Ischemic heart disease (IHD) is a leading cause of death worldwide.[1,2] Reduction of total cholesterol (TC) has been an integral part of public health campaigns. TC has also been a major part of cardiovascular disease (CVD) risk prediction and prevention models.[3–5] For IHD prevention, “the lower, the better” cholesterol hypothesis has been accepted in the medical community for persons with a high risk of heart disease, especially men with manifest CVD, in whom heart disease mortality constituted approximately 50% of all deaths. In the Korean general population with a relatively low risk of heart disease, however, it is less clear whether the lower the cholesterol levels are, the lower the mortality risks of IHD are.
Through a large prospective cohort study including 503,340 Korean participants, we set out to obtain detailed information on the association between TC and mortality from IHD and overall CVD, and to estimate the cholesterol levels associated with the lowest mortality. In Koreans, IHD mortality constituted approximately 5% of all-cause mortality. Further, detailed estimates of the relative risk associated with TC levels <140 mg/dL could help inform decision-making in the clinical and public health settings for CVD prevention and management.
2.1 Data statement
Access to the data analyzed in this study is available from the National Health Insurance Service (NHIS) of Korea for researchers, upon review and approval of their study protocol by the NHIS.[8,9]
2.2 Study population and follow-up
Compulsory health insurance is provided by the NHIS to 97% of the population of Korea. The cohort of this study (n = 514,795) comprised a random sample of 10% of the 5.15 million individuals covered by the NHIS who were 40–79 years of age in 2002 and underwent health examinations in 2002 or 2003. From this sample, 11,455 subjects were excluded because information was missing on TC (n = 1710), body mass index (BMI), systolic blood pressure (SBP), or fasting glucose; because they had an extremely high BMI (≥50 kg/m2, n = 52); or because they had a history of heart disease and stroke (n = 9693). Record linkage with national death records was used to identify CVD deaths among the remaining 503,340 subjects through December 31, 2013. International Classification of Diseases-10th Revision codes were used to define cause-specific death: I00-I99 for CVD, I20-I25 for IHD, and I21 for acute myocardial infarction (AMI). This study was approved by the Institutional Review Board of Catholic Kwandong University, Republic of Korea. The requirement for informed consent was waived due to the use of anonymized data provided by the NHIS following thorough confidentiality protocols.
2.3 Data collection
Enzymatic methods were used to assay fasting serum TC and glucose levels. SBP was measured using a standard mercury sphygmomanometer, with the participant seated. Height was measured to the nearest centimeter and weight was measured to the nearest kilogram. BMI was calculated as weight (kilograms) divided by the square of height (meters) (kg/m2). Physical activity, alcohol use, and smoking history were assessed via a questionnaire. All health examinations and the data collection process followed a standard protocol officially registered by the Ministry of Health and Welfare. The Korean Association of Quality Assurance for Clinical Laboratory supervised external quality assessment for clinical chemistry, such as TC measurements, in participating hospitals, with regular assessments of assay quality.
2.4 Statistical analysis
Participants were subdivided into 8 groups by TC levels (<140, 140–159, 160–179, 180–199 [reference], 200–219, 220–239, 240–259, and ≥260 mg/dL). The reference category was defined as the category with the lowest CVD mortality. Log risk was regressed on TC as a continuous variable within the ranges of <200 mg/dL (the lower range), 200–349 mg/dL (the upper range), and <350 mg/dL (the full range), resulting in HRs per 39 mg/dL (1 mmol/L) increase in TC within each range. The data were also analyzed using a restricted cubic spline transformation analysis of TC with 3 knots (150, 190, and 230 mg/dL).
Cox proportional hazard models were used to calculate the HRs for CVD mortality stratified by baseline age (40–44, 45–54, 55–64, 65–74, and 75–80 years) after adjusting for baseline age (continuous variable; within each age group), sex, alcohol use (frequency; monthly or less, 2 days/month to 2 days/week, 3–7 days/week, and missing information [n = 9458]), smoking status (current smoker, former smoker, never smoker, and missing information [n = 21,282]), physical activity (at least once a week; yes or no), beneficiary income status (deciles; below 4 [low income], 4–7, 8–10 [high income]), SBP (continuous variable), fasting glucose (continuous variable), and BMI (continuous variable). In a sensitivity analysis, further adjustment was made for baseline lipid-lowering medication use (yes [n = 7742] or no).
The nonlinear associations of TC with CVD mortality were assessed with a likelihood ratio test. In this test, the model with only the linear term was compared to the corresponding model with both the linear and the cubic spline terms. Sensitivity analyses were conducted using subgroups with varying categories of TC.
All P values were 2-sided. All analyses used SAS version 9.4 (SAS Institute Inc., Cary, NC).
During a mean 10.5 years of follow-up, 2770 women and 4206 men died from CVD. The number of deaths from IHD, AMI, total stroke, hemorrhagic stroke, and ischemic stroke were 1313, 1008, 1927, 631, and 622, respectively, in men and 606, 467, 1456, 553, and 422, respectively, in women. At baseline the mean TC level was 200.4 ± 38.7 mg/dL (202.5 ± 39.4 mg/dL in women; 198.7 ± 38.1 mg/dL in men) and the mean age was 52.9 ± 9.7 years (Table 1). Individuals with higher TC values tended to be older, female, never smokers, and infrequent alcohol users, and more likely to have higher blood pressure, fasting glucose, and BMI (Table S1, http://links.lww.com/MD/D210).
3.2 TC and CVD mortality
The associations between TC and CVD mortality were generally nonlinear. In the sex- and age-adjusted analysis, U-curve associations between TC and overall CVD mortality were found, with a nadir at 180–199 mg/dL (Fig. 1). Both above and below this range, progressive excess mortality from CVD was observed. Among the CVD subtypes, mortality from IHD (including AMI) increased with cholesterol ≥200 mg/dL, but it was not associated with TC < 200 mg/dL (a reverse L-curve).
After adjustment for other confounders, the associations generally did not substantially change (Table S2, http://links.lww.com/MD/D210). Further adjustment for lipid-lowering medication use at baseline yielded no substantial change (Table S3, http://links.lww.com/MD/D210).
In the restricted cubic spline analysis, the patterns of associations and TC values associated with the lowest mortality were generally same as in the categorical analysis of TC, and the nonlinear associations were statistically confirmed (Fig. 2).
Assuming a linear association below 200 mg/dL, TC was inversely associated with CVD mortality (Fig. 3; HR per 39 mg/dL higher level = 0.90). Assuming a linear association in the upper range, TC was positively associated with CVD mortality, largely due to a strong association with IHD (HR = 1.19), especially AMI (HR = 1.23). In the full range (<350 mg/dL), assuming linear relationships, TC was positively associated with mortality from overall CVD, including IHD.
3.3 Subgroup analysis
Women had lower HRs for overall CVD mortality in the upper range of cholesterol (Fig. 4). For IHD mortality, HRs in the upper range were generally lower in women than in men, but with overlapping confidence intervals (Figs. 5 and 6). In a linear analysis in the upper range of cholesterol, middle-aged adult did not have significantly stronger associations than elderly adults (Figs. S1 and S2, http://links.lww.com/MD/D210), while men did not have significantly higher HRs than women for IHD mortality (Figs. S3 and S4, http://links.lww.com/MD/D210; e.g., for AMI mortality, the HR in men was 1.27 [95% CI = 1.12–1.43], whereas it was 1.16 in women [1.00–1.35], Pheterogeneity = 0.374). In the categorical analyses, except for the highest (≥260 mg/dL) and lowest (<140 mg/dL) TC categories, TC categories were not associated with statistically significantly higher mortality from CVD or IHD compared to the reference category (Table S2, http://links.lww.com/MD/D210).
Middle-aged (40–64 years old) and non-hypertensive persons (SBP < 140 mm Hg) had higher estimated relative risks for mortality from CVD or IHD than elderly (aged ≥60 years) and hypertensive (SBP ≥140 mm Hg) persons both in the lower and upper range, although the confidence intervals overlapped between comparison groups (Figs. 4–6). In the elderly, the HRs were 1.11 (95% CI = 1.04–1.18) for overall CVD, 1.16 (1.04–1.31) for IHD, and 1.19 (1.05–1.36) for AMI, in the upper range.
3.4 Sensitivity analysis using standard classification of TC
High TC levels (≥240 mg/dL), but not borderline levels (200–239 mg/dL), were associated with greater CVD mortality, compared to the desirable level of <200 mg/dL (Tables S4 and S5, http://links.lww.com/MD/D210). For IHD, both borderline and high levels were associated with greater mortality.
TC had nonlinear associations with CVD mortality, U-curve associations for overall CVD, and reverse-L-curve associations for IHD, including AMI. CVD mortality was lowest at TC levels of 180–199 mg/dL, while above and below this range, progressive excess mortality from CVD was found. For IHD mortality, a positive graded association was found in the upper range (≥200 mg/dL), while no association was found in the lower range (<200 mg/dL).
4.1 ischemic heart disease (IHD)
Many previous studies have reported a continuous graded association of cholesterol levels with IHD. Surprisingly, however, few studies have provided information on the risks of IHD associated with detailed TC categories covering the range <180 mg/dL,[13–16] perhaps due to the small number of individuals with TC<180 mg/dL in European-origin populations. Upon closer scrutiny, previous studies have provided evidence that TC is strongly and directly related to IHD mortality, without evidence of a threshold, down to around 180–200 mg/dL but not below this range,[13–16] and our study yielded further supporting evidence of this finding. This nonlinear association of TC with IHD is a novel, but not improbable, finding given the evidence of a nonlinear association of SBP with CVD, including IHD . Several studies found similar results: below around 180–200 mg/dL, there were no positive associations between cholesterol levels and IHD.[13,15,18,19] These studies, however, have not considered non-linear associations. Statin trials showed that more intensive statin use reduced IHD mortality more than less intensive use, although statin trials among Asian populations have shown no clear advantage against IHD mortality.[20,21] However, the lack of trials examining whether lower cholesterol targets have more beneficial effects indicates that evidence from clinical trials may not be definitive enough to confirm that “the lower, the better.” Additionally, the achieved cholesterol levels had reverse-L-curve, or even U-curve, associations with vascular morbidity and mortality in some,[7,23,24] but not all, observational studies in individuals taking statins. Overall, in Korean adults, the lowest cholesterol levels were not associated with the lowest mortality, and no positive association was found in the levels <200 mg/dL, whereas a positive association was found in the range ≥200 mg/dL.
Our estimated relative risks associated with 1 mmol/L increases in TC (HR = 1.19, in the upper range) were comparable to those in a recent systematic review, but were lower than those reported in a large US study including men aged 35–57 years.
4.2 Overall CVD
U-curve associations were found between TC and CVD in our study. A recent review claimed that cholesterol had no association or an inverse association with CVD mortality in the elderly. Their conclusion, however, did not consider subtypes of CVD or nonlinear associations between cholesterol and CVD. The shape of the association of TC with mortality was different across CVD subtypes: a reverse L-curve for IHD was found in the current study; U-curves for total stroke, and an L-curve for hemorrhagic stroke, especially intracerebral hemorrhage, were found in our previous study of the same study population. The negative association in the lower range is mostly explained by hemorrhagic stroke, while the positive association in the upper range is largely accounted for by IHD. In our study, in the upper range, each 1 mmol/L higher TC was associated with 11% higher mortality from CVD (95% CI, 4%–18%), 16% for IHD (4%–31%), and 19% for AMI (5%–36%) in the elderly (Figs. 4–6). The corresponding risk associated with stroke mortality was 11% (2%–21%) in the elderly, as shown in our previous research. Additionally, in the upper range, no CVD subtypes had negative associations with TC, while in the lower range, mortality from CVD, especially ICH, was inversely associated with TC. These inverse associations of TC remained stable after adjusting for high-density lipoprotein cholesterol in Korean and Japanese populations.[28,29] The higher mortality from CVD associated with lower cholesterol levels might be the effect of a long-term period of lower levels, or reflect the effect of this factor in individuals with a low risk of heart disease, which most randomized trials cannot reflect. Additionally, the results from a recent systematic review, stating that a greater reduction of cardiovascular mortality was not found in groups with more intensive cholesterol-lowering regimens compared those with less intensive regimens when the baseline low-density lipoprotein cholesterol (LDL-C) level was less than 100 mg/dL, agrees with the current study to some degree.
Both cubic spline and categorical analysis plots suggested that women had a higher TC range associated with the lowest CVD mortality (around 200–220 mg/dL; approximately equivalent to 125–140 mg/dL of LDL-C in Koreans) than men (around 180–200 mg/dL; roughly equivalent to 110–125 mg/dL of LDL-C), and that women might have weaker positive associations in the upper range than men. However, women had no materially weaker associations with overall CVD or any subtype. Upon closer scrutiny, combining the impacts of a modestly lower HR in the upper range in women than in men with the larger proportion of hemorrhagic stroke deaths and lower proportion of IHD deaths among CVD mortality in women than in men in Korea during the follow-up period [9,31] would explain this finding. This result suggests that for overall CVD, the pattern of associations and the TC range associated with the lowest risk may differ across regional and ethnic populations or different time periods in a population, since the distribution of CVD subtypes varies by time period, region, and ethnicity.[32,33]
The patterns of associations for CVD and its subtypes were generally similar between middle-aged (40–64 years) and elderly (≥65 years) adults, as well as non-hypertensive and hypertensive persons, which may be in discordance with a collaborative study in which age and blood pressure seemed to affect the strength and the direction of the association. That collaborative study, however, collected information from 61 prospective studies with different regions, ethnicities, and time periods.[9,14] An analysis of 65,594 Japanese adults reported that the associations between TC levels and deaths from IHD were similar in each sex and age subgroup, except for elderly women, while the confidence interval associated with each TC category generally overlapped across sex and age subgroups. The Evidence for Cardiovascular Prevention from Observational Cohorts in Japan (EPOCH–JAPAN) and the Asia Pacific Cohort Studies Collaboration (APCSC) suggested a significant interaction between SBP and TC for CVD; however, the suggested patterns of interaction were different between those studies.[35,36] The EPOCH-JAPAN study reported that the HRs associated with TC were greater in persons with higher SBP, whereas the associations were greater in persons with lower SBP in the APCSC study.
4.3 Strengths and limitations of the study
The main strengths of this study include its prospective design, large sample size, and complete follow-up for mortality. Another strength is the fact that our study population consisted of ethnically homogeneous individuals covered by the same health care system and living in a similar environment. A further strength is that our study estimated the risk associated with TC levels including those lower than 140 mg/dL. Nonetheless, this study has several limitations. First, the fact that the study population was homogenously Korean may affect the generalizability of our findings. The cholesterol-lowering effects of statins have reported to be stronger in Asians than in Western populations. The beneficial effects of intensive statin therapy over moderate-dose therapy have not been confirmed in Asian populations. The genetic factors related to cholesterol biosynthesis, cholesterol transport, and statin metabolism may be different across ethnic groups to some degree. Considering these differences between ethnic groups, some results, such as the magnitude of relative risk associated with TC for CVD and IHD deaths and the TC range associated with the lowest mortality, may need further assessment in other populations with varying TC levels, health care utilization, distribution of CVD subtypes, and distribution of environmental and individual risk factors. Second, cholesterol-lowering therapy use during follow-up was unaccounted for in the analysis. However, due to relatively lower statin use in the Korean population (<10% in persons with hypercholesterolemia), the effects of not considering lipid-lowering medication during follow-up were assumed to be modest. Third, this study used a single measurement of TC at baseline. The estimated relative risk may underestimate the true association, due to a regression dilution effect. Fourth, information on the cause of death obtained from death certificates may be imperfect due to misclassifications, but a comparison of death certificates with medical records in Korea showed them to be reasonably valid. Additionally, the pattern of potential misclassification tends not to vary according to TC levels, so this factor is not likely to have resulted in an overestimation of the risk.
U-curve and reverse L-curve associations were found between TC and mortality from overall CVD and IHD, respectively, in Korean adults who had a relatively low risk of heart disease mortality. For overall CVD, TC levels of 180–200 mg/dL were associated with the lowest mortality. TC had no positive associations with IHD mortality in the lower range. The shape of the association for overall CVD, however, might be different across ethnic groups, particularly due to the distribution of CVD subtypes in a population. The associations for CVD and IHD mortality were generally similar in men versus women and in middle-aged (40–64 years) versus elderly (≥65 years) adults.
The authors thank the staff at the Big Data Steering Department at the NHIS for providing the data and support.
Conceptualization: Sang-Wook Yi.
Data curation: Sang-Wook Yi.
Formal analysis: Daeho Kwon, Sang-Wook Yi, Jee-Jeon Yi.
Investigation: Sang-Wook Yi.
Methodology: Daeho Kwon, Sang-Wook Yi, Jee-Jeon Yi, Heechoul Ohrr.
Visualization: Jee-Jeon Yi.
Writing – original draft: Daeho Kwon, Sang-Wook Yi.
Writing – review & editing: Daeho Kwon, Sang-Wook Yi, Jee-Jeon Yi, Heechoul Ohrr.
Sang-Wook Yi orcid: 0000-0002-6656-6205.
. Benjamin EJ, Blaha MJ, Chiuve SE, et al. Heart Disease
and Stroke Statistics-2017 Update: A report from the American Heart Association. Circulation 2017;135:e146–603.
. DALYs Hale GBD Collaborators. Global, regional, and national disability-adjusted life-years (DALYs) for 315 diseases and injuries and healthy life expectancy (HALE), 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016;388:1603–58.
. Conroy RM, Pyorala K, Fitzgerald AP, et al. Estimation of ten-year risk of fatal cardiovascular disease
in Europe: the SCORE project. Eur Heart J 2003;24:987–1003.
. D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation 2008;117:743–53.
. Hippisley-Cox J, Coupland C, Robson J, et al. Derivation, validation, and evaluation of a new QRISK model to estimate lifetime risk of cardiovascular disease
: cohort study using QResearch database. BMJ 2010;341:c6624.
. Baigent C, Blackwell L, et al. Cholesterol Treatment Trialists Collaboration. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet 2010;376:1670–81.
. Leibowitz M, Karpati T, Cohen-Stavi CJ, et al. Association between achieved low-density lipoprotein levels and major adverse cardiac events in patients with stable ischemic heart disease
taking statin treatment. JAMA Intern Med 2016;176:1105–13.
. Seong SC, Kim YY, Park SK, et al. Cohort profile: the National Health Insurance Service-National Health Screening Cohort (NHIS-HEALS) in Korea. BMJ Open 2017;7:e016640.
. Yi SW, Shin DH, Kim H, et al. Total cholesterol and stroke mortality in middle-aged and elderly adults: a prospective cohort study. Atherosclerosis 2018;270:211–7.
. Yi SW, Ohrr H, Shin SA, et al. Sex-age-specific association of body mass index with all-cause mortality among 12.8 million Korean adults: a prospective cohort study. Int J Epidemiol 2015;44:1696–705.
. Seong SC, Kim YY, Khang YH, et al. Data resource profile: The National Health Information Database of the National Health Insurance Service in South Korea. Int J Epidemiol 2017;46:799–800.
. Min WK, Kim KD, Kim DJ, et al. Annual report on external quality assessment in clinical chemistry in Korea (2002). J Lab Med Qual Assur 2003;25:1–4.
. Choi JS, Song YM, Sung J. Serum total cholesterol and mortality in middle-aged Korean women. Atherosclerosis 2007;192:445–7.
. Lewington S, Whitlock G, Clarke R, et al. Prospective Studies Collaboration. Blood cholesterol
and vascular mortality by age, sex, and blood pressure: a meta-analysis of individual data from 61 prospective studies with 55,000 vascular deaths. Lancet 2007;370:1829–39.
. Song YM, Sung J, Kim JS. Which cholesterol level is related to the lowest mortality in a population with low mean cholesterol level: a 6.4-year follow-up study of 482,472 Korean men. Am J Epidemiol 2000;151:739–47.
. Stamler J, Wentworth D, Neaton JD. Is relationship between serum cholesterol and risk of premature death from coronary heart disease
continuous and graded? Findings in 356,222 primary screenees of the Multiple Risk Factor Intervention Trial (MRFIT). JAMA 1986;256:2823–8.
. Yi SW, Mok Y, Ohrr H, et al. Low systolic blood pressure and vascular mortality among more than 1 million Korean adults. Circulation 2016;133:2381–90.
. Goldbourt U, Yaari S. Cholesterol and coronary heart disease
mortality. A 23-year follow-up study of 9902 men in Israel. Arteriosclerosis 1990;10:512–9.
. Nago N, Ishikawa S, Goto T, et al. Low cholesterol is associated with mortality from stroke, heart disease
, and cancer: the Jichi Medical School Cohort Study. J Epidemiol 2011;21:67–74.
. Nakamura H, Arakawa K, Itakura H, et al. Primary prevention of cardiovascular disease
with pravastatin in Japan (MEGA Study): a prospective randomised controlled trial. Lancet 2006;368:1155–63.
. Zhao SP, Yu BL, Peng DQ, et al. The effect of moderate-dose versus double-dose statins on patients with acute coronary syndrome in China: results of the CHILLAS trial. Atherosclerosis 2014;233:707–12.
. Chou R, Dana T, Blazina I, et al. Statins for prevention of cardiovascular disease
in adults: evidence report and systematic review for the US Preventive Services Task Force. JAMA 2016;316:2008–24.
. Matsuzaki M, Kita T, Mabuchi H, et al. Large scale cohort study of the relationship between serum cholesterol concentration and coronary events with low-dose simvastatin therapy in Japanese patients with hypercholesterolemia. Circ J 2002;66:1087–95.
. LaRosa JC, Grundy SM, Kastelein JJ, et al. Safety and efficacy of Atorvastatin-induced very low-density lipoprotein cholesterol levels in Patients with coronary heart disease
(a post hoc analysis of the treating to new targets [TNT] study). Am J Cardiol 2007;100:747–52.
. Wiviott SD, Cannon CP, Morrow DA, et al. Can low-density lipoprotein be too low? The safety and efficacy of achieving very low low-density lipoprotein with intensive statin therapy: a PROVE IT-TIMI 22 substudy. J Am Coll Cardiol 2005;46:1411–6.
. Peters SA, Singhateh Y, Mackay D, et al. Total cholesterol as a risk factor for coronary heart disease
and stroke in women compared with men: a systematic review and meta-analysis. Atherosclerosis 2016;248:123–31.
. Ravnskov U, Diamond DM, Hama R, et al. Lack of an association or an inverse association between low-density-lipoprotein cholesterol and mortality in the elderly: a systematic review. BMJ Open 2016;6:e010401.
. Bae JM, Yang YJ, Li ZM, et al. Low cholesterol is associated with mortality from cardiovascular diseases: a dynamic cohort study in Korean adults. J Korean Med Sci 2012;27:58–63.
. Hirata T, Sugiyama D, Nagasawa SY, et al. A pooled analysis of the association of isolated low levels of high-density lipoprotein cholesterol with cardiovascular mortality in Japan. Eur J Epidemiol 2017;32:547–57.
. Navarese EP, Robinson JG, Kowalewski M, et al. Association between baseline LDL-C level and total and cardiovascular mortality after LDL-C lowering: a systematic review and meta-analysis. JAMA 2018;319:1566–79.
. Lee SW, Kim HC, Lee HS, et al. Thirty-year trends in mortality from cerebrovascular diseases in Korea. Korean Circ J 2016;46:507–14.
. Mortality Causes of Death GBD Collaborators. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016;388:1459–544.
. Kim BJ, Kim JS. Ischemic stroke subtype classification: an Asian viewpoint. J Stroke 2014;16:8–17.
. Nagasawa SY, Okamura T, Iso H, et al. Relation between serum total cholesterol level and cardiovascular disease
stratified by sex and age group: a pooled analysis of 65 594 individuals from 10 cohort studies
in Japan. J Am Heart Assoc 2012;1:e001974.
. Satoh M, Ohkubo T, Asayama K, et al. Combined effect of blood pressure and total cholesterol levels on long-term risks of subtypes of cardiovascular death: evidence for cardiovascular prevention from observational cohorts in Japan. Hypertension 2015;65:517–24.
. Asia Pacific Cohort Studies
C. Joint effects of systolic blood pressure and serum cholesterol on cardiovascular disease
in the Asia Pacific region. Circulation 2005;112:3384–90.
. Ko DT, Alter DA, Guo H, et al. High-density lipoprotein cholesterol and cause-specific mortality in individuals without previous cardiovascular conditions: the CANHEART study. J Am Coll Cardiol 2016;68:2073–83.
. Naito R, Miyauchi K, Daida H. Racial differences in the cholesterol-lowering effect of statin. J Atheroscler Thromb 2017;24:19–25.
. Lee YH, Lee SG, Lee MH, et al. Serum cholesterol concentration and prevalence, awareness, treatment, and control of high low-density lipoprotein cholesterol in the Korea National Health and Nutrition Examination Surveys 2008-2010: Beyond the tip of the iceberg. J Am Heart Assoc 2014;3:e000650.
. Won TY, Kang BS, Im TH, et al. The study of accuracy of death statistics. J Korean Soc Emerg Med 2007;18:256–62.