Failure of negative dipyridamole thallium scans to predict perioperative myocardial ischaemia and infarction

A Meta-Analytic Comparison of Preoperative Stress Echocardiography and Nuclear Scintigraphy Imaging

In this meta-analysis we compared thallium imaging (TI) and stress echocardiography (SE) in patients at risk for myocardial infarction (MI) scheduled for elective noncardiac surgery. Two searches of published articles were used to identify relevant articles. We included all studies that stated the criteria for a positive test and detailed the frequency of postoperative MI and in-hospital death. Data were abstracted by two authors and captured preoperative patient characteristics, study design, blinding, and outcome adjudication. We defined a positive test as a test with a reversible defect and, where possible, quantified the size of the defects in each study. MI and/or death were the only postoperative outcomes of interest. We calculated the sensitivity, specificity, and likelihood ratio (LR) and, where possible, the Receiver Operating Characteristic (ROC) curve of a cardiac event in each study. The LR and ROC were combined by meta-analyses using the random effects model. Heterogeneity was assessed using the I2 test. The search revealed 68 studies of 10,049 patients. There were 25 SE studies and 50 TI studies. There were 7 studies with a direct comparison of the two methodologies. The quality of studies differed; routine screening for MI was used more frequently in SE studies (47.8% versus 21.2%; P = 0.008) and screening dictated treatment more often after TI (72.1%) than after SE (46.3%) (P = 0.027). The LR for SE was more indicative of a postoperative cardiac event than TI (LR, 4.09; 95% CI, 3.21-6.56 versus 1.83; 1.59-2.10; P = 0.001). This difference was attributable to fewer false-negative SEs. There was no difference in the cumulative ROC curves from qualitative studies (SE, 0.80; 95% CI, 0.76-.84 versus TI, 0.75; 95% CI, 0.70-081). Again, the LR for a negative SE was less (0.23; 95% CI, 0.17-0.32 versus 0.44; 95% CI, 0.36-0.54). A moderate-to-large defect, seen in 14% of patients, by either method predicts a postoperative cardiac event (LR, 8.35; 95% CI, 5.6-12.45). This meta-analysis possesses the statistical power to demonstrate that SE has better negative predicative characteristics than TI. A moderate-to-large perfusion defect by either SE or TI predicts postoperative MI and death. We conclude the SE is superior to TI in predicting postoperative cardiac events.

Use of positron emission tomography for prediction of perioperative and late cardiac events before vascular surgery

The efficacy of myocardial perfusion imaging for cardiac-risk stratification of patients undergoing vascular surgery has been disputed recently. In comparison with conventional techniques, positron emission tomography (PET) has the benefit of permitting a true resting scan, allows accurate measurement of the extent of ischemia, and is highly specific for the diagnosis of coronary disease. We therefore investigated the use of PET for risk stratification at the time of vascular surgery and subsequent follow-up in 78 patients (aged 67 +/- 11 years, 52 men), selected for testing before the performance of extensive surgery or because of one or more clinical risk factors. Perfusion images were obtained by using a standard rubidium 82 protocol before and after dipyridamole-handgrip stress. With use of a quantitative color scale in a 24-segment model of the left ventricle, scans were reported as showing normal perfusion, resting defects, or stress-induced defects (deterioration > 15% with stress). After exclusion of 6 patients referred for myocardial revascularization, 72 patients were followed up in the perioperative period and for 18 +/- 12 months for late cardiac death, myocardial infarction, or unstable angina. Perioperative events occurred in 14 patients (5 with myocardial infarction and 9 with unstable angina), 10 of whom had ischemia at PET (sensitivity, 71%; predictive value of a positive test, 45%). Isolated resting perfusion defects were not associated with events. The presence of extensive ischemia (more than five segments) had a positive predictive value of 64%, and its absence gave a negative predictive value of 89%.(ABSTRACT TRUNCATED AT 250 WORDS)

Current practice in the preoperative evaluation of patients undergoing major vascular surgery: a survey of cardiovascular anesthesiologists

Various cardiovascular testing modalities have been proposed as a means of reducing perioperative risk in patients undergoing vascular surgery. It is currently unknown with what frequency and under what conditions these tests are used in clinical practice. In an attempt to determine the spectrum of current clinical practice, the authors surveyed anesthesiologists who attended the 1992 Society of Cardiovascular Anesthesiologists Annual Meeting. A total of 77 of 100 surveys were returned, representing 58 institutions in the United States: 26 university hospitals, 29 private hospitals, 2 Veterans Administration hospitals, and 1 military hospital. The incidence of preoperative testing was not significantly different between private and university institutions, and ranged from a mean of 54% of patients undergoing lower extremity procedures to 73% of patients undergoing aortic surgery, with a wide disparity among institutions. Dipyridamole or exercise thallium imaging was the most common test used, and Holter monitoring was the least common. The testing led to revascularization in 52% of university and 72% of private hospitals. Testing was thought to result in modification of perioperative monitoring in 85% of the institutions and in modification of other aspects of care in a significant percentage of institutions. In conclusion, in this random population, preoperative testing was reported to be used in a significant percentage of patients, and similar practices were found between university and private institutions. The practices at the individual institutions and types of tests ordered varied greatly, and suggest the need for further research to determine optimal preoperative strategies.