Myocardial Perfusion Imaging vs. Troponin I in Acute MI


Synopsis: It appears that we still are searching for the optimal tests (and their timing) to predict the major ACS in ED patients with chest pain.

Source: Kontos MC, et al. Circulation 1999;99:2073-2078.

Investigators from the medical college of virginia in Richmond reviewed data from 721 patients who were admitted to the coronary care unit from the emergency department (ED) and who were at low-to-moderate risk for acute coronary syndrome (ACS). The study cohort consisted of patients who had had both early gated rest myocardial perfusion imaging (PI) with 99mTc sestamibi and serial serum cardiac troponin I (cTnI) sampling. The purpose of the study was to compare the sensitivity and specificity of these two diagnostic tests for predicting MI within one week of admission, significant coronary disease on angiography within six weeks of admission, and performance of revascularization (CABG or PTCA).

Of the final cohort of 620 patients, MI was diagnosed in 59 patients (9%) using CK-MB and relative index as the gold-standard criteria. Sensitivity for detecting MI was not significantly different between perfusion imag-ing (92%) and serial cTnI (90%), and both were signifi-cantly higher than the initial cTnI (39%). The specificity for MI was 67% for PI and 96% for serial cTnI. PI iden-tified many more patients than cTnI who subsequently underwent revascularization or who had significant coronary disease. PI had a lower specificity for all end points. Lowering the cutoff value of cTnI from 2.0 ng/mL to 1.0 ng/mL did not significantly change the results.

Comment by Stephanie Abbuhl, MD

This is essentially a retrospective study with several design flaws. Only 140 of 620 patients had angiography, leaving the potential for significant bias in conclusions about the end points of significant coronary disease and revascularization. In addition, 71 patients were excluded from the final cohort because only initial cTnI sampling had been done and, therefore, the final sensitivity and specificity calculations to predict all end points may be in error. Despite these limitations, there are some points to be garnered from this study. The low sensitivity to detect MI of the initial cTnI (39%) is a cautious reminder of why we must avoid the temptation to use this single determination to make decisions about admission or discharge for potential MI patients. Kontos and colleagues point out that troponin values may be negative in patients with MI for a number of reasons, including size of the infarct, timing of the sampling (optimal sensitivity is 8-12 hours after onset of necrosis), the threshold for abnormality, and the particular "gold standard" used to define MI.

At first glance, early PI appears promising, with a sensitivity of 92% to predict MI, and the test results were available about three hours after the initial ED evaluation. However, there were many false positives and the cost of PI was not addressed. While Kontos et al conclude that the information from these two tests is complementary, it appears that we still are searching for the optimal tests (and their timing) to predict the major ACS in ED patients with chest pain. (Dr. Abbuhl is Medical Director, Department of Emergency Medicine, The Hospital of the University of Pennsylvania; Associate Professor of Emergency Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA.)

All of the following were true about early perfusion imaging (PI) in the Kontos study of ED patients at low-to-moderate risk for acute coronary syndromes except:

a. PI had a sensitivity of 92% to predict MI.

b. PI had a sensitivity very similar to the initial cardiac troponin I.

c. PI had a lower specificity for predicting MI than serial cardiac troponin I.

d. PI results were obtained within about three hours of initial ED evaluation.