravenously. Initial image recording was obtained 10 minutes later. We used a gamma camera with a low-energy, high-resolution collimator. Images were obtained during a 180” circular movement, starting from the 50” left posterior oblique projection using a 6” increment (32 projections). Transaxial sections were obtained by filtered back projection and reoriented to the 3 standard cardiac planes (short, horizontal long, and vertical axes). The left ventricle was divided into 6 segments (apical, anterior, septal, lateral, inferior, and posterior). Qualitative visual analysis was performed and consensus was obtained from 2 experienced observers who had no knowledge of the clinical data. Four hours after the initial scan, a reinjection study was performed 30 minutes after reinjection of I mCi (37.0 MBq) of thallium. Each segment was classified as follows: (1) normal, (2) with reversible (transient) defect in case of partial or total normalization, or (3) with fixed (persistent) defect in case of persistent abnormality at 4 hours. Localization of abnormal defects was also considered (Figure 1). Segments 2 and 3 were considered as anterior, the other segments as not anterior. Surgical management,

Vertical

1 n

4 m

Apex

long axis

Lateral

2 =Anterior

5 m

Inferior

3 FlSeptal

6 m

Posterior

FKYJRE 1. Left ventricle Each segment is scored

144

Horizontal

long axis

THE AMFRICAN

segmentation as normal,

JOURNAL

according to the 3 arthagonal reversible, or fixed defect.

Oi

CARDIOLOGY”

VOL. 77

planes.

JAI\IUAR”

and peri- and post-

operative care: Our study group included patients who underwent surgical procedures involving the abdominal aorta (abdominal aortic aneurysm resection. n = 89; aottoiliac reconstruction for occlusive disease, n = 45). The average duration was 40 minutes for aortic clamping and 3.6 hours for the entire surgical procedure. Anti-ischemic therapy was not modified or discontinued. Clinicians managing peri- and postoperative care and anesthetic procedures were blinded from initial thallium results. Hemodynamic monitoring using radial and Swan-Ganz catheters was performed in all patients during the perioperative period. Postoperatively all patients were kept 224 hours in the intensive cam unit with continuous 3-lead electrocardiographic monitoring. A 1Zlead electrocardiogram was recorded twice daily and each time abnormalities were observed during monitoring. Particular attention was made to avoid perioperative fluid overload or loss. Creatine kinase enzymes were monitored twice daily for 3 days with creatine kinase-MB measurement when total creatine kinasc exceeded 240 IU. The mean hospitalization period was 14 days in our study group versus 10 days for low-risk patients (i.e., with <2 risk factors, p cO.05). ‘5,

1994

TABLE I Univariate

Relation

Between

Variables

Analyzed

and

Occurrence

of Cardiac

Events No

Variable Age 1~4 Sex ratio (men/women) A ngina pectoris Myocardial infarction Congestive heart failure Systemic hypertension Diabetes mellitus Smoking history Hypercholesterolemia (~240 mg/dl) Q waves Anterior Not anterior ST ischemic abnormalities Anterior Not anterior Abnormal TI scan Reversible TI defects Fixed TI defects Number of TI defects Number of TI reversible defects Number of TI fixed defects

No Event (n = 104)

Any Event (n = 30)

65

65* 10 28/2 8 (271 14 (47) 8 (27) 16 (53) 3 IlO) 13 (43) 15 (50) 16 (53) 7 (231 11 (37) 1s (50) 13 (43) 7 (231 26 (87) 24 (80) 14 (47) 1.8 f 1.1 1.1 rt 0.8 0.7 f 0.9

i 10 98/6 23 (22) 25 (24) 20 (19) 62 (60) 15 (14) 40 (38) 45 (43) 32 (31) 5 (51 26 (25) 24 (23) 16 (15) 11 (11) 57 (55) 24 (23) 43 (41) 0.9 zt 1.0 0.3 f 0.5 0.6 iz 0.9

p Value*

Major Event (n = 122)

Major Event (n = 12)

NS NS NS 0.029 NS NS NS NS NS 0.039 0.005 NS 0.008 0.002 NS 0.003
65* 10 115/7 28 (23) 32 (26) 24 (20) 69 (57) 16 (13) 49 (40) 54 (44) 42 (34) 1 1 (91 32 (26) 35 (29) 25 (20) 17 (14) 72 (59) 37 (30) 51 (42) 1.0 f 1.1 0.4 f 0.6 0.6 f 0.Y

67* 10 1 l/l 3 (251 7 (581 4 (331 9 (751 2 (171 4 (331 6 (501 6 (50) 1 181 5 (421 4 (331 4 (33) 1 PI 11 (92) 11 (92) 6 150) 1.9 f 0.9 1.2 * 0.7 0.7 f 0.8

p Valuet NS NS NS 0.045 NS NS NS NS NS NS

NS NS NS NS

NS 0.055
*No events versus any events. YMojor events versus no major events. Values are expressed as number [X) or mean + SD. TI = thallium.

End points: Major cardiac end points were defined as cardiac death or nonfatal myocardial infarction during the hospitalization period. Cardiac death was defined as sudden death or death caused by myocardial infarction or congestive cardiac failure. A diagnosis of myocardial infarction was made when serum creatine kinase levels were >240 IU/L, confirmed by significant elevation of MB isoenzymes (>1.5 IU) and associated with the appearance of a new Q wave on the 1Zlead electrocardiogram (duration 20.04 second), or ST-segment depression >I mm 60 ms after the j point (non-Q-wave myocardial infarction). Besides the 2 major end points. the other end points were unstable angina, congestive heart failure. or severe ventricular tachyarrhythmias. Unstable angina was defined as the appearance of reversible ischemic ST changes (i.e., ST-segment depression or elevation of 21 mm on the 12-lead electrocardiogram) with clinical symptoms. Congestive heart failure was defined as clinical and radiologic evidence of acute pulmonary edema, pulmonary wedge pressure elevation during hemodynamic monitoring, or need for major inotropic support. Ventricular tachyarrhythmias were defined as the presence of ventricular fibrillation or ventricular tachycardia demanding immediate therapeutic intervention. If >I cardiac event was present in I patient, only the most severe events were considered for statistical purposes. Data collection and statistical analysis: Data collection for perioperative events was prospectively performed by an investigator blinded to SPECT imaging results. Clinical variables taken into account were: age, gender. history of angina pcctoris, history of myocardial infarction. history of heart failure (New York Heart Association class III or IV), hypertension, diabetes mellitus, smokCCRONARY

ARTERY

ing, and hypercholesterolcmia (serum cholesterol >6.47 mmol/L). Electrocardiographic analysis included presence of necrotic Q waves and ischemic ST abnormalities (ST elevation or depression 21 mm from baseline). Electrocardiographic abnormalities were assigned as anterior if localized in leads 1 to 5: and not anterior if localized in other leads. Continuous variables are presented as mean + 1 SD. Thallium results were assessed as both continuous and categorical variables. The univariate relation between the dill‘erent variables and end points was assessed using Student’s t test or Mann-Whitney rank-sum test for continuous variables, and the chisquare test with continuity correction for categorical variables. Stepwise logistic regression models were iitted to identify independent predictors of cardiac events (SPSS-PC + 4.0. SPSS Inc.: Chicago, Illinois). A p value ~0.05 was considered statistically significant.

DISEASE,;

RESULTS Clinical baseline data: There were no differences bctween the 134 selected patients and the other 317 patients who underwent abdominal aortic surgery, except for the 2 following parameters: age (65 f IO vs 63 f 11 years, p ~0.05) and the number ol‘ clinical risk factors (4.1 f 0.6 vs 0.7 f 0.4, p
SPtC[

A,NC

A?T~IXX~INAL

40PTIC

:SURGiRY

145

Single predictors of cardiac events (p ~0.05) were history of myocardial inNo Major, Major farction, presence of a necrotic Q wave No Event Any Event Event Event on the electrocardiogram, presence of a [n = 104) (n = 30) p Value* (n = 122) [n = 12) p Valuet Variable Q wave in anterior leads, ischemic ST Anterior defects 6 (6) 9 (30) co.oo1 1 1 (9) 4 (33) 0.038 abnormalities, anterior ischemic ST abReversible 3 (31 6 (2’3) 0.003 6 PI 3 (25) 0.040 normalities, and the presence and mean 4 (4) 5 (17) 0.039 7 (6) 2 (17) NS Fixed number of segments with thallium and No anterior defects 55 (53) 20 (67) .O!& 66 (54) 9 (75) NS reversible defects (Table I). Univariate 21 (20) 19 (63) 31 (25) 9 (75) 0.001 Reversible Fixed 42 (40) 9 (30) NS 47 (39) 4 (33) NS analysis showed the following factors to be predictive for the occurrence of ‘No events versus any events. tMojor events versus no ma@ events. major cardiac events: history of myoValues are expressed CIS number YL). cardial infarction, mean number of segments with thallium defects, presence of reversible thallium abnormalities, and the number of TABLE Ill Predictive Value of Thallium SPECT According to Number segments with reversible defects. When localization of of Clinical Risk Factors: Occurrence of Cardiac Events thallium defects was taken into consideration, reversible defects with either anterior or no anterior localization were significantly associated with an adverse cardiac outcome (Table II). Combined fixed and reversible anterior defects Two to Four Risk Factors (n = 76) also appeared to be predictors of cardiac events. Fixed ann = 23 n = 53 terior defects only were significant when all cardiac events Maior events I%) 0 0.04 3 I131 were taken into consideration. Any events (%) 2 (3.8) 10 j43j 0.0001 In multivariate analysis, the history of myocardial Five or More Risk Factors (n = 58) infarction (p <0.05), anterior Q wave on the electrocardiogram (p <0.05), anterior ischemia on the electrocardiogram (p <0.05), and the number of segments with reversible thallium defects (p ~0.001) were significant independent predictors for occurrence of any cardiac SPECT = single-photon emission computed tomography; TI = thallium. event. History of myocardial infarction (p ~0.05) and the number of segments with reversible thallium defects (p Thallium single-photon emission computed tomograph
146

THE

AMERICAN

Results

According

JOURNAL

to localization

OF

CARDIOLOGY’

of Abnormal

VOL.

Segments

77

JANUARY

I

15,

1996

DISCUSSION In 1985 Boucher et al9 for the first time stressed the importance of thallium imaging in assessing the risk of perioperative ischemic events before vascular surgery. Eagle,i2 Levinson,i4 Lette,i5 and their co-workers, and, more recently, Brown and Rowen16 noted that both the presence and extent of a reversible thallium defect were highly predictive of the occurrence of severe perioperative cardiac events. However, most investigators used heterogeneous groups of surgically treated patients. 9-17 A homogenous group of patients, all of whom underwent abdominal aortic procedures, has been used in 3 studies.19.20,22McEnroe et all9 and Cutler and Leppo*O stressed the good negative predictive value of the test, and found that the presence of a reversible defect on 22 myocardial segments significantly increases the probability of the occurrence of perio rative myocardial infarction. Interestingly, Baron et al F2 did not find any prognostic value for thallium SPECT in a large series of 457 consecutive patients. An original factor in our study is that it is a prospective one. In particular, the surgeon and anesthesiologist were blinded to the SPECT results. This point is important, especially because contradictory results have recently been published. 21*22In most studies, with the exception of Baron’s work, the most important single predictor remains the presence of a reversible defect on the redistribution scan.‘-17~19~20 The presence of a fixed defect had no predictive value9-16*20with the exception of the study of McFalls et al.‘* Shaw et al” found that the presence of a combined (fixed and reversible) defect was predictive of subsequent cardiac events. All of these studies were performed using redistribution imaging. In our series, a reinjection study was performed to allow a better discrimination between ischemic and necrotic tissue.24,25 However, a concomitant redistribution study was only performed in a few patients, and comparison of these 2 techniques is not possible. No quantification Thallium

imaging:

PROGNOSTIC VALUE:

FIGURE 2. Probability of perioperative cardiovascukr events when the variables studied are considered in an hierarchical order: First bar represents he probabil’ of events for all patients, no varia “b le being taken into consideration; second bar represents he additive prediaive value of a history of myocardial infarction (Ml) (p = 0.037 for all events, p = 0.04 for major events); hifd bar represents he additive edictive val& of the presence or an abnormal ehxtrocardiiram (ECG) (ST abnormalities or Q wave) in patients with a history of myocardial infarction (p = 0.017 for all events, p = NS for major evanfs); bst 3 bars indicate the addiive value of thallium single-photon emission computed tomogmphy (SPECT)in tients with previous myocardial in fr rction and bbnormal &ctrocordi rams when 1, 2. and 3 reversible “il( ev) defects. r&p&ively, were preSent ip d.odl for all events and major events).

Probability

All

COROKARY

patients

has been performed in our study as well as in previous studies (to the extent that its usefulness has not been demonstrated for such prognostic evaluations). We found that fixed defects were generally not predictive of perioperative events. However, fixed defects may be of greater pronostic value in late follow-up, as recently reported by Hendel et a1.26 Only 3 recent reports used thallium SPECT 17q18.22 Thallium SPECT imaging increases the sensitivity of detection of coronary lesions,*’ and in particular, improves the quality of segmental analysis of the left ventricle. In our study, we demonstrated that the presence of an anterior (reversible or fixed) defect was predictive of a potential cardiovascular event. Additive predictive value of thallium single-photon emission computed tomography: The main identified pre-

dictors for perioperative complications are age >70 years, a history of myocardial infarction or angina pectoris, diabetes mellitus, severe hypertension, a previous episode of congestive heart failure, or an abnormal electrocardiogram (Q-wave or ST-T ischemic abnormalities).4-8 Conversely, Eagle et al,‘* in a retrospective observational study, found that when no clinical variables with prognostic values were present, the probability of perioperative complications was low; when some of these clinical variables were present, thallium redistrib ution had a high predictive value for perioperative events. Because thallium SPECT is expensive and timeconsuming, we prospectively decided to select only patients considered high risk. These patients represented 34% of the patients referred for elective abdominal aortic surgery during this period. Major event rates (death or myocardial infarction) were significantly higher for the selected patients than for patients with no or only 1 clinical variable. Importantly, among selected patients, thallium SPECT accurately discriminated between highand low-risk patients: the major event rate was 23% and 1% when a reversible defect was present or absent. The

of

Patients prewous

ARTERY DISEASF/THAllIUM

cardiovascular

with MI

events

Patients with previous MI and abnormal ECG

SPECl AND

ABDOMINA!

(%)

Patients with previous MI, abnormal ECG. and rev. SPECT defects

AORTIC

SURGERY

147

additive predictive value of thallium SPECT was demonstrated both in patients with a very high perioperative risk (25 clinical risk factors) and in those with an intermediate perioperative risk (2 to 4 clinical risk factors), as already pointed out by Eagle et aLI In multivariate analysis, only the number of abnormal reversible segments, history of myocardial infarction, and the presence of a Q wave and/or ST abnormalities on the electrocardiogram were significant predictors. The divergent results reported by Baronet a12’ can be explained by different selection criteria: our patients were high-risk, selected patients. whereas those studied by Baron et al were nonselected patients. Thus, the previous infarction rate was 39% in our patients versus 16% for Baron’s study. Furthermore, methods of thallium SPECT analysis could have been difl‘erent and could have interfered with the results.28 Patient management: -In light of our findings, coronary angiography should be considered when SPECT thallium shows 1 of the following abnormalities: reversible defect whatever the localization (especially when >l abnormal segment is present) or when a large abnormality is found on the anterior wall. According to these criteria, coronary angiography would have been performed in 55 of 134 (41%) of our clinically selected highrisk patients. The suitability of such a strategy will have to be cordirmed by analyzing the prognosis in patients who have been subjected to myocardial revascularization before abdominal aortic surgery.*’ Our results suggest that thallium evaluation could be focused on selected high-risk patients. In fact, in our series, as in others, the perioperative complication rate is low in patients with <2 clinical risk factors, and in these low-risk patients, the mortality of a myocardial revascularization procedure could exceed the mortality rate due to the vascular procedure itself.‘” Thirty-four percent of the patients were scheduled for thallium evaluation in our series, whereas Eagle et alI2 estimated the subset of patients needing thallium preoperative study at ~50%.

1. Ilertzcr NI’. Bcvcn EC;. Young JR. O’IIara PJ, Ruschlaupt WF III, Graor RA, Dewolti: VG. Maljovcc LC. Coronary artery tliscaw in peripheral vascular patients. A classification of IO(K) coronary angiograms and results of surgical management Ann Surfi 1984: 199:223--233. 2. Crawford ES. Bornbcrger KA, Glacser Dll. Salch SA. Row1 WL. Aortoiliac occlusive disease: factorb inllucncing sur*i*al and Function following rcconstructive operation over a wcnty-five-year period. Surges 1981;1N):1055-1067. 3. DcBakey ME. i.awrie GM. Combined coronary artery and peripheral vascular disease: recognition and Irearment. / VUW Surg lYX4;5:Ml5dO7. 4. Calvin JE. Kieser TM. Walley WM, McPhail NV, Barber GG. Scobie TK. Cardiac mortality and morhitlity after vascular surgery. Gun J Surg 19X6:2%93-97, 5. Detsky .AS, Abrams HB. McLaughlin JR, Dnskcr DJ. Sasson Z, Johnston N, Scolt JG Forbath N, Hilliard JR. Predicting cardiac complications in patients undergoing non-cardiac surgery. .I Gen lnmrrr Med 19X6,1:21 l-219. 6. Coopwnan M. Ptlug B, Martin EW Jr. Evans WE. Cardiovascular risk factors in patients with lwiphcral vascular disease. Surgc~ 197X$4:505-SO9. 7. Jamieson WR, Janusz MI. Miyagishima RT, &rein AN. influence orischemic hean discase on early and late mortaliry after burgcry for peripheral occlusive vascular disease. Circularion 19X2:66:1 92-197. 8. Goldman L, Caldcra DL. ~ussbaum SR, Soulhwick FS, Krogstad D, Murray B.

Burke DS, O’Malley TA, Gon,ll AH. &plan CH. Nolan J, Carabello B, Slater EE. Multifactorial index of cardiac risk in noncardiac surgical procedures. !v Ennpl .I Med 1977;297:845-X50. 9. Boucher CA, Brewster DC, Darling RC. Okada RD. Strauss IIW. Pohort GM. Dctcrmination of cardiac risk by dipyridamoie-thallium imaging before peripheral vascular surpy. N &$ .I Med 1985;312:389-394. IO. Leppo 1, Piaja I. Gionet M. Tumolo J, Paraskos IA, Cutler BS. Noninvasive evaluation of cardiac risk before clcctive vascular surgery. JAm Coil Cwdiol 19X7: 9:26Y-276. 11. Eagle KA, Singer DE, Brewslcr DC, Darling RC, Mullcy 4G. Bouchcr CA. Dipyridamole-thallium scannina in palients undergoing vascular surgery: olxirnizing prc-operative evaluation OT cardiac risk. JAMA 1987:257:21X5-21 X9. 12. Eagle KA, Coley CM, Newell JB, Brewster DC, Darling KC. Straw HW, Guiney TE: Boucher CA. Combining clinical and thallium data oplimizes pre-operativc assessment of cardiac risk berorc major vascular surgry. Ann Inrerrz Ma/ 19X9:1 10:8.59-X66. 13. Lane SE. Lewis SM, Pippin JJ, Kosinki FJ, Campbell D, Neslo RW. Hill T. Prcdicdw value ofqwntivative dipyridamolc-thallium xintigraphy in assessing cardiovascular risk after vascular surgery in diabetes mellhus. Aw ./ Curdiol 1989;&lz 1239-I 275. 14. Levinson JR, Boucher CA. Coiey CM, Guiney TE. Strauss I-IW. Eagle KA. Usefulness of semi-quantitative analysis of dipyridamoic-thallium-201 redistribution for improving risk stratification before vascular surgery. Am J Curdiol 1990:66: 4w10. 15. L.ettc J, Waters D. Lapointe J, Gagnon A, Picard M. Cerino M, Kerouac M. llsefulncss of the scvctity and extent of reversible perfusion dcfccls during thaliiurn-dipyridamole imaging [or cardiac risk asscssmenl before noncardiac surpe~. Am .I Curdiol I989;64:276-28 1. 16. Brown KA. Rowcn M. Extent of jeopardized viahie myocardium dctcrmined by myocardial perfusion imaging best predicts pcri-operalivc cardiac ercnts in paicnts undergoing noncardiac surgery. J Am CON h-dial lYY3:21:32.5~--330. 17. Shaw L. Miller DD. Kong BA, Hilton T, Steiken A, Stocke K. Chairman BR. Dctcrmination of peri-opualivc cardiac risk by adenosinc thallium-201 imaging. 4m Hean .I 1992:124:X61-X69. 18. McFallh EO, Doliwny KM, Grund F. Chute I;, Chesler E. Angina and persisLen1 exercise thallium defects: indcpcndent risk factors in elective vascular sorgcry. J Am Co// Curdiol 1993:2 I : 1347-l 352. 19. McEnroe CS, O’Donnell TP. Yeagcr A, Konalarn M, Mackey WC. Comparison of ejection fraction and Goldman risk factor analysis to dipyridamolc-thallium 201 studies in the evaluation of cardiac morbidity allcr aortic aneurysm surgery. J ihc surg 1990: I I :497-x11. 20. Cutler BS, Leppo JA. Dipyridamolc thallium 201 scintigraphy to detect corenary artcry disease lxfore abdominal aortic surgery. J VQSC Surg 19X7:5:91-100. 21. Mangano DT, London MJ. Tubau JF, Browner WS, Hollenberg M, Krupski W. Layug EL. Massie B, and the Study of Perioperativc Ischemia Kcscarch Croup. Dipyridamoie thallium-201 scintigraphy as a prc-operative scrcenin: test: a reexamination of its predictive potential. Circulahm I99 I :X4:493-502. 22. Baron JF, Mundler 0, Bertrand M, Vicaul E. Barr6 E, Godet G. Samarna CM, Coriat P, Kieffer E. Viars P. Dipyridamoie-thallium scintigraphy and gated radionuclidc angiography to assess cardiac risk before abdominal aortic surgery. N Engl J Med 1994;330:663-669. 23. Machecourt J. Longere P. Fagret D, Van~cllo ti, Wolf JE. Polidori C. Comet M. Denis B. Prognostic value of thallium-201 sin&photon emission computed tomographic myocardiai perfusion imaging according to ~xlcnl of myocardial defect. Study in 1,926 patients with follow-up at 33 months. J Am Co// Curdiol 19Y4;23: 1096-I 106. 24. Dilsizian V, Rooco T’P, Freedman NM. Leon .MB. Bonow RO. Enhanced detection of ischcmic hut viable myocardiurn by the rein.jection of thallium nltcr stressredistribution imaging. N &,$ J Med 1090;323:141-146. 25. Belier GA. New directions in mytrardial perfusion imaging. C/br Cardioio[ 1093: 16:8&94. 26. Hcndei KC, Chcn MH, L’Italien GJ, Newell JB, Paul SD, Eagb KA. Leppo JA. Sex differerences in periopcrative and long-twn cardiac evcnl-free suwival in vascular surgery patients. An analysis of clinical and scintigraphic variahkx Circulurion 1995;91:1014-1051, 27. Fintcl DJ, Links JM. Brinker JA, Frank TL, Parker hl, Becker LC. Improved diagnostic perCorniance of exercise thallium 201 single photon emission computed romography over planar imaging in the diagnosis or coronary artery disease: a receiver operating charactcdstic analysis. J Am Coil Curdiol 1989: I3:600-612. 28. Brown KA. Prognostic value of thallium-201 myocardial perlusion imaging: a d&nostic tool comes of age. Circulation 199 I $3:363-3X I. 29. Domanski M, Ellis S. Eagle K. Does preopcrativc ctronary rcvascularization before noncxdiac surgery reduce the risk of coronary events in patients with known comnary artery disease’? Am J Curdiol 1999;75:820-83 I. 30. Hertxr NR, Young JR, Beven EG. O’Hara PJ. Late results of coronary bypass in patients with infrarenal aortic ancur);sm. ‘The Cleveland Clinic Study. Ann Surg 1987;205:36&367i7.

148

JANUARY

THE AMERICAN

JOUZNAL

OF CARDIOLOGY’+

VOL. 77

15,

1996