Coronary artery calcium score prediction of all cause mortality and cardiovascular events in people with type 2 diabetes: systematic review and meta-analysis
BMJ 2013; 346 doi: https://doi.org/10.1136/bmj.f1654 (Published 25 March 2013) Cite this as: BMJ 2013;346:f1654
All rapid responses
In Seth Robert’s comments about our paper, there was a misunderstanding of our results. In Figure 5, the left hand graph shows the Bayes nomogram for the combined outcome (death + cardiovascular events) (event A) while the right graph shows the nomogram for only cardiovascular events (event B). The percentages shown in our graph are indeed correct because the pre-test probabilities for these two graphs were not calculated based on the same population. As stated in the paper, eight studies were included in analyses of event A (n = 6521; 802 events = pretest probability of ~12%) and four studies evaluated cardiovascular events as the outcome (n = 1805; 351 events = pretest probability of ~19%). We hope that this clarification will be helpful to readers in evaluating these data.
Competing interests: RR has served as a consultant for Merck and Novo Nordisk,received grants from Merck and Novo Nordisk, and received payment for lectures for Eli Lilly; no other relationships or activities that could appear to have influenced the submitted work.
The results in Figure 5 (Bayes nomogram, which the caption misspells as "normogram") are impossible, in the sense of inconsistent. The left-hand figure shows the probability of Event A (death) OR Event B (cardiovascular event). The right-hand figure shows the probability of Event B alone. The probability of {A or B} must be equal to or more than the probability of just {B}, yet the probabilities in the left-hand graph of Figure 5 are LESS than the probabilities in the right-hand graph. For example, the left-hand graph of Figure 5 shows that the pre-test probability of {A or B} is about 12% and the right-hand graph shows that the pretest probability of {B} is about 18%. That's impossible -- both percentages cannot be correct.
Competing interests: No competing interests
Re: Coronary artery calcium score prediction of all cause mortality and cardiovascular events in people with type 2 diabetes: systematic review and meta-analysis
Dear Editors,
We read with interest the meta-analysis by Kramer et al(1) regarding the use of coronary artery calcium (CAC) scoring for the prediction of all-cause mortality and cardiovascular events in people with diabetes. Of particular importance is their conclusion that the risk of events among individuals with diabetes is heterogeneous and that CAC scoring has the potential to identify individuals within this traditionally high risk group that are in fact low risk.
The authors identified a CAC score < 10 as the threshold identifying the lowest risk. Although individuals with a CAC score of < 10 may be at lower risk than those with CAC > 10, there is heterogeneity among individuals within the 0-10 range. Previous data has demonstrated a significant difference in risk between CAC = 0, versus CAC 1-10.(2,3) Blaha et al previously demonstrated in a cohort of asymptomatic patients referred for CAC scoring that the multivariable adjusted hazard ratio for all-cause mortality among individuals with CAC 1-10 versus CAC 0 was 1.99 (1.44 – 2.75).(2) Budoff et al demonstrated similar results from the Multi-Ethnic Study of Atherosclerosis (MESA), a cohort of asymptomatic individuals free of baseline cardiovascular disease. The multivariable adjusted hazard ratio for hard CHD events among individuals with CAC 1-10 versus CAC 0 was 3.23 (1.17 – 8.95).(3)
We conducted a secondary analysis of our cohort of 2,384 asymptomatic individuals with diabetes referred for CAC testing.(4) We found that the multivariable adjusted hazard ratio for all-cause mortality among individuals with CAC 1-10 versus CAC 0 was 2.7 (1.12-6.51). Thus, in the general population, as well as in individuals with diabetes, CAC 0 carries a distinct risk from CAC 1-10. In fact CAC=0 stands out as the most potent negative risk measure amongst all populations studied to date.(5)
Many of the studies included in the meta-analysis did not report event rates for CAC 0 separately from CAC < 10. We contend that CAC 0 is a more ideal number to identify individuals who are truly low risk given the heterogeneity seen with CAC < 10. Hopefully future studies will report events associated with CAC 0 as a distinct CAC score category.
References
1. Kramer CK, Zinman B, Gross JL, Canani LH, Rodrigues TC, Azevedo MJ, Retnakaran R. Coronary artery calcium score prediction of all cause mortality and cardiovascular events in people with type 2 diabetes: systematic review and meta-analysis. BMJ. 2013 Mar 25;346:f1654.
2. Blaha M, Budoff MJ, Shaw LJ, Khosa F, Rumberger JA, Berman D, Callister T, Raggi P, Blumenthal RS, Nasir K. Absence of coronary artery calcification and all-cause mortality. JACC Cardiovasc Imaging. 2009 Jun;2(6):692-700.
3. Budoff MJ, McClelland RL, Nasir K, Greenland P, Kronmal RA, Kondos GT, Shea S, Lima JA, Blumenthal RS. Cardiovascular events with absent or minimal coronary calcification: the Multi-Ethnic Study of Atherosclerosis (MESA). Am Heart J. 2009 Oct;158(4):554-61.
4. Silverman MG, Blaha MJ, Budoff MJ, Rivera JJ, Raggi P, Shaw LJ, Berman D, Callister T, Rumberger JA, Rana JS, Blumenthal RS, Nasir K. Potential implications of coronary artery calcium testing for guiding aspirin use among asymptomatic individuals with diabetes. Diabetes Care. 2012 Mar;35(3):624-6.
5. Blaha MJ, Blumenthal RS, Budoff MJ, Nasir K. Understanding the utility of zero coronary calcium as a prognostic test: a Bayesian approach. Circ Cardiovasc Qual Outcomes. 2011 Mar;4(2):253-6.
Competing interests: No competing interests