Abstract
Background: We examined whether slow heart rate recovery (HRR) after exercise testing as an estimate of impaired autonomic function is related to coronary artery calcification (CAC), an emerging marker of coronary atherosclerosis.
Methods: We evaluated 2088 men who participated in a health-screening program that included measures of CAC and peak or symptom-limited cardiopulmonary exercise testing. HRR was calculated as the difference between peak heart rate (HR) during exercise testing and the HR at 2 min of recovery after peak exercise. We measured CAC using multidetector computed tomography to calculate the Agatston coronary artery calcium score. Advanced CAC was defined as a mean CAC >75th percentile for each age group.
Results: HRR was negatively correlated with CAC (r = −.14, p < .01). After adjusting for conventional risk factors, participants in the lowest quartile of HRR (<38 bpm) were 1.59 times (95% CI: 1.17–2.18; p < .05) more likely to have advanced CAC than their counterparts in the highest quartile of HRR (>52 bpm). Each 1 bpm decrease in HRR was associated with 1% increase in advanced CAC after adjusting for potential confounders.
Conclusions: An attenuated HRR after exercise testing is associated with advanced CAC, independent of coronary risk factors and other related hemodynamic response.
Slow heart rate recovery (HRR) after maximal exercise testing, indicating decreased autonomic function, is associated with an increased risk of cardiovascular event and mortality.
Slow HRR has been linked with the occurrence of malignant ventricular arrhythmias, but it remains unclear whether slow HRR is associated with an increased risk of coronary artery calcification (CAC), an emerging marker of coronary atherosclerosis.
An attenuated HRR after exercise testing was associated with advanced CAC, independent of coronary risk factors and other potential hemodynamic confounder, supporting the hypothesis that slow HRR is related to the burden of atherosclerotic coronary artery disease.
KEY MESSAGES
Introduction
Although a slow heart rate recovery (HRR) after peak or symptom-limited exercise testing is associated with a heightened risk of cardiovascular events and mortality (Citation1,Citation2), the underlying pathophysiological mechanisms remain unclear. Slow HRR after maximal exercise testing, which suggests decreased vagal activity (Citation3), has been linked with the occurrence of malignant ventricular arrhythmias (Citation4); however, it remains unclear whether the increased cardiovascular risk is due to decreased autonomic function, increased coronary atherosclerotic burden, or both. Several studies have reported that slow HRR is related to impaired arterial function (Citation5,Citation6), the presence of carotid atherosclerosis (Citation7) and coronary artery disease (Citation8,Citation9). Thus, there remains the need to clarify the mechanisms by which slow HRR is associated with an increased risk of atherosclerotic cardiovascular disease.
Coronary artery calcification (CAC), an emerging marker of coronary atherosclerosis, has been shown to be a strong predictor of future cardiovascular events and mortality (Citation10,Citation11). To our knowledge, only two studies have examined the association between HRR and CAC (Citation12,Citation13) in western populations, and the results are conflicting, potentially due to confounding variables, differing methodologies for HRR assessment, and age differences in the study populations. In addition, the relation between slow HRR and CAC has not been previously reported in Asian populations. Therefore, we investigated the decrement in (HR) after peak or symptom-limited treadmill exercise testing as categorical and/or continuous variables and CAC in middle-aged and older Asian men, adjusting for potential confounders. We tested the hypothesis that slow HRR after exercise testing, signifying impaired autonomic (parasympathetic) function, is associated with a higher prevalence of advanced CAC in Korean men, independent of coronary risk factors. Because HRR after exercise testing may be influenced by the peak attained HR (Citation14,Citation15), we additionally examined whether the association between HRR and the presence of advanced CAC persisted after adjusting for peak HR and other potential confounders.
Materials and methods
Study participants
Participants were ∼14,000 men who underwent a general health examination as part of a comprehensive health-screening program at Samsung Medical Center, Seoul, South Korea, from January 2010 to December 2010. For early detection of chronic disease, this routine health screening program provided a physical examination, anthropometric measurements, blood pressure assessments, electrocardiography, and blood chemistry analysis; however, some screening programs also included adjunctive coronary artery computed tomography (n = 2200) and peak or symptom-limited treadmill exercise testing (n = 3400) with concomitant metabolic gas analysis. After excluding 102 participants who demonstrated positive exercise tests and an additional 112 subjects with missing and/or incomplete data, the remaining 2088 men (age 53 ± 6 years, range 41–78 years) with no known history of coronary heart disease, who underwent additional assessments of CAC and cardiopulmonary exercise testing, served as the study population for this cross-sectional investigation. Although participants were primarily self-referred, some were enrolled by their employers. Most participants lived in Seoul and surrounding areas and were well educated Asians of middle-to-upper socioeconomic status. The health examination program administered to individual clients largely depended on self-selection and the type of exam recommended by their physician. Accordingly, our evaluation was not based on a systemic research protocol but rather on the above-referenced variables, which were likely to be more comprehensive in middle-aged and older adults. Information regarding cigarette smoking was obtained by a self-reported questionnaire. Diabetes was defined as a fasting glucose level ≥126 mg/dl and/or self-reported by participants. Hypertension was defined as a systolic and/or diastolic blood pressure ≥140/90 mmHg or hypertension diagnosed by a physician. Written informed consent was obtained from all subjects before participating in the health-screening program, and the study was approved by the medical center institutional review board.
Coronary artery calcification measurements
CAC was measured using the multi-detector computed tomography system, Brilliance 40 (Phillips Medical Systems, Cleveland, OH) or VCT LightSpeed 64 (GE Healthcare, Milwaukee, WI). The total CAC score was computed by summing the CAC scores of all foci in the epicardial coronary system and systematically quantitated using the Agatston score (Citation16). The prevalence of advanced CAC was defined as a mean CAC >75th percentile for each age group (Citation17).
Measurements of heart rate recovery
Resting HR was measured in the supine position using an electrocardiogram (Hewlett-Packard ECG M 1700A, Hewlett Packard Enterprises, Madison, WI) following at least 5 min of quiet rest. Participants underwent peak or symptom-limited cardiopulmonary exercise testing using the conventional Bruce treadmill protocol. Expired gases were collected breath-by-breath using a one-way valve and analyzed for volume and composition. Cardiorespiratory fitness was directly measured using an automated metabolic measurement system (Jaeger Oxycon Delta, Eric Jaeger, Hoechberq, Germany) and defined as the highest or peak attained oxygen consumption (VO2peak), expressed as ml/kg/min, during progressive testing to volitional fatigue. Exercise HR was measured in the standing position, during each 3 min stage of exercise and throughout a 3 min recovery. An electrocardiogram was monitored continuously by oscilloscope (Quinton Q-4500, Bothell, WA) with 3-channel recordings obtained throughout the exercise test and 12-lead electrocardiograms recorded at the end of each stage, during peak or maximal exercise, and in recovery. Exercise tests were terminated for any of the following reasons: a rating of perceived exertion >17 (very hard work); achievement of >90% of age-predicted maximal HR; participant request, secondary to volitional fatigue; systolic blood pressure >250 mm/Hg; increasing chest discomfort (≥2/4); threatening ventricular arrhythmias; or, >1 mm of horizontal or downsloping ST segment depression. Peak HR was defined as the highest HR achieved during the test. HRR was calculated as the difference between the peak HR during the test and the HR at 2 min of recovery after peak exercise (Citation18). The recovery protocol included 1 min of slow, level walking (1.2 mph at 0% grade) immediately after peak exercise testing followed by 3 min of seated rest. HRR was assessed 2 min after cessation of peak exercise as previously described (Citation19,Citation20).
Other measurements
Body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). Resting blood pressure was obtained on all subjects in the seated position following ≥5 min of quiet rest using an automated sphygmomanometer (Dinamap PRO 100, Milwaukee, WI). Blood samples were collected following a 12 h overnight fast and analyzed. Total cholesterol (TC), low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), and triglycerides (TG) were analyzed by enzymatic colorimetric and liquid selective detergent methods, respectively, using a Hitachi 7600 (Hitachi Co. Tokyo, Japan) analyzer. Fasting glucose levels were determined using the Hexokinase, UV method (Hitachi-7600, Tokyo, Japan). High-sensitivity C-reactive protein (hsCRP) was measured using a CRP (II) Latax X2 turbidimetric method (Hitachi Corporation, Tokyo, Japan).
Statistical analysis
Data are presented as mean ± SD and proportion (%) for continuous with normal distribution and categorical variables, respectively. Continuous variables with non-normal distribution were expressed as a median with interquartile range. Participants were divided into quartiles by HRR. For group comparisons by quartiles of HRR, we used a one-way analysis of variance (ANOVA) for continuous variables with normal distribution or the Kruskal–Wallis test for continuous ones with non-normal distribution and the chi-square (χ2) test for categorical ones. The relation between HRR and CAC was evaluated using a Spearman’s rank correlation. We calculated odds ratios (OR) with 95% confidence interval (CI) from logistic regression to determine the relation of the HRR using categorical (quartiles) and continuous variables (per 1 bpm) to advanced CAC. An unadjusted model and multivariable models included age, BMI, systolic blood pressure, HDL-C, TG, glucose, hsCRP, cigarette smoking, hypertension, and diabetes as model 1, and peak HR as an additional potential confounder in model 2. The lowest quartile of HRR was used as a reference. Statistical significance was set at p < .05, and analyses were conducted using the SPSS 22.0 (SPSS, Armonk, NY).
Results
Demographic characteristics and cardiovascular risk factors of the participants, categorized by HRR quartiles, are presented in . Participants within the lowest quartile of HRR had greater cardiovascular disease risk factors, a higher prevalence of cigarette smoking, diabetes and hypertension, a higher resting HR, and lower peak HR and VO2peak than participants within the highest quartile of HRR. Participants within the lowest quartile of HRR also had significantly higher CAC scores than those within the highest quartile of HRR (p < .001).
Table 1. Characteristics of the participants according to heart rate recovery quartiles (n = 2088).
The median CAC score was 3.05 (interquartile range, 0–58) with 56.6% (n = 1183) of the cohort having a CAC score >0. The presence of CAC (>0) was inversely associated with HRR quartiles (Q1 65.3%, Q2 56.4%, Q3 52.1%, and Q4 51.0%; p < .001). CAC was weakly correlated with age (r = .34, p < .01), SBP (r = .09, p < .01), peak HR (r= −.23, p < .01), HR reserve (r = −.10, p < .01), VO2peak (r = −.20, p < .01), and HRR (r = −.14, p < .01), but not resting HR (r = .03, p = .137) in the univariate analysis.
shows that the prevalence of advanced CAC was inversely associated with increasing HRR quartiles of HRR (Q1 [lowest, <38 bpm] 31.5, Q2 25.4, Q3 22.6, and Q4 [highest, >52 bpm] 18.6%; p < .01 for trend). shows the OR and 95% CI of HRR quartiles with the prevalence of advanced CAC in a multivariable, adjusted logistic regression analysis. After adjusting for age, BMI, systolic blood pressure, HDL-C, TG, hsCRP, glucose, cigarette smoking, hypertension, and diabetes, participants in the lowest quartile of HRR were 1.68 times (95% CI: 1.24–2.28; p < .01) more likely to have advanced CAC than participants in the highest quartile (model 1). After adjusting for these variables and for peak HR as an additional potential confounder, participants in the lowest quartile of HRR were 1.59 times (95% CI: 1.17–2.18; p < .05) more likely to have advanced CAC than participants in the highest quartile of HRR (model 2). Each 1 bpm decrease in HRR as a continuous variable was associated with a 1% increase in the prevalence of advanced CAC after adjusting for conventional risk factors and peak HR (OR 1.01, 95% CI 1.00–1.02, p = .018).
Figure 1. Prevalence of advanced coronary artery calcification by quartiles of heart rate recovery (linear trend p < .001).
Table 2. Odd ratios and 95% CI of advanced coronary artery calcification by heart rate recovery quartiles.
Discussion
In this cross-sectional study, we found that participants in the lowest quartile of HRR after peak or symptom-limited treadmill exercise testing were more likely to have advanced CAC (>75th percentile based on age) than their counterparts in the highest quartile of HRR, that is, those who demonstrated the greatest decrement in HR after peak exercise. This association persisted even after adjusting for conventional risk factors as well as the peak attained HR, an important confounder in the HRR trend.
Although slow HRR is associated with an increased risk of cardiovascular events and mortality, the underlying pathophysiological mechanisms for these heightened risks have remained unclear. Because slow HRR, an estimate of impaired autonomic function (Citation3), is associated with sudden cardiac death (Citation1), it is important to clarify whether this relation is due to malignant ventricular arrhythmias and/or the presence or extent of atherosclerosis per se (Citation4). The present findings indicate that participants with impaired or slow HRR are more likely to have advanced CAC than their counterparts with faster HRR, supporting the hypothesis that this response is likely related to the burden of atherosclerotic coronary artery disease, as previously suggested (Citation8,Citation9).
We found that peak HR was related to the presence of advanced CAC in our univariate analysis – findings that are consistent with a previous study regarding the relation between low chronotropic responses (i.e., reduced peak HR), another indicator of autonomic imbalance, and an increased risk of CAC in men (Citation13). These findings further suggest that the prevalence of CAC is associated with impaired parasympathetic reactivation as well as sympathetic derangements. Admittedly, slow HRR after exercise testing may be influenced by a decreased peak HR (Citation14,Citation15). In this study, peak HR was lower in the lowest quartiles of HRR as compared with the highest quartiles of HRR. Therefore, the relation between HRR and the presence of advanced CAC required additional clarification whether this association persisted after adjusting for peak HR, and it did. Accordingly, slow HRR may have even greater utility as a harbinger of the burden of subclinical coronary artery disease, manifested as advanced CAC. However, additional studies are needed to elucidate whether slow HRR is independently associated with adverse cardiovascular outcomes after adjusting for peak HR and other potential confounders such as racial/ethnic differences.
Although the association between slow HRR and CAC is not fully understood, an attenuated HRR after exercise has been linked with an increased risk of subclinical atherosclerosis, suggesting that the possible underlying mechanisms may include abnormal risk factors (Citation21,Citation22), arterial stiffening (Citation5), endothelial dysfunction (Citation6), and inflammation (Citation23), all of which may accelerate atherosclerosis and CAC. In addition, impaired HRR is also related to traditional cardiovascular risk factors such as insulin resistance and hypertension, and these may, over time, increase levels of CAC. Slow HRR after peak or symptom-limited exercise testing may further reflect impaired parasympathetic reactivation, along with reduced sympathetic withdrawal (Citation3). Cardiac autonomic dysfunction is also associated with an increased risk of coronary atherosclerosis via decreased arterial compliance, increased vasomotor tone, and smooth muscle hypertrophy, fibrosis, and reduced elastin syndrome (Citation24–26). Collectively, the present results and aforementioned studies support the biologic plausibility for a relation between slow HRR and increased CAC.
Our results are consistent with these findings in that we observed associations between slow HRR and an increased risk of subclinical atherosclerosis in men. Others have reported that an abnormal 1 min HRR (≤15 bpm) in Caucasian women (mean age, 60 years) was independently associated with a heightened CAC burden (Citation13). Similarly, we demonstrated that slow HRR after exercise testing was associated with carotid atherosclerosis, another surrogate marker of subclinical atherosclerosis, independent of established risk factors in middle-aged men (Citation7). Another investigation reported a greater reduction in parasympathetic activity in response to a standardized psychological stressor that was associated with increased CAC in postmenopausal women (Citation27). Thus, associations between impaired HRR as categorical and/or continuous variables and advanced CAC have been observed in both Caucasian women and Asian men. In contrast, the coronary artery risk development in young adults (CARDIA) study reported that slow HRR in young adults was not associated with increased measures of CAC over a 15-year follow-up in Caucasian and African–American men and women (Citation12). These conflicting data may be attributed, at least in part, to differing methodologies for HRR assessment, racial/ethnic and sex or age differences in the study populations (Citation22). Participants in the CARDIA study were healthy young adults aged 18–30 years. In contrast, our subjects were middle-old aged and older Asian men (mean age 53 ± 6 years, range 41–78 years), who are more likely to demonstrate increased levels of CAC and decreased parasympathetic activity (Citation28). Younger populations and women are generally less likely to demonstrate CAC than older adults and men (Citation29,Citation30). In addition, the magnitude of the difference in CAC measures between Caucasian and Asian men appeared to be greater among those aged 45–74 years, and more modest in the younger age groups (Citation31). We believe that these methodologic differences may, at least in part, be responsible for the conflicting data in the literature, highlighting the need to further clarify the association between slow HRR and CAC after accounting for potential confounders.
We acknowledge some limitations when interpreting these results. Our study population included only men; thus, we are unable to determine whether this association extends to women. Due to the cross-sectional nature of our investigation, we cannot infer a cause-and-effect relation between HRR and increased CAC. Thus, additional longitudinal studies are needed to further clarify the relation between HRR and the progression of CAC across varied racial/ethnic and/or age cohorts. Although we adjusted for potential confounders, it is possible that residual variables that we did not measure may have influenced these associations. Because we measured CAC with two different scanners, discrepancies in Agatston scores may have occurred, based on the testing methodology employed. However, these discrepancies would need to vary systematically to produce the significant associations we observed. Finally, we did not control for dietary intake and prescribed cardioprotective medications such as β-blockers, which may potentially influence the relation between HRR and CAC.
The present results suggest that slow HRR to peak or symptom-limited treadmill exercise testing is associated with advanced CAC in middle-aged and older Korean men, independent of established risk factors and the peak attained HR. Therefore, advanced CAC may contribute to the higher incidence of cardiovascular disease in individuals with slow HRR immediately following exercise testing.
Disclosure statement
The authors have no conflicts of interest to declare.
References
- Jouven X, Empana JP, Schwartz PJ, Desnos M, Courbon D, Ducimetière P. Heart-rate profile during exercise as a predictor of sudden death. N Engl J Med. 2005;352:1951–8.
- Morshedi-Meibodi A, Larson MG, Levy D, O'Donnell CJ, Vasan RS. Heart rate recovery after treadmill exercise testing and risk of cardiovascular disease events (The Framingham Heart Study). Am J Cardiol. 2002;90:848–52.
- Imai K, Sato H, Hori M, Kusuoka H, Ozaki H, Yokoyama H, et al. Vagally mediated heart rate recovery after exercise is accelerated in athletes but blunted in patients with chronic heart failure. J Am Coll Cardiol. 1994;24:1529–35.
- Chaitman BR. Abnormal heart rate responses to exercise predict increased long-term mortality regardless of coronary disease extent: the question is why? J Am Coll Cardiol. 2003;42:839–41.
- Fei DY, Arena R, Arrowood JA, Kraft KA. Relationship between arterial stiffness and heart rate recovery in apparently healthy adults. Vasc Health Risk Manag. 2005;1:85–9.
- Huang PH, Leu HB, Chen JW, Cheng CM, Huang CY, Tuan TC, et al. Usefulness of attenuated heart rate recovery immediately after exercise to predict endothelial dysfunction in patients with suspected coronary artery disease. Am J Cardiol. 2004;93:10–13.
- Jae SY, Carnethon MR, Heffernan KS, Choi YH, Lee MK, Park WH, et al. Slow heart rate recovery after exercise is associated with carotid atherosclerosis. Atherosclerosis. 2008;196:256–61.
- Diaz LA, Brunken RC, Blackstone EH, Snader CE, Lauer MS. Independent contribution of myocardial perfusion defects to exercise capacity and heart rate recovery for prediction of all-cause mortality in patients with known or suspected coronary heart disease. J Am Coll Cardiol. 2001;37:1558–64.
- Lipinski MJ, Vetrovec GW, Froelicher VF. Importance of the first two minutes of heart rate recovery after exercise treadmill testing in predicting mortality and the presence of coronary artery disease in men. Am J Cardiol. 2004;93:445–9.
- Kramer CK, Zinman B, Gross JL, Canani LH, Rodrigues TC, Azevedo MJ, et al. 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:f1654.
- Detrano R, Guerci AD, Carr JJ, Bild DE, Burke G, Folsom AR, et al. Coronary calcium as a predictor of coronary events in four racial or ethnic groups. N Engl J Med. 2008;358:1336–45.
- Kizilbash MA, Carnethon MR, Chan C, Jacobs DR Jr, Lloyd-Jones DM, Sidney S, et al. The association of heart rate recovery immediately after exercise with coronary artery calcium: the coronary artery risk development in young adults study. Clin Auton Res. 2007;17:46–9.
- Möhlenkamp S, Lehmann N, Schmermund A, Roggenbuck U, Moebus S, Dragano N, Heinz Nixdorf Recall Study Investigators, et al. Association of exercise capacity and the heart rate profile during exercise stress testing with subclinical coronary atherosclerosis: data from the Heinz Nixdorf Recall study. Clin Res Cardiol. 2009;98:665–76.
- Zaim S, Schesser J, Hirsch LS, Rockland R. Influence of the maximum heart rate attained during exercise testing on subsequent heart rate recovery. Ann Noninvasive Electrocardiol. 2010;15:43–8.
- Cincin A, Sari I, Oğuz M, Sert S, Bozbay M, Ataş H, et al. Effect of acute sleep deprivation on heart rate recovery in healthy young adults. Sleep Breath. 2015;19:631–6.
- Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827–32.
- Raggi P, Callister TQ, Cooil B, He ZX, Lippolis NJ, Russo DJ, et al. Identification of patients at increased risk of first unheralded acute myocardial infarction by electron-beam computed tomography. Circulation. 2000;101:850–5.
- Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery after submaximal exercise testing as a predictor of mortality in a cardiovascularly healthy cohort. Ann Intern Med. 2000;132:552-5.
- Shetler K, Marcus R, Froelicher VF, Vora S, Kalisetti D, Prakash M, et al. Heart rate recovery: validation and methodologic issues. J Am Coll Cardiol. 2001;38:1980–7.
- Peçanha T, Silva-Júnior ND, Forjaz CL. Heart rate recovery: autonomic determinants, methods of assessment and association with mortality and cardiovascular diseases. Clin Physiol Funct. Imaging 2014;34:327–39.
- Jae SY, Carnethon MR, Ahn ES, Heffernan KS, Choi YH, Lee MK, et al. Association between heart rate recovery after exercise testing and plasminogen activator inhibitor 1, tissue plasminogen activator, and fibrinogen in apparently healthy men. Atherosclerosis. 2008;197:415–19.
- Arena R, Arrowood JA, Fei D, Shelar S, Helm S, Kraft KA. The influence of sex on the relationship between heart rate recovery and other cardiovascular risk factors in apparently healthy subjects. Scand J Med Sci Sports. 2010;20:291–7.
- Jae SY, Ahn ES, Heffernan KS, Woods JA, Lee MK, Park WH, et al. Relation of heart rate recovery after exercise to C-reactive protein and white blood cell count. Am J Cardiol. 2007;99:707–10.
- Huikuri HV, Jokinen V, Syvänne M, Nieminen MS, Airaksinen KE, Ikäheimo MJ, et al. Heart rate variability and progression of coronary atherosclerosis. Arterioscler Thromb Vasc Biol. 1999;19:1979–85.
- Colhoun HM, Francis DP, Rubens MB, Underwood SR, Fuller JH. The association of heart-rate variability with cardiovascular risk factors and coronary artery calcification: a study in type 1 diabetic patients and the general population. Diabetes Care. 2001;24:1108–14.
- Whelton SP, Blankstein R, Al-Mallah MH, Lima JA, Bluemke DA, Hundley WG, et al. Association of resting heart rate with carotid and aortic arterial stiffness: multi-ethnic study of atherosclerosis. Hypertension. 2013;62:477–84.
- Gianaros PJ, Salomon K, Zhou F, Owens JF, Edmundowicz D, Kuller LH, et al. A greater reduction in high-frequency heart rate variability to a psychological stressor is associated with subclinical coronary and aortic calcification in postmenopausal women. Psychosom Med. 2005;67:553–60.
- Ferrari AU, Radaelli A, Centola M. Invited review: aging and the cardiovascular system. J Appl Physiol (1985). 2003;95:2591–7.
- Devries S, Wolfkiel C, Fusman B, Bakdash H, Ahmed A, Levy P, et al. Influence of age and gender on the presence of coronary calcium detected by ultrafast computed tomography. J Am Coll Cardiol. 1995;25:76–82.
- Shisen J, Leung DY, Juergens CP. Gender and age differences in the prevalence of coronary artery calcification in 953 Chinese subjects. Heart Lung Circ. 2005;14:69–73.
- Fujiyoshi A, Miura K, Ohkubo T, Kadowaki T, Kadowaki S, Zaid M, et al. Cross-sectional comparison of coronary artery calcium scores between Caucasian men in the United States and Japanese men in Japan: the multi-ethnic study of atherosclerosis and the Shiga epidemiological study of subclinical atherosclerosis. Am J Epidemiol. 2014;180:590–8.