Alcohol-independent beneficial cardiometabolic profile of individuals with hyper-HDL cholesterolemia in Japanese men and women

Journal of Clinical Lipidology (2015) 9, 684–691

Alcohol-independent beneficial cardiometabolic profile of individuals with hyper-HDL cholesterolemia in Japanese men and women Ichiro Wakabayashi, MD, PhD*, Takashi Daimon, PhD Department of Environmental and Preventive Medicine, Hyogo College of Medicine, Nishinomiya, Hyogo, Japan (Dr Wakabayashi); and Division of Biostatistics, Hyogo College of Medicine, Nishinomiya, Hyogo, Japan (Dr Daimon) KEYWORDS: Alcohol; Cardiovascular disease; Diabetes mellitus; Dyslipidemia; Hyper-HDL cholesterolemia; Metabolic syndrome; Obesity

BACKGROUND: There is limited information on characterization of individuals with hyper–high-density lipoprotein (HDL) cholesterolemia. OBJECTIVES: The purpose of this study was to investigate the cardiometabolic profile of individuals with hyper-HDL cholesterolemia in comparison with the profile of individuals with normo-HDL cholesterolemia. METHODS: The subjects were Japanese men and women who had hyper-HDL cholesterolemia ($100 mg/dL) and their control subjects who had normal HDL cholesterol levels ($40 and ,80 mg/dL) and were matched for age or age and alcohol consumption. The cardiometabolic profiles were compared between the hyper- and normo-HDL cholesterolemic groups. RESULTS: Both in men and women, body mass index, waist-to-height ratio, triglycerides, low-density lipoprotein cholesterol, and hemoglobin A1c were significantly lower in subjects with hyper-HDL cholesterolemia than in subjects with normo-HDL cholesterolemia, whereas systolic and diastolic blood pressure levels were not significantly different between the 2 groups. In generalized estimating equation with adjustment for smoking and regular exercise, odds ratios of the hyper- vs normo-HDL cholesterolemic groups were significantly lower than the reference level of 1.00 for high body mass index, high waist-to-height ratio, hypertriglyceridemia, hyper–low-density lipoprotein cholesterolemia, high lipid accumulation product, and metabolic syndrome. The previously mentioned results were obtained both in age-matched analysis and in age- and alcohol intake–matched analysis, although the percentage of regular drinkers was significantly higher in the hyper-HDL cholesterolemic group than in the age-matched control group. CONCLUSIONS: Hyper-HDL cholesterolemia was inversely associated with obesity, dyslipidemia, and metabolic syndrome in the analysis using alcohol intake–matched subject groups. Therefore, the association of hyper-HDL cholesterolemia with lower cardiometabolic risk is thought to be independent of habitual alcohol drinking. Ó 2015 National Lipid Association. All rights reserved.

Introduction * Corresponding author. Department of Environmental and Preventive Medicine, Hyogo College of Medicine, Mukogawa-cho 1-1, Nishinomiya, Hyogo 663-8501, Japan. E-mail address: [email protected] Submitted March 4, 2015. Accepted for publication July 4, 2015.

1933-2874/Ó 2015 National Lipid Association. All rights reserved. http://dx.doi.org/10.1016/j.jacl.2015.07.002

There is limited information on characterization of persons with very high blood concentrations of high-density lipoprotein (HDL) cholesterol. In Japan, hyper-HDL cholesterolemia is defined as blood HDL cholesterol levels of 100 mg/dL or

Wakabayashi and Daimon

Hyper-HDL cholesterolemia and cardiovascular risk

higher,1 and 57% of hyper-HDL cholesterolemic individuals have been reported to have mutations of the cholesterol ester transfer protein (CETP) gene.2 However, it is not clear whether and how cardiovascular risk is altered in persons with hyperHDL cholesterolemia. The risk of cardiovascular disease is known to be lower in light-to-moderate drinkers than in nondrinkers.3 A dose-response positive relationship has been reported between alcohol consumption and blood HDL cholesterol level.4 Although blood HDL cholesterol level is known to be potently influenced by the status of habitual alcohol drinking,4 it also remains to be clarified whether the status of habitual drinking modifies the relationship between hyper-HDL cholesterolemia and cardiovascular risk. The purpose of this study was therefore to investigate the cardiometabolic profiles of individuals with hyper-HDL cholesterolemia in comparison with the profiles of individuals who have normal blood HDL cholesterol levels and to determine whether the differences in their cardiometabolic profiles can be explained by alcohol consumption. Common cardiovascular risk factors, such as adiposity, blood pressure, blood lipids, glycemic status, and metabolic syndrome, were compared in male and female hyper-HDL cholesterolemic groups and the corresponding control (normo-HDL cholesterolemic) groups matched for age or for both age and alcohol consumption.

Methods Subjects A cross-sectional study was performed using a local population-based database. The subjects in the original database of health checkups included 32,105 men and 17,900 women, aged from 35 to 60 years, who had received periodic health examinations at their workplaces in

Table 1

685

Yamagata Prefecture in Japan. The original database used in this study was a collection of the results of annual health checkup examinations performed from April 2005 to March 2006 for workers in a district of Japan. Subjects who were receiving treatment for any illness were requested to state the names of the diseases in a questionnaire at the health checkup. Those receiving treatment for dyslipidemia were excluded from the subjects of this study. From overall subjects in the original database, subjects showing hyper-HDL cholesterolemia (373 men and 387 women) were selected. Hyper-HDL cholesterolemia was defined as HDL cholesterol levels of 100 mg/dL or higher.1 The prevalences of hyper-HDL cholesterolemia were 1.16% in men and 2.16% in women. Then subjects with normal blood HDL cholesterol levels ($40 and ,80 mg/ dL), who were matched with the extracted hyper-HDL cholesterolemic subjects, were further extracted from the overall subjects. These one-to-one matchings were based on age and on both age and degree of alcohol consumption and, consequently, 2 control groups were obtained as ‘‘control 1’’ and ‘‘control 2,’’ respectively (Tables 1–4). This study was approved by the Ethics Committee of Yamagata University School of Medicine (no. 112 from April 2005 to March 2006, approved on March 13, 2006). In a questionnaire at the health checkup, subjects were required to identify any conditions for which they were receiving treatment. The questionnaire also surveyed the subjects’ histories of alcohol consumption, cigarette smoking, and habitual exercise. The subjects were divided into 3 groups based on average daily cigarette consumption (nonsmokers; light smokers, less than 20 cigarettes per day; heavy smokers, 20 cigarettes or more but less than 40 cigarettes per day; very heavy smokers, 40 cigarettes or more per day). Subjects who exercised almost every day for 30 minutes or longer per day were defined as those with a history of regular exercise habit.

Characteristics of subjects with hyper-HDL cholesterolemia and normal blood HDL cholesterol levels

Variable Number Age (y) HDL-C (mg/dL) Therapy for HT (%) Therapy for DM (%) Drinkers (%) Non Occasional Regular Smokers (%) Exercise (%)

Men Hyper-HDL-C

Control 1

Control 2

Women Hyper-HDL-C

Control 1

Control 2

373 48.9 6 7.4 109.5 6 10.9 8.3 1.9

373 48.9 6 7.4 54.8 6 10.1** 11.8 3.2

373 48.9 6 7.4 56.7 6 10.5 12.3 3.2

387 48.3 6 6.6 108.2 6 8.1 7.0 0.5

387 48.3 6 6.6 62.7 6 10.2 6.7 0.5

387 48.3 6 6.6 63.1 6 9.8 10.3 0.3

4.8 21.5 73.7 55.8 10.2

22.5** 29.2* 48.3** 54.4 9.1

4.8 21.5 73.7 56.8 9.1

35.9 35.2 28.9 19.6 8.0

63.8** 29.2 7.0** 20.4 6.5

35.9 35.2 28.9 28.4** 6.2

DM, diabetes mellitus; HDL-C, high-density lipoprotein cholesterol; hyper-HDL-C, hyper–high-density lipoprotein cholesterolemia; HT, hypertension. Means with standard deviations or percentages of each variable are shown. Control 1 and control 2 indicate age-matched and age- and alcohol intake– matched control groups with normal blood HDL cholesterol levels, respectively. Asterisks denote significant differences from the hyper-HDL cholesterolemic group (*P , .05; **P , .01).

686

Journal of Clinical Lipidology, Vol 9, No 5, October 2015

Table 2 Comparison of means of each variable related to cardiovascular risk between subjects with hyper-HDL cholesterolemia and normal blood HDL cholesterol levels Men Hyper-HDL-C

Variable

Control 1

Women Hyper-HDL-C

Control 2

Control 1

Control 2

20.9 6 2.3 23.7 6 3.1** 23.7 6 3.0** 20.2 6 2.3 22.6 6 3.7** 22.7 6 3.6** BMI (kg/m ) WHtR 0.453 6 0.041 0.497 6 0.051** 0.499 6 0.047** 0.462 6 0.048 0.504 6 0.065** 0.506 6 0.065** SBP (mm Hg) 130.7 6 15.6 130.2 6 17.5 130.5 6 17.6 120.9 6 18.2 121.0 6 16.8 123.0 6 16.8 DBP (mm Hg) 79.1 6 11.0 78.5 6 12.1 79.1 6 12.3 72.2 6 11.5 72.0 6 11.6 73.1 6 11.1 Log TG (mg/dL) 1.84 6 0.21 2.10 6 0.27** 2.10 6 0.28** 1.77 6 0.18 1.92 6 0.22** 1.93 6 0.23** LDL-C (mg/dL) 87.1 6 25.8 120.4 6 31.9** 116.7 6 33.2** 96.8 6 28.3 117.2 6 30.1** 115.6 6 31.8** Hemoglobin A1c (%) 5.30 6 0.78 5.46 6 0.64** 5.46 6 0.75** 5.25 6 0.35 5.34 6 0.31** 5.36 6 0.55** 2

BMI, body mass index; DBP, diastolic blood pressure; HDL, high-density lipoprotein; hyper-HDL-C, hyper-HDL cholesterolemia; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; TG, triglycerides; WHtR, waist-to-height ratio. Means with standard deviations of each variable are shown. Control 1 and control 2 indicate age-matched and age- and alcohol intake–matched control groups with normal blood HDL cholesterol levels, respectively. Asterisks denote significant differences from the hyper-HDL cholesterolemic group (**P , .01).

Classification of drinker groups The frequency of habitual alcohol drinking was assessed using the following questionnaire item: ‘‘How frequently do you drink alcohol?’’ The frequency of weekly alcohol drinking was categorized as ‘‘every day’’ (regular drinkers), ‘‘sometimes’’ (occasional drinkers) and ‘‘never’’ (nondrinkers). The usual daily alcohol consumption was calculated in terms of the equivalent number of ‘‘go,’’ a traditional Japanese unit of sake (rice wine). The amounts of other alcoholic beverages, including beer, wine, whisky, and shochu (a traditional Japanese distilled spirit), were converted and expressed as units of ‘‘go.’’ One go corresponds to approximately 180 mL of sake, 500 mL of beer, 240 mL of wine, 60 mL of whisky, and 80 mL of shochu. The amount of alcohol consumed daily was categorized as ‘‘null’’ (nondrinkers), ‘‘less than 1 go per day’’ (light drinkers), ‘‘1 go or more but less than 2 go per day’’ (moderate drinkers), ‘‘2 go or more but less than 3 go

per day’’ (heavy drinkers), and ‘‘3 go or more per day’’ (very heavy drinkers). One ‘‘go’’ contains about 22 g of ethanol, and this amount was used to separate moderate drinkers from light drinkers because it is generally accepted that alcohol intake should be reduced to less than 20 to 30 g per day from the viewpoint of preventing hypertension.5,6 Thus, 9 categories of alcohol drinking defined by the above frequency and amount of alcohol consumption were used to prepare the alcohol-matched normo-HDL cholesterolemic control groups of men and women.

Measurements Height and body weight were measured with the subjects wearing light clothes without a jacket or coat at the health checkup. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Waist circumference (WC) was measured at the navel level according to the recommendation of the

Table 3 Comparison of prevalence of each cardiovascular risk factor between the groups of subjects with hyper-HDL cholesterolemia and normal blood HDL cholesterol levels Variable High BMI (%) High WHtR (%) Hypertension (%) High TG (%) High LDL-C (%) High LAP (%) Hyperglycemia (%) Diabetes mellitus (%) MetS (%)

Men Hyper-HDL-C

Control 1

Control 2

3.2 10.7 34.9 8.0 1.6 2.1 8.8 2.9 2.4

31.9** 46.1** 35.9 36.2** 23.6** 35.4** 16.6** 5.9 17.4**

29.0** 48.0** 36.7 38.1** 21.2** 35.7** 16.6** 6.2 18.0**

Women Hyper-HDL-C

Control 1

Control 2

3.1 20.4 18.6 3.6 6.7 9.3 6.7 0.5 1.3

20.9** 48.6** 18.6 12.9** 21.2** 42.6** 7.8 0.3 5.2**

22.5** 51.2** 22.5 14.7** 20.2** 43.7** 9.6 1.0 6.7**

BMI, body mass index; HDL, high-density lipoprotein; hyper-HDL-C, hyper-HDL cholesterolemia; LAP, lipid accumulation product; LDL-C, low-density lipoprotein cholesterol; MetS, metabolic syndrome; TG, triglycerides; WHtR, waist-to-height ratio. Prevalences of each variable are shown. Control 1 and control 2 indicate age-matched and age- and alcohol intake–matched control groups with normal blood HDL cholesterol levels, respectively. Asterisks denote significant differences from the hyper-HDL cholesterolemic group (**P , .01).

Wakabayashi and Daimon

Hyper-HDL cholesterolemia and cardiovascular risk

687

Table 4 Odds ratios of subjects with hyper-HDL cholesterolemia vs those with normal blood HDL cholesterol levels for high body mass index, high waist-to-height ratio, hypertension, high triglycerides, high LDL cholesterol, high lipid accumulation product, hyperglycemia, or metabolic syndrome Variable Men High BMI High WHtR Hypertension High TG High LDL-C High LAP Hyperglycemia MetS Women High BMI High WHtR Hypertension High TG High LDL-C High LAP Hyperglycemia MetS

vs control 1 (univariate)

vs control 2 (univariate)

vs control 1 (multivariate)

vs control 2 (multivariate)

0.06 0.13 0.96 0.15 0.06 0.05 0.47 0.14

(0.03–0.13)** (0.08–0.21)** (0.71–1.29) (0.10–0.25)** (0.02–0.14)** (0.03–0.11)** (0.30–0.75)** (0.07–0.28)**

0.09 0.13 0.92 0.16 0.06 0.04 0.48 0.11

(0.05–0.18)** (0.08–0.20)** (0.68–1.24) (0.11–0.26)** (0.03–0.16)** (0.02–0.09)** (0.31–0.76)** (0.05–0.24)**

0.07 0.12 1.12 0.17 0.09 0.06 0.75 0.21

(0.04–0.13)** (0.08–0.18)** (0.78–1.59) (0.11–0.27)** (0.04–0.21)** (0.03–0.12)** (0.44–1.28) (0.09–0.47)**

0.08 0.13 1.30 0.17 0.08 0.06 0.67 0.19

(0.04–0.15)** (0.09–0.19)** (0.93–1.81) (0.11–0.27)** (0.04–0.20)** (0.03–0.13)** (0.40–1.12) (0.09–0.41)**

0.12 0.26 1.00 0.25 0.26 0.13 0.85 0.25

(0.06–0.23)** (0.19–0.38)** (0.68–1.47) (0.13–0.47)** (0.16–0.43)** (0.08–0.21)** (0.49–1.49) (0.09–0.67)**

0.11 0.23 0.77 0.25 0.25 0.14 0.68 0.19

(0.05–0.21)** (0.16–0.33)** (0.54–1.11) (0.14–0.44)** (0.15–0.42)** (0.09–0.22)** (0.40–1.15) (0.07–0.50)**

0.11 0.25 1.19 0.30 0.41 0.18 1.39 0.36

(0.05–0.21)** (0.18–0.35)** (0.79–1.81) (0.15–0.61)** (0.24–0.68)** (0.11–0.28)** (0.73–2.63) (0.11–1.19)

0.11 0.24 1.04 0.32 0.35 0.19 0.83 0.31

(0.06–0.20)** (0.18–0.33)** (0.72–1.50) (0.16–0.62)** (0.22–0.56)** (0.12–0.29)** (0.48–1.44) (0.10–0.94)*

BMI, body mass index; HDL, high-density lipoprotein; LAP, lipid accumulation product; LDL-C, low-density lipoprotein cholesterol; MetS, metabolic syndrome; TG, triglycerides; WHtR, waist-to-height ratio. Odds ratios with their confidence intervals for each variable are shown. Control 1 and Control 2 indicate age-matched and age- and alcohol intake– matched control groups with normal blood HDL cholesterol levels, respectively. In multivariate analyses, smoking, alcohol drinking, and regular exercise were used as explanatory variables. In the analysis for the variables except for high body mass index and high waist-to-height ratio, body mass index was also used as an explanatory variable. Asterisks denote significant differences from the reference level of 1.00 (*P , .05; **P , .01).

definition of the Japanese Committee for the Diagnostic Criteria of Metabolic Syndrome.7 Cutoff values of BMI and waist-to-height ratio (WHtR) used were 25 kg/m2 and 0.5, respectively.8,9 Blood pressure was measured by trained nurses, who were part of the local health checkup company, with a mercury sphygmomanometer once on the day of the health checkup after each subject had rested quietly for at least 5 minutes in a sitting position. Korotkoff phase V was used to define diastolic pressure. Hypertension was defined as systolic blood pressure of $140 mm Hg and/ or diastolic blood pressure of $90 mmHg. Subjects receiving drug therapy for hypertension were also included in the hypertensive group. Fasted blood was sampled from each subject in the morning, and serum triglyceride (TG), HDL cholesterol, and low-density lipoprotein (LDL) cholesterol levels were measured by enzymatic methods using commercial kits, pureauto S TG-N, cholestest N-HDL, and cholestest LDL (Sekisui Medical Co, Ltd, Tokyo, Japan), respectively. The coefficients of variation for the reproducibility of measurement were #3% for TGs, #5% for HDL cholesterol, and #5% for LDL cholesterol. Lipid accumulation product (LAP) was determined by using TG level and WC as follows: LAP 5 TG (mmol/L) ! (WC [cm] 2 65) for men and LAP 5 TG (mmol/L) ! (WC [cm] 2 58) for women.10 The cutoff values of high LAP were defined as 37.2 for men and 21.1 for women.11 Hemoglobin A1c was measured by the National Glycohemoglobin Standardization Program

(NGSP)-approved technique using the latex cohesion method with a commercial kit (Determiner HbA1c; Kyowa Medex, Tokyo, Japan). The coefficient of variation for reproducibility of hemoglobin A1c measurement was #5%. Because the standards of hemoglobin A1c used for measurement are different in the NGSP method and the Japan Diabetes Society (JDS) method, hemoglobin A1c values were calibrated by using a formula proposed by the JDS12: hemoglobin A1c (NGSP; %) 5 1.02 ! hemoglobin A1c (JDS; %) 1 0.25%. Subjects with hyperglycemia and diabetes were defined as those showing hemoglobin A1c levels of $5.7% and $6.5%, respectively, according to the criteria for diagnosis of diabetes by the American Diabetes Association13 and/or having a current history of drug therapy for diabetes. The category of subjects with diabetes was not further analyzed in this study because the proportion of the subjects with diabetes by the previously mentioned definition was too small for statistical analysis in the subjects of this study.

Criteria of metabolic syndrome Metabolic syndrome was defined, according to the criteria by the International Diabetes Federation14 with a slight modification, as the presence of 2 or more risk factors in addition to visceral obesity diagnosed as high WHtR. Risk factors included in the criteria are visceral obesity, high blood pressure, dyslipidemia (high TGs and/or low HDL cholesterol), and diabetes. The criterion for each

688 risk factor was defined as follows: visceral obesity, WHtR $0.5; high blood pressure, systolic blood pressure $130 mm Hg and/or diastolic blood pressure $85 mm Hg; low HDL cholesterol, HDL cholesterol ,40 mg/dL; high TGs, TGs $150 mg/dL; and diabetes, hemoglobin A1c, $6.5%. Subjects receiving drug therapy for hypertension and diabetes were also included in the previously mentioned definitions of high blood pressure and diabetes, respectively. All the subjects showed normal HDL cholesterol levels ($40 mg/dL).

Statistical analysis Continuous and categorical variables are summarized as means with standard deviations and frequencies (percentages), respectively. Continuous variables were compared between the matched high and normal HDL cholesterol groups using Student paired t test. Because TG levels did not show a normal distribution, they were used after log transformation. LAP levels also did not show a normal distribution. However, LAP level was analyzed only as a categorical variable because there were some negative values of LAP, which were unable to be log-transformed. Categorical variables were compared using the Mantel–Haenszel test with each matched pair as a separate stratum. Categorized cardiometabolic variables were also compared using the generalized estimating equation approach with a working unstructured correlation matrix for the matched pair, adjusting for histories of smoking, regular exercise, alcohol drinking, and BMI. The adjustment did not include history of alcohol drinking for comparisons between the 2 groups matched with respect to degree of alcohol consumption and age and did not include BMI for comparisons of BMI and WHtR between the groups. The results are shown as crude and adjusted odds ratios with their 95% confidence intervals and P values, which were estimated using the Mantel–Haenszel procedure as a univariate analysis and using the generalized estimating equation approach as a multivariate analysis, respectively. All P values are 2-sided, and values of P less than .05 were considered to indicate statistical significance. Statistical analyses were performed with the use of SPSS software, version 22 (IBM) and R software, version 3.1.1 (www.r-project.org).

Results Characteristics of the subject groups Table 1 shows characteristics of the hyper-HDL cholesterolemic group and the age-matched and age- and alcohol intake–matched control groups with normal HDL cholesterol levels. Both in men and women, the percentages of regular drinkers and nondrinkers were significantly higher and lower, respectively, in the hyper-HDL cholesterolemic group than in the age-matched control group. In women, the percentage of smokers was significantly lower in the

Journal of Clinical Lipidology, Vol 9, No 5, October 2015 hyper-HDL cholesterolemic group than in the age- and alcohol intake–matched control group but was not significantly different in the hyper-HDL cholesterolemic and age alone–matched control group. In men, there was no significant difference in the percentage of smokers between the hyper-HDL cholesterolemic group and the age- or age- and alcohol intake–matched control group. Both in men and women, the percentages of subjects having a habit of regular exercise and receiving therapy for hypertension or diabetes were not significantly different in the hyper-HDL cholesterolemic group and the corresponding control group matched for age or age and alcohol intake.

Comparison of means of each variable related to cardiovascular risk between the hyper- and normo-HDL cholesterolemic groups Means of each variable related to cardiovascular risk were compared in the groups of subjects with hyper- and normo-HDL cholesterolemia (Table 2). Both in the analyses using age-matched and age- and alcohol-matched control groups, BMI, WHtR, log-transformed TGs, LDL cholesterol, and hemoglobin A1c in men and women were significantly lower in the hyper-HDL cholesterolemic group than in the control groups. On the other hand, systolic and diastolic blood pressure levels in men and women were not significantly different in the hyper- and normo-HDL cholesterolemic groups.

Comparison of prevalences of each variable related to cardiovascular risk between the hyper- and normo-HDL cholesterolemic groups Prevalences of each cardiovascular risk factor were compared in the hyper- and normo-HDL cholesterolemic groups (Table 3). Both in men and women, the percentages of subjects with high BMI, high WHtR, high TGs, high LDL cholesterol, high LAP, and metabolic syndrome were markedly lower in the hyper-HDL cholesterolemic group than in the control groups matched for age and for age and alcohol. In men, the percentage of subjects with hyperglycemia was also significantly lower in the hyperHDL cholesterolemic group than in the control groups matched for age and for age and alcohol, whereas the prevalence of hyperglycemia was not significantly different in the hyper- and normo-HDL cholesterolemic groups of women. Both in men and women, the percentages of subjects with hypertension were not significantly different in the hyper-HDL cholesterolemic group and the control groups matched for age and for age and alcohol.

Odds ratios of the hyper- vs normo-HDL cholesterolemic groups for each cardiovascular risk factor Crude and adjusted odds ratios of the hyper- vs normoHDL cholesterolemic groups for each cardiovascular risk

Wakabayashi and Daimon

Hyper-HDL cholesterolemia and cardiovascular risk

factor are shown in Table 4. In multivariate analysis, age and histories of smoking, alcohol drinking, and regular exercise were adjusted. In some analyses, BMI was also used as an explanatory variable. Both in the analyses using agematched and age- and alcohol-matched controls and both in univariate and multivariate analyses, odds ratios of hyper- vs normo-HDL cholesterolemic groups for high BMI, high WHtR, high TGs, high LDL cholesterol, high LAP, and metabolic syndrome were significantly lower than the reference level of 1.00 except for the odds ratio for metabolic syndrome in multivariate analysis using the age alone-matched control group in women. On the other hand, in all these analyses, odds ratios for hypertension were not significantly different from the reference level. The odds ratio for hyperglycemia was significantly lower than the reference level in univariate analysis using male subjects but not significantly different from the reference level in multivariate analysis using male subjects and in the univariate and multivariate analyses using female subjects.

Discussion The mean levels of each variable of cardiovascular risk factors, including obesity, dyslipidemia and metabolic syndrome, and their prevalences were markedly lower in hyper-HDL cholesterolemic men and women than in their corresponding control groups. The percentage of drinkers was higher in the hyper-HDL cholesterolemic group than in the age-matched control group (Table 1), and light-to-moderate alcohol intake has been shown to be associated with lower risks of obesity, dyslipidemia, diabetes, and metabolic syndrome.15 However, the previously mentioned results for the differences between the hyper- and normo-HDL cholesterolemic groups were found in both analyses using the age-matched control group and the age- and alcoholmatched control group. Thus, individuals with hyper-HDL cholesterolemia have more beneficial cardiometabolic profiles than do those with normo-HDL cholesterolemia independently of alcohol consumption. This is, to the best of our knowledge, the first study showing alcohol-independent relationships between hyper-HDL cholesterolemia and cardiovascular risk factors. In addition to alcohol drinking, smoking and exercise are known to influence blood HDL cholesterol level, which is lower in smokers than in nonsmokers and is higher in individuals with a habit of regular exercise than in those without the habit.16 In the present study, the percentages of smokers and subjects with a habit of regular exercise were not significantly different in the hyper-HDL cholesterolemic group and the age-matched control groups, and in addition, inverse associations of hyper-HDL cholesterolemia with obesity, dyslipidemia, and metabolic syndrome were found in the analyses with adjustment for habits of smoking and regular exercise. Therefore, the beneficial cardiometabolic profile of the hyper-HDL cholesterolemic subjects is thought to be also independent of histories of smoking and regular exercise.

689

Metabolic syndrome is a cluster of cardiometabolic abnormalities including abdominal obesity, hypertension, dyslipidemia, and glucose intolerance.14 LAP is calculated by using WC and TGs10 and has been shown to be associated with risk of cardiovascular events.17 Because abdominal obesity and hypertriglyceridemia were inversely associated with hyper-HDL cholesterolemia, it is reasonable that prevalences of metabolic syndrome and high LAP were lower in the hyper-HDL cholesterolemic subjects than in the normo-HDL cholesterolemic subjects. There is an integral linking in lipoprotein metabolism, and HDL cholesterol is inversely correlated with TGs.18 Therefore, it is reasonable that TG level and prevalence of hypertriglyceridemia were considerably lower in the subjects with hyper-HDL cholesterolemia than in those with normo-HDL cholesterolemia. More than half of Japanese with hyper-HDL cholesterolemia have been reported to have mutations in the CETP gene,2 although results of CETP activity were not available in this study. CETP has been shown to be a determinant of serum LDL cholesterol19; and thus, CETP deficiency may partly explain the lower LDL cholesterol levels in the hyper-HDL cholesterolemic group than in the normo-HDL cholesterolemic group. A causal relationship has been shown between obesity and low HDL cholesterol: impaired cholesterol efflux to HDL from adipocytes and downregulation of apolipoprotein A1 and ATP-binding cassette transporter A1 in the liver by alterations in adipokines, including increases in tumor necrosis factor a and interleukin-1b and a decrease in adiponectin, have been proposed as mechanisms for obesity-induced decrease in HDL cholesterol.20 In fact, there were marked differences in the level of adiposity and prevalence of obesity between the hyper- and normoHDL cholesterolemic groups in the present study. Obesity is known to be a risk factor for diabetes, hypertension, and dyslipidemia. Thus, obesity is a central risk factor for these cardiovascular risk factors and metabolic syndrome. Then, the relationships of hyper-HDL cholesterolemia with other blood lipids and metabolic syndrome were investigated by generalized linear estimating equation, with adjustment for BMI. As shown in Table 4, odds ratios of the hyper- vs normo-HDL cholesterolemic groups for hypertriglyceridemia, hyper-LDL cholesterolemia, high LAP, and metabolic syndrome were significantly lower than the reference level of 1.00 after adjustment for BMI. Therefore, the inverse associations between these cardiometabolic risks and hyper-HDL cholesterolemia are thought to be independent of obesity. Obesity, which is inversely associated with HDL cholesterol level, is known to be an important risk factor for hypertension. In addition, HDL has been shown to increase NO synthase activity in the vascular endothelium, resulting in greater production of NO, a potent vasodilator.21 However, in the present study, blood pressure levels and prevalence of hypertension were not significantly different in the hyper- and normo-HDL cholesterolemic groups. Interestingly, a previous study, in which data for

690 Japanese hyper- and hypo-HDL cholesterolemic subjects were analyzed, also showed that blood pressure was not significantly different between the subject groups.22 Although the reason for no relationship between blood pressure and HDL cholesterol remains to be clarified, there is a possibility of the existence of a factor(s) elevating blood pressure in individuals with hyper-HDL cholesterolemia. Both in the analyses using age-matched and age- and alcohol-matched control subjects, hyper-HDL cholesterolemia showed no associations with blood pressure and hypertension. Therefore, habitual alcohol drinking, which is known to be a risk factor for hypertension23,24 and was more prevalent in hyper-HDL cholesterolemic subjects than in age-matched normo-HDL cholesterolemic subjects in the present study (Table 1), is unlikely to explain cancelation of the previously mentioned potential difference in blood pressure of hyper- and normo-HDL cholesterolemic subjects. Obesity is also a major risk factor for glucose intolerance. Hyper-HDL cholesterolemia showed a significant association with hyperglycemia in the univariate analysis for men but not in that for women and not in the multivariate analyses for men and women (Tables 3 and 4). A recent study has demonstrated that prevalence of diabetes was lower in individuals with high HDL cholesterol levels (greater than 90th percentile) than in those with low HDL cholesterol levels (less than 10th percentile).25 However, the relationship between hyper-HDL cholesterolemia and diabetes could not be determined because the number of subjects with diabetes was too small in the subjects in the present study (Table 1). Thus, further studies are needed to clarify whether the risk of diabetes is lowered in individuals with hyper-HDL cholesterolemia. This study has some limitations. Blood lipid levels are influenced by daily diet, for which information was not available. Nine categories of drinkers defined by frequency and amount of drinking were used to match alcohol consumption of the hyper-HDL cholesterolemic and control groups. Thus, a completely alcohol-matched control group could not be prepared, and there is a possibility that the mean amounts of alcohol consumption in the same categories of drinkers were still larger in the hyper-HDL cholesterolemic group than in the age- and alcohol-matched control group. In addition, information on polymorphisms of alcoholmetabolizing enzymes, including acetaldehyde dehydrogenase 2, which are known to influence effects of alcohol drinking, especially in Asians,26,27 was not included in the database used. Information on mutations of the CETP gene was also unavailable in this study. Premenopausal women generally have lower levels of LDL cholesterol and higher levels of HDL cholesterol than those in age-matched men and a correspondingly lower risk of coronary artery disease,28 although cardiovascular risk is increased among certain racial and/or ethnic groups and in women with polycystic ovary syndrome.29 Menopause is often believed to increase the risk of atherosclerotic cardiovascular disease. However, women have an atherosclerotic cardiovascular

Journal of Clinical Lipidology, Vol 9, No 5, October 2015 disease risk that increases linearly with aging. Menopause may be a surrogate for age, and women often gain body fat and experience worsening dyslipidemia during and after the menopause transition, similar to aging men.29 Nonetheless, no available information on menopause and sex hormone use was an important limitation of the present study. All the subjects of this study were of Japanese origin, and there is a possibility of racial and/or ethnic differences in the relationships between hyper-HDL cholesterolemia and cardiometabolic risk factors and in modification of the relationships by alcohol. Therefore, the findings of this study may not necessarily be extended to all populations. Finally, this study was a case–control study, and further prospective studies are needed to elucidate the causal relationship between hyper-HDL cholesterolemia and cardiometabolic risk.

Conclusion Individuals with hyper-HDL cholesterolemia tend to have beneficial cardiometabolic profiles, including lower prevalences of obesity, dyslipidemia, and metabolic syndrome, compared with individuals with normal HDL cholesterol levels. The relationships of hyper-HDL cholesterolemia with lower cardiometabolic risk are thought to be independent of habitual alcohol drinking.

Acknowledgment This work was supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (no. 24390171).

References 1. Investigating Committee of Guideline for Diagnosis and Treatment of Hyperlipidemias, Japan Atherosclerosis Society. Guideline for diagnosis and treatment of hyperlipidemias in adults. J Jpn Atheroscler Soc. 1997;25:1–34 (in Japanese). 2. Maruyama T, Sakai N, Ishigami M, et al. Prevalence and phenotypic spectrum of cholesteryl ester transfer protein gene mutations in Japanese hyperalphalipoproteinemia. Atherosclerosis. 2003;166:177–185. 3. Corrao G, Rubbiati L, Bagnardim V, Zambon A, Poikolainen K. Alcohol and coronary heart disease: a meta-analysis. Addiction. 2000;95:1505–1523. 4. Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ. 1999;319:1523–1528. 5. Mancia G, De Backer G, Dominiczak A, et al. 2007 Guidelines for the Management of Arterial Hypertension: the Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2007;25:1105–1187. 6. Ogihara T, Kikuchi K, Matsuoka H, et al, Japanese Society of Hypertension Committee. The Japanese Society of Hypertension Guidelines for the Management of Hypertension (JSH 2009). Hypertens Res. 2009;32:3–107. 7. Anonymous: Metabolic syndrome-definition and diagnostic criteria in Japan. J Jpn Soc Int Med. 2005;94:794–809 (in Japanese).

Wakabayashi and Daimon

Hyper-HDL cholesterolemia and cardiovascular risk

8. Anonymous: Examination Committee of Criteria for ’Obesity Disease’ in Japan; Japan Society for the Study of Obesity. New criteria for ’obesity disease’ in Japan. Circ J. 2002;66:987–992. 9. Hsieh SD, Muto T. Metabolic syndrome in Japanese men and women with special reference to the anthropometric criteria for the assessment of obesity: proposal to use the waist-to-height ratio. Prev Med. 2006; 42:135–139. 10. Kahn HS. The ‘‘lipid accumulation product’’ performs better than the body mass index for recognizing cardiovascular risk: a populationbased comparison. BMC Cardiovasc Disord. 2005;5:26. 11. Wakabayashi I, Daimon T. A strong association between lipid accumulation product and diabetes mellitus in Japanese women and men. J Atheroscler Thromb. 2014;21:282–288. 12. Kashiwagi A, Kasuga M, Araki E, et al, Committee on the Standardization of Diabetes Mellitus-Related Laboratory Testing of Japan Diabetes Society (JDS). International clinical harmonization of glycated hemoglobin in Japan: from Japan Diabetes Society to National Glycohemoglobin Standardization Program values. Diabetol Int. 2012;3:8–10. 13. Anonymous; American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2010;33(Suppl 1):S62–S69. 14. Alberti KG, Zimmet P, Shaw J, IDF Epidemiology Task Force Consensus Group. The metabolic syndrome–a new worldwide definition. Lancet. 2005;366:1059–1062. 15. Wakabayashi I. Alcohol. In: Wakabayashi I, Groschner K, editors. Interdisciplinary Concepts in Cardiovascular Health., Volume I: Primary risk factors. Wien: Springer-Verlag, 2013. p. 145–166. 16. Link JJ, Rohatgi A, de Lemos JA. HDL cholesterol: physiology, pathophysiology, and management. Curr Probl Cardiol. 2007;32:268–314. 17. Wehr E, Pilz S, Boehm BO, M€arz W, Obermayer-Pietsch B. The lipid accumulation product is associated with increased mortality in normal weight postmenopausal women. Obesity (Silver Spring). 2011;19: 1873–1880. 18. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in

19.

20.

21. 22.

23.

24.

25.

26.

27.

28. 29.

691

Adults (Adult Treatment Panel III). Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002;106:3143–3421. Kinoshita M, Teramoto T, Shimazu N, et al. CETP is a determinant of serum LDL-cholesterol but not HDL-cholesterol in healthy Japanese. Atherosclerosis. 1996;120:75–82. Wang H, Peng D-Q. New insights into the mechanism of low high-density lipoprotein cholesterol in obesity. Lipids Health Dis. 2011;10:176. http://dx.doi.org/10.1186/1476-511X-10-176. Mineo C, Shaul PW. Novel biological functions of high-density lipoprotein cholesterol. Circ Res. 2012;111:1079–1090. Hata Y, Yamamoto M, Oikawa T, et al. [Clinical significance of hyper-HDL-cholesterolemia. II. Characteristics in clinical features]. Nihon Ronen Igakkai Zasshi. 1986;23:413–421 (in Japanese). Klatsky AL. Alcohol and cardiovascular disease–more than one paradox to consider. Alcohol and hypertension: does it matter? Yes. J Cardiovasc Risk. 2003;10:21–24. Taylor B, Irving HM, Baliunas D, et al. Alcohol and hypertension: gender differences in dose-response relationships determined through systematic review and meta-analysis. Addiction. 2009;104:1981–1990. Juren AJ, Sarwal G, Al-Sarraf A, et al. Low prevalence of type 2 diabetes mellitus among patients with high levels of high-density lipoprotein cholesterol. J Clin Lipidol. 2013;7:194–198. Mizoi Y, Tatsuno Y, Adachi J, et al. Alcohol sensitivity related to polymorphism of alcohol-metabolizing enzymes in Japanese. Pharmacol Biochem Behav. 1983;18(Suppl 1):127–133. Suddendorf RF. Research on alcohol metabolism among Asians and its implications for understanding causes of alcoholism. Public Health Rep. 1989;104:615–620. Collins P. HDL-C in post-menopausal women: an important therapeutic target. Int J Cardiol. 2008;124:275–282. Bays HE, Jones PH, Brown WV, Jacobson TA. National Lipid Association Annual Summary of Clinical Lipidology 2015. J Clin Lipidol. 2014;8(6 Suppl):S1–S36.