The article by Detrano et al 1 concerning the predictive value of coronary artery calcium contains fundamental flaws in methodology that may invalidate their conclusions. Their protocol differs significantly from the original protocol, which was designed to be sensitive to the presence of small amounts of calcium and to accurately reflect the quantity of calcium present. 2 Multiple studies have shown that these goals were met surprisingly well. 3 The differences in protocol can account for the

more »

... ack of discriminative power found in their article, without criticizing the high-risk composition of their study group. 4 Using 6-mm-thick slices, instead of 3 mm, increases the volume averaging inherent in CT scanning. Small calcified lesions may not meet the density threshold criteria for a calcified lesion with 6-mm-thick slices. There is also a 2-fold decrease in the number of lesions seen in 6-mm slices versus 3-mm slices, which decreases the calcium score by a factor of 2. The use of an 8.16 mm 3 threshold for defining a calcified lesion also significantly reduces sensitivity. This is 5ϫ the original threshold. Lesions, which could have a potential contribution of 10 to the total calcium, could have a score of 0 with Detrano et al's modifications. In combination with increased volume averaging, this higher volume threshold will have an even greater effect. It seems that a total calcium score in the range of 80 to 160 defines a high-risk group. 5,6 The potential magnitude of score reduction using Detrano et al's modifications are of the same order of magnitude. The data presented support the assertion that the calcium scores are significantly lower than would be expected. With a mean age of 66Ϯ8 years, a median score of 44 and a tertile 1 score range of 0 to 3.4 were seen. Janowitz et al 7 published similar data; in their 60 to 70 year age group, comparable numbers were a median score of 88, with the first tertile ranging from 0 to 24. This group would be expected to have lower scores than Detrano et al's high-risk group. There is also no way to determine if patients classified by Detrano et al's technique would fall into the same tertile groups using the standard protocol. By selectively eliminating smaller lesions, which may be a better predictor of future events than larger, more heavily calcified lesions, it seems that the superior predictive value of electron beam CT for future events is diminished. Attempting to improve reproducibility, Detrano et al have sacrificed the sensitivity and utility of the technique. Unless one is familiar with coronary calcium quantification, it is difficult to distinguish the differences in protocols and the fact that the calcium scores reported are significantly different than those that would have been obtained by the Agatston-Janowitz scoring system. The data in this article should not be compared with the large body of existing literature concerning coronary artery calcium quantification by electron beam CT. At best, the conclusion should be that the scoring system used was unable to provide additional information over conventional risk factors in their high-risk population. 1. Detrano RC, Wong ND, Doherty TM, et al. Coronary calcium does not accurately predict near-term future coronary events in high-risk adults. Circulation. 1999;99:2633-2638. 2. Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827-832. 3. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: pathophysiology, epidemiology, imaging methods, and clinical implications: a statement for health professionals from the American Heart Association. al. Predictive value of electron beam computed tomography of the coronary arteries: 19 month follow-up of 1173 asymptomatic subjects. Circulation. 1996;93: 1951-1953. 6. Agatston AS, Janowitz WR, Kaplan GS, et al. Electron beam CT coronary calcium predicts future coronary events. Circulation. 1996; 94(suppl I):I-360. Abstract. 7. Janowitz WR, Agatston AS, Kaplan G, et al. Differences in prevalence and extent of coronary artery calcium detected by ultrafast computed tomography in asymptomatic men and women. Am J Cardiol. 1993;72:247-254. Response Our prospective investigation 1 demonstrated that the assessment of coronary calcium with electron beam computed tomography (EBCT) was equivalent to standard risk-factor stratification in determining the risk of a future coronary heart disease event in subjects with coronary risk factors. 1 Both EBCT and risk-factor assessment were only moderately accurate for predicting future coronary events. In our study, we used a 6-mm tomographic slice thickness image-acquisition protocol. Dr Janowitz states correctly that the 6-mm protocol we used is less sensitive for detecting coronary calcium than is the more commonly used 3-mm protocol. However, Dr Janowitz's speculation that this reduced sensitivity will lead to reduced predictive accuracy is not supported by the literature or by the data we present below. A major problem with EBCT involves its high retest variability. 2 The 6-mm protocol we used demonstrates lower retest variability compared with the more commonly used 3-mm protocol. 3 For this reason, we used the 6-mm protocol in our investigation. Like Dr Janowitz, we were concerned that improved retest reproducibility might come at the expense of prognostic accuracy. For this reason, we compared the prognostic accuracy of these 2 protocols in a cohort of 326 patients (derived from our full cohort), and we found that the prognostic accuracy of the 2 protocols was equivalent. 4 Nevertheless, it would be helpful to determine whether our previous findings demonstrating an equivalent prognostic accuracy of the more reproducible 6-mm protocol and the more sensitive 3-mm protocol are confirmed in patients with a longer follow-up duration and more coronary events. We have now followed this subset of 326 subjects for 44 months. They have experienced a total of 16 hard coronary events (coronary death or nonfatal myocardial infarction) and 25 total coronary events, when revascularizations (coronary bypass or angioplasty) are included. All 326 subjects underwent, within 15 minutes, consecutive EBCT coronary calcium assessments using both the 3-mm and the 6-mm protocols. Subjects were divided into equal tertiles according to calcium score. The distribution of hard coronary events for the 3-mm scans was as follows: 4 in the lowest tertile (zero score), 4 in the middle tertile, and 8 in the highest tertile. The distribution of hard events for the 6-mm scans was as follows: 3, 6, and 7 in the 3 tertiles. These distributions are not significantly different from one another. When all events (including revas-1 by guest on July 21, 2018 http://circ.ahajournals.org/ Downloaded from