Coronary Calcium: 'Bad News' for Men, but Even 'Worse News' for Women?

Stephan Achenbach


Eur Heart J. 2018;39(41):3736-3738. 

Cardiovascular disease differs between men and women not only regarding its incidence and prevalence, but also regarding the type of clinical manifestation.[1,2] For example, stroke and heart failure are more frequently the initial presentation of cardiovascular disease in women than in men, while myocardial infarction is a more frequent first manifestation of cardiovascular disease in men than in women.[3] At the same time, survival for many types of heart disease is worse in women than in men, including atrial fibrillation,[4] hypertrophic cardiomyopathy,[5] and myocardial infarction.[6] In acute coronary syndromes, female patients more frequently present with atypical symptoms, and women suffer from a longer delay between symptom onset and treatment.[7–10] Furthermore, women seem to have a different distribution of mechanisms leading to intracoronary thrombus formation than men, especially when young.[11–13] These findings nurture the notion that atherosclerotic plaque composition and mechanisms underlying acute coronary events in women may be substantially different from that in men (a hypothesis that is not unanimously supported by data).[14,15] However, as a consequence, this would lead to the conclusion that if cardiovascular risk factors or imaging findings are used for refined targeting of preventive therapies, risk assessment algorithms, target variables, and treatment thresholds likely need to be different between male and female individuals.

It is with this background that Shaw and colleagues present an intriguing analysis in this issue of the European Heart Jounal.[16] A very large cohort of 63 215 asymptomatic individuals who underwent physician-referred coronary calcium imaging by computed tomography (CT) for risk stratification was created by pooling patients from four centres in the US (two centres in California, one in Ohio, and one in Minnesota) between 1991 and 2010, creating a mean follow-up period of almost 13 years. Traditional cardiovascular risk factors were determined at baseline to the allow calculation of cardiovascular event risk according to Framingham and the pooled cohort equation. Patients were followed not only for overall mortality but also (a particular strength of this cohort) for cardiovascular mortality, with a total of 919 cardiovascular disease deaths recorded (approximately one-third of all-cause mortality). Coronary calcium was quantified in two ways. On one hand, the well-established 'Agatston Score' was calculated,[17] which assigns a value determined from plaque area and peak density to each calcified lesion visible in the cross-sectional CT images, and then summarizes all plaque-specific values for the entire coronary tree. While the score is easy to calculate and easily transferrable from one scanner to another, information on the number of lesions, individual plaque size, and density is lost. On the other hand, Shaw and colleagues performed and carried along a very detailed analysis of calcified plaques in all 63 000 participants; the number of lesions, number of calcified arteries, the calcium volume in mm3, and average lesion size were documented for every enrolled individual. Furthermore, mean calcium density was documented in a subgroup of 10 374 individuals.

Shaw et al. report that 20 508 of the enrolled asymptomatic individuals were women who smoked less frequently but who were significantly older than men (56 vs. 54 years), and more frequently had a family history of premature cardiovascular disease (53 vs. 43%), but who, in spite of a similar prevalence of dyslipidaemia (56% for both men and women), used statins significantly less frequently (16 vs. 23%) or aspirin (38 vs. 45%). This fits the known pattern of less-stringent risk factor modification in women than in men[18,19] (and adds the new observation that different thresholds seem to exist for referring female vs. male patients to refined risk assessment, e.g. by coronary calcium).

As expected and in keeping with prior findings, the presence of coronary calcium significantly impacted all-cause and cardiovascular mortality, with an Agatston Score ≥ 400 present in 11% of the population (5% of women and 14% of men) that was associated with a hazard ratio of approximately 4.2 for overall mortality, 9.1 for cardiovascular mortality, and 2.8 for cancer mortality.

The manuscript emphasizes the gender-related aspects regarding the influence of calcium on cardiovascular mortality. It is very interesting to observe that for a calcium score of zero, cardiovascular mortality was very low and virtually equal between men and women (0.4% over the mean follow-up period). However, two observations highlighted important differences between women and men. First, if any calcium was present, this had a significantly stronger impact on cardiovascular mortality in women than in men. The differential effect of calcium on cardiovascular mortality in women and in men increased with increasing calcium scores: a calcium score ≥ 400 had an almost two times higher effect on cardiovascular mortality in women than it had in men.

The second observation concerns the distribution of calcium. The study results are reported in a way that suggests that, if calcium of an equal overall amount (Agatston Score) is present, women have fewer lesions and a lower overall calcified volume, but a larger mean lesion size than men, while calcium density is not significantly different. Technical factors (so-called 'partial volume effects') rather strongly influence the relationship between calcium density and measured lesion size in CT, to the extent that this editorialist would interpret information on individual lesion size and density with a great deal of caution. Also, the observation that calcium is concentrated in fewer lesions in women could partly be explained by their smaller arteries.[20,21] Calcium of the same amount is more likely to 'merge' into one single lesion if the artery is smaller. Furthermore, the authors observed and reported a significantly lower number of involved coronary arteries in women vs. men, and yes indeed, the difference was significant, but it was almost negligibly small: as an example, for an Agatston Score between 1 and 100, 1.5 vessels were calcified in women and 1.6 vessels in men. Clinically, this hardly makes a difference.

Hence, the most robust finding of this publication is the following: coronary calcium exerts a stronger relationship on cardiovascular mortality (and, by the way, overall mortality) in women than in men. In short: if coronary calcium is 'bad news' for men, it may be even 'worse news' for women.

It is the task of an editorialist to point out some weaknesses of a published study, so here are some small drawbacks of the work under discussion, not to diminish the value of this first publication coming out of the 'Coronary Artery Calcium Consortium'. Mainly, we are not looking at an analysis of a population-based random sample. As detectable from the baseline characteristics, women of different risk profiles were referred to calcium imaging as compared with men. Also, imaging results were reported to the patients, which will have influenced further management, and this in turn may again have been entirely different between women and men. Calcium density measurements were available only in a small subgroup and, finally, the methods paper published elsewhere indicates an unexplained, rather strong difference in age-adjusted mortality rates between the four sites that contributed patients.[22]

So what can we learn from this admirable analysis? First of all, coronary calcium works as a risk stratification tool. Once again, and in addition to previously published, large population-based cohorts such as the Heinz–Nixdorff Recall Study[23] and the Multi-Ethnic Study of Atherosclerosis (MESA) study,[24] coronary calcium turns out to be strongly associated with mortality risk (and, as reported here, in particular with cardiovascular mortality risk). Second, atherosclerotic plaque patterns in women may be somewhat different from men, a finding that is not entirely new and, as pointed out above, likely requires more refined analyses than calcium-based parameters alone. Third, and this is most important and somewhat different to previously reported for slightly different endpoints,[24,25] the impact of calcium on cardiovascular mortality in this study was markedly stronger in women than in men. While randomized outcome trials that use coronary calcium as the decision maker for statin therapy and other risk modification are still missing, the findings of the analysis reported here seem conclusive enough to let clinicians address risk factors aggressively, particularly in women with high calcium scores.