Abstract
Aims: There is no community-based cohort study to examine the effect of very high level of high-density lipoprotein cholesterol (HDL-C) on coronary heart disease (CHD) and other cause-specific mortality. Therefore, we investigated the relationship between HDL-C including very high level and cause-specific mortality in a 20-year cohort study of the representative sample of Japanese.
Methods: We followed 7,019 individuals from the Japanese general population (2,946 men and 4,073 women). We defined HDL-C levels as follow: low (HDL-C < 1.04 mmol/L), reference (1.04–1.55 mmol/L), high (1.56–2.06 mmol/L), very high (≥ 2.07 mmol/L). The multivariate adjusted hazard ratio (HR) for all-cause or cause-specific mortality was calculated using a Cox proportional hazards model adjusted for other traditional risk factors.
Results: During follow-up, we observed 1,598 deaths. No significant association was observed between HDL-C and all-cause mortality. Serum HDL-C also showed no association with stroke. In contrast, the risk for CHD among high HDL-C was lower than reference, HRs were 0.51 [95% confidence interval (CI): 0.21–1.23] in men, 0.33 (95% CI: 0.11–0.95) in women, and 0.41 (95% CI: 0.21–0.81) when men and women were combined. However, very high HDL-C did not show significant association with CHD and other cause-specific mortality.
Conclusions: HDL-C was not associated with all-cause and stroke mortality. In contrast, high serum HDL-C levels, at least up to 2.06 mmol/L, were protective against CHD, although further high levels were not. However, sample size of cause-specific death in very high HDL-C group was not enough even in this 20-year follow-up of 7,019 Japanese; larger cohort studies should be warranted.
Keywords: High-density lipoprotein cholesterol, Mortality, Risk factors, Cohort studies
See editorial vol. 23: 771–772
Introduction
Serum high-density lipoprotein cholesterol (HDL-C) is known as a protective factor for coronary heart disease (CHD)1–9). In some cohort studies, serum HDL-C is also inversely associated with all-cause mortality10–12). However, it is still unclear whether very high HDL-C levels are beneficial for community dwellers or not. To date, there are no community-based cohort studies concerning the relationship between very high serum levels of HDL-C and CHD or cause-specific mortality. It is very difficult to set a very high HDL-C group in a cohort study13). Although our previous study tried to examine the relationship between very high HDL-C levels and cause-specific mortality14), we did not set a very high HDL-C group because of small number of events.
Accordingly, we investigated the relationship between serum HDL-C categories that included very high HDL-C group and cause-specific and all-cause mortality using data of population samples with long-term follow-up period. This study extended the follow-up period of our previous study14) from 10 to 20 years. Because Japanese population has more individuals with very high HDL-C levels than Western populations15), it is suitable to investigate the effect of very high HDL-C levels on cause-specific mortality.
Methods
Study Participants
National Integrated Project for Prospective Observation of Non-communicable Disease and Its Trends in the Aged (NIPPON DATA) are the cohort studies that are based on the National Survey on Circulatory Disorders, Japan. NIPPON DATA included two long-term cohort studies. The baseline surveys were performed in 1980 and in 1990 (NIPPON DATA80 and NIPPON DATA90), and the details of these cohorts have been reported14, 16–19). In this study, we analyzed data from NIPPON DATA90 because the baseline survey of NIPPON DATA80 did not include the measurement of serum HDL-C.
A total of 8,383 residents (3,504 men and 4,879 women, aged ≥ 30 years) from 300 randomly selected districts participated in the survey and were followed until 15 November 2010. The participation rate in this survey was 76.5%. Of the 8,383 participants, 1,364 were excluded. The reasons were as follows: history of CHD or stroke (n = 248), the use of lipid-lowering agents (n = 267), information missing at the baseline survey (n = 662), and failure to access because of incomplete residential access information at the second survey (n = 304). The remaining 7,019 participants (2,946 men and 4,073 women) were included in the analysis.
Follow-Up Survey
The underlying cause of death in the National Vital Statistics was coded according to the ICD-9 until the end of 1994 and the ICD-10 from the start of 1995 until the end of 2010. Details of these classifications are described elsewhere14, 16–19). This study was approved by the Institutional Review Board of Shiga University of Medical Science (No. 12–18, 2000; No. 17–21-1, 2010).
Baseline Examination
Nonfasting blood samples were obtained from all participants, and blood samples were shipped to the commissioned clinical laboratory (SRL, Tokyo, Japan) for blood measurement after they were preprocessed. Serum total cholesterol and triglyceride (TG) were measured by enzymatic methods. HDL-C was measured by the precipitation method using heparin-calcium. Lipid measurement was standardized by the Centers for Disease Control/National Heart, Lung, and Blood Institute Lipid Standardization Program20). Plasma glucose was measured enzymatically.
Public health nurses obtained information about lifestyle factors such as the current use of medication, the extent of smoking and alcohol consumption, physical activity, and history of cardiovascular disease (CVD). The body mass index (BMI) was calculated as the weight (kg) divided by the height squared (m2). Baseline blood pressure was measured by trained observers using a standard mercury sphygmomanometer on the right arm of seated subjects.
We defined the metabolic syndrome based on the criteria defined by the joint committee of eight Japanese medical societies; visceral obesity (waist ≥ 85 cm in men, ≥ 90 cm in women) plus ≥ 2 items of the following three components: (1) HDL-C < 1.04 mmol/L (40 mg/dl) or TG ≥ 1.70 mmol/L (150 mg/dl) or medication for hyperlipidemia, (2) blood pressure ≥ 130/85 mmHg or taking an antihypertensive, (3) fasting blood glucose ≥ 6.10 mmol/L (110 mg/dl) or medication for diabetes. In this study, we substitute BMI (≥ 25 kg/m2) for waist and nonfasting blood glucose (≥ 7.77 mmol/L, 140 mg/dl) for fasting blood glucose. In addition, we defined diabetes as serum glucose ≥ 11.1 mmol/L, a history of diabetes, or both and hypertension as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, the use of antihypertensive agents, or any combination of these.
Statistical Analysis
We divided all participants into four HDL-C categories according to the definition in a large previous cross-sectional study in Japan21), which was as follows: low [HDL-C < 1.04 mmol/L (< 40 mg/dl)], reference [1.04–1.55 mmol/L (40–59 mg/dl)], high [1.56–2.06 mmol/L (60–79 mg/dl)], very high [≥ 2.07 mmol/L (≥ 80 mg/dl)]. Then, we performed analysis of variance to compare multiple group means and the χ2 test to compare frequencies.
The multivariate adjusted hazard ratio (HR) of each HDL-C category for all-cause or cause-specific mortality, i.e., CVD, CHD, stroke, cerebral infarction, cerebral hemorrhage, was calculated using the Cox proportional hazards model. In model 1, we adjusted for age, in model 2, we adjusted for age, BMI, hypertension, diabetes, smoking category (never-smoked, ex-smoker, current smoker; ≤ 20 and ≥ 21 cigarettes/day), drinking category (never-drinker, ex-drinker, and current drinker), non-HDL cholesterol and TG (log-transformed). Sex was also adjusted in sex-combined analysis. Wald's tests were used to test for significance of HR in each HDL-C categories compared with the reference group22). The model with continuous serum HDL-C values instead of HDL-C categories was also examined to clarify the linear trend between HDL-C and mortality. The proportional hazard assumption was checked by the schoenfeld residuals plot. In addition, we performed analysis using the same cut points as our former study14): very low [HDL-C < 0.91 mmol/L (< 35 mg/dl)], low [0.91–1.03 mmol/L (35–39 mg/dl)], reference [1.04–1.55 mmol/L (40–59 mg/dl)], high [1.56–1.81 mmol/L (60–69 mg/dl)], and very high [≥ 1.82 mmol/L (≥ 70 mg/dl)].
We performed additional analysis stratified by hypertension (with/without), smoking (with/without), drinking (with/without), diabetes (with/without), high serum levels of non-HDL-C (with/without, defined by median; ≥ 3.62, < 3.62), hypertriglyceridemia (with/without, defined by median; ≥ 1.21, < 1.21) examine certainty of the results for CVD and all-cause mortality. Sensitivity analysis for each cause of death was not performed due to small numbers of each cause-specific event.
All CIs were estimated at the 95% level. P-value < 0.05 was considered significant. Statistical analysis was performed with IBM SPSS Statistics Ver. 22.
Results
Total person-years were 126,481 (51,356 in men and 75,122 in women), and the mean follow-up period was 18.0 years. During follow-up, 1,598 participants died, of which 450 died due to CVD. Table 1 shows the baseline characteristics of the study participants according to HDL-C category.
Table 1. Baseline characteristics stratified by HDL cholesterol level at the baseline survey in 1990, NIPPON DATA90.
| Risk charasteristics | Baseline HDL-C range, mmol/L (mg/dl) |
P-values | |||
|---|---|---|---|---|---|
| Low < 1.04 (< 40) |
Reference 1.04–1.55 (40–59) |
High 1.56–2.06 (60–79) |
Very high 2.07+ (80+) |
||
| Men | |||||
| Number of subjects | 702 | 1554 | 554 | 136 | |
| HDL-cholesterol (mmol/L) | 0.87 ± 0.12 | 1.26 ± 0.14 | 1.74 ± 0.14 | 2.32 ± 0.28 | < 0.001 |
| Age (years) | 52.5 ± 13.7 | 52.7 ± 13.4 | 53.5 ± 13.3 | 517 ± 13.0 | 0.048 |
| BMI (kg/m2) | 24.2 ± 2.9 | 22.9 ± 3.0 | 21.7 ± 2.6 | 21.3 ± 3.0 | < 0.001 |
| Total cholesterol (mmol/L) | 512 ± 1.04 | 5.06 ± 0.92 | 5.21 ± 0.88 | 5.43 ± 0.91 | < 0.001 |
| Non-HDL cholesterol (mmol/L) | 4.24 ± 1.04 | 3.81 ± 0.93 | 3.47 ± 0.87 | 3.10 ± 0.93 | < 0.001 |
| Triglycerides (mmol/L)§ | 2.07 | 1.34 | 1.04 | 0.92 | < 0.001 |
| Hypertension (%) | 46.0 | 46.5 | 48.6 | 42.6 | 0.612 |
| Diabetes (%) | 9.4 | 6.2 | 6.7 | 8.8 | 0.040 |
| Metabolic syndrome (%) | 26.9 | 11.5 | 2.5 | 2.9 | < 0.001 |
| Smoking | |||||
| Never-smoker (%) | 20.5 | 20.1 | 22.4 | 22.1 | < 0.001 |
| Ex-smoker (%) | 19.2 | 23.9 | 25.6 | 23.5 | |
| Current smoker (≦ 20 cigarettes/day)(%) | 34.6 | 37.0 | 37.0 | 44.1 | |
| Current smoker (≧ 21 cigarettes/day)(%) | 25.6 | 18.9 | 15.0 | 10.3 | |
| Alcohol drinking | |||||
| Never-drinker (%) | 52.1 | 32.1 | 25.1 | 11.8 | < 0.001 |
| Ex-drinker (%) | 8.4 | 5.7 | 3.8 | 3.7 | |
| Current drinker (%) | 39.5 | 62.2 | 71.1 | 84.6 | |
| Women | |||||
| Number of subjects (proportion, %) | 425 | 2009 | 1341 | 298 | |
| HDL-cholesterol (mmol/L) | 0.89 ± 0.11 | 1.30 ± 0.15 | 1.75 ± 0.14 | 2.29 ± 0.22 | < 0.001 |
| Age (years) | 57.7 ± 13.9 | 52.4 ± 13.8 | 49.4 ± 13.1 | 49.7 ± 12.9 | < 0.001 |
| BMI (kg/m2) | 24.4 ± 3.1 | 23.3 ± 3.3 | 21.9 ± 3.1 | 21.2 ± 2.6 | < 0.001 |
| Total cholesterol (mmol/L) | 5.27 ± 1.09 | 5.24 ± 0.99 | 5.34 ± 0.88 | 572 ± 0.87 | < 0.001 |
| Non-HDL cholesterol (mmol/L) | 4.38 ± 1.09 | 3.94 ± 1.00 | 3.59 ± 0.89 | 3.43 ± 0.82 | < 0.001 |
| Triglyceride (mmol/L)§ | 1.99 | 1.25 | 0.93 | 0.77 | < 0.001 |
| Hypertensions (%) | 47.1 | 45.0 | 45.9 | 43.0 | 0.694 |
| Diabetes (%) | 6.1 | 4.4 | 2.1 | 1.7 | < 0.001 |
| Metabolic syndrome (%) | 30.8 | 9.2 | 2.4 | 1.3 | < 0.001 |
| Smoking | |||||
| Never-smoker (%) | 84.7 | 88.4 | 90.2 | 88.3 | 0.005 |
| Ex-smoker (%) | 2.1 | 2.4 | 2.8 | 2.7 | |
| Current smoker (≦ 20 cigarettes/day)(%) | 10.8 | 8.4 | 6.6 | 8.4 | |
| Current smoker (≧ 21 cigarettes/day)(%) | 2.4 | 0.8 | 0.4 | 0.7 | |
| Drinking | |||||
| Never-drinker (%) | 95.1 | 94.0 | 91.1 | 85.6 | < 0.001 |
| Ex-drinker (%) | 1.2 | 0.8 | 1.2 | 0.3 | |
| Current drinker (%) | 3.8 | 5.2 | 7.8 | 14.1 | |
geometric mean
The value were expressed as the means ± srandard deviation (SD) for continuous variables, and as percentages for categorical variables. Analysis of variance was used for comparisons of multiple group means and the χ2-test was used to compare proportions.
Table 2 shows the number of deaths; multiple adjusted HRs; and 95% CIs for all-cause, CVD, and CHD mortality. In age-adjusted model, we did not observe any linear association between HDL-C categories and each cause of death. The risk for all-cause mortality did not show statistical difference among the four HDL-C categories. As a result of testing for the proportional hazard assumption using the schoenfeld residuals plot, the assumption was not violated for any variables. In addition, we observed an inverse relationship between HDL-C and all-cause mortality similar to that of our previous study in the same cut points as our former study (data not shown).
Table 2. The number of death and multivariate-adjusted HR (95% CIs) for all-cause and cardiovascular deaths according to serum HDL cholesterol.
Table 2A. [Model 1].
| Baseline HDL cholesterol level, mmol/L (mg/dl) | No. of persons | Person-years | All-cause |
Cardiovascular disease |
Coronary heart disease |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of deaths |
HR (95% CI) |
P | P¶ | No. of deaths |
HR (95% CI) |
P | P¶ | No. of deaths |
HR (95% CI) |
P | P¶ | |||
| Men | ||||||||||||||
| < 1.04 (< 40) | 702 | 12152 | 217 | 1.10 (0.94, 1.30) | 0.24 | 0.34 | 60 | 1.12 (0.82, 1.53) | 0.46 | 0.26 | 18 | 1.31 (0.73, 2.34) | 0.37 | 0.52 |
| 1.04–1.55 (40–59) | 1554 | 27119 | 443 | 1.00 | 120 | 1.00 | 31 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 554 | 9677 | 160 | 0.95 (0.80, 1.14) | 0.61 | 36 | 0.79 (0.54, 1.15) | 0.21 | 6 | 0.51 (0.21, 1.23) | 0.14 | |||
| 2.07+ (80+) | 136 | 2408 | 38 | 1.11 (0.79, 1.54) | 0.56 | 10 | 1.09 (0.57, 2.08) | 0.79 | 2 | 0.80 (0.19, 3.36) | 0.77 | |||
| Women | ||||||||||||||
| < 1.04 (< 40) | 425 | 7483 | 122 | 1.00 (0.81, 1.22) | 0.97 | 0.82 | 46 | 1.27 (0.89, 1.79) | 0.19 | 0.79 | 5 | 0.62 (0.24, 1.64) | 0.34 | 0.71 |
| 1.04–1.55 (40–59) | 2009 | 36829 | 377 | 1.00 | 106 | 1.00 | 24 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1341 | 25255 | 195 | 1.00 (0.84, 1.19) | 0.97 | 56 | 1.07 (0.77, 1.48) | 0.68 | 4 | 0.33 (0.11, 0.95) | 0.04 | |||
| 2.07+ (80+) | 298 | 5555 | 46 | 1.14 (0.84, 1.54) | 0.42 | 16 | 1.46 (0.86, 2.47) | 0.16 | 5 | 1.94 (0.74, 5.09) | 0.18 | |||
| Men and women combined§ | ||||||||||||||
| < 1.04 (< 40) | 1127 | 19636 | 339 | 1.07 (0.94, 1.21) | 0.32 | 0.51 | 106 | 1.21 (0.96, 1.52) | 0.11 | 0.43 | 23 | 1.07 (0.66, 1.74) | 0.79 | 0.71 |
| 1.04–1.55 (40–59) | 3563 | 63949 | 820 | 1.00 | 226 | 1.00 | 55 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1895 | 34933 | 355 | 0.98 (0.86, 1.11) | 0.69 | 92 | 0.92 (0.72, 1.17) | 0.50 | 10 | 0.41 (0.21, 0.81) | 0.01 | |||
| 2.07+ (80+) | 434 | 7963 | 84 | 1.12 (0.90, 1.40) | 0.32 | 26 | 1.28 (0.85, 1.92) | 0.23 | 7 | 1.39 (0.63, 3.06) | 0.41 | |||
HR: hazard ratio, 95% CI: 95% confidence interval
The HR was calculated using a Cox proportional hazard model. The Wald's test was used to examine the difference in the HR of each HDL-C category compared with the reference group.
Continuous serum HDL-C value used in the proportional hazard model.
Model 1: adjusted for age (+sex§)
Table 2B. [Model 2].
| Baseline HDL cholesterol level, mmol/L (mg/dl) | No. of persons | Person-years | All-cause |
Cardiovascular disease |
Coronary heart disease |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of deaths |
HR (95% CI) | P | P¶ | No. of deaths |
HR (95% CI) | P | P¶ | No. of deaths |
HR (95% CI) | P | P¶ | |||
| Men | ||||||||||||||
| < 1.04 (< 40) | 702 | 12152 | 217 | 1.13 (0.94, 1.35) | 0.19 | 0.16 | 60 | 1.12 (0.79, 1.59) | 0.51 | 0.38 | 18 | 1.70 (0.89, 3.27) | 0.11 | 0.43 |
| 1.04–1.55 (40–59) | 1554 | 27119 | 443 | 1.00 | 120 | 1.00 | 31 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 554 | 9677 | 160 | 0.90 (0.74, 1.10) | 0.25 | 36 | 0.78 (0.53, 1.14) | 0.20 | 6 | 0.46 (0.19, 1.12) | 0.09 | |||
| 2.07+ (80+) | 136 | 2408 | 38 | 1.02 (0.73, 1.43) | 0.91 | 10 | 1.08 (0.55, 2.09) | 0.83 | 2 | 0.77 (0.18, 3.34) | 0.73 | |||
| Women | ||||||||||||||
| < 1.04 (< 40) | 425 | 7483 | 122 | 1.09 (0.87, 1.35) | 0.46 | 0.12 | 46 | 1.40 (0.97, 2.04) | 0.07 | 0.55 | 5 | 0.68 (0.25, 1.88) | 0.46 | 0.97 |
| 1.04–1.55 (40–59) | 2009 | 36829 | 377 | 1.00 | 106 | 1.00 | 24 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1341 | 25255 | 195 | 0.93 (0.78, 1.12) | 0.46 | 56 | 1.01 (0.72, 1.42) | 0.95 | 4 | 0.31 (0.11, 0.93) | 0.04 | |||
| 2.07+ (80+) | 298 | 5555 | 46 | 0.96 (0.69, 1.32) | 0.79 | 16 | 1.20 (0.68, 2.11) | 0.52 | 5 | 1.72 (0.60, 4.90) | 0.31 | |||
| Men and women combined§ | ||||||||||||||
| < 1.04 (< 40) | 1127 | 19636 | 339 | 1.12 (0.98, 1.29) | 0.11 | 0.06 | 106 | 1.27 (0.99, 1.63) | 0.07 | 0.25 | 23 | 1.29 (0.76, 2.20) | 0.35 | 0.47 |
| 1.04–1.55 (40–59) | 3563 | 63949 | 820 | 1.00 | 226 | 1.00 | 55 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1895 | 34933 | 355 | 0.93 (0.81, 1.05) | 0.24 | 92 | 0.89 (0.69, 1.14) | 0.36 | 10 | 0.38 (0.19, 0.75) | 0.01 | |||
| 2.07+ (80+) | 434 | 7963 | 84 | 1.01 (0.80, 1.27) | 0.95 | 26 | 1.14 (0.74, 1.74) | 0.55 | 7 | 1.23 (0.54, 2.79) | 0.62 | |||
HR: hazard ratio, 95% CI: 95% confidence interval
The HR was calculated using a Cox proportional hazard model. The Wald's test was used to examine the difference in the HR of each HDL-C category compared with the reference group.
Continuous serum HDL-C value used in the proportional hazard model.
Model 2: adjusted for age + bmi + triglyceride (log-transformed) + non-hdl + hypertension + diabetes + smoking + drinking (+ sex§)
The risks for CHD among high HDL-C category (1.56–2.06 mmol/L) were lower than those among reference category (1.01–1.55 mmol/L). HRs were 0.33 (95% CI: 0.11–0.95) in women and 0.41 (95% CI: 0.21–0.81) when men and women were combined, and the risk in men was also lower (HR: 0.51, HR 0.21–1.23) although it did not reach statistical significance (p = 0.14). We did not observe lower risk for CHD in very high HDL-C category compared with the reference group. The results were almost similar after multivariable adjustment.
Table 3 shows the number of deaths, multiple adjusted HRs, and 95% CIs for stroke and its subtypes (cerebral infarction and intracerebral hemorrhage). We did not observe any significant difference in the risk for each subtype of stroke among the four HDL-C categories. We did not also observe any linear association between HDL-C categories and stroke mortality.
Table 3. The number of deaths and multivariate-adjusted HR (95% CIs) for subgroups of stroke death according to serum HDL cholesterol level.
Table 3A. [Model 1].
| Baseline HDL cholesterol level, mmol/L (mg/dl) | No. of persons | Person-years | Stroke |
Cerebral Infarction |
Cerebral Hemorrhage |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of deaths |
HR (95% CI) |
P | P¶ | No. of deaths |
HR (95% CI) |
P | P¶ | No. of deaths |
HR (95% CI) |
P | P¶ | |||
| Men | ||||||||||||||
| < 1.04 (< 40) | 702 | 12152 | 18 | 0.86 (0.50, 1.48) | 0.59 | 0.22 | 12 | 0.95 (0.49, 1.87) | 0.89 | 0.96 | 3 | 0.48 (0.14, 1.68) | 0.25 | 0.28 |
| 1.04–1.55 (40–59) | 1554 | 27119 | 47 | 1.00 | 29 | 1.00 | 14 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 554 | 9677 | 20 | 1.12 (0.67, 1.89) | 0.67 | 11 | 1.00 (0.50, 2.01) | 0.99 | 4 | 0.76 (0.25, 2.32) | 0.64 | |||
| 2.07+ (80+) | 136 | 2408 | 7 | 1.98 (0.89, 4.39) | 0.09 | 4 | 2.09 (0.73, 6.00) | 0.17 | 2 | 1.68 (0.38, 7.39) | 0.49 | |||
| Women | ||||||||||||||
| < 1.04 (< 40) | 425 | 7483 | 19 | 1.34 (0.77, 2.31) | 0.30 | 0.96 | 12 | 1.22 (0.62, 2.41) | 0.57 | 0.71 | 5 | 2.52 (0.76, 8.32) | 0.13 | 0.83 |
| 1.04–1.55 (40–59) | 2009 | 36829 | 41 | 1.00 | 27 | 1.00 | 6 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1341 | 25255 | 21 | 1.06 (0.62, 1.80) | 0.83 | 14 | 1.16 (0.60, 2.22) | 0.66 | 5 | 1.61 (0.49, 5.30) | 0.43 | |||
| 2.07+ (80+) | 298 | 5555 | 6 | 1.44 (0.61, 3.39) | 0.41 | 2 | 0.76 (0.18, 3.21) | 0.71 | 2 | 3.11 (0.63, 15.4) | 0.17 | |||
| Men and women combined§ | ||||||||||||||
| < 1.04 (< 40) | 1127 | 19636 | 37 | 1.07 (0.73, 1.57) | 0.74 | 0.44 | 24 | 1.07 (0.67, 1.73) | 0.77 | 0.81 | 8 | 1.05 (0.46, 2.38) | 0.92 | 0.42 |
| 1.04–1.55 (40–59) | 3563 | 63949 | 88 | 1.00 | 56 | 1.00 | 20 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1895 | 34933 | 41 | 1.07 (0.74, 1.55) | 0.73 | 25 | 1.07 (0.67, 1.71) | 0.79 | 9 | 0.99 (0.45, 2.19) | 0.99 | |||
| 2.07+ (80+) | 434 | 7963 | 13 | 1.69 (0.94, 3.04) | 0.08 | 6 | 1.34 (0.58, 3.13) | 0.49 | 4 | 2.07 (0.71, 6.07) | 0.19 | |||
HR: hazard ratio, 95% CI: 95% confidence interval
The HR was calculated using a Cox proportional hazard model. The Wald's test was used to examine the difference in the HR of each HDL-C
Continuous serum HDL-C value used in the proportional hazard model.
Model 1: adjusted for age (+sex§)
Table 3B. [Model 2].
| Baseline HDL cholesterol level, mmol/L (mg/dl) | No. of persons | Person-years | Stroke |
Cerebral Infarction |
Cerebral Hemorrhage |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of deaths |
HR (95% CI) | P | P¶ | No. of deaths |
HR (95% CI) | P | P¶ | No. of deaths |
HR (95% CI) | P | P¶ | |||
| Men | ||||||||||||||
| < 1.04 (< 40) | 702 | 12152 | 18 | 0.82 (0.45, 1.48) | 0.50 | 0.29 | 12 | 0.72 (0.34, 1.53) | 0.40 | 0.58 | 3 | 0.67 (0.17, 2.61) | 0.57 | 0.79 |
| 1.04–1.55 (40–59) | 1554 | 27119 | 47 | 1.00 | 29 | 1.00 | 14 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 554 | 9677 | 20 | 1.15 (0.66, 1.97) | 0.63 | 11 | 1.06 (0.52, 2.17) | 0.88 | 4 | 0.69 (0.22, 2.19) | 0.53 | |||
| 2.07+ (80+) | 136 | 2408 | 7 | 1.93 (0.84, 4.44) | 0.12 | 4 | 2.47 (0.82, 7.46) | 0.11 | 2 | 1.39 (0.29, 6.71) | 0.68 | |||
| Women | ||||||||||||||
| < 1.04 (< 40) | 425 | 7483 | 19 | 1.42 (0.79, 2.55) | 0.24 | 0.64 | 12 | 1.31 (0.63, 2.74) | 0.48 | 0.43 | 5 | 2.50 (0.70, 8.94) | 0.16 | 0.87 |
| 1.04–1.55 (40–59) | 2009 | 36829 | 41 | 1.00 | 27 | 1.00 | 6 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1341 | 25255 | 21 | 0.96 (0.56, 1.66) | 0.89 | 14 | 1.01 (0.52, 1.99) | 0.97 | 5 | 1.54 (0.45, 5.31) | 0.49 | |||
| 2.07+ (80+) | 298 | 5555 | 6 | 1.26 (0.52, 3.07) | 0.62 | 2 | 0.67 (0.15, 2.90) | 0.59 | 2 | 2.81 (0.50, 15.7) | 0.24 | |||
| Men and women combined§ | ||||||||||||||
| < 1.04 (< 40) | 1127 | 19636 | 37 | 1.10 (0.73, 1.67) | 0.65 | 0.76 | 24 | 0.99 (0.59, 1.68) | 0.98 | 0.84 | 8 | 1.25 (0.52, 3.04) | 0.62 | 0.89 |
| 1.04–1.55 (40–59) | 3563 | 63949 | 88 | 1.00 | 56 | 1.00 | 20 | 1.00 | ||||||
| 1.56–2.06 (60–79) | 1895 | 34933 | 41 | 1.03 (0.70, 1.52) | 0.87 | 25 | 1.05 (0.64, 1.70) | 0.86 | 9 | 0.91 (0.40, 2.05) | 0.81 | |||
| 2.07+ (80+) | 434 | 7963 | 13 | 1.53 (0.84, 2.81) | 0.17 | 6 | 1.33 (0.56, 3.16) | 0.53 | 4 | 1.62 (0.52, 5.02) | 0.41 | |||
HR: hazard ratio, 95% CI: 95% confidence interval
The HR was calculated using a Cox proportional hazard model. The Wald's test was used to examine the difference in the HR of each HDL-C category compared with the reference group.
Continuous serum HDL-C value used in the proportional hazard model.
Model 2: adjusted for age + bmi + triglyceride (log-transformed) + non-hdl + hypertension + diabetes + smoking + drinking (+ sex§)
The results of stratification by the existence of hypertension, smoking, alcohol drinking, diabetes, high serum level of non-HDL-C, hypertriglyceridemia for CVD, and all-cause mortality showed substantially similar results shown in Table 2 (data not shown).
Discussion
In our 20-year follow-up study, serum HDL-C levels did not show significant association with all-cause and stroke mortality in this Japanese general population. Furthermore, the high HDL-C category (1.56–2.06 mmol/L) showed significant inverse association with CHD mortality; however, the very high HDL-C category (≥ 2.07 mmol/L) did not show significant relationship with CHD. Consequently, there is no inverse relationship between HDL-C levels and cardiovascular mortality because major subtype of CVD in Japan is stroke13, 23); in fact, the proportion of stroke within CVD (39%) was twice as large as that of CHD (21%) in the present study.
In this same cohort, we have already reported an inverse relationship between HDL-C and all-cause mortality at 10-year follow-up14). In the present study, we followed up same individuals for 20 years, which was twice the follow-up period of our previous study. As a result, HDL-C levels did not show significant association with all-cause mortality. This result may seem to show different relationship from our previous study14). However, the result of the present study does not show variation in these relationships for 20 years because we observed an inverse relationship between HDL-C and all-cause mortality similar to that of our previous study although we used the same cut points, and the proportional hazard assumption was not violated for variables. These findings indicate that the discrepancy among our studies is apparently caused by the difference of cut points, especially because of adding very high HDL-C category. Moreover, because we assessed the effect of HDL-C only based on single measurement at baseline survey, the longer follow-up period was prolonged, the more potential confounding factors, especially during follow-up period, affected the results. For example, statins, which had been launched in 1989, have not expanded yet when baseline surveys started in 1990, however the use of statin has been very popular year by year. Thus, prolongation of follow-up period may strengthen the confounding effect of statin therapy. In addition, change of lifestyle of participants during follow-up period may be another important confounding factor.
In the present study, HDL-C did not show significant association with stroke, even with cerebral infarction. As mentioned above, in Japan, stroke mortality is much higher than CHD mortality13, 23), and to our knowledge, almost all Japanese cohort studies did not show any positive relationship between TC or LDL-C and cerebral infarction24, 25). Even in a large worldwide pooled analysis26), a weak positive relationship between TC or LDL-C and stroke was only observed. Usually, main biological background of protective effect of HDL-C is explained by reverse-cholesterol transport, which acts in an opposing manner against cholesterol accumulation to artery because of LDL-C. Thus, it is not surprising that there is no association between low HDL-C and cerebral infarction in a population where there is no relationship between high LDL-C and cerebral infarction. A previous study in Japan showed a significant association between low HDL-C and risk of cerebral infarction; however, the definition of low HDL-C was too low [< 0.78 mmol/L (30 mg/dl)], which suggested other metabolic background such as diabetes or visceral obesity27).
We found that the high HDL-C category (1.56–2.06 mmol/L) significantly reduces the risk of CHD, whereas the very high HDL-C category (≥ 2.07 mmol/L) does not. It has been controversial whether very high HDL-C levels prevent CHD. It has been reported that high levels of serum HDL-C per se protect against CHD regardless of the cause of HDL-C elevation when high HDL-C was defined as ≥ 1.56 mmol/L13, 21). In contrast, pooled analysis of six cohort studies has shown that very high HDL-C defined as ≥ 2.07 mmol/L increased the risk for CHD events28). Abnormal HDL-C function raised by cholesterol ester transfer protein (CETP) deficiency leads to lowering of cholesterol efflux, then reverse cholesterol transport was inhibited29). A clinical trial using CETP inhibitor has failed to show the protective effect for CHD, although participants in this study showed an extremely HDL-C elevation almost equivalent to very high HDL-C level in our study30). Some previous studies have shown mild inverse or no association between HDL-C concentrations and plasma CETP activity31–33). Moreover, plasma CETP activity showed inverse correlation with CVD risk independent of HDL-C levels34). Furthermore, cholesterol efflux capacity that reflects HDL-C function has shown the inverse relationship with some status of atherosclerosis35). These findings indicate that the function as well as quantity of HDL-C is carefully evaluated for an individual with very high HDL-C level. Further studies with information about HDL-C function should be warranted.
Several studies reported that heavy alcohol consumption or alcoholism can be confounding factors of excess death in high HDL-C group36). In addition, in the present study, the proportion of drinkers between both men and women in very high HDL-C category was the highest in the four categories. However, the proportions of population with hypertension, diabetes, smoker, and obesity among very high HDL-C categories are not higher compared with those among high HDL-C categories, hence they are not considered to be unhealthy population. In addition, a small percentage of women among very high HDL-C category, which was 14%, were drinker. Therefore, we consider that alcohol consumption was not a confounding factor among very high HDL-C category in the present study.
There are some limitations in this study. First, we did not have any information about participants concerning change in risk characteristics including use of lipid lowering agents during follow-up. Second, causes of death may be misclassified because “heart failure” may include CHD. Accordingly, underestimation of CHD deaths is possible37). This should make it more difficult to show an association between HDL-C and death due to CHD; accordingly, we believe our results may be conservative and not to be over-stated. Third, the number of events among very high HDL-C category was not enough even in our extended 20-year cohort study. In future, larger population study is needed to investigate population with very high HDL-C levels. This result indicates a potential that the function as well as the quantity of HDL-C might be evaluated for individuals with very high HDL-C level.
Conclusions
A 20-year follow-up study did not show the significant association between serum HDL-C levels and all-cause mortality and stroke. In contrast, high HDL-C level (1.56–2.06 mmol/L) significantly reduced risk for CHD death. However, very high HDL-C level (≥ 2.07 mmol/L) did not. Because sample size of cause-specific death in very high HDL-C group was not enough even in this 20-year follow-up of 7,019 Japanese, a larger community-based cohort study or a pooled analysis of a number of cohort studies is needed in the future.
Acknowledgements
This study was supported by the grant-in-aid of the Ministry of Health, Labor and Welfare under the auspices of Japanese Association for Cerebro-cardiovascular Disease Control, the Research Grant for Cardiovascular Diseases (7A-2) from the Ministry of Health, Labor and Welfare, and the Health and Labor Sciences Research Grant, Japan (Comprehensive Research on Aging and Health [H11-Chouju-046, H14-Chouju-003, H17-Chouju-012, H19-Chouju-Ippan-014] and Comprehensive Research on Life-Style Related Diseases including Cardiovascular Diseases and Diabetes Mellitus [H22-Jyunkankitou-Seisyu-Sitei-017, H25-Junkankitou-Seishuu-Sitei-022]. We appreciate the members of the NIPPON DATA80/90 Research Group which is listed in the Appendix.
Appendix
The NIPPON DATA80/90 Research Group
Chairpersons: Hirotsugu Ueshima (Center for Epidemiologic Research in Asia, Shiga University of Medical Science, Otsu, Shiga), Akira Okayama (Research Institute of Strategy for Prevention, Tokyo) for the NIPPON DATA80; Hirotsugu Ueshima, Tomonori Okamura (Department of Preventive Medicine and Public Health, Keio University, Tokyo) for the NIPPON DATA90.
Research members: Shigeyuki Saitoh (School of Health Sciences, Sapporo Medical University, Sapporo, Hokkaido), Kiyomi Sakata (Department of Hygiene and Preventive Medicine, Iwate Medical University, Morioka, Iwate), Atsushi Hozawa (Department of Preventive Medicine and Epidemiology, Tohoku Medical Megabank Organization, Tohoku University, Sendai, Miyagi), Takehito Hayakawa (Department of Hygiene and Preventive Medicine, Fukushima Medical University, Fukushima), Yoshikazu Nakamura (Department of Public Health, Jichi Medical University, Shimotsuke, Tochigi), Nobuo Nishi (Center for International Collaboration and Partnership, National Institutes of Biomedical Innovation, Health and Nutrition, Tokyo), Nagako Okuda (Department of Health and Nutrition, University of Human Arts and Sciences, Saitama), Takayoshi Ohkubo (Department of Hygiene and Public Health, Teikyo University School of Medicine, Tokyo), Fumiyoshi Kasagi (Institute of Radiation Epidemiology, Radiation Effects Association, Tokyo), Yoshitaka Murakami (Department of Medical Statistics, Toho University, Tokyo), Toru Izumi (Kitasato University, Sagamihara, Kanagawa), Yasuhiro Matsumura (Faculty of Health and Nutrition, Bunkyo University, Chigasaki, Kanagawa), Toshiyuki Ojima (Department of Community Health and Preventive Medicine, Hamamatsu University School of Medicine, Hamamatsu, Shizuoka), Koji Tamakoshi (Department of Public Health and Health Information Dynamics, Nagoya University Graduate School of Medicine, Nagoya, Aichi), Hideaki Nakagawa (Kanazawa Medical University Medical Research Institute, Kanazawa, Ishikawa), Yoshikuni Kita (Faculty of Nursing Science, Tsuruga Nursing University, Tsuruga, Fukui), Katsuyuki Miura (Director of the Center for Epidemiologic Research in Asia, Department of Public Health, Shiga University of Medical Science, Otsu, Shiga), Aya Kadota (Center for Epidemiologic Research in Asia, Shiga University of Medical Science, Otsu, Shiga), Akira Fujiyoshi (Department of Public Health, Shiga University of Medical Science, Otsu, Shiga), Naomi Miyamatsu (Department of Clinical Nursing, Shiga University of Medical Science, Otsu, Shiga), Yasuyuki Nakamura (Department of Food Science and Human Nutrition, Ryukoku University, Otsu, Shiga), Katsushi Yoshita (Osaka City University Graduate School of Human Life Science, Osaka), Yoshihiro Miyamoto (Department of Preventive Cardiology, National Cerebral and Cardiovascular Center, Suita, Osaka), Kazunori Kodama (Radiation Effects Research Foundation, Hiroshima) and Yutaka Kiyohara (Department of Environmental Medicine, Kyushu University, Fukuoka).
Conflicts of Interest
There are no conflicts of interest in the present study.
Reference
- 1). Castelli WP, Garrison RJ, Wilson PW, Abbott RD, Kalousdian S, Kannel WB: Incidence of coronary heart disease and lipoprotein cholesterol levels. The Framingham Study. JAMA, 1986; 256: 2835-2838 [PubMed] [Google Scholar]
- 2). Després JP, Lemieux I, Dagenais GR, Cantin B, Lamarche B: HDL-cholesterol as a marker of coronary heart disease risk: The Québec cardiovascular study. Atherosclerosis, 2000; 153: 263-272 [DOI] [PubMed] [Google Scholar]
- 3). Assman G, Schulte H. Relation of high-density lipoprotein cholesterol and triglycerides to incidence of atherosclerotic coronary disease (the PROCAM experience). Am J Cardiol, 1992; 70: 733-737 [DOI] [PubMed] [Google Scholar]
- 4). Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR: High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study. Am J Med, 1977; 62: 707-714 [DOI] [PubMed] [Google Scholar]
- 5). Wilson PW, Abbott RD, Castelli WP: High density lipoprotein cholesterol and mortality. The Framingham Heart Study. Arteriosclerosis, 1988; 8: 737-741 [DOI] [PubMed] [Google Scholar]
- 6). Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD, Jacobs DR, Jr, Bangdiwala S, Tyroler HA: High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation, 1989; 79: 8-15 [DOI] [PubMed] [Google Scholar]
- 7). Kitamura A, Iso H, Naito Y, Iida M, Konishi M, Folsom AR, Sato S, Kiyama M, Nakamura M, Senkai T: High-density lipoprotein cholesterol and premature coronary heart disease in urban Japanese men. Circulation, 1994; 89: 2533-2539 [DOI] [PubMed] [Google Scholar]
- 8). Goldbourt U, Yaari S, Medalie JH: Isolated low HDL cholesterol as a risk factor for coronary heart disease mortality. A 21-year follow-up of 8000 men. Arterioscler Thromb Vasc Biol, 1997; 17: 107-113 [DOI] [PubMed] [Google Scholar]
- 9). Matsuzaki M, Kita T, Mabuchi H, Matsuzawa Y, Nakaya N, Oikawa S, Saito Y, Sasaki J, Shimamoto K, Itakura H, J-LIT Study Group : 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: 1087-1095 [DOI] [PubMed] [Google Scholar]
- 10). Norrish A, North D, Yee RL, Jackson R: Do cardiovascular disease risk factors predict all-cause mortality? Int J Epidemiol, 1995; 25: 908-914 [DOI] [PubMed] [Google Scholar]
- 11). Rywik SL, Manolio TA, Pajak A, Piotrowski W, Davis CE, Broda GB, Kawalec E: Association of lipids and lipoprotein level with total mortality and mortality caused by cardiovascular and cancer diseases (Poland and United States collaborative study on cardiovascular epidemiology). Am J Cardiol, 1999; 84: 540-548 [DOI] [PubMed] [Google Scholar]
- 12). Perova NV, Oganov RG, Williams DH, Irving SH, Abernathy JR, Deev AD, Shestov DB, Zhukovsky GS, Davis CE, Tyroler HA: Association of high-density-lipoprotein cholesterol with mortality and other risk factors for major chronic noncommunicable diseases in samples of US and Russian men. Ann Epidemiol, 1995; 5: 179-185 [DOI] [PubMed] [Google Scholar]
- 13). Curb JD, Abbott RD, Rodriguez BL, Masaki K, Chen R, Sharp DS, Tall AR: A prospective study of HDL-C and cholesteryl ester transfer protein gene mutations and the risk of coronary heart disease in the elderly. J Lipid Res, 2004; 45: 948-953 [DOI] [PubMed] [Google Scholar]
- 14). Okamura T, Hayakawa T, Kadowaki T, Kita Y, Okayama A, Ueshima H, NIPPON DATA90 Research Group : The inverse relationship between serum high-density lipoprotein cholesterol level and all-cause mortality in a 9.6-year follow-up study in the Japanese general population. Atherosclerosis, 2006; 184: 143-150 [DOI] [PubMed] [Google Scholar]
- 15). Yano Y, Irie N, Homma Y, Tsushima M, Takeuchi I, Nakaya N, Goto Y: High density lipoprotein cholesterol levels in the Japanese. Atherosclerosis, 1980; 36: 173-181 [DOI] [PubMed] [Google Scholar]
- 16). Okamura T, Hayakawa T, Kadowaki T, Kita Y, Okayama A, Elliott P, Ueshima H: A combination of serum low albumin and above-average cholesterol level was associated with excess mortality. J Clin Epidemiol, 2004; 57: 1188-1195 [DOI] [PubMed] [Google Scholar]
- 17). Sugiyama D, Okamura T, Watanabe M, Higashiyama A, Okuda N, Nakamura Y, Hozawa A, Kita Y, Kadota A, Murakami Y, Miyamatsu N, Ohkubo T, Hayakawa T, Miyamoto Y, Miura K, Okayama A, Ueshima H, NIPPON DATA 80/90 Research Group : Risk of hypercholesterolemia for cardiovascular disease and the population attributable fraction in a 24-year Japanese cohort study. J Atheroscler Thromb, 2015; 22: 95-107 [DOI] [PubMed] [Google Scholar]
- 18). Nakamura K, Okamura T, Hayakawa T, Hozawa A, Kadowaki T, Murakami Y, Kita Y, Okayama A, Ueshima H, NIPPON DATA90 Research Group : The proportion of individuals with alcohol-induced hypertension among total hypertensives in a general Japanese population: NIPPON DATA90. Hypertens Res, 2007; 30: 663-668 [DOI] [PubMed] [Google Scholar]
- 19). Takashima N, Miura K, Hozawa A, Kadota A, Okamura T, Nakamura Y, Hayakawa T, Okuda N, Fujiyoshi A, Nagasawa SY, Kadowaki T, Murakami Y, Kita Y, Okayama A, Ueshima H, Takashima N, Miura K: Population attributable fraction of smoking and metabolic syndrome on cardiovascular disease mortality in Japan: a 15-year follow up of NIPPON DATA90. BMC Public Health, 2010; 10: 306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20). Nakamura M, Sato S, Shimamoto T: Improvement in Japanese clinical laboratory measurements of total cholesterol and HDL-cholesterol by the US Cholesterol Reference Method Laboratory Network. J Atheroscler Thromb, 2003; 10: 145-153 [DOI] [PubMed] [Google Scholar]
- 21). Moriyama Y, Okamura T, Inazu A, Doi M, Iso H, Mouri Y, Ishikawa Y, Suzuki H, Iida M, Koizumi J, Mabuchi H, Komachi Y: A low prevalence of coronary heart disease among subjects with increased high-density lipoprotein cholesterol levels, including those with plasma cholesteryl ester transfer protein deficiency. Prev Med, 1998; 27: 659-667 [DOI] [PubMed] [Google Scholar]
- 22). Lee KL, Harrell FE, Jr, Tolley HD, Rosati RA: A comparison of test statistics for assessing the effects of concomitant variables in survival analysis. Biometrics, 1983; 39: 341-350 [PubMed] [Google Scholar]
- 23). Keys A, Menotti A, Aravanis C, Blackburn H, Djordevic BS, Buzina R, Dontas AS, Fidanza F, Karvonen MJ, Kimura N, Mohacek I, Nedelikovic S, Puddu V, Punsar S, Taylor HL, Conti A, Kromhout D, Toshima H: The seven countries study: 2,289 deaths in 15 years. Prev Med, 1984; 13: 141-154 [DOI] [PubMed] [Google Scholar]
- 24). Tanaka T, Okamura T: Blood cholesterol level and risk of stroke in community-based or worksite cohort studies: a review of Japanese cohort studies in the past 20 years. Keio J Med, 2012; 61: 79-88 [DOI] [PubMed] [Google Scholar]
- 25). Satoh M, Ohkubo T, Asayama K, Murakami Y, Sakurai M, Nakagawa H, Iso H1, Okayama A, Miura K, Imai Y, Ueshima H, Okamura T, Evidence for Cardiovascular Prevention From Observational Cohorts in Japan (EPOCH-JAPAN) Research Group : 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: 517-524 [DOI] [PubMed] [Google Scholar]
- 26). Prospective Studies Collaboration. Lewington S, Whitlock G, Clarke R, Sherliker P, Emberson J, Halsey J, Qizilbash N, Peto R, Collins R: 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: 1829-1839 [DOI] [PubMed] [Google Scholar]
- 27). Soyama Y, Miura K, Morikawa Y, Nishijo M, Nakanishi Y, Naruse Y, Kagamimori S, Nakagawa H, Oyabe Study : High-density lipoprotein cholesterol and risk of stroke in Japanese men and women: the Oyabe Study. Stroke, 2003; 34: 863-868 [DOI] [PubMed] [Google Scholar]
- 28). Wilkins JT, Ning H, Stone NJ, Criqui MH, Zhao L, Greenland P, Lloyd-Jones DM: Coronary heart disease risks associated with high levels of HDL cholesterol. J Am Heart Assoc, 2014; 3: e000519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29). Barkowski RS, Frishman WH: HDL metabolism and CETP inhibition. Cardiol Rev, 2008; 16: 154-162 [DOI] [PubMed] [Google Scholar]
- 30). Barter PJ, Caulfield M, Eriksson M, Grundy SM, Kastelein JJ, Komajda M, Lopez-Sendon J, Mosca L, Tardif JC, Waters DD, Shear CL, Revkin JH, Buhr KA, Fisher MR, Tall AR, Brewer B, ILLUMINATE Investigators : Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med, 2007; 357: 2109-2122 [DOI] [PubMed] [Google Scholar]
- 31). Tato F, Vega GL, Tall AR, Grundy SM: Relation between cholesterol ester transfer protein activities and lipoprotein cholesterol in patients with hypercholesterolemia and combined hyperlipidemia. Arterioscler Thromb Vasc Biol, 1995; 15: 112-120 [DOI] [PubMed] [Google Scholar]
- 32). Kinoshita M, Teramoto T, Shimazu N, Kaneko K, Ohta M, Koike T, Hosogaya S, Ozaki Y, Kume S, Yamanaka M: CETP is a determinant of serum LDL-cholesterol but not HDL-cholesterol in healthy Japanese. Atherosclerosis, 1996; 120: 75-82 [DOI] [PubMed] [Google Scholar]
- 33). Smaoui M, Hammami S, Attia N, Chaaba R, Abid N, Kilani N, Kchaou H, Mahjoub S, Abid M, Hammami M: Modulation of plasma cholesteryl ester transfer protein activity by unsaturated fatty acids in Tunisian type 2 diabetic women. Nutr Metab Cardiovasc Dis, 2006; 16: 44-53 [DOI] [PubMed] [Google Scholar]
- 34). Vasan RS, Pencina MJ, Robins SJ, Zachariah JP, Kaur G, D'Agostino RB, Ordovas JM: Association of circulating cholesteryl ester transfer protein activity with incidence of cardiovascular disease in the community. Circulation, 2009; 120: 2414-2420 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35). Khera AV, Cuchel M, de la Llera-Moya M, Rodrigues A, Burke MF, Jafri K, French BC, Phillips JA, Mucksavage ML, Wilensky RL, Mohler ER, Rothblat GH, Rader DJ: Cholesterol efflux capacity, high-density lipoprotein function, and atherosclerosis. N Engl J Med, 2011; 364: 127-135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36). Paunio M, Heinonen OP, Virtamo J, Klag MJ, Manninen V, Albanes D, Comstock GW: HDL cholesterol and mortality in Finnish men with special reference to alcohol intake. Circulation, 1994; 90: 2909-2918 [DOI] [PubMed] [Google Scholar]
- 37). Saito I, Folsom AR, Aono H, Ozawa H, Ikebe T, Yamashita T: Comparison of fatal coronary heart disease occurrence based on population surveys in Japan and the USA. Int J Epidemiol, 2000; 29: 837-844 [DOI] [PubMed] [Google Scholar]