Correlation between clinical outcomes and appropriateness grading for referral to myocardial perfusion imaging for preoperative evaluation prior to non-cardiac surgery

Background. In a previous study, we have found that referral to myocardial perfusion imaging (MPI) for preoperative evaluation of patients before non-cardiac surgery was the most common cause of inappropriate referral based on AHA/ACC Appropriate Use Guidelines, though 40% of scans graded as inappropriate had abnormal MPI results. The aim of this study was to correlate appropriateness grading with (1) the outcome of MPI scans, and (2) the clinical outcome of patients after surgery, so as to determine if the predictive value of MPI was related to appropriateness grading. Methods. All consecutive patients referred to the MPI laboratory of our center from March 2009 to July 2009 for preoperative risk stratification were prospectively studied. Patients' medical records and stress data were collected, and all imaging results were recorded. Based on appropriate use criteria (AUC), MPI studies were classified into appropriate, inappropriate, uncertain or unclassified. MPI studies were classified on the basis of their results into normal or abnormal scans. Primary clinical outcome measured was the occurrence of any major cardiac and cerebrovascular event during follow-up up to 90 days. Results. There were 176 referrals for preoperative evaluation. 39.8% (n = 70) of these referrals were graded as inappropriate. Based on AUC, referrals for MPI in intermediate-and high-risk groups with poor functional class were graded as appropriate, while referrals for MPI in low-risk and intermediate risk groups with normal functional class were graded as inappropriate referrals. The overall event rate was 6.25%. Cumulative death rate at 90 days was 1.7%, non-fatal MI 4%, and occurrence of stroke occurred in 0.6%. The primary outcome was higher in the intermediate group with poor functional class (13%) and high-risk group (64%) than the low-risk and intermediate risk groups with good functional class (4%). In the high-risk group, an abnormal MPI scan was associated with an extremely high event rate (50%) that was significantly greater than the event rate in patients with a normal MPI result (14%) (P = .01). Although 40% of preoperative low-risk and intermediate risk patients had an abnormal MPI result, their overall event rates were low (0 and 4%, respectively), with no difference in the rate of events between patients with normal and abnormal MPI scans in the inappropriate group. Conclusions. In conclusion, in correlating the relationship between appropriateness grading by AUC with the outcome of MPI scans and subsequent event rates, we found that MPI results predicted outcome in appropriately tested patients, but not in patients whose tests were classified as inappropriate, in whom event rates were low, regardless of the results of testing. Our findings support the AUC recommendations for the selective role of testing in preoperative risk stratification, which use the type of surgery and functional class to determine the appropriateness of referral. (J Nucl Cardiol 2012;19:277-84.)