By Michael H. Crawford, MD, Editor

SYNOPSIS: In a study of women undergoing routine mammographic screening for breast cancer, mammographic features, such as microcalcifications and breast density, were associated with the risk of developing cardiovascular disease.

SOURCE: Grassmann F, Yang H, Eriksson M, et al. Mammographic features are associated with cardiometabolic disease risk and mortality. Eur Heart J 2021;42:3361-3370.

The identification of prevalent cardiovascular disease is challenging, especially in women, for whom the incidence is lower than men until later in life. Although controversial, mammographic screening programs have been shown to reduce mortality from breast cancer. Mammography not only detects early cancer, but other features of the images have been shown to be associated with an increased risk of cardiovascular disease (CVD).

Grassmann et al aimed to further develop the relationship between mammography features and CVD in a large prospective study of mammographic screening: the Karolinska Mammography Project for Risk Prediction of Breast Cancer, or KARMA. (Learn more at Between 2011 and 2013, the KARMA creators enrolled more than 70,000 women in Sweden, then followed this group through 2018. Those with breast cancer (BC) diagnosed before or during the trial period were excluded, as were any subjects who underwent breast surgery before enrollment or for whom there were missing data. This resulted in a total study population of 57,867 women who underwent computer-aided detection of microcalcifications (MC) and breast density (BD) on routine mammographic images.

Those with BD greater than the mean value were considered to have dense breasts. Other data collected included family history of BC and a genetic risk score for BC. Various comorbidities were recorded, such as risk factors for CVD (identified by ICD-10 codes), and death caused by CVD was ascertained. They found increases in MC or BD were associated with the development of BC, as expected. Increases in MC were associated with the development of CVD, especially in women with a preexisting CVD. In women without a prior CVD, MC was associated with hypertension and heart failure. By contrast, higher BD in women with a pre-existing CVD was not associated with CVD. In women without a prior CVD, it was associated with a lower rate of CVD independent of MC, reproductive history, and lifestyle factors.

By the end of 2018, 233 women had died. Increased MC was associated with a CVD death (HR, 1.46; 95% CI, 1.10-1.94; P = 0.01) in the total study population, but not in women without prior CVD. Increased BD in the total study population was associated with a lower CVD mortality (HR, 0.64; 95%CI, 0.46-0.90; P = 0.01). Also, a family history of BC and a genetic risk score for BC were associated with a lower risk of CVD. The authors concluded MC and BD on mammography are associated with CVD and CVD mortality. These features on routine screening mammography may be of value for fine-tuning the risk of CVD in women.


The work by Grassmann et al was a well-conducted observational study that amplifies the associations between mammographic features and CVD that has been seen in smaller studies and strengthens the concept that mammogram results should be considered when risk-stratifying women.

Grassmann et al defined CVD rather broadly and included diseases that might be considered risk factors themselves, such as diabetes and renal failure. In addition, there was considerable granularity in that conditions such as cardiac conduction disorders and cardiac dysrhythmias were included. Another strength was the fact the Swedish health system provided the investigators with detailed information about all participants.

Standard formulae for estimating the risk of developing CVD to fine-tune prevention measures fail to consider important known risk factors, such as family history of early CVD, the detection of vascular calcium on CT scans, and prior radiation therapy. It seems we should add the presence of MC on routine mammograms to the list of other considerations when deciding whether a young woman should take a statin.

The association of low-density breasts with a higher risk of CVD is more problematic since low-density breasts contain more fat; obesity could be a confounder here. Another issue is the current algorithm for quantifying MC cannot distinguish between breast tissue and vascular calcification. The former is associated with the development of BC, but there was no difference in the study results in those who developed BC.

Another strength of the Grassmann et al study, other than the use of a large cohort of subjects, is the fact that the investigators identified and studied the sisters of the KARMA participants. Doing so led to the discovery of similar results. Whether this was because of like environments or family genetics is unclear. Nevertheless, all this indicates this was a robust study.