|
|
| |||||||||||||||||||||||||||||||||||||||
Background Because many patients with atherosclerotic disease of the abdominal aorta also have coronary artery disease, assessment of cardiac risk before abdominal aortic surgery has received much attention. Our prospective study was designed to identify predictors of cardiac risk in consecutive patients evaluated preoperatively with dipyridamole-thallium single-photon-emission computed tomography (SPECT) to assess myocardial perfusion and radionuclide angiography to measure left ventricular ejection fraction.
Methods Clinical and scintigraphic data were collected prospectively during hospitalization in 457 consecutive patients undergoing elective abdominal aortic surgery. Adverse cardiac outcomes were predicted with multivariate analyses.
Results Eighty-six patients (19 percent) had one or more of the following postoperative complications: prolonged myocardial ischemia (61 patients), myocardial infarction (22), congestive heart failure (20), and severe ventricular tachyarrhythmia (2). Twenty patients died postoperatively (4.4 percent), half of them from cardiac causes. Information about myocardial perfusion obtained from dipyridamole-thallium SPECT did not accurately predict adverse cardiac outcomes. The best correlates of cardiac complications were definite clinical evidence of coronary artery disease (odds ratio, 2.6; 95 percent confidence interval, 1.6 to 4.3) and age greater than 65 years (odds ratio, 2.3; 95 percent confidence interval, 1.4 to 3.6). Measurement of the ejection fraction was useful only in the prediction of left ventricular failure. Age greater than 65 years was the only predictor of death (odds ratio, 26.4; 95 percent confidence interval, 3.5 to 200.0).
Conclusions The presence of definite clinical evidence of coronary artery disease and older age were the most important preoperative predictors of an adverse cardiac outcome after abdominal aortic surgery. These results suggest that the routine use of dipyridamole-thallium SPECT and radionuclide angiography for screening before abdominal aortic surgery may not be justified.
Given the controversy over dipyridamole-thallium scintigraphy, our prospective study was designed to determine the predictors of cardiac risk before elective abdominal aortic surgery. We were particularly interested in evaluating the usefulness of dipyridamole-thallium SPECT (to assess myocardial perfusion) and radionuclide angiography (to measure left ventricular ejection fraction) in the prediction of postoperative cardiac complications.
Methods
Patients
Over a three-year period, 513 patients referred to the Department of Vascular Surgery of the Hopital Pitie-Salpetriere for elective abdominal aortic reconstructive surgery were prospectively included in the study. Informed consent was obtained, and the protocol was approved by the local ethics committee.
Every patient was seen by a staff anesthesiologist, who performed a clinical examination. Previous myocardial infarction was diagnosed on the basis of the patient's history or from electrocardiographic evidence of myocardial infarction. Angina was defined as chest pain with at least three of the following characteristics: substernal location, precipitation by stress, duration of less than 15 minutes, and resolution after rest or nitroglycerin. We paid special attention to identifying ischemic ST-T abnormalities on the preoperative electrocardiogram recorded while the patient was at rest. We explicitly excluded nonspecific ST-T abnormalities. Definite coronary artery disease was indicated by a history of myocardial infarction or angina or by ischemic ST-T abnormalities on the electrocardiogram.
Owing to scheduling problems and technical failures, neither dipyridamole-thallium SPECT nor radionuclide angiography was performed in 45 patients, who were therefore excluded from the study. Indications for coronary angiography were based on clinical findings: new or medically unstable angina, disabling angina, recent myocardial infarction, or persistent angina with a previous myocardial infarction. Thirty-seven patients underwent coronary angiography, which revealed clinically important lesions in 32. In 11 patients, coronary artery disease was considered to be severe and correctable. Percutaneous transluminal coronary angioplasty was performed in five patients, and coronary artery bypass grafting in six. These 11 patients were also excluded from the study, leaving a total of 457 patients in the final analysis.
Dipyridamole-Thallium SPECT
Dipyridamole-thallium SPECT17 was performed 1 to 10 days before surgery. With the patient in the supine position while the electrocardiogram and blood pressure were monitored, dipyridamole was infused intravenously at a rate of 0.14 mg per kilogram of body weight per minute for four minutes. Three minutes after the infusion, 2.5 mCi of intravenous thallium-201 was injected. The initial images were recorded five minutes after the injection, and the delayed images three hours later. Myocardial uptake of thallium was assessed in 32 projections with a 180-degree tomographic acquisition gamma camera (Gammatome II, Sopha Medical, Buc, France). Background noise was automatically subtracted, and the images were filtered with a Butterworth 6/16 filter (Sopha Medical). Sagittal, coronal, and transverse sections were obtained after reconstruction by backward projection, with approximately 10 sections on each axis. The images were interpreted qualitatively by two independent observers. Each section was interpreted with a semiquantitative scoring system using three grades: grade 2 denoted normal uptake; grade 1, moderately decreased uptake; and grade 0, uptake reduced to the background level. Initial thallium scans showing moderate or total perfusion defects (grades 1 and 0, respectively) on two consecutive sections and two different projections were considered abnormal. From the three-dimensional cross sections, thallium abnormalities were classified according to a topographic system with eight areas. Perfusion defects were further categorized as showing redistribution if the defect appeared filled on the delayed images (i.e., if there was improvement by one grade or more) or as being persistent if they did not fill in. Differences in interpretation between observers were resolved by consensus. The two physicians involved in the interpretation were unaware of any preoperative and postoperative clinical information regarding the patients.
Gated Radionuclide Angiography
Radionuclide angiography was performed to measure left ventricular ejection fraction18. After a bolus injection of red cells labeled in vitro, data were collected with a large-field camera when the labeled cells were uniformly distributed in the circulation. Approximately 250 cardiac cycles were recorded with a temporal resolution obtained by gating the registered activity from the electrocardiogram. The cardiac cycles were cumulated to obtain a composite cycle in 12 frames. The end-systolic and end-diastolic frames were selected, and the left ventricular edges outlined. The ejection fraction was derived from end-diastolic and end-systolic counts by a standard formula. The measurements were repeated twice by two experienced physicians, and the mean results were used in our analysis.
Anesthetic and Postoperative Management
The research data were collected in parallel with the clinical data. The physicians caring for the patients had access to the research data and were aware of the clinical findings, the results of dipyridamole-thallium scanning, and any preoperative cardiac evaluation. Preoperative cardiac treatment was continued until two hours before surgery (Table 1). ST-T segments were continuously monitored (with leads DII, CS5, and V4) (Marquette 7010 Monitor, Milwaukee) in only 323 patients during the preoperative and intraoperative period, depending on the availability of the monitor. Standard electrocardiographic monitoring was used for the remaining patients. The unblinding of the data did not induce a selection bias, since no major difference in the use of ST-segment analysis was observed in relation to clinical symptoms or scintigraphic findings (Table 2). In all 457 patients, hemodynamic measures obtained from radial-artery and pulmonary-artery catheters were continuously monitored perioperatively.
|
|
Surgical Management
All patients underwent inclusion graft procedures involving the abdominal aorta. Suprarenal aortic cross-clamping was required in 74 patients. The prosthetic graft was a straight tube in 94 patients and a bifurcated graft in 363 patients. Associated vascular procedures are described in Table 2. The mean (±SD) duration of the surgical procedure was 3.8 ±1.6 hours.
Analysis of Clinical Outcomes
The major clinical outcome variables prospectively analyzed were mortality and major cardiac morbidity. Adverse outcomes were reported by the physicians caring for the patients and were validated separately by two investigators. Disagreements were resolved by consensus.
Mortality was defined as death occurring within 30 days after the surgical procedure. A postoperative cardiac complication was defined as the appearance of prolonged myocardial ischemia, myocardial infarction, congestive heart failure, or ventricular tachyarrhythmias. Prolonged myocardial ischemia was defined as a new ST-T abnormality (ST depression >1 mm at 60 msec after the J point) on at least two successive daily 12-lead electrocardiograms. We also considered patients with persistent T-wave inversion after an episode of ST-segment depression as having prolonged myocardial ischemia.
The diagnosis of postoperative myocardial infarction required diagnosis at autopsy, new Q waves on the electrocardiogram that were at least 0.04 second in duration and 1 mm in depth, or persistent ST-segment depression associated with an elevation of the serum creatine kinase MB isoenzyme (to >20 ng per milliliter). Congestive heart failure was defined as the need for sympathomimetic support (dobutamine or dopamine) during the postoperative period associated with hemodynamic and pulmonary findings consistent with the diagnosis: classic chest-film changes, pulmonary-capillary wedge pressure persistently greater than 18 mm Hg, or both. An impairment in global left ventricular function was assessed by transesophageal echocardiography as standard clinical practice to clarify all postoperative hemodynamic disorders. Ventricular tachyarrhythmia was defined as a documented ventricular tachycardia or fibrillation with hemodynamic compromise.
Statistical Analysis
Potential univariate correlates of adverse cardiac outcomes were identified by chi-square analysis or Fisher's exact test. Multivariate analysis was performed with stepwise logistic regression (Biomedical Data Processing Package, UCLA, Los Angeles). To avoid overestimating the number of predictive variables,21,22 we selected these variables with conservative criteria, as follows: limits to enter or remove variables in the regression equation must have had a 5 percent probability value; the ratio between the corresponding regression coefficient and its standard error must have been greater than 223; and results were verified by two numerical procedures, an asymptotic covariance estimate and the maximum-likelihood method. Univariate and multivariate odds ratios and their 95 percent confidence intervals were calculated from the model.
In addition, for the best estimate of the predictive powers of dipyridamole-thallium SPECT and radionuclide angiography, a complementary multivariate procedure was carried out. Fixed-defect scans, scans showing redistribution, and those with an ejection fraction of less than 50 percent were forced into the multivariate model, and adjusted odds ratios were recalculated from the resulting models.
Results
Characteristics of the Patients
A total of 468 patients (mean [±SD] age, 63 ±11 years) were initially enrolled in the study (11 patients with severe but correctable coronary artery disease were subsequently excluded). Of these, 202 were more than 65 years old. Definite coronary artery disease was observed in 163 patients (Table 1). Thallium defects were observed in 254 patients and were categorized as persistent in 94 and as showing redistribution in 160. The relation between clinical findings and data obtained by dipyridamole-thallium SPECT is shown in Table 1. The ejection fraction was below 50 percent in 75 patients and below 35 percent in 7. Intraoperative anesthetic and surgical management was not influenced by preoperative unblinded cardiac assessment with dipyridamole-thallium SPECT and radionuclide angiography (Table 2).
Clinical Outcomes
Eighty-six patients (19 percent) had at least one postoperative complication: prolonged myocardial ischemia (61 patients), myocardial infarction (22), congestive heart failure (20), or severe ventricular tachyarrhythmia (2). Twenty patients (4.4 percent) died postoperatively, five of myocardial infarction and five of left ventricular failure. Multiple-organ failure was the cause of death in the remaining 10 patients.
Univariate Analysis
The preoperative variables significantly associated with an adverse cardiac outcome in the univariate analysis (Table 3) were an age greater than 65 years, previous myocardial infarction, a history of angina, ischemic ST-T abnormalities, definite coronary artery disease, hypertension, and an ejection fraction of less than 50 percent. The variables obtained from dipyridamole-thallium SPECT -- a scan with a fixed defect, or one showing thallium redistribution -- were not univariate correlates of adverse cardiac outcomes. Univariate analyses of each cardiac complication and of postoperative mortality are shown in Table 3, Table 4, and Table 5.
|
|
|
The preoperative predictors of cardiac complications were definite coronary artery disease and age greater than 65 years (Table 3). In patients 65 or younger who did not have definite coronary artery disease (169 patients, or 37 percent), the rate of cardiac complications was 10 percent. In patients with one of the two predictors (288 patients, or 63 percent), the cardiaccomplication rate was 24 percent. In patients with both predictors (77 patients, or 17 percent), the cardiac-complication rate was 39 percent. Definite coronary artery disease was the only characteristic in the multivariate analysis that was associated with the occurrence of prolonged myocardial ischemia (Table 4). The preoperative predictors of postoperative myocardial infarction were definite coronary artery disease and age greater than 65 years (Table 4). Age greater than 65 years, previous myocardial infarction, and an ejection fraction below 50 percent were the predictors of postoperative congestive left ventricular failure (Table 4). Age greater than 65 years was the only characteristic associated with postoperative mortality in the multivariate analysis (Table 5).
The values of adjusted odds ratios when these variables were inserted into the model to obtain the best estimate demonstrate the weak predictive power of thallium scintigraphic data and ejection fraction (Table 3, Table 4, and Table 5).
Discussion
The principal finding of this study is that definite clinical evidence of coronary artery disease and age are the two most important variables correlated with adverse cardiac outcomes in patients undergoing abdominal aortic surgery. In contrast, the results of dipyridamole-thallium SPECT were not predictive of adverse cardiac outcomes.
The preoperative clinical characteristics and rates of cardiac complications in our patients were very similar to those in previous reports11,12,13,14,15,24. Patients with severe and correctable coronary artery disease were excluded from the analysis. However, indications for coronary angiography were based on predefined clinical criteria and did not take scintigraphic data into account. A total of 37 patients underwent coronary angiography, and 11 patients had myocardial revascularization procedures. The inclusion of these 11 patients did not influence the results of the statistical analysis.
Another issue is that unblinded preoperative cardiac evaluation could have influenced anesthetic management. Pulmonary-artery catheters were inserted in all patients, and intraoperative ST-T analysis was performed in most. The type of anesthetic technique used, either balanced general anesthesia or thoracic epidural anesthesia associated with light general anesthesia, was not dictated by the protocol. We recently showed in a large randomized study that these two techniques have similar influences on postoperative outcome19. High-dose narcotic anesthesia, which is frequently used elsewhere in patients with severe coronary artery disease or ventricular dysfunction, was not used in this study. The analgesic technique was not chosen according to the results of the preoperative cardiac evaluation. Although postoperative care was identical in all patients with regard to other treatments, we cannot rule out the possibility that the analgesic techniques could themselves have influenced the postoperative outcome.
The incidence of postoperative unstable angina could have been underestimated in our study. Since the diagnosis of prolonged myocardial ischemia was based only on the persistence of new ST-T abnormalities on at least two successive daily electrocardiograms, we cannot rule out the possibility that shorter episodes of unstable angina may have occurred. The incidence of ventricular dysrhythmias may also have been underestimated because of the absence of a continuous recording technique24.
Many studies have identified preoperative clinical markers that correlate with postoperative adverse cardiac outcomes. A history of myocardial infarction,25,26,27 angina pectoris,26 abnormal base-line electrocardiography,25,26,27,28 advanced age,28 hypertension,28 diabetes mellitus,28 congestive heart failure,25,28 and cardiac arrhythmias25,27 are the most frequently reported markers. Several clinical scoring systems27,29,30,31 have been identified with multivariate discriminant analysis. Although this point was not investigated in this study, it has recently been stressed that intraoperative and postoperative variables could be major predictors of postoperative adverse cardiac outcomes. In the study by Mangano et al.,24 postoperative ischemia was a more powerful correlate. In a study of 254 patients scheduled for vascular surgery, Eagle and coworkers15 used multivariate analysis to identify five clinical factors that were independent correlates of postoperative cardiac events. These markers were advanced age (>70 years), Q waves on the electrocardiogram, diabetes mellitus, a history of ventricular arrhythmias, and a history of angina. Except for diabetes mellitus, these clinical markers are very similar to those we identified. The higher prevalence of diabetes mellitus in the study by Eagle et al. -- 18 percent, as compared with 8.5 percent in our study -- is probably responsible for the difference. Their inclusion of one third of patients with peripheral vascular surgery (femoropopliteal bypass) could explain the higher prevalence of diabetes mellitus in their study.
Several studies have suggested that dipyridamole-thallium scintigraphy is the ideal preoperative screening test for evaluating the risk of postoperative myocardial infarction11,12,13,14,15. Thallium redistribution has been shown to be the most important predictor. In comparison, we enrolled more patients and did not confirm the ability of thallium scanning to predict the risk of cardiac complications. These results support those of Mangano et al.,16 who reported in a recent prospective study of 60 patients undergoing vascular procedures that there was no association between redistribution defects and adverse cardiac outcomes. Some aspects of this controversy have been discussed by Mangano et al. Most previous studies did not enroll consecutive patients. Another issue is the nonstandardization of monitoring and of the use of anesthetic and surgical techniques, mainly because physicians were not blinded to the scintigraphic data. Although Mangano et al. overcame these limitations, they studied a limited number of patients and used planar dipyridamole-thallium scintigraphy, not SPECT. Some technical aspects of planar dipyridamole-thallium scintigraphy and its reliability in detecting coronary artery disease have already been questioned32,33,34. Better identification of artifactual defects can be expected with SPECT, although it does not perfectly distinguish fixed defects from those with redistribution35. These limitations are likely to explain the weak relation observed between clinical findings and dipyridamole-thallium scans and may partly explain the lack of prediction of cardiac outcome.
Several studies have demonstrated a higher risk of postoperative morbidity and mortality in patients with impaired preoperative left ventricular function9,36,37. Kazmers et al.37,38 reported conflicting results with regard to postoperative mortality. In our study, left ventricular ejection fraction as assessed by radionuclide angiography was a significant predictor only of postoperative left ventricular failure.
Conclusions
As compared with most recent studies, our study does not confirm that dipyridamole-thallium SPECT is an accurate screening test for the evaluation of cardiac risk before abdominal aortic surgery. The value of measuring the ejection fraction is limited to the prediction of postoperative left ventricular failure. Definite clinical evidence of coronary artery disease and age greater than 65 years are the most important preoperative predictors of adverse cardiac outcomes. On the basis of these data, we question whether the substantial cost of preoperative assessment with dipyridamole-thallium scintigraphy and radionuclide angiography is warranted.
Source Information
From the Departments of Anesthesiology (J.-F.B., M.B., E.B., G.G., C.M.S., P.C., P.V.) and Vascular Surgery (E.K.), Hopital Pitie-Salpetriere; the Laboratory of Nuclear Medicine, Hopital Lariboisiere (O.M.); and the Laboratory of Biophysics and INSERM, Unite 141, Hopital Fernand Widal (E.V.) -- all in Paris.
Address reprint requests to Dr. Baron at the Departement d'Anesthesie-Reanimation, Hopital Broussais, 96, rue Didot, 75014 Paris, France.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Assessment of Cardiac Risk before Abdominal Aortic Surgery
Eagle K. A., Cambria R., Coley C., Abbott W., Baron J.-F.
Extract |
Full Text
N Engl J Med 1994;
331:480, Aug 18, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
