Prediction of major cardiac events after peripheral vascular surgery using dipyridamole echocardiography

Prediction of Major Cardiac Events After Peripheral Vascular Surgery Using Dipyridamole Echocardiography Marc D. Tischler, MD, Thomas H. Lee, MD, Alan T. Hirsch, MD, Christopher P. Lord, BA, Lee Goldman, MD, Mark A. Creager, MD, and Richard T. Lee, MD

Patients undergoing peripheral vascular surgery are at increased risk of postoperative cardiac complications. To evaluate the role of dipyridamole echocardiography in predicting major cardiac events, 109 unselected patients undergoing elective peripheral vascular surgery were prospectively studied. Preoperative dipyridamole echocardiograms were interpreted by an echocardiographer unaware of all clinical data. Patients were followed up until hospital discharge by research physicians without knowledge of dipyridamole echocardiography results. Outcomes were classified using strict predefined criteria by reviewers unaware of other clinical and echocardiographic data. Of the 109 patients, 9 (8%) had positive studies defined as development of new regional wall motion abnormalities or worsening of preexistent wall motion abnormalities. Of these 9 patients, 7 had postoperative events, including 3 cardiac deaths, 1 nonfatal myocardial infarction, 2 with unstable angina, and 1 with pulmonary edema. Only 1 event occurred among the 100 patients with negative studies. The sensitivity and specificity of dipyridamole echocardiography for predicting cardiac events after vascular surgery were 88 and 9S%, respectively; the positive and negative predictive values were 78 and 99%. The relative risk of having a cardiac event if dipyridamole echocardiography was abnormal was 78 (95% confidence interval, 11

to 594; p
atients undergoing peripheral vascular surgery are at increasedrisk for postoperativecomplications.I-4 Preoperative assessmentof cardiovascular risk is difficult in these patients becauseof the high frequency of conditions that preclude conventional exercise testing. Specialized preoperative tests such as dipyridamole thallium scintigraphy3-6 and ambulatory electrocardiographicmonitoring 2,7have been highly sensitive in predicting perioperative cardiac events. However, becauseof the relatively low specificity and positive predictive value of these tests, a significant number of patients may have positive test results and not have a postoperativeevent. Such patients may be referred for unnecessaryinvasive testing and revascularization procedures,which may carry high risks becauseof coexistentvascular and renal disease.In a population with a relatively low predicted event rate, a screeningtest should have, in addition to high sensitivity, sufficient specificity and positive predictive value to minimize the number of false-positivestudies. Dipyridamole echocardiography has been reported to be a highly specific test for detecting coronary artery disThe aim of this study was to test the hypotheFrom the Cardiovascular Division and the Divisions of Clinical Epide- ease.8-13 sis that dipyridamole echocardiography,by virtue of its miology and General Medicine, Brigham and Women’sHospital, Harvard Medical School, Boston, Massachusetts.Dr. Richard T. Lee is a high specificity for coronary artery disease,is an inderecipient of Physician Scientist Award HL-01835 and Dr. Mark A. pendent predictor of perioperative cardiac eventsin paCreager is a recipient of Research Career Development Award HL01768 from the National Heart, Lung, and Blood Institute, Bethesda, tients undergoing noncardiac vascular surgery. Maryland. Dr. Thomas H. Lee is a recipient of an EstablishedInvestigator award (900119) from the American Heart Association, Dallas, Texas. Dr. Alan Hirsch is a recipient of an individual National Research Service award (F32-HL-07702), Bethesda,Maryland. Manuscript receivedMarch 14,199l; revisedmanuscriptreceivedand accepted April 29, 1991. Address for reprints: Richard T. Lee, MD, Cardiovascular Division, Brigham and Women’s Hospital, Harvard Medical School,Boston, Massachusetts02115.

P

METHODS Patient population: One hundred thirty-one patients scheduledto have vascular operations- excluding thoracic, venous and emergency procedures- from September 1989 to July 1990 were prospectively screened. Fifteen patients could not be enrolled becausedipyrida-

DIPYRIDAMOLE ECHOCARDIOGRAPHY

593

TABLE

I Clinical

Correlates

of Postoperative Event Occurred (n = 8)

No Event (n = 101,93%)

pValue

69 + 8 513

67 2 9 61/40

NS NS

1 4 1 2 3 3 2 4 6 5 2 0 5 2 1 0 7 4

35 (32%) 42 (39%) 10 (9%) 14 (13%) 31(28%) 29 (27%) 49 (45%) 37 (34%) 64 (59%) 27 (25%) 12 (11%) 16 (15%) 32 (29%) 16 (15%) 8 (7%) 2 (2%) 67 (62%) 70 (64%)

NS NS NS NS NS NS NS NS NS SO.05 NS NS NS NS NS NS NS NS

Correlate Age (mean ? SD) Men/women Surgery Aortic Femoral Carotid Other p blocker Calcium antagonist Past smoking Present smoking Systemic hypertension Diabetes mellitus Angina pectoris Congestive heart failure History of myocardial infarction History of CABG or PTCA Rales S3 gallop Abnormal ECG Q wave on KG

Events

Exceptfor age, all variables expressedas absolute number (percent of total). CABG = coronary artery bypass graft; ECG = electrocardiogram; NS = not significant; PTCA = percutaneoustransluminal coronaty angioplasty; SD = standard deviation.

mole was not available at the time of the scheduled operation. The protocol excluded patients with angina at rest (n = 2) myocardial infarction within 2 weeksof screeningand failure to give informed consent (n = 5). The remaining 109 patients comprisedthe study group. Each patient had a detailed history and physical examination by research physicians. Each patient was independently evaluated preoperatively by a staff cardiologist who judged the surgical risk to be acceptable.Caffeine and theophylline preparations were withheld for a minimum of 24 hours before testing. Dipyridamole echocardiograms: All patients had baseline 2-dimensional echocardiograms immediately after peripheral angiography. Dipyridamole was infused intravenously through an antecubital catheter at a rate of 0.14 mg/kg/min for 4 minutes with continuous blood pressure and electrocardiographic monitoring. Six minutes after the beginning of the infusion, a repeat 2-dimensional echocardiogram was recorded. If the patient remained asymptomatic and there was no evidenceof ischemia, an additional 0.14 mg/kg/min of dipyridamole was infused for 2 minutes, bringing the total doseto 0.84 mg/kg. At this time, a third 2-dimensional echocardiogram was recorded. Two-dimensional echocardiograms were recorded with a Hewlett-Packard 77020 AC/AR phased-array ultrasonoscopedevice using a 2.5-MHz transducer. Images were obtained from the parasternal short-axis, parasternal long-axis, apical %-chamberand apical 4-chamber views. In cases in which parasternal short-axis images were not ade594

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

68

quate, short-axis images were obtained from the subcostal view. All echocardiogramswere evaluated preoperatively by a single echocardiographer unaware of the clinical and electrocardiographic responsesto dipyridamole and all other clinical information. All echocardiogramswere then analyzed postoperatively by a second blinded echocardiographer in order to examine interobserver variability. In keeping with current recommendations,14 each echocardiographerhad previous experienceinterpreting > 100 stressechocardiographic studies. Systolic ventricular function was qualitatively graded as normal, mildly reduced, moderately reduced, or severely reduced. The presenceof left ventricular hypertrophy was assessedfrom the 2-dimensional echocardiogram using American Society of Echocardiography standards.l 5 Left ventricular segmental anatomy was identified using standard nomenclature.l6 Segmentswere scoredfor qualitative segmentalwall motion (0 = normal; 1 = hypokinesia; 2 = akinesia; 3 = dyskinesia). Studies were classified as positive based only on the development of ( 1) a new transient regional wall motion abnormality in a region normal at rest; or (2) marked worsening of a previous regional dyssynergy by either an increase in severity (hypokinesia to dyskinesia) or extension of a baseline abnormality to adjacent segments.13~*‘-igThe clinical and electrocardiographic responseto dipyridamole was recorded but was not a factor in determining test positivity. Minor worsening of a previous regional dyssynergy (hypokinesia to akinesia or akinesia to dyskinesia) or equivocal degreesof hypokinesia were not recorded. Postoperative follow-up: After surgery, patients were followed until hospital discharge by researchphysicians unaware of echocardiographic results. Patients underwent 12-lead electrocardiography before their operations and on postoperativedays 1 through 3. Serum creatine kinase levels were measuredwith MB fractionation20every 8 to 12 hours on postoperativedays 1 and 2. Classification of postoperativeevents was performed using strict predelined criteria by an independentphysician unaware of other clinical and echocardiographic data. Cardiac eventswere defined as: death from cardiac causes;nonfatal myocardial infarction (new Q waves >0.03 second, or creatine kinase >2.7 umol/s/liter with MB >5%); unstable angina (characteristic chest pain and >_1 mm of ST-segment depression in >2 leads); or pulmonary edema confirmed by chest radiographic findings. For unstable angina and pulmonary edema, management in an intensive care unit or a change in medication was required. Statistical analysis: In the univariate analysis, associations betweenmajor cardiac eventsand possibleclinical and test predictors were examined using the chi-

SEPTEMBER

1, 1991

I

TABLE III Dipyridamole Echocardiographic Postoperative Events

TABLE II Clinical Response to Dipyridamole Tests

Mean increase in heart rate (beats/min) Mean decrease in systolic blood pressure (mm Hg) Chest pain ST depression Aminophylline administered

Positive Negative (n = 9, 8%) (n = 100,92%)

pValue

ECG Variables

12 rt 6

13 -f 8

NS

17 f 7

17 i- 9

NS

3 6 9/9

3 (3.0%) 3 (3.0%) 121100

SO.05 10.01 NS

Left ventricular hypertrophy Left ventricular dysfunction Resting regional wall motion abnormality Inducible regional wall motion abnormality Dipyridamole responses Chest pain ST depression

Chest pain and ST depression expressed as absolute number (% of column), NS = not significant.

Correlates of

Event Occurred (n = 8)

No Event (n = 101, 93%)

5 3 6

55 (54.5%) 19 (18.8%) 36 (35.6%)

I-----2 3

pValue

NS NS 10.05

2 (2.0%)

s 0.0001

4 (4.0%) 9 (8.9%)

5 0.05 so.01

All variables expressedas absolute number (% of column). ECG = echocardiographic; NS = not srgnificant.

square test with appropriate degreesof freedom for categoric variables and the Student’s t test for continuous variables. The clinical variables consideredare listed in Table I. Variables from dipyridamole echocardiography included the presence of systolic left ventricular dysfunction, baseline regional wall motion abnormalities, development of chest pain, or 11 mm of ST depressionsafter dipyridamole infusion, and the parameter of interest, inducible regional wall motion abnormalities. The relative risk of an event was assessedfor patients with or without positive results on dipyridamole echocardiogramsand comparedby chi-square statistic (all p values are 2-tailed). A kappa statistic was calculated to assessthe statistical significance of interobserver agreement. RESULTS

Regional wall motion abnormalities were evident on the baseline 2-dimensional examination of 42 patients, 6 of whom had cardiac events (p CO.05). Baselineleft ventricular systolic function was normal in 87 patients, mildly reduced in 8, moderately reduced in 7 and severely reduced in 14. Of the 8 patients with postoperative cardiac events, 5 had normal systolic function and 1 had mild and 2 had moderately reduced function (p = not significant). Dipyridamole echocardiographywas abnormal in 9 patients, in each caseonly after the seconddosing interval. Six patients (5.5%) complained of chest pain during dipyridamole infusion, 3 of whom had dipyridamole-induced wall motion abnormalities (Table II). Nine patients (8.3%) had ST depressionsof L 1 mm in 1 or more leads during dipyridamole infusion; 6 of these patients had dipyridamole-induced wall motion abnormalities. Patients experiencedan averageincrease in heart rate of 13 f 8 beats/min and an average decrease in systolic blood pressure of 17 f 9 mm Hg (Table II). Twenty-one patients received intravenous aminophylline, 12 because of noncardiac side effects and 9 becauseof evidenceof myocardial ischemia. Postoperative cardiac events: Postoperative events occurred in 8 patients. Seven of 8 cardiac events occurred within 48 hours of surgery; 1 event occurred 77 hours after surgery. There were 3 cardiac deaths, 1 nonfatal myocardial infarction, 3 casesof unstable angina, and 1 case of pulmonary edema. Of the clinical variables examined (Table I), only diabetes had a statistically significant associationwith the occurrence of postoperative cardiac events (p
Patient characteristics (Table I): Patients were aged 67.6 f 8.9 years (mean f standard deviation) and 66 (61%) were men. Ninety-two patients (84%) reported past or current cigarette use, 32 (29%) were diabetic, and 70 (64%) had a history of hypertension. Thirtyseven patients (34%) had a history of myocardial infarction and 18 (17%) had had percutaneoustransluminal coronary angioplasty or coronary artery bypass surgery. Thirty-four patients (3 1%) were taking a 0 blocker and 32 (29%) were taking a calcium antagonist at the time of operation. There were 36 (33%) abdominal aortic procedures, 46 (42%) femoral artery procedures, 11 ( 10%) carotid endarterectomiesand 16 ( 15%) other vascular procedures. Dipyridamde echocardiograms: Adequate acoustic windows were obtained in all patients. Interobserver reproducibility for interpretation of dipyridamole echocardiograms was excellent, with agreement as to the presence or absence of significant new regional wall motion abnormalities in 108 of 109 studies (99%; kappa statistic = 0.94). The only disagreementoccurred on Correlation of dipyridamde echocardiography with a study that was interpreted as negative by the primary postoperative cardiac events: Postoperative cardiac reader and positive by the secondaryreader. The index events were found more often in patients with abnorsubject had an uneventful postoperative course. mal dipyridamole echocardiograms(p
595

pain during dipyridamole infusion (p <0.05), ST depression during dipyridamole infusion (p
2.8%) and a relatively low risk of seriouscardiac events (8 of 109; 7.3%). These rates are similar to those previously reported in several large series.1T2,4 Preoperative dipyridamole echocardiography was both sensitive (88%) and specific (98%). Several other test variables were also found to have statistically significant correlations with the occurrence of postoperative cardiac events, including baseline regional wall motion abnormalities (p
THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 68

SEPTEMBER 1, 1991

merit.Ann Surg 1984;199:223-233. 2. Raby KE, Goldman L, Creager MA, Cook EF, Weisberg MC, Whittemore AD, Selwyn AP. Correlation between preoperative ischemia and major cardiac events after peripheral vascular surgery. N Engl J Med 1989;321:1296-1301. 3. Eagle KA, Coley CM, Newell JB, Brewster DC, Darling RC, Strauss HW, Guiney TE, Boucher CA. Combining clinical and thallium data optimizes preoperative assessment of cardiac risk before major vascular surgery. Ann Intern Med 1989;l l&859-866. 4. Lane SE, Lewis SM, Pippin JJ, Kosinski EJ, Campbell D, Nesto RW, Hill T. Predictive value of quantitative dipyridamole-thallium scintigraphy in assessing cardiovascular risk after vascular surgery in diabetes mellitus. Am J Cardiol 1989;64:1275-1279. 5. Eagle KE, Singer DE, Brewster DC, Darling RC, Mulley AG, Boucher CA. Determination of cardiac risk by dipyridamole-thallium imaging before peripheral vascular surgery. N Engl J Med 1985;312:389-394. 6. Hendel RC, Layden JJ, Leppo JA. Prognostic value of dipyridamole thallium scintigraphy for evaluation of ischemic heart disease. J Am Co11 Cardiol 1990;15:109-116. 7. Pasternack PF, Grossi EA, Baumann FG, Riles TS, Lamparello PJ, Giangola G, Primis LK, Mintzer R, Imparato AM. The value of silent myccardial ischemia monitoring in the prediction of perioperative myocardial infarction in patients undergoing peripheral vascular surgery. J Vase Surg 1989;10:617-625. 8. Picano E, Lattanzi F, Masini M, D&ante A, L’Abbate A. High dose dipyridamole echocardiography test in effort angina pectoris. J Am Coil Cardiol 1986;8:848-854. 9. Picano E. Dipyridamole-echocardiography test: historical background and physiologic basis. Eur Heart J 1989;10:3655375. 10. Picano E, Lattanzi F, Masini M, Distante A, L’Abbate A. Comparison of the high-dose dipyridamole echocardiography test and exercise 2-dimensional echocardiography for diagnosis of coronary artery disease. Am J Cardiol 1987; 59~539-542. II. Picano E, Severi S, Michelassi C, Lattanzi F, Masini M, Orsini E, Distante A, L’Abbatte A. Prognostic importance of dipyridamole-echocardiography test in coronary artery disease. Circulation 1989;80:450-457. 12. Picano E, Lattanzi F, Masini M, Distante A, L’Abbate A. Different degrees of ischemic threshold stratified by the dipyridamole-echocardiography test. Am J Cardiol 1987;59:71-73.

13. Bologncse L, Sarasso G, Aralda D, Bongo A, Rossi L. High dose dipyridamole echocardiography early after uncomplicated myocardial infarction: correlation with exercise testing and coronary angiography. J Am Co11 Cardiol 1989;14:357-363. 14. Picano E, Lattanzi F, Orlandini A, Marini C, L’Abbate A. Stress echocardiography and the human factor: the importance of being expert. J Am Co/l Cardiol 1991;17:666-669. 15. Sahn DJ, DeMaria A, Kisslo J, Weyman AE. Recommendations regarding quantification in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978;58:1072-1083. 16. Weyman AE. A modified segmental system for describing left ventricular function. In: Weyman AE, ed. Cross-Sectional Echocardiography. Philadelphia: Lea & Febiger, 1982493-496. 17. Mason SJ, Wiss JL, Weisfeldt M, Garrison JB, Fortuin NJ. Exercise echocardiography: detection of wall motion abnormalities during ischemia. Circulation 1979;59:50-59. 18. Wann LS, Faris JV, Childress RW, Dillon JC, Weyman AE, Feigenbaum H. Exercise cross-sectional echocardiography in ischemic heart disease. Circulation 1979;60:1300-1308. 19. Labovite AJ, Pearson AC, Chaitman BR. Doppler and Z-dimensional echocardiographic assessment of left ventricular function before and after intravenous dipyridamole stress testing for detection of coronary artery disease. Am J Cardiol 1988;62:1180-1185. 20. Rosalki SB. An improved procedure for serum creatine phosphokinase determination. J Lab Clin Med 1967;69:696-705, 21. Boucher CA, Brewster DC, Darling RC, Okada RD, Strauss HW, Pohost GM. Determination of cardiac risk by dipyridamole-thallium imaging before peripheral vascular surgery. N Engl J Med 1985;312:389-394. 22. Leppo J, Plaja J, Gionet M, Tumolo J, Paraskos JA, Cutler BS. Noninvasive evaluation of cardiac risk before elective vascular surgery. J Am Co11 Cardiol 1987;9:269-276. 23. Eagle KE, Singer DE, Brewster DC, Darling RC, Mulley AG, Boucher CA. Dipyridamole-thallium scanning in patients undergoing vascular surgery: optimizing preoperative evaluation of cardiac risk. JAMA 1987;257:2185-2189. 24. Tischler MD, Lee RT, Lee TH, Creager MA, Lord C, Hirsch AT, Raby K. Dipyridamole echocardiography versus ambulatory ischemia monitoring in the assessment of perioperative risk (abstr). J Am Co/l Cardiol 1991;17:264A.

DIPYRIDAMOLE ECHOCARDIOGRAPHY

597