By Michael H. Crawford, MD, Editor

SYNOPSIS: Researchers analyzed the MESA study to determine the comparative value of the pooled risk equation vs. CT coronary artery calcium score for determining which aspirin-naïve patients < age 70 years without overt coronary artery disease would benefit most from primary prevention with aspirin. They found a calcium score of 100 was superior for this determination.

SOURCE: Cainzos-Achirica M, Miedema MD, McEvoy JW, et al. Coronary artery calcium for personalized allocation of aspirin in primary prevention of cardiovascular disease in 2019: The MESA study (Multi-Ethnic Study of Atherosclerosis). Circulation 2020;141:1541-1553.

The results of recent studies have challenged the relative value of aspirin for primary prevention of cardiovascular disease (CVD) and highlighted the increased risk of bleeding. Guidelines now list aspirin as a class IIb recommendation for primary prevention. However, selecting the right patients for aspirin prophylaxis is unclear.

Investigators from the MESA study sought to assess the potential value of coronary artery calcium (CAC) for guiding aspirin allocation for primary prevention. MESA is an ongoing community-based, multi-ethnic study of 6,814 subjects ≥ age 45 years who were free of overt CVD at entry. After excluding those with missing data, aspirin use at baseline, age > age 70 years, or high bleeding risk 3,540 subjects remained. All of the remaining subjects had a CT-determined CAC score at baseline and data for estimating CVD risk by the pooled cohort equation (PCE). The primary outcomes were coronary heart disease events, stroke, CV death, and major bleeding events. The identification of subjects who could be considered for aspirin therapy was defined as a PCE 10-year risk of CVD > 20%. Overall, 4.9% were in this high-risk category at baseline; 79% were men, and 83% were between age 60 and 69 years. CAC scores were > 100 in 42%, 1-99 in 26%, and zero in 32% of subjects. Only 10% of all the subjects were on statins at baseline. In general, the higher the CAC score, the older the age, the higher the proportion of men, and the higher the PCE risk.

Overall rates per 1,000 patient-years were 3.52 for CVD, 2.34 for coronary heart disease, and 1.34 for major bleeding. The higher the CAC, the higher the risk, with the highest risk observed in those with CAC scores > 100 (12.4 per 1,000 patient-years). Based on previous primary prevention trials with aspirin, a five-year number needed to treat (NNT) to prevent one event and a number needed to harm (NNH; bleeding) was determined. The overall NNT with aspirin was 476, and the NNH was 355. These numbers did not vary much based on the PCE risk score. However, in the CAC score > 100 subgroup, the NNT was lower than NNH (NNT = 140; NNH = 518). In the CAC = 0 subgroup, NNT was much higher than the NNH (1,190 vs. 567). The authors concluded that determining CAC scores may be superior to the PCE for deciding whom to treat with aspirin for primary prevention.


Using the PCE failed to identify those most likely to benefit from aspirin among the aspirin-naïve subjects < age 70 years at baseline in MESA. In fact, at all PCE 10-year risk strata (< 5%, 5-20%, > 20%), the NNT was equal to or greater than the NNH at five years. On the other hand, CAC score was predictive of who might benefit from aspirin. At a CAC score > 100, the NNT dropped well below the NNH. At a CAC score < 100, the NNT was consistently higher than the NNH.

The estimated efficacy of aspirin for preventing CVD events and risk of bleeding were derived from a meta-analysis that included only studies published after 2000 (statin era) and the excluded patients with peripheral artery disease.

This meta-analysis revealed a relative risk reduction for primary prevention of CVD events of 9% and a relative risk of bleeding of 39%, which clearly did not support aspirin for general primary prevention. These results influenced the new guidelines (i.e., that only in subjects at high risk of CVD and low-bleeding risk should aspirin be considered).

The MESA study participants were a relatively low-risk population (mean age, 50 years; 10% diabetics) and the NNT was relatively high. Even at a CAC score > 400, the NNT was 97. In contrast to the meta-analysis, using a CAC score > 100 to prescribe aspirin improved the relative risk reduction from 9% to 12%, which still is modest.

Although MESA is a multiethnic population, it does not include all ethnicities nor does it mimic the populations of other large studies used to estimate the potential benefits of aspirin. MESA did not assess bleeding risk on the initial visit; this was added later, and it was not designed to capture all bleeding events (only hospital records were used). Subgrouping by statin use was not possible because of low use at baseline in the population. Finally, this was an observational study; as such, it should be considered hypothesis-generating.

For the likely modest net benefit of using aspirin in primary prevention, should a relatively expensive test that involves some radiation exposure be used to decide? It does not seem reasonable right now for general application, but could be valuable in selected subjects, such as those with a bad family history, which is not included in the PCE. Also, bleeding risk rises with the CAC score, so finding equipoise here could be problematic. However, if a CT scan performed for another reason shows calcium in the coronary arteries, we should take that into consideration.