Abstract
Aims: The associations between low-density lipoprotein cholesterol (LDL-C) and the risk of cardiovascular disease (CVD) subtypes are not well established among the Japanese population. This study used longitudinal data from the Aichi Workers’ Cohort Study to explore the association between LDL-C levels and the risk of coronary heart disease (CHD) and stroke subtypes.
Methods: Pooled data of 8966 adults (7093men and 1903 women) who were recruited between (2002) and (2008) were used for the current analysis. Propensity scores for the LDL-C categories were generated using multinomial logistic regression. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated from the inverse probability weighted Cox proportional hazards model for LDL-C category associations with risks of CHD, stroke subtypes, and CVD.
Results: During a median follow-up of 12 years, 122 strokes (57 ischemic strokes, 25 intracerebral hemorrhage, and 40 unknown subtypes) and 82 cases of CHD were observed. LDL-C 160– mg/dL compared to LDL-C 100–119 mg/dL was positively and significantly associated with the risk of CHD (HR: 4.56; 95% CI: 1.91–10.9) but not with ischemic stroke (HR: 0.99; 95% CI: 0.44–2.22). LDL-C was inversely associated with the risk of intracerebral hemorrhage (P for trend=0.009).
Conclusion: In middle-aged Japanese workers, LDL-C was significantly and positively associated with CHD, but not with ischemic stroke. LDL-C was inversely significantly associated with intracerebral hemorrhage.
See editorial vol. 30: 432-433
Introduction
Low-density lipoprotein cholesterol (LDL-C) is an established causal factor for coronary heart disease (CHD) and ischemic stroke1, 2). Reducing the blood level of LDL-C has been the main target of lipid management for the prevention of atherosclerotic cardiovascular diseases (CVDs)3-7). A linear positive association between LDL-C and risk of CHD or ischemic stroke has been confirmed in Western populations4, 8, 9), but it remains less evident in Japanese population, particularly with regard to the risk of ischemic stroke3, 10). Indeed, the results of previous prospective studies are inconsistent10-12). The null association for ischemic stroke in Japan may be due to differences in the subtype mix11). Lacunar stroke is still more prevalent in Japan than in Western countries, where ischemic stroke resulting from large artery atherosclerosis is more common13).
Alternatively, although the incidence rate of CHD among the Japanese has been much lower than those in the Western population8, 14, 15), previous long-term Japanese prospective studies initiated in the 1980s already found a positive association between LDL-C categories and the risk of CHD10-12). Newer studies also demonstrated positive association of LDL-C levels with CHD in men16, 17) and in people younger than 50 years18). However, it would be noticeable that the Japanese had experienced an increase in the blood level of total cholesterol from 1980s to the 1990s 19, 20), and the level was reportedly higher among urban populations than among their rural counterparts21), which warrants further investigation on this issue in urban contemporary populations.
Aim
The aim of this study is to investigate the association of LDL-C with CHD and stroke subtypes in middle-aged workers from a metropolitan area in Japan recruited more recently, who manifested higher LDL-C levels (mean LDL-C = 124.6 mg/dL) than some cohorts with early baseline settings12, 16, 17). We employed a propensity score method, or inverse probability of weighting (IPW) technique, which has increasingly been used in observational studies to control for confounding since traditional multivariable regression methods may result in bias when the number of events is small due to overfitting22).
Methods
Study Population
The Aichi Workers’ Cohort Study is an ongoing prospective epidemiological study of noncommunicable diseases, including diabetes and CVD. It was initially established in 1997 and includes local government workers in Aichi Prefecture, which consists of urban and suburban areas and is located in central Japan. Most workers in the cohort are engaged in clerical work. The present analysis included participants who were recruited in 2002, 2005 and 2008. From the eligible participants (N=10217), those without LDL-C (N=985) at baseline, with a history of CVD (N=102) and with follow-up less than one year (N=164) were excluded, leaving 8966 participants for the analyses. Written informed consent was obtained for their participation, including our use of the results of mandatory annual health checkups provided by the worksites. The study protocol was approved by the Ethics Review Committees of Nagoya University School of Medicine (2007-0504) and Fujita Health University (HM17-470 and HM19-018).
LDL-C Data and the Use of Dyslipidemia Medication
In 2002, we stored residual blood samples used for the annual health checkup, for which participants were instructed to fast overnight or for at least eight hours. Even though LDL-C levels in 2002 were directly measured enzymatically in a commercial laboratory (BML, Inc.) using the stored blood samples, we first used LDL-C values estimated by the following Friedwald formula: total cholesterol − HDL-C − (triglycerides/5). If triglycerides ≥ 400 mg/dL, we used directly measured LDL-C (N=72). We used total cholesterol, triglycerides and HDL cholesterol values obtained for the annual health checkup in 2002 for the calculation instead of those measured in a laboratory since the number of participants who agreed to provide health checkup data was much higher than that of the participants who agreed to provide residual blood samples (N=5596 vs. 4213). Nevertheless, the agreement between directly measured LDL-C and calculated LDL-C in 2002 evaluated by the intraclass correlation coefficient was high (0.98, 95% confidence interval (CI): 0.97–0.98). LDL-C values in participants recruited in 2005 were estimated using the Friedwald formula only with total cholesterol, triglycerides and HDL cholesterol obtained for the annual health checkup in 2005. In participants recruited in 2008, only directly-measured LDL-C values were available, which were obtained for the annual health checkup in 2008. The same instructions were given to the participants regarding the duration of fasting during the annual health checkups. LDL-C values were then grouped into five categories: ＜ 100, 100–119 (reference), 120–139, 140–159, and 160– mg/dL. Dyslipidemia medication use was self-reported. We previously validated the accuracy of the self-report of medical history of several diseases among the Japanese population and the accuracy of self-report of hyperlipidemia was 95%23).
Incidence of CHD and Stroke
Incidence was ascertained from the following three sources. First, we conducted a biennial self-administered questionnaire survey on the medical history of specific conditions including CHD and stroke subtypes. The participants who reported their medical history were asked to provide the names of physicians and hospitals where their conditions were treated as well as the consent for us to inquire about the details of their medical history. We then asked these physicians to provide the participant’s medical history using a standardized form so that we could validate and determine the details of the self-report (type, location, severity, procedure and treatment, etc.). In this study, CHD included myocardial infarction and unstable angina followed by percutaneous coronary intervention. A diagnosis of myocardial infarction was made according to the modified World Health Organization MONICA Project (monitoring trends and determinants in cardiovascular disease)24) criteria, in which we also included cardiac troponin as a marker of cardiac injury. In addition, confirmation of acute coronary artery occlusion by angiography or of abnormal cardiac wall motion that correspond to the clinical findings were employed as the definition. The incidence of stroke was confirmed by a physician by using the definition of sudden diffuse or focal neurological deficit due to cerebrovascular disorder, which persists for a minimum of 24 hours. A definite diagnosis must have computed tomography or magnetic resonance imaging findings that are consistent with the signs and symptoms.
Second, the incidence was also ascertained from the worksite’s healthcare division by using information from the workers when they apply for a sick leave. These cases were also classified as definite because they are usually accompanied by medical certificates. Fatal cases were also determined using this channel. Finally, we used histories of myocardial infarction and stroke that were self-reported in the form used for annual mandatory health checkups. These cases were classified as probable.
Retirees were followed up using a biennial self-administered questionnaire only. The next of kin of the participants was asked to provide details on whether the participants were deceased or unable to respond. Details of the histories of the retirees who self-reported were ascertained by their physicians when it is possible.
This study included both definite and probable cases. Nevertheless, we performed sensitivity analyses excluding probable CVD cases by censoring them at the date of the occurrence. Total stroke included ischemic stroke, intracerebral hemorrhage, subarachnoid hemorrhage, and stroke of unknown subtypes. Types of endpoints used in the analyses were CHD, total stroke, ischemic stroke, intracerebral hemorrhage, and CVD (CHD ＋ total stroke).
Follow-Up
We conducted surveillance for the occurrence of the endpoints through March 2018. Participants were censored when they developed the aforementioned endpoints, at the date of our last valid access to them, or the date of their retirement unless they agreed to provide their postal address for us to periodically contact about their health statuses after retirement (approximately 52.1 % of the retired participants). The date of our last valid access for the retirees was set as the date of response to the biennial self-administered questionnaire. The date of the last access for those who were in the worksite was set as the end of March, 2018. Person-years were calculated from the baseline date to the date of censoring or the ascertainment of incident CVD. We performed a sensitivity analysis excluding participants with less than three years of follow-up.
Definition of Covariates
Body mass index (BMI) was calculated as body weight (kg) divided by the square of height (m2). Height was measured in the standing position to the nearest 0.1 cm, and body weight was measured to the nearest 0.1 kg at the annual health checkup. Blood pressure was obtained in the seated position. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, and/or taking hypertension medication. Diabetes was defined as fasting blood glucose ≥ 126 mg/dL and/or taking diabetes medication. The use of blood pressure lowering medication and/or glucose lowering medication was self-reported. Alcohol consumption was estimated on the basis of the type of alcohol, the amount of each type, and the frequency of drinking in 2002. In 2005 and 2008, the participants were given an instruction to convert the amount of their alcohol intake into Japanese standard unit (go) instead of assessing the type of alcohol. The participants were then grouped into the low consumption group (＜ 40 g/day for men and ＜ 20 g/day for women) and high consumption group (≥ 40 g/day for men and ≥ 20 g/day for women). Smoking was categorized as current smoker or nonsmoker. Physical activity was defined as the number of times that a participant engaged in vigorous or moderate exercise for ≥ 60 min/month. A self-administered questionnaire was used at each survey (2002, 2005, and 2008) to assess smoking, alcohol consumption, and physical activity.
Statistical Analysis
All the statistical analyses were performed using R software and the following packages: mice, miceadds, MatchThem, Cobalt, Survival, Survey, rms. A P-value ＜ 0.05 was considered to indicate statistical significance. Missing data were imputed using multiple imputation with the R package “mice”. This method generates 10 datasets with imputed missing values. The results from each dataset were combined to pool parameter estimates. Differences in continuous and categorical variables across LDL-C categories were tested by analysis of variance and chi-squared test, respectively.
In examining the associations of LDL-C with the incidence of CHD and stroke, the IPW of propensity score was used to control for the baseline confounders22). Namely, multinominal logistic regression was applied to calculate the probability of each individual being in a certain LDL-C category with the following covariates: age, sex, smoking, alcohol consumption, physical activity, BMI, diabetes, hypertension, HDL-C, triglycerides, antihypertensive medication, and survey year. The resulting probability was used to generate weight for each individual using R package “MatchThem”25). We trimmed the weight to the 1st and 99th percentile considering the potential influence of extreme weights. The balance among LDL-C categories after applying the weight was checked by the absolute standardized mean difference (SMD) with the cutoff of ＜ 0.1 using the R package “cobalt”26). Cox proportional hazard model was performed with R package “Survival”27) after incorporating weights into the imputed datasets using R package “Survey” (Model 1)28). In addition, different propensity score was generated with an additional covariate (dyslipidemia medication) and used in the analysis (Model 2).
For the assessment of non-linear association, we performed multivariable-adjusted restricted cubic spline analyses with four knots at the 5th, 35th, 65th, and 95th percentiles of LDL-C distribution with the R package “rms”29, 30) for each endpoint. In the spline analyses, we excluded participants with extreme LDL-C at the 1st and 99th percentile in an attempt to avoid potential influence of outliers.
Results
Age, the proportion of men, and that of high alcohol intake differed significantly according to LDL-C categories (Table 1). BMI, triglycerides, HDL-C, and the proportions of hypertension, diabetes and dyslipidemia medication use and physical activity were also associated with LDL-C categories.
Table 1.
Baseline characteristics of the participants according to the LDL-C categories, Aichi Workers’ Cohort, 2002–2007 (N = 8966)
|
LDL-C mg/dL groups |
|
Variables
N
|
All
8966
|
＜100
1946
|
100–119
2213
|
120–139
2214
|
140–159
1491
|
160
1132
|
P-value
|
| Mean, mg/dL |
124.6 |
84.6 |
110.0 |
129.6 |
148.7 |
180.2 |
＜0.001 |
| Age, years |
45.6 |
43.9 |
45.1 |
45.9 |
47.0 |
47.2 |
＜0.001 |
| Men (%) |
79.0 |
72.9 |
76.7 |
81.8 |
83.0 |
81.5 |
＜0.001 |
| Physical activity, days/month＊
|
8.7 |
8.9 |
8.8 |
8.6 |
8.8 |
7.8 |
0.036 |
| Current smoker, (%) |
31.4 |
32.2 |
30.0 |
30.0 |
31.9 |
35.1 |
0.115 |
| Alcohol intake, (High risk %)＊＊
|
10.6 |
16.6 |
10.5 |
10.2 |
8.1 |
7.3 |
＜0.001 |
| Body mass index, kg/m2
|
22.9 |
22.1 |
22.6 |
23.2 |
23.6 |
23.9 |
＜0.001 |
| Triglycerides, mg/dL |
120.7 |
115.4 |
114.4 |
118.0 |
126.8 |
139.6 |
＜0.001 |
| HDL cholesterol, mg/dL |
58.5 |
61.9 |
59.4 |
57.7 |
56.3 |
55.7 |
＜0.001 |
| Hypertension, (Yes %) |
23.1 |
18.9 |
18.6 |
22.3 |
25.5 |
27.1 |
＜0.001 |
| Diabetes, (Yes %) |
5.3 |
5.5 |
4.4 |
5.1 |
5.3 |
7.1 |
0.027 |
| Hypertension medication, (Yes %) |
6.8 |
5.7 |
6.5 |
7.2 |
7.8 |
7.0 |
0.119 |
| Dyslipidemia medication, (Yes %) |
3.0 |
2.1 |
2.5 |
2.6 |
2.7 |
6.4 |
＜0.001 |
| Baseline survey year, (2002 %) |
52.8 |
60.3 |
56.0 |
50.7 |
49.0 |
42.3 |
＜0.001 |
Abbreviations: BMI, Body Mass Index; LDL-C, Low Density Lipoprotein Cholesterol; HDL, High Density Lipoprotein. Data are presented as the mean for continues variables and the percentage (%) for categorical variables.
P-values are from analysis of variance for continuous variables and chi–squared test for categorical variables.
＊Physical activity: number of days per one month engaging in ≥ 60 minutes vigorous or moderate exercise.
＊＊High risk: Alcohol intake ≥ 40 g/day for men and ≥ 20 g/day for women.
The means of all covariates according to LDL-C categories and SMD before and after weighting (Model 2) are shown in Table 2, which showed that age, sex, the proportion of high alcohol intake, BMI, triglycerides, HDL cholesterol, the proportions of hypertension became balanced after weighting with SMD ＜ 0.1.
Table 2.
Pre and post-inverse probability weighting means for covariates and standardized mean differences for LDL-C Groups, Aichi Workers’ Cohort, 2002–2007 (N = 8966)
|
Unweighted |
Weighted |
|
LDL-C mg/dL groups
N
Variables
|
＜100 |
100–119 |
120–139 |
140–159 |
160– |
SMD |
＜100 |
100–119 |
120–139 |
140–159 |
160– |
SMD |
| 1946 |
2213 |
2214 |
1491 |
1132 |
|
9261 |
9055 |
8978 |
8893 |
8741 |
|
| Age, years |
43.9 |
45.1 |
45.9 |
47.0 |
47.2 |
0.233 |
45.9 |
45.7 |
45.6 |
45.8 |
46.0 |
0.027 |
| Men (%) |
72.9 |
76.7 |
81.8 |
83.0 |
81.5 |
0.123 |
79.6 |
79.1 |
78.9 |
79.5 |
81.9 |
0.033 |
| Physical activity, days/month＊
|
8.9 |
8.8 |
8.6 |
8.8 |
7.8 |
0.050 |
8.7 |
8.7 |
8.7 |
8.7 |
8.9 |
0.008 |
| Current smoker, (%) |
32.2 |
30.0 |
3.0 |
31.9 |
35.1 |
0.031 |
31.6 |
31.5 |
31.5 |
32.5 |
30.2 |
0.013 |
| Alcohol intake, (High risk %)＊＊
|
16.6 |
10.5 |
9.2 |
8.1 |
7.3 |
0.139 |
10.4 |
10.7 |
10.5 |
10.7 |
10.0 |
0.008 |
| Body mass index, kg/m2
|
22.1 |
22.6 |
23.2 |
23.6 |
23.9 |
0.331 |
23.2 |
23.0 |
22.9 |
23.1 |
23.2 |
0.051 |
| Triglycerides, mg/dL |
115.4 |
114.4 |
118.0 |
126.8 |
139.6 |
0.157 |
122.2 |
124.4 |
119.6 |
124.3 |
128.4 |
0.048 |
| HDL cholesterol, mg/dL |
61.9 |
59.4 |
57.7 |
56.3 |
55.7 |
0.201 |
57.8 |
58.4 |
58.6 |
58.3 |
57.6 |
0.035 |
| Hypertension, (Yes %) |
18.9 |
18.6 |
22.3 |
25.5 |
27.1 |
0.113 |
22.6 |
22.4 |
21.8 |
22.6 |
24.1 |
0.026 |
| Diabetes, (Yes %) |
5.5 |
4.4 |
5.1 |
5.3 |
7.1 |
0.050 |
5.1 |
5.6 |
5.2 |
5.7 |
5.3 |
0.012 |
| Hypertension medication, (Yes %) |
5.7 |
6.5 |
7.2 |
7.8 |
7.0 |
0.040 |
6.8 |
7.1 |
6.8 |
7.1 |
8.0 |
0.022 |
| Dyslipidemia medication, (Yes %) |
2.1 |
2.5 |
2.6 |
2.7 |
6.4 |
0.090 |
2.9 |
3.3 |
2.9 |
3.1 |
3.2 |
0.010 |
| Baseline survey year, (2002 %) |
60.3 |
56.0 |
50.7 |
49.0 |
42.3 |
0.121 |
50.0 |
52.5 |
52.6 |
51.8 |
50.5 |
0.016 |
Abbreviations: LDL-C, Low Density Lipoprotein Cholesterol; BMI, Body Mass Index; HDL, High Density Lipoprotein; SMD, Standardized Mean Differences. ＊Physical activity: number of days per one month engaging in ≥ 60 minutes vigorous or moderate exercise.
＊＊High risk: Alcohol intake ≥ 40 g/day for men and ≥ 20 g/day for women.
During the follow-up (median: 12 years), there were 204 CVDs including 82 CHDs and 122 strokes that consisted of 57 ischemic strokes, 25 intracerebral hemorrhage, 13 subarachnoid hemorrhage and 27 strokes of unknown subtypes, of which 112 (55.0%) CVDs, 41 (50.0%) CHDs, 71 (58.2%) total strokes, 41 (71.9%) ischemic strokes, and 20 (80.0%) intracerebral hemorrhages were classified as definite. LDL-C was positively associated with risk of CVD (Model 1 HR of 160– mg/dL compared to 100–119 mg/dL: 1.51; 95% CI: 0.96–2.37) (Table 3) (Fig.1A). Further adjustments for dyslipidemia medication did not materially change the association (Model 2 HR: 1.53; 95% CI: 0.97–2.41). LDL-C was positively and significantly associated with CHD, and the HRs compared to LDL-C 100–119 mg/dL were statistically significant in those with LDL-C levels of 120–139 mg/dL (HR: 2.40; 95% CI: 1.00–5.77), 140–159 mg/dL (HR: 3.05; 95% CI: 1.25–7.41) and 160– mg/dL (HR: 4.56; 95% CI: 1.91–10.9). Spline analyses also indicate that the risk of CHD gradually increased from the LDL-C level of 120 mg/dL (Fig.1B). LDL-C was not associated with total and ischemic strokes (Fig.1C). However, LDL-C was significantly and inversely associated with intracerebral hemorrhage (Model 2 P for trend =0.009) (Fig.1D). Similar results were observed even after excluding participants with ＜ 3 years of follow-up (Supplementary Table 1). Also, the analyses that excluded probable cases yielded essentially the same findings (Supplementary Table 2).
Table 3.
Hazard ratios and the 95% confidence intervals for the association between LDL-C and risk of CVD, Aichi Workers’ Cohort, 2002–2018 (N = 8966)
|
|
|
Model 1 |
Model 2 |
| Outcome |
LDL-C mg/dL |
n/N＊
|
HR (95% CI) |
P-trend
|
HR (95% CI) |
P-trend
|
| Cardiovascular disease |
＜100 |
31/1946 |
0.95 (0.57–1.58) |
0.054 |
0.94 (0.56–1.56) |
0.044 |
|
100–119 |
39/2213 |
Ref |
|
Ref |
|
|
120–139 |
52/2214 |
1.10 (0.72–1.68) |
|
1.09 (0.71–1.66) |
|
|
140–159 |
38/1491 |
1.15 (0.72–1.83) |
|
1.14 (0.72–1.81) |
|
|
160– |
44/1132 |
1.51 (0.96–2.37) |
|
1.53 (0.97–2.41) |
|
| Coronary heart diseases |
＜100 |
7/1946 |
1.16 (0.39–3.47) |
＜0.001 |
1.13 (0.37–3.39) |
＜0.001 |
|
100–119 |
7/2213 |
Ref |
|
Ref |
|
|
120–139 |
22/2214 |
2.46 (1.03–5.85) |
|
2.40 (1.00–5.77) |
|
|
140–159 |
20/1491 |
3.08 (1.27–7.42) |
|
3.05 (1.25–7.41) |
|
|
160– |
26/1132 |
4.43 (1.88–10.5) |
|
4.56 (1.91–10.9) |
|
| Total stroke |
＜100 |
24/1946 |
0.91 (0.51–1.61) |
0.502 |
0.89 (0.50–1.59) |
0.503 |
|
100–119 |
32/2213 |
Ref |
|
Ref |
|
|
120–139 |
30/2214 |
0.78 (0.47–1.30) |
|
0.78 (0.47–1.29) |
|
|
140–159 |
18/1491 |
0.70 (0.38–1.29) |
|
0.69 (0.37–1.28) |
|
|
160– |
18/1132 |
0.82 (0.45–1.52) |
|
0.82 (0.44–1.52) |
|
| Ischemic stroke |
＜100 |
9/1946 |
0.71 (0.29–1.72) |
0.581 |
0.69 (0.28–1.67) |
0.608 |
|
100–119 |
12/2213 |
Ref |
|
Ref |
|
|
120–139 |
14/2214 |
0.86 (0.44–1.69) |
|
0.85 (0.43–1.67) |
|
|
140–159 |
11/1491 |
0.83 (0.39–1.74) |
|
0.81 (0.38–1.71) |
|
|
160– |
11/1132 |
1.03 (0.47–2.24) |
|
0.99 (0.44–2.22) |
|
| Intracerebral hemorrhage |
＜100 |
9/1946 |
1.33 (0.48–3.67) |
0.008 |
1.33 (0.48–3.65) |
0.009 |
|
100–119 |
8/2213 |
Ref |
|
Ref |
|
|
120–139 |
4/2214 |
0.39 (0.11–1.32) |
|
0.39 (0.11–1.31) |
|
|
140–159 |
3/1491 |
0.46 (0.11–1.84) |
|
0.46 (0.11–1.83) |
|
|
160– |
1/1132 |
0.16 (0.02–1.35) |
|
0.17 (0.02–1.37) |
|
Abbreviations: LDL-C, Low Density Lipoprotein Cholesterol; CVD, Cardiovascular Diseases; Ref, Reference category.
＊N = number of subjects; n = number of events.
Model 1: Propensity score adjusted for age, sex, smoking, alcohol drinking, physical activity, body mass index, high- density lipoprotein cholesterol, triglycerides, history of diabetes and hypertension, antihypertensive medication and survey year
Model 2: Propensity score adjusted for variables in model 1＋ dyslipidemia medication.
Supplementary Table 1.
Hazard ratios and the 95% confidence intervals for the association between LDL-C and risk of CVD for participants with follow-up ＜ 3 years (N = 8199), Aichi Workers’ Cohort, 2002–2018
|
|
|
Model 1 |
Model 2 |
| Outcome |
LDL-C mg/dL |
n/N＊
|
HR (95% CI) |
P-trend
|
HR (95% CI) |
P-trend
|
| Cardiovascular disease |
＜100 |
27/1805 |
0.91 (0.53–1.57) |
0.051 |
0.90 (0.52–1.55) |
0.048 |
|
100–119 |
33/2007 |
Ref |
|
Ref |
|
|
120–139 |
41/2020 |
1.01 (0.63–1.62) |
|
1.00 (0.62–1.60) |
|
|
140–159 |
30/1334 |
1.08 (0.65–1.81) |
|
1.07 (0.63–1.79) |
|
|
160– |
39/1033 |
1.54 (0.95–2.50) |
|
1.55 (0.95–2.53) |
|
| Coronary heart diseases |
＜100 |
7/1805 |
1.35 (0.43–4.23) |
＜0.001 |
1.32 (0.40–4.18) |
＜0.001 |
|
100–119 |
6/2007 |
Ref |
|
Ref |
|
|
120–139 |
15/2020 |
1.94 (0.73–5.13) |
|
1.88 (0.70–5.01) |
|
|
140–159 |
17/1334 |
3.05 (1.17–7.95) |
|
3.02 (1.15–7.96) |
|
|
160– |
22/1033 |
4.36 (1.70–11.1) |
|
4.58 (1.77–11.8) |
|
| Total stroke |
＜100 |
20/1805 |
0.81 (0.44–1.50) |
0.704 |
0.80 (0.43–1.48) |
0.651 |
|
100–119 |
27/2007 |
Ref |
|
Ref |
|
|
120–139 |
26/2020 |
0.79 (0.45–1.37) |
|
0.79 (0.45–1.37) |
|
|
140–159 |
13/1334 |
0.62 (0.30–1.25) |
|
0.60 (0.30–1.23) |
|
|
160– |
17/1033 |
0.86 (0.45–1.62) |
|
0.84 (0.44–1.59) |
|
| Ischemic stroke |
＜100 |
7/1805 |
0.57 (0.21–1.51) |
0.396 |
0.55 (0.20–1.46) |
0.460 |
|
100–119 |
11/2007 |
Ref |
|
Ref |
|
|
120–139 |
11/2020 |
0.89 (0.41–1.91) |
|
0.89 (0.41–1.91) |
|
|
140–159 |
7/1334 |
0.72 (0.30–1.73) |
|
0.69 (0.29–1.67) |
|
|
160– |
10/1033 |
1.06 (0.46–2.47) |
|
1.00 (0.42–2.35) |
|
| Intracerebral hemorrhage |
＜100 |
7/1805 |
1.06 (0.35–3.19) |
0.074 |
1.07 (0.36–3.21) |
0.068 |
|
100–119 |
7/2007 |
Ref |
|
Ref |
|
|
120–139 |
3/2020 |
0.51 (0.11–2.23) |
|
0.50 (0.11–2.21) |
|
|
140–159 |
2/1334 |
0.52 (0.11–2.33) |
|
0.52 (0.11–2.33) |
|
|
160– |
1/1033 |
0.20 (0.02–1.58) |
|
0.19 (0.02–1.46) |
|
Abbreviations: LDL-C, Low Density Lipoprotein Cholesterol; CVD, Cardiovascular Diseases; Ref, Reference category.
＊N = number of subjects; n = number of events.
Model 1: Propensity score adjusted for age, sex, smoking, alcohol drinking, physical activity, body mass index, high- density lipoprotein cholesterol, triglycerides, history of diabetes and hypertension, antihypertensive medication and survey year
Model 2: Propensity score adjusted for variables in model 1＋ dyslipidemia medication
Supplementary Table 2.
Hazard ratios and the 95% confidence intervals for the association between LDL-C and risk of definite CVD (N = 8966), Aichi Workers’ Cohort, 2002–2018
|
|
|
Model 1 |
Model 2 |
| Outcome |
LDL-C mg/dL |
n/N＊
|
HR (95% CI) |
P-trend
|
HR (95% CI) |
P-trend
|
| Cardiovascular disease |
＜100 |
19/1946 |
1.23 (0.64–2.36) |
0.432 |
1.21 (0.63–2.34) |
0.412 |
|
100–119 |
20/2213 |
Ref |
|
Ref |
|
|
120–139 |
27/2214 |
1.15 (0.64–2.06) |
|
1.15 (0.64–2.07) |
|
|
140–159 |
26/1491 |
1.52 (0.84–2.75) |
|
1.53 (0.85–2.77) |
|
|
160– |
20/1132 |
1.36 (0.72–2.58) |
|
1.38 (0.73–2.62) |
|
| Coronary heart diseases |
＜100 |
4/1946 |
1.84 (0.40–8.40) |
0.009 |
1.83 (0.40–8.34) |
0.007 |
|
100–119 |
3/2213 |
Ref |
|
Ref |
|
|
120–139 |
10/2214 |
2.75 (0.75–10.0) |
|
2.77 (0.76–10.1) |
|
|
140–159 |
12/1491 |
4.57 (1.28–16.3) |
|
4.61 (1.29–16.4) |
|
|
160– |
12/1132 |
5.16 (1.44–18.5) |
|
5.39 (1.50–19.3) |
|
| Total stroke |
＜100 |
15/1946 |
1.11 (0.53–2.30) |
0.265 |
1.10 (0.53–2.28) |
0.255 |
|
100–119 |
17/2213 |
Ref |
|
Ref |
|
|
120–139 |
17/2214 |
0.84 (0.43–1.67) |
|
0.85 (0.43–1.68) |
|
|
140–159 |
14/1491 |
0.95 (0.46–1.96) |
|
0.95 (0.46–1.97) |
|
|
160– |
8/1132 |
0.65 (0.27–1.58) |
|
0.63 (0.26–1.55) |
|
| Ischemic stroke |
＜100 |
6/1946 |
1.00 (0.34–2.96) |
0.869 |
0.99 (0.33–2.92) |
0.922 |
|
100–119 |
8/2213 |
Ref |
|
Ref |
|
|
120–139 |
11/2214 |
1.17 (0.47–2.93) |
|
1.19 (0.47–2.97) |
|
|
140–159 |
11/1491 |
1.58 (0.63–3.97) |
|
1.59 (0.63–4.01) |
|
|
160– |
5/1132 |
0.85 (0.27–2.72) |
|
0.80 (0.24–2.61) |
|
| Intracerebral hemorrhage |
＜100 |
6/1946 |
1.20 (0.35–4.06) |
0.056 |
1.19 (0.35–4.02) |
0.058 |
|
100–119 |
6/2213 |
Ref |
|
Ref |
|
|
120–139 |
4/2214 |
0.50 (0.14–1.82) |
|
0.50 (0.14–1.81) |
|
|
140–159 |
3/1491 |
0.60 (0.14–2.51) |
|
0.59 (0.14–2.50) |
|
|
160– |
1/1132 |
0.21 (0.02–1.80) |
|
0.21 (0.02–1.83) |
|
Abbreviations: LDL-C, Low Density Lipoprotein Cholesterol; CVD, Cardiovascular Diseases; Ref, Reference category.
＊N = number of subjects; n = number of events.
Model 1: Propensity score adjusted for age, sex, smoking, alcohol drinking, physical activity, body mass index, high- density lipoprotein cholesterol, triglycerides, history of diabetes and hypertension, antihypertensive medication and survey year
Model 2: Propensity score adjusted for variables in model 1＋ dyslipidemia medication
Discussion
The present prospective cohort study assessed the associations of LDL-C categories with the risk of CHD, stroke subtypes and CVD among a middle-aged Japanese working population with no history of CVD at baseline. We observed that participants with LDL-C levels of 120–139, 140–159 and 160– mg/dL were 2.4, 3.1 and 4.6 times more likely to develop CHD than those with LDL-C of 100–119 mg/dL. Although not statistically significant, LDL-C ＜ 100 mg/dL tended to be positively associated with risk of intracerebral hemorrhage compared to subjects with LDL-C of 100–119 mg/dL. These results did not change even after excluding individuals with less than three years of follow-up. To the best of our knowledge, this is the first study to use IPW to investigate the association between LDL-C level and risk of CHD and stroke among the Japanese population.
Our findings on CHD are supported by previous studies on the Japanese population. A study (LDL-C mean = 115 mg/dL) among young adults aged ＜ 50 years observed higher risk of incident myocardial infraction (HR: 1.83; 95% CI: 1.65–2.02) in individuals in the highest category (LDL-C ≥ 140 mg/dL) compared with LDL-C ＜ 140 mg/dL18). The CIRCS study (LDL-C mean = 105.5 mg/dL) found LDL-C ≥ 140 mg/dL was associated with increased risk of total CHD (HR: 2.80; 95% CI: 1.59–4.92) and risk of myocardial infarction (HR: 3.83; 95% CI: 1.78–8.23) compared to LDL-C ＜ 80 mg/dL among adults aged 40 to 69 years12). Furthermore, the Iwate KENKO study (LDL-C mean = 118.1 mg/dL) showed that the highest quartile (LDL-C ≥ 140 mg/dL) was associated with higher risk of acute myocardial infarction (HR: 2.50; 95% CI: 1.02–6.09) compared with the reference quartile (LDL-C ＜ 100 mg/dL) among adults aged ≥ 18 years16). However, another Japanese study with baseline years of 1995–2000 conducted among individuals without clinical treatment for cardiovascular risk factors did not identify an excessive risk of CHD incidence or mortality from CHD in the highest levels of LDL-C (≥ 140 mg/dL) compared with ＜ 80 mg/dL31).
By relying on more recent data on the middle-aged urban working population, we confirmed that LDL-C level linearly and independently accounted for the risk of CHD, which was compatible with observations from the US and Chinese populations14, 15, 32-34). The strong association found in the present study may be due to the relatively higher incidence rate of CHD, which is almost comparable with that of total stroke. This was supported by previous observations in Japan that the incidence of CHD among urban male employees35, 36) was higher than that in the rural population37, 38). Previous studies have also reported a higher risk of CHD among Japanese Americans compared with their counterparts in Japan, thus indicating the role of high blood LDL-C levels39). The latest statistics from a national survey in Japan40) revealed that the prevalence of dyslipidemia (LDL-C 160– mg/dL) deviated from the targeted goal for the prevention of onset and progression of cardiovascular disease in Health Japan 21 (the second term)41), and the findings of the present study and of the national survey imply possible increase in the future burden of CHD among the Japanese population.
Our study showed no association between LDL-C 160– mg/dL and the risk of total and ischemic stroke compared with LDL-C 100–119 mg/dL. This finding is partially consistent with previous studies from Japan and Western countries10, 11, 42, 43). The Iwate KENKO study reported a non-significant negative association between the highest quartile (LDL-C ≥ 140 mg/dL) and risk of ischemic stroke (HR: 0.73; 95% CI: 0.42–1.27) compared with the reference quartile (LDL-C ＜ 100 mg/dL)16). The Suita study (LDL-C mean = 129.5 mg/dL) reported no association between higher quartile (men: LDL-C ≥ 179.4 mg/dL, women: LDL-C ≥ 188.7 mg/dL) and the risk of ischemic stroke (HR: 0.99; 95% CI: 0.48–2.03) compared with the reference quartile (men: LDL-C ＜ 123.0 mg/dL, women: LDL-C ＜ 143.1 mg/dL)10). The Hisayama study also reported a non-significant association between the higher quartile (LDL-C ≥ 150.4 mg/dL) with ischemic stroke (HR: 1.35; 95% CI: 0.85–2.14) compared with the reference quartile (LDL-C ＜ 102.5 mg/dL)11). The AMORIS study in Sweden (LDL-C mean = 150 mg/dL), reported a positive association between each change in standard deviation of LDL-C and risk of ischemic stroke (men, HR: 1.12; 95% CI: 1.08–1.17; women, HR: 1.10; 95% CI: 1.05–1.16)42). The Cardiovascular Health Study in the United States (LDL-C mean = 135.7 mg/dL) reported a positive and significant association between each change in standard deviation of LDL-C and ischemic stroke (HR: 1.12; 95% CI: 1.01–1.25)43).
We observed a non-significant positive association between LDL-C ＜ 100 mg/dL with the risk of intracerebral hemorrhage compared with LDL-C 100–119 mg/dL. Furthermore, the trend test showed significant inverse association between LDL-C and risk of intracerebral hemorrhage. These findings are partially in agreement with the previous Japanese studies. The Ibaraki Prefectural Health study (LDL-C mean = 117.6 mg/dL) reported significant negative association between LDL-C ≥ 140 mg/dL and risk of death due to intracerebral hemorrhage (HR: 0.45; 95% CI: 0.30–0.69) compared with reference group (LDL-C ＜ 80 mg/dL)44). The Kailuan study in China reported significantly increased risk of intracerebral hemorrhage in subjects with LDL-C 50–69 mg/dL (HR: 1.65; 95% CI: 1.32–2.05) compared with LDL-C 70–99 mg/dL45).
The strength of this study is that the study population represented the recent situation of LDL-C in Japan and manifested a higher incidence of CHD than previous population-based studies in Japan, thus enabling us to capture the effects of relatively higher levels of LDL-C on CHD risk. In the present study, we used IPW technique to adjust for baseline confounders since traditional multivariable regression methods in a study with relatively small number of events may yield biased estimates because of overfitting46). In the application of this method, we truncated large weights at 1st and 99th percentiles47) in order to avoid putting large weights in unexposed individuals with very low propensity score. This study has some limitations. First, we used a single measurement at baseline; therefore, the relationships between LDL-C, CHD, and stroke might be underestimated owing to random error resulting from the measurement of LDL-C48). Second, there was a small number of incident cases of CHD and subtypes of stroke, thus leading to wide CIs for the HRs of the associations of LDL-C. Third, our study used self-reported of lipid lowering medication. Even though we showed accuracy of the positive self-report in a different study23), possible under-reporting might have weakened the real associations in the present study. bias might be introduced due to error in the year of dyslipidemia diagnosis
Conclusion
This long-term prospective study among middle-aged Japanese workers that used IPW to adjust for a number of confounding variables showed that LDL-C was significantly and positively associated with the risk of CHD and inversely with intracerebral hemorrhage.
Acknowledgements
The authors would like to thank the participants and healthcare personnel of the local government office. We also thank the technical assistance of Mr. Hidetoshi Ichimura.
Conflict of Interest
All authors declare no competing interests.
Grants and/or Financial Support
This work was supported in part by MEXT/JSPS KAKENHI (grant Nos. 13470087 and 17390185 to Professor Emeritus Hideaki Toyoshima; grant Nos. 17790384, 22390133, 23659346, 26293153, 18H03057 and 22H03349 to HY; grant Nos. 16590499, 18590594, 20590641, 23590787, and 15K08802 to KT; and grant Nos. 25893088 and 16K19278 to YL); Health and Labor Sciences research grants for Comprehensive Research on Cardiovascular and Life-Style Related Diseases (H26-Junkankitou [Seisaku]-Ippan-001, H29-Junkankitou [Seishuu]-Ippan-003 and 20FA1002) from the Ministry of Health, Labor and Welfare; and research grants from the Japan Atherosclerosis Prevention Fund (to HY), the Aichi Health Promotion Foundation (to HY), and the Uehara Memorial Fund (to HY) and the Noguchi Memorial Research Institute (to HY).
References
- 1) Labarthe DR: Adverse blood lipid profile. In: Epidemiology and prevention of cardiocascular diseases: A Global Challenge 2nd Ed, pp 269-304, Jones and Bartlett Publishers, Inc, Massachusetts, USA, 2011
- 2) Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Bruckert E, Hegele RA, Krauss RM, Raal FJ, Schunkert H, Watts GF, Borén J, Fazio S, Horton JD, Masana L, Nicholls SJ, Nordestgaard BG, van de Sluis B, Taskinen MR, Tokgözoglu L, Landmesser U, Laufs U, Wiklund O, Stock JK, Chapman MJ, Catapano AL. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. Eur Heart J, 2017; 38: 2459-2472
- 3) Kinoshita M, Yokote K, Arai H, Iida M, Ishigaki Y, Ishibashi S, Umemoto S, Egusa G, Ohmura H, Okamura T, Kihara S, Koba S, Saito I, Shoji T, Daida H, Tsukamoto K, Deguchi J, Dohi S, Dobashi K, Hamaguchi H, Hara M, Hiro T, Biro S, Fujioka Y, Maruyama C, Miyamoto Y, Murakami Y, Yokode M, Yoshida H, Rakugi H, Wakatsuki A, Yamashita S: Comprehensive risk assessment. In: Committee for Epidemiology and Clinical Management of Atherosclerosis. Japan Atherosclerosis Society (JAS) Guidelines for Prevention of Atherosclerotic Cardiovascular Diseases 2017. J Atheroscler Thromb, 2018; 25: 859-860
- 4) National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Third Report of the NCEP Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation, 2002; 106: 3167-3172. https://www.ahajournals.org/doi/10.1161/circ.106.25.3143 (Access date: May 10, 2022)
- 5) François M, Colin B, Alberico LC, Konstantinos CK, Manuela C, Lina B, M JC, Guy GDB, Victoria D, Brian AF, Ian MG, Alison H, Ulf L, Borislava M, Terje RP, Gabriele R, Dimitrios JR, Marc SS, Marja-Riitta T, Lale T, Olov W, ESC Scientific Document Group. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk: The Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and European Atherosclerosis Society (EAS). Eur Heart J, 2020; 41: 111-188
- 6) Honda T, Chen S, Hata J, Yoshida D, Hirakawa Y, Furuta Y, Shibata M, Sakata S, Kitazono T, Ninomiya T. Development and Validation of a Risk Prediction Model for Atherosclerotic Cardiovascular Disease in Japanese Adults: The Hisayama Study. J Atheroscler Thromb, 2022; 29: 345-361
- 7) Nishimura K, Okamura T, Watanabe M, Nakai M, Takegami M, Higashiyama A, Kokubo Y, Okayama A, Miyamoto Y. Predicting coronary heart disease using risk factor categories for a Japanese urban population, and comparison with the framingham risk score: the suita study. J Atheroscler Thromb, 2014; 21: 784-798
- 8) Liu J, Sempos CT, Donahue RP, Dorn J, Trevisan M, Grundy SM. Non-high-density lipoprotein and very-low-density lipoprotein cholesterol and their risk predictive values in coronary heart disease. Am J Cardiol, 2006; 98: 1363-1368
- 9) Wilson PWF, Polonsky TS, Miedema MD, Khera A, Kosinski AS, Kuvin JT. Systematic Review for the 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation, 2019; 139: e1144-e1161
- 10) Okamura T, Kokubo Y, Watanabe M, Higashiyama A, Miyamoto Y, Yoshimasa Y, Okayama A. Low-density lipoprotein cholesterol and non-high-density lipoprotein cholesterol and the incidence of cardiovascular disease in an urban Japanese cohort study: The Suita study. Atherosclerosis, 2009; 203: 587-592
- 11) Imamura T, Doi Y, Arima H, Yonemoto K, Hata J, Kubo M, Tanizaki Y, Ibayashi S, Iida M, Kiyohara Y. LDL cholesterol and the development of stroke subtypes and coronary heart disease in a general Japanese population: the Hisayama study. Stroke, 2009; 40: 382-388
- 12) Imano H, Noda H, Kitamura A, Sato S, Kiyama M, Sankai T, Ohira T, Nakamura M, Yamagishi K, Ikeda A, Shimamoto T, Iso H. Low-density lipoprotein cholesterol and risk of coronary heart disease among Japanese men and women: the Circulatory Risk in Communities Study (CIRCS). Prev Med, 2011; 52: 381-386
- 13) Kitamura A, Nakagawa Y, Sato M, Iso H, Sato S, Imano H, Kiyama M, Okada T, Okada H, Iida M, Shimamoto T. Proportions of stroke subtypes among men and women ＞ or =40 years of age in an urban Japanese city in 1992, 1997, and 2002. Stroke, 2006; 37: 1374-1378
- 14) Pischon T, Girman CJ, Sacks FM, Rifai N, Stampfer MJ, Rimm EB. Non-high-density lipoprotein cholesterol and apolipoprotein B in the prediction of coronary heart disease in men. Circulation, 2005; 112: 3375-3383
- 15) Shai I, Rimm EB, Hankinson SE, Curhan G, Manson JE, Rifai N, Stampfer MJ, Ma J. Multivariate assessment of lipid parameters as predictors of coronary heart disease among postmenopausal women: potential implications for clinical guidelines. Circulation, 2004; 110: 2824-2830
- 16) Yokokawa H, Yasumura S, Tanno K, Ohsawa M, Onoda T, Itai K, Sakata K, Kawamura K, Tanaka F, Yoshida Y, Nakamura M, Terayama Y, Ogawa A, Okayama A. Serum low-density lipoprotein to high-density lipoprotein ratio as a predictor of future acute myocardial infarction among men in a 2.7-year cohort study of a Japanese northern rural population. J Atheroscler Thromb, 2011; 18: 89-98
- 17) Noda H, Iso H, Irie F, Sairenchi T, Ohtaka E, Ohta H. Gender difference of association between LDL cholesterol concentrations and mortality from coronary heart disease amongst Japanese: the Ibaraki Prefectural Health Study. J Intern Med, 2010; 267: 576-587
- 18) Kaneko H, Itoh H, Kiriyama H, Kamon T, Fujiu K, Morita K, Michihata N, Jo T, Takeda N, Morita H, Yasunaga H, Komuro I. Lipid Profile and Subsequent Cardiovascular Disease among Young Adults Aged ＜50 Years. Am J Cardiol, 2021; 142: 59-65
- 19) Ministry of Health, Labour and Welfare of Japan. The National Health and Nutrition Survey. Trends in the serum total and LDL cholesterol in Japan, 1989-2019. https: //www.nibiohn.go.jp/eiken/kenkounippon21/eiyouchousa/keinen_henka_shintai.html. (Access date: April 15, 2022)
- 20) Sekikawa A, Miyamoto Y, Miura K, Nishimura K, Willcox BJ, Masaki KH, Rodriguez B, Tracy RP, Okamura T, Kuller LH. Continuous decline in mortality from coronary heart disease in Japan despite a continuous and marked rise in total cholesterol: Japanese experience after the Seven Countries Study. Int J Epidemiol, 2015; 44: 1614-1624
- 21) Nishi N, Sugiyama H, Kasagi F, Kodama K, Hayakawa T, Ueda K, Okayama A, Ueshima H. Urban-rural difference in stroke mortality from a 19-year cohort study of the Japanese general population: NIPPON DATA80. Soc Sci Med, 2007; 65: 822-832
- 22) Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behav Res, 2011; 46: 399-424
- 23) Wada K, Yatsuya H, Ouyang P, Otsuka R, Mitsuhashi H, Takefuji S, Matsushita K, Sugiura K, Hotta Y, Toyoshima H, Tamakoshi K. Self-reported medical history was generally accurate among Japanese workplace population. J Clin Epidemiol, 2009; 62: 306-313
- 24) The World Health Organization MONICA Project (monitoring trends and determinants in cardiovascular disease): a major international collaboration. WHO MONICA Project Principal Investigators. J Clin Epidemiol, 1988; 41: 105-114
- 25) Pishgar F, Greifer N, Leyrat C, Stuart E. Package MatchThem: Matching and weighting multiply imputed datasets. R Top Doc, 2020: 19-21. https: //cran.r-project.org/web/packages/MatchThem/MatchThem.pdf
- 26) Greifer N. cobalt: Covariate Balance Tables and Plots. R Top Doc, 2022: 6-11. https: //ngreifer.github.io/cobalt/ (Access date: April 15, 2022)
- 27) Therneau T, Lumley T, Elizabeth A, Cynthia C. R Package ‘survival’. R Top Doc, 2022: 129-131. https: //cran.r-project.org/web/packages/survival/survival.pdf. (Access date: April 15, 2022)
- 28) Lumley T. R package ‘survey’. R Top Doc, 2021: 82-84. https: //cran.r-project.org/web/packages/survey/survey.pdf (Access date: April 15, 2022)
- 29) Harrell F.E: General aspects of fitting regression models. In: Regression modeling strategies 2nd Ed, pp 24-28, Springer International Publishing, Switzerland, 2015
- 30) Harrell F.E: The rms Package. In: Regression modeling strategies 2nd Ed, pp 130-140, Springer International Publishing, Switzerland, 2015
- 31) Matsumura T, Sankai T, Yamagishi K, Kubota Y, Hayama-Terada M, Muraki I, Umesawa M, Cui R, Imano H, Ohira T, Kitamura A, Okada T, Kiyama M, Iso H. Impact of Major Cardiovascular Risk Factors on the Incidence of Cardiovascular Disease among Overweight and Non-Overweight Individuals: The Circulatory Risk in Communities Study (CIRCS). J Atheroscler Thromb, 2022; 29: 422-437
- 32) Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation, 1998; 97: 1837-1847
- 33) Mora S, Otvos JD, Rifai N, Rosenson RS, Buring JE, Ridker PM. Lipoprotein particle profiles by nuclear magnetic resonance compared with standard lipids and apolipoproteins in predicting incident cardiovascular disease in women. Circulation, 2009; 119: 931-939
- 34) Ren J, Grundy SM, Liu J, Wang W, Wang M, Sun J, Liu J, Li Y, Wu Z, Zhao D. Long-term coronary heart disease risk associated with very-low-density lipoprotein cholesterol in Chinese: the results of a 15-Year Chinese Multi-Provincial Cohort Study (CMCS). Atherosclerosis, 2010; 211: 327-332
- 35) Kitamura A, Sato S, Kiyama M, Imano H, Iso H, Okada T, Ohira T, Tanigawa T, Yamagishi K, Nakamura M, Konishi M, Shimamoto T, Iida M, Komachi Y. Trends in the incidence of coronary heart disease and stroke and their risk factors in Japan, 1964 to 2003: the Akita-Osaka study. J Am Coll Cardiol, 2008; 52: 71-79
- 36) Iso H. Changes in coronary heart disease risk among Japanese. Circulation, 2008; 118: 2725-2729
- 37) Rumana N, Kita Y, Turin TC, Murakami Y, Sugihara H, Morita Y, Tomioka N, Okayama A, Nakamura Y, Abbott RD, Ueshima H. Trend of increase in the incidence of acute myocardial infarction in a Japanese population: Takashima AMI Registry, 1990-2001. Am J Epidemiol, 2008; 167: 1358-1364
- 38) Cui R, Iso H, Yamagishi K, Saito I, Kokubo Y, Inoue M, Sawada N, Tsugane S; JPHC Study Group. Trends in the proportions of stroke subtypes and coronary heart disease in the Japanese men and women from 1995 to 2009. Atherosclerosis, 2016; 248: 219-223
- 39) Ueshima H, Okayama A, Saitoh S, Nakagawa H, Rodriguez B, Sakata K, Okuda N, Choudhury SR, Curb JD; INTERLIPID Research Group. Differences in cardiovascular disease risk factors between Japanese in Japan and Japanese-Americans in Hawaii: the INTERLIPID study. J Hum Hypertens, 2003; 17: 631-639
- 40) Ministry of Health, Labour and Welfare of Japan. The National Health and Nutrition Survey. https: //www.e-stat.go.jp/stat-search/files?page=1&layout=datalist&toukei=00450171&tstat=000001041744&cycle=7&tclass1=000001148507&tclass2val=0. (Access date: May 19, 2022)
- 41) Ministry of Health, Labour and Welfare of Japan. Health Japan 21 (the second term). https: //www.nibiohn.go.jp/eiken/kenkounippon21/en/kenkounippon21/genjouchi.html (Accessdate: April 15, 2022)
- 42) Holme I, Aastveit AH, Hammar N, Jungner I, Walldius G. Relationships between lipoprotein components and risk of ischaemic and haemorrhagic stroke in the Apolipoprotein MOrtality RISk study (AMORIS). J Intern Med, 2009; 265: 275-287
- 43) Psaty BM, Anderson M, Kronmal RA, Tracy RP, Orchard T, Fried LP, Lumley T, Robbins J, Burke G, Newman AB, Furberg CD. The association between lipid levels and the risks of incident myocardial infarction, stroke, and total mortality: The Cardiovascular Health Study. J Am Geriatr Soc, 2004; 52: 1639-1647
- 44) Noda H, Iso H, Irie F, Sairenchi T, Ohtaka E, Doi M, Izumi Y, Ohta H. Low-density lipoprotein cholesterol concentrations and death due to intraparenchymal hemorrhage: the Ibaraki Prefectural Health Study. Circulation, 2009; 119: 2136-2145
- 45) Ma C, Gurol ME, Huang Z, Lichtenstein AH, Wang X, Wang Y, Neumann S, Wu S, Gao X. Low-density lipoprotein cholesterol and risk of intracerebral hemorrhage: A prospective study. Neurology, 2019; 93: e445-e457
- 46) Peduzzi P, Concato J, Feinstein AR, Holford TR. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol, 1995; 48: 1503-1510
- 47) Lee BK, Lessler J, Stuart EA. Weight trimming and propensity score weighting. PLoS One, 2011; 6: e18174
- 48) Hutcheon JA, Chiolero A, Hanley JA. Random measurement error and regression dilution bias. BMJ, 2010; 340: c2289
