Comparing coronary artery cross-sectional area among asymptomatic South Asian, White, and Black participants: the MASALA and CARDIA studies

Numerous studies have demonstrated that people of South Asian background (those with origins from India, Pakistan, Bangladesh, Nepal, Sri Lanka, Bhutan, and the Maldives) experience a higher incidence of atherosclerotic cardiovascular disease (ASCVD) compared to individuals from other racial/ethnic groups [1, 2]. The World Health Organization reported that South Asians account for 60 percent of patients with coronary artery disease globally, while comprising only 23 percent of the world’s population [3]. South Asian populations also tend to develop ASCVD and acute myocardial infarction at younger ages than individuals from other racial/ethnic groups [4], a disparity that persists even after accounting for differences in hypertension, type 2 diabetes, obesity, smoking, and lipid levels [5].

Understanding the multilevel factors that account for higher rates of ASCVD in South Asian populations is important for prevention and treatment. Some experts have suggested that a smaller coronary artery size may be responsible. While a small body of literature comparing coronary artery size between symptomatic South Asians and White individuals exists, these studies were limited by small sample sizes, differences in covariate adjustment, and use of coronary angiography, which primarily measures lumen size [6, 7]. No studies have compared coronary cross-sectional area (including the lumen and arterial wall) across multiple diverse race/ethnic groups among asymptomatic individuals with no clinical evidence of ASCVD.

We utilized two epidemiologic cohorts for the current study: South Asian men and women enrolled in the Mediators of Atherosclerosis in South Asians Living in America (MASALA) and Black and White participants in the Coronary Artery Risk Development in Young Adults study (CARDIA). We compared left anterior descending artery cross-sectional area (LAD CSA) using non-contrast coronary CT images among South Asian, White, and Black participants. We also examined associations of various cardiovascular risk factors on LAD CSA. We hypothesized that South Asian participants would have smaller LAD CSAs than both Black and White participants, and that this association would be attenuated after adjusting for body size (BMI and height) and cardiovascular risk factors.

We compared key demographic, anthropometric. and cardiovascular risk factors among 513 South Asians participants in the MASALA cohort and 2,840 Black and White participants from CARDIA. Table 1 displays these data stratified by gender.

South Asian participants were, on average, approximately 6 years older and had 2–3 more years of education than White and Black participants. Lifetime smoking was lower among South Asians; notably, 71% and 96% of South Asian men and women participants, respectively had never smoked. Black women had the highest prevalence of hypertension (49.5%). South Asian men had the highest use of cholesterol-lowering medication use (40%) of any group. Anti-hypertensive medication use varied among groups, with Black women having the highest use (40%) and White women with the lowest (17%). South Asian men and women had lower anthropometric measures (weight, height, BMI, and waist circumference) than Black and White participants in CARDIA. After adjusting for age, South Asian men had the highest prevalence of any CAC (58% in South Asian men, 48% in White men, and 38% in Black men). South Asian and Black women had similar prevalences of CAC (approximately 20%) higher than among White women (17%).

In unadjusted analyses, South Asian men and women had smaller average LAD CSA than both Black and White participants in CARDIA (Fig. 1).

We used BMI and height as distinct measures of body size (r < 0.20), and did not adjust for weight, waist circumference, or body surface area as they were highly correlated with BMI and height in both cohorts. There was no significant difference in LAD CSA between South Asian men and women compared to White men and women after adjusting for age and body size. Further adjustment for ASCVD risk factors showed that South Asian men had somewhat smaller LAD CSA compared to White men (22.3 mm² (99% CI 21.4 – 23.2) vs. 23.6 mm² (99% CI 23.0 – 24.2), p = 0.009), but there remained no significant difference between South Asian and White women. Adjustment for body size and other risk factors did not substantially attenuate the differences in LAD CSA between South Asian and Black men and women.

We found a positive, graded association between greater CAC burden and LAD CSA for both men and women and across all three racial/ethnic groups after adjusting for age, BMI, and height (Fig. 2).

Across all three racial/ethnic subgroups, CAC scores of 1–99 were associated with a 1.0 mm² higher LAD CSA, and CAC scores of 100 or greater were associated with 2.3 mm² higher LAD CSA compared to those with CAC of zero after adjusting for age, gender, BMI and height.

In multivariable models adjusted for age, gender, BMI, and height we found that South Asian participants had smaller LAD CSA than Black participants, but there was no difference between South Asian and White participants (Table 2).

On average, LAD CSAs were 2.3 mm² larger for Black individuals compared to South Asian participants. Adjusting for body size, cardiovascular risk factors, and CAC amplified the difference between LAD CSA among racial/ethnic groups, with greater LAD CSA among Black (2.7 mm²) and White participants (0.8 mm²) compared to South Asian individuals. Height, male gender, hypertension, CAC 1–99, CAC ≥ 100, BMI, and age were all associated with higher LAD CSA, whereas current and former smoking were inversely associated.

In a sensitivity analysis restricted to 1,401 participants (61 South Asian men, 119 South Asian women, 279 White men, 496 White women, 161 Black men, and 285 Black women) with no history of smoking, hypertension, or detectable CAC, we found no significant differences in LAD CSA comparing South Asian and White men and women, respectively, after adjusting for age, BMI and height. Black men and women had greater LAD CSA compared to South Asian men and women (Supplemental Figure 3).

In another sensitivity analysis using BSA instead of BMI and height as a surrogate for body size yielded similar results (Supplemental Table, Supplemental Figure 4). In models using BSA for body size adjustment, South Asian participants had significantly smaller LAD CSAs compared to Black participants while there was no significant difference between South Asians and White participants in both minimally and fully adjusted models (Supplemental Table).

South Asian men and women have smaller left anterior descending coronary artery cross-sectional area compared to other groups, but the differences between South Asian and White individuals were attenuated by adjustment for age and body size. Black men and women had greater LAD CSA than both South Asian and White men and women, respectively, even after accounting for differences in body size and cardiovascular risk factors. This study includes participants without clinical coronary heart disease from two large community-based cohorts, extending what has been previously reported in symptomatic clinical populations.

We identified seven observational studies from 1995 to 2017 comparing LAD diameters between South Asian and White patients [6, 18‐23]. Five of these studies found that South Asians had smaller diameters after adjusting for body surface area [6, 18, 20, 21, 23]; two studies found no difference between groups, which is consistent with the current study [19, 22]. Regardless, these existing studies have similar limitations: small sample sizes, inclusion of only symptomatic patients, and requirement of invasive coronary angiography. Additionally, although these studies adjusted coronary measurements for body size, they did not adjust for cardiovascular risk factors. Therefore, existing knowledge of this topic is limited. Our study is novel in that it (1) utilized non-invasive methods for coronary artery size characterization, (2) considered the associations of subclinical coronary artery disease as measured by CAC and other traditional risk factors, and (3) directly compared South Asians to both Black and White participants from two large, well characterized epidemiologic cohorts.

Coronary artery calcium score is a strong independent predictor of ASCVD [24, 25]. Multiple studies have demonstrated the direct association between CAC burden and cardiovascular disease, fatal and non-fatal myocardial infarction, and mortality [26‐30]. As expected, we observed that individuals with higher CAC had larger LAD CSA, regardless of gender or race/ethnic background. Additionally, LAD CSA was similar in all groups of women across CAC categories, but differed primarily between South Asian and Black men. However, adjusting for CAC level did not explain the smaller coronary size seen in South Asian participants compared to Black individuals. Furthermore, as both South Asian and Black populations have been shown to have higher incidence of ASCVD compared to White populations, differences in LAD CSA and CAC will not explain these disparities in ASCVD. Makaryus et al. have posited that higher CAD burden among South Asians may be due to the lesser atheroma mass required to create unstable plaques [23]. Nonetheless, this relationship between CAC and coronary size is consistent with the findings of Glagov et al. in 1987, who showed that coronary arteries with larger plaque burden tend to have larger diameter, owing to compensatory “positive” or “outward” remodeling such that luminal area is initially preserved despite plaque growth [31]. Our findings suggest this mechanism is preserved across race/ethnicity and gender. Few studies have examined the relationship between CAC score and coronary cross-sectional area. In 2014, Hamirani et al. used coronary CT data from 140 symptomatic patients to compare the associations of CAC, gender, smoking, BMI, and hypertension with left main and right coronary artery diameters. They also found a direct relationship between CAC and vessel diameter [32]. In 2022, using the same methods as the current study, Werede et al. observed that, in persons with HIV, larger LAD CSA was associated with a lower CD4 T-cell count and lower circulating interleukin-10 levels [14]. Our study extends these results to community-based samples among participants from three different racial/ethnic groups.

We observed an unexpected association between current smoking and lower LAD CSA in the fully adjusted model. Smoking is been associated with greater coronary atherosclerosis; CAC score and incidence of ASCVD events are higher among individuals who smoke [33]. In accordance with our findings that CAC score and coronary area were directly proportional, we expected smoking to be associated with larger coronary size. Interestingly, in a recent review, Salehi et al. reinforced that smoking was associated with greater severity of atherosclerosis, but found that location of plaques and number of coronary arteries involved were not reliably associated with smoking [34]. Nicotine is known to expedite atherosclerosis, impair coronary vasodilation and promote vasoconstriction at the level of the endothelium [35, 36]; perhaps these endothelial changes outweigh the natural compensatory mechanism for stenosed coronaries to dilate, as described by Glagov et al. Further, Quillen et al. describe temporary vasoconstriction within thirty minutes of smoking, which may have impacted our results [37]. Additional research is needed to further elucidate these relationships.

A potential limitation of the current study is that it draws samples from two different study cohorts: MASALA and CARDIA. Although both cohorts included asymptomatic community-dwelling participants and used similar cardiac CT protocols and reading center for LAD CSA measurements, differences in samples may exist, such as temporal, geographic, and CT acquisition factors. Our study’s cross-sectional design limits causal inference. Additionally, while our analysis focuses on the LAD artery, it is possible that differences exist in atherosclerosis distribution among racial/ethnic groups. Finally, self-reported race is a social, not a biological construct, though we operationalized South Asian ethnicity as a distinct ancestral group from White and Black race for this analysis. Our analysis did not account for measures of socioeconomic status, birth and current home environment, geographic location in the U.S., and other structural factors which may be associated with self-identified race/ethnicity and may also affect coronary size.

South Asian men and women have smaller coronary artery cross-sectional area than White and Black individuals, but the differences between South Asian and White individuals may be explained by differences in body size. CAC burden, a correlate of overall plaque burden, was associated with larger LAD CSA among all groups, but plaque burden did not account for the smaller coronary size seen in South Asian participants. These results represent a novel finding in our attempt to understand South Asian risk for cardiovascular disease. Future studies following these cohorts should determine whether LAD CSA is associated with cardiovascular event incidence, and whether this may help explain the higher ASCVD risk among South Asians.