To the Editor:
As avid football fans, we read with great interest the recent study by Schmermund et al1 in Circulation (“Coronary Artery Calcium in Acute Coronary Syndromes: A Comparative Study of Electron-Beam Computed Tomography, Coronary Angiography, and Intracoronary Ultrasound in Survivors of Acute Myocardial Infarction and Unstable Angina”).
In the introduction, Dr Schmermund and colleagues suggest, as have we,2 the distinct possibility that the association of coronary calcium with subsequent events may be a consequence of higher numbers of rupture-prone lesions that may not be calcified in patients who have coronary lesions elsewhere in the coronary vasculature that are calcified. Any association of calcium with subsequent events may thus actually be a spurious one. We hasten to add, however, that as yet there is no direct evidence that addresses this question, and the study by Schmermund et al would seem to shed little light on the subject.
The authors conclude that “… EBCT identifies calcified plaques in the vast majority of patients with vulnerable atherosclerotic plaques” and that EBCT “seems a very promising method to identify patients at risk for future atherosclerosis-related coronary unstable events.” Unfortunately, however, this study was not only retrospective but examined patients with EBCT up to 18 months after their events. “Future” coronary events were simply not addressed. Yet this is perhaps the least of the difficulties encountered by Schmermund and colleagues; there are profound methodological deficiencies in both the design and analysis of their study that render such conclusions untenable. Some of these difficulties are familiar to us from our own investigations.
Some years ago, while pondering the relationship between calcium and the probability of plaque rupture and subsequent clinical events, we conducted a pilot study to attempt to elucidate this relationship. We assessed coronary calcium with digital cinefluoroscopy in 24 patients referred for angiography within 2 or 3 days of acute Q-wave myocardial infarction. We identified the segment of the coronary artery responsible for the infarction and found that this segment had no detectable calcium in 16 (66.7%) of the cases, but nearly all patients had calcium somewhere in their coronary vasculature. Although intriguing, in the end we found these data meaningless. The primary reason was the obvious and overwhelming selection bias. A large number of patients who in fact had plaque rupture were simply excluded from the study, for a variety of reasons: they died before reaching the hospital, they died before angiography, they were not referred for angiography, they refused angiography, they did not have clear evidence either by ECG or symptoms of a myocardial infarction or unstable angina, the culprit lesion could not be identified with confidence, the symptoms of plaque rupture were unrecognized or misdiagnosed, or plaque rupture was symptomatically silent and they never sought medical attention. Given this enormous exclusion of patients who in fact had ruptured coronary lesions, no conclusion could possibly be reached regarding the relationship of calcium to plaque rupture. When it’s fourth and long and you are deep in your own territory with plenty of time left in the game, prudence dictates you punt; accordingly, our results were never submitted for publication.
Ah, but the punt has taken a favorable bounce, rolls into the end zone, and is about to roll harmlessly past the end line, when Schmermund et al field the punt 9 yards deep in their own end zone and attempt to return it, with defenders converging rapidly. The crowd can only gasp; Schmermund et al are mercilessly gang tackled by selection bias.
For it is precisely the selection bias encountered by us that plagues the study by Schmermund et al and similarly renders their data all but meaningless. The fact that 28 months was required to recruit 118 patients (an average of only 4 patients per month) likely is testimony to enormous exclusion. Therefore, their conclusion that “[t]he vast majority of patients with acute coronary syndromes and at least moderate angiographic disease have identifiable coronary calcium by EBCT” is neither supported nor excluded by their study; no statement whatsoever can be made regarding the “vast majority” because these were largely excluded from enrollment.
But the methodological difficulties do not end here. Although the authors do not explicitly state this one way or another, one can only surmise that those reading the EBCT scans were not blinded to angiographic or clinical data, introducing yet another source of bias (if one was needed): test-review bias.3 Furthermore, it is not known what change in coronary calcium may have occurred in the interval between the acute event and the EBCT scan, which was as long as 11/2 years. Finally, given that a major problem with EBCT calcium scanning is not lack of sensitivity but rather lack of specificity,4 5 6 where was the control group? How would the sensitivity and specificity have been affected by including a large group of patients who, for instance, were admitted to rule out myocardial infarction, subsequently did have myocardial infarction ruled out, and were found to have angiographically normal coronary arteries? Plays may be executed to perfection during the pregame warm-up, but how effective they are when the opposing team is on the field determines whether points are scored.
What, then, might be the clinical implications of the results of the study by Schmermund et al? Patients in their study had all suffered at least 1 acute coronary event already (mostly myocardial infarction) and were referred for angiography “for further diagnosis and/or treatment.” EBCT scans in these patients, performed as long as 18 months later, were positive in 90% of these patients; that is to say, 90% of these patients had detectable calcium somewhere in their coronary vasculature. Thus, in reference to “clinical implications,” the message to referring physicians would appear to be, “Doctor, if you have a patient who you know has coronary artery disease, who has had a myocardial infarction, and who you refer to us for ‘further diagnosis and treatment,’ if we perform an EBCT scan within 18 months, we can provide you with the following information with ≈90% accuracy—your patient has coronary artery disease and has at some time in the past 18 months had an acute coronary event.”
We applaud Schmermund and colleagues for their attempts, however ill fated, to elucidate the relationship between calcium in a lesion and the probability of that lesion to rupture and cause a clinical event. However, although we can well empathize with the methodological difficulties encountered by Schmermund et al, their conclusions are nevertheless completely unjustified on the basis of their data. Eventually, as the recent American Heart Association position paper has indicated,4 EBCT may well be shown to be of definite usefulness, in at least some patients with known or suspected coronary artery disease. For the time being, however, if EBCT calcium evaluation is “a very promising method,” one would never know it from the study by Schmermund et al. A sound strategy, in both science and football, is punt on fourth and long.
- Copyright © 1998 by American Heart Association
Schmermund A, Baumgart D, Görge G, Seibel R, Grönemeyer D, Ge J, Haude M, Rumberger J, Erbel R. Coronary artery calcium in acute coronary syndromes: a comparative study of electron-beam computed tomography, coronary angiography, and intracoronary ultrasound in survivors of acute myocardial infarction and unstable angina. Circulation. 1997;96:1461–1469.
Wexler L, Brundage B, Crouse J, Detrano R, Fuster V, Maddahi H, Rumberger J, Stanford W, White R, Taubert K. Coronary artery calcification: pathophysiology, epidemiology, imaging methods, and clinical implications: a statement for health professionals from the American Heart Association. Circulation. 1996;94:1175–1192.
Wong ND, Detrano RC, Abrahamson D, Tobis JM, Gardin JM. Coronary artery screening by electron beam computed tomography: facts, controversy, and future. Circulation. 1995;91:632–636.
Detrano RC. Why the controversy about electron beam computed tomographic screening for coronary atherosclerosis? Br Heart J. 1994;72:313–314.
Although Drs Doherty and Detrano emphasize epidemiological questions regarding our study, we addressed a completely different issue. We presented the first reportR1 on the detection of coronary calcium with respect to coronary plaque morphology in patients with acute coronary syndromes based on clinical, coronary angiographic, and intracoronary ultrasound data. Contrary to Doherty and Detrano’s statement, we provided direct (morphological) evidence pertaining to the association of coronary calcium and acute coronary syndromes, even though we agree that many details remain to be learned. We used intracoronary ultrasound to examine coronary anatomy of culprit lesions. Intracoronary ultrasound allows classification of plaque formation corresponding to the recommendations by the American Heart AssociationR2 R3 R4 R5 and visualization of plaque rupture even at very early stages of the atherosclerotic disease process.R6 R7 In view of the diagnostic and clinical implications of negative calcium scans by electron-beam CT (EBCT) in some patients with acute coronary syndromes, it was critical to explain the mechanisms of acute ischemia in these patients. We state in our article that “… patients with low coronary plaque burden or mechanisms of unstable events that are not specifically related to atherosclerosis may be missed by EBCT calcium scanning… . [However,] … EBCT reliably identified coronary calcific plaques in those patients who had moderate to severe coronary atherosclerotic disease.”R1
We never denied that there may have been an inherent selection bias, and in fact we clarify this in the article by stating “Most of [our] patients were referred by their physicians or other medical centers for further diagnosis and/or treatment at the University Clinic Essen.”R1 One will never be able to examine a truly unselected population of patients with acute coronary syndromes due to the problems Doherty and Detrano explain in their letter. Our report included only patients with index myocardial infarctions or a first episode of unstable angina and no further events. They were a random sample of all eligible patients whom we were able to scan at a remote EBCT site, with scanner availability as the limiting factor. We sought to explain where calcium scanning works and where it may not work and why and to discuss the clinical implications. We found that coronary calcium was reliably detected in patients with acute coronary syndromes and moderate to severe plaque burdens. It has been known for more than a century that patients with unstable angina or acute myocardial infarction usually have extensive coronary disease. In this respect, the selection of patients in our study tended to play against calcium scanning, because a comparably large percentage (15 of 118 patients, or 13%) had no angiographically obstructive coronary lesions.
Our conclusions are therefore fully justified that “… EBCT seems a very promising method to identify patients at risk for future atherosclerosis-related coronary unstable events.”R1 Our article clearly focused on sensitivity considerations. We purposely did not consider the issue of specificity, which is indeed very complex and needs to be addressed by much larger prospective studies in the general population.
Schmermund A, Baumgart D, Görge G, Seibel R, Grönemeyer D, Ge J, Haude M, Rumberger JA, Erbel R. Coronary artery calcium in acute coronary syndromes: a comparative study of electron beam computed tomography, coronary angiography, and intracoronary ultrasound in survivors of acute myocardial infarction and unstable angina. Circulation. 1997;96:1461–1469.
Stary HC, Chandler AB, Dinsmore RE, Fuster V, Glagov S, Insull W, Rosenfeld ME, Schwartz CJ, Wagner WD, Wissler RW. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis: a report from the Committee on Vascular Lesions of the Council on Arteriosclerosis, American Heart Association. Circulation. 1995;92:1355–1374.
Stary HC. The sequence of cell and matrix changes in atherosclerotic lesions of coronary arteries in the first forty years of life. Eur Heart J. 1990;11(suppl E):3–19.
Erbel R, Ge J, Bockisch A, Kearney P, Görge G, Haude M, Schümann D, Zamorano J, Rupprecht HJ, Meyer J. Value of intracoronary ultrasound and Doppler in the differentiation of angiographically normal coronary arteries: a prospective study in patients with angina pectoris. Eur Heart J. 1996;17:880–889.
Ge J, Haude M, Görge G, Liu F, Erbel R. Silent healing of spontaneous plaque disruption demonstrated by intracoronary ultrasound. Eur Heart J. 1995;16:1149–1151.