Electron Beam Computed Tomography for the Diagnosis and Prognosis of Coronary Artery Disease
To the Editor:
It seems that the committee responsible for the American College of Cardiology/American Heart Association Expert Consensus Document on Electron Beam Computed Tomography for the Diagnosis and Prognosis of Coronary Artery Disease1 conducted its deliberations and reached its conclusions about the prognostic accuracy of coronary calcium screening on the basis of an honest but grave misunderstanding of the St Francis Hospital data. At 1.6 years of follow-up, the odds ratio for the prediction of cardiac death or nonfatal myocardial infarction was 22 (95% CI, 2.7 to 170; P=0.004) for a calcium score threshold ≥100, not 2, and 30 (95% CI, 3.7 to 245; P<0.002) for a calcium score threshold ≥160.2 At 3.6 years of follow-up, the odds ratio for a calcium score threshold ≥160 was 22.2 (95% CI, 6.4 to 77.4).3 4
Given the Committee’s stated reluctance to extrapolate from autopsy, angiographic, and other indirectly supportive evidence, the argument for or against electron beam CT coronary calcium screening boils down to a comparison of the St Francis Hospital and South Bay Heart Watch data. The inadvertent misrepresentation of the St Francis Hospital experience with electron beam CT scanning is compounded by a failure to discuss the limitations of the South Bay Heart Watch. To qualify for entry into the South Bay Heart Watch, subjects had to be at or above the 75th percentile for coronary artery disease risk by Framingham criteria. The observed range of risk was, therefore, limited by design, and the resulting homogeneity of the study population has made it difficult to demonstrate the prognostic power of standard coronary disease risk factors, let alone the coronary calcium score.5
- Copyright © 2001 by American Heart Association
O’Rourke RA, Brundage BH, Froelicher VF, et al. American College of Cardiology/American Heart Association expert consensus document on electron beam computed tomography for the diagnosis and prognosis of coronary artery disease. Circulation. 2000;102:126-140.
Arad Y, Guerci AD. Predictive value of electron beam computed tomography. Circulation. 1997;95:535–536. Letter.
Arad Y, Spadaro L, Goodman K, et al. 3.6 Years follow-up of 1136 asymptomatic adults undergoing electron beam CT (EBCT) of the coronary arteries. J Am Coll Cardiol. 1998;31:210A. Abstract.
Detrano RC, Wong ND, Doherty TM, et al. Coronary calcium does not accurately predict near term coronary events in high risk adults. Circulation. 1999;99:2633–2638.
Recently, Circulation published the American College of Cardiology (ACC)/American Heart Association (AHA) Expert Consensus Document on Electron Beam Computed Tomography (EBT),R1 which detailed a technology assessment for the use of coronary calcium screening for (1) diagnosis, (2) risk stratification, and (3) serial therapeutic monitoring. The letter by Drs Arad and Guerci now provides data that was previously unavailable to the EBT committee. To understand whether additional data could have affected the results of the ACC/AHA document, it is imperative to understand the principles of technology assessment. A technology evaluation includes a systematic, unbiased evaluation of peer-reviewed literature. We used rules for the identification of relevant literature, excluding abstracted data and letters to the editor. This policy is consistent with prior evidence-based evaluations from the ACC, AHA, and other medical societies.R2 R3
This committee performed a detailed review of diagnostic and prognostic data from 19 publications. Meta-analytic methods were used to recalculate prognostic odds ratios (OR) for death or myocardial infarction (2.0; 95% CI, 0.5 to 8.2; P>0.20) from the St Francis Hospital data. The frequency statistics for events in low- and high-risk calcium scores were reanalyzed using a meta-analytic calculation of this odds ratio (95% CI). Because the EBT data are from observational series, we used a conservative, random-effects model. This type of model evaluates unadjusted outcomes considering additional confounders (ie, clinical and historical factors) affecting the study outcome. The results of a random-effects model will have a wider range of uncertainty about the odds ratio (ie, wider confidence intervals). Using combined end points for risk stratification with coronary calcium scores, point estimates for EBT odds ratios range from 1.05 to 60.6. It is expected given the wide range of values that weighted average odds ratios would be nonsignificant (95% CI, 0.71 to 101.2). These results indicate extreme diversity in the published findings, with the overall results revealing a nonsignificant elevation in risk.
A critical component to any risk evaluation is the use of rigorously evaluated, statistically powered samples. It was the purpose of this committee to evaluate the rigor with which the EBT data had been evaluated, as well as the application of current epidemiologic and risk-adjusted statistical methodologies.R4 Currently, prognostication in low-risk populations requires substantially larger populations than are currently available. We await the results from statistically powered samples such as the National Institutes of Health’s Multi-Ethnic Study of Atherosclerosis or registries being planned by the Society for Atherosclerotic Imaging.
We hope that this clarifies any perceived discrepancies within the ACC/AHA document and the published literature.
O’Rourke RA, Brundage BH, Froelicher VF, et al. American College of Cardiology/American Heart Association expert consensus document on electron beam computed tomography for the diagnosis of coronary artery disease. Circulation. 2000;102:126–140.
Braunwald E, Mark DB, Jones RH, et al. Clinical Practice Guideline Number 10: Evaluation Instruments for the Care of Patients With Unstable Angina. Washington, DC: Department of Health and Human Services, 1995. AHCPR Publication No. 95-0602.