- Email: [email protected]
Influence of mild-to-moderate alcohol consumption on cardiovascular diseases in men from the general population
Atherosclerosis 224 (2012) 222e227
Contents lists available at SciVerse ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Influence of mild-to-moderate alcohol consumption on cardiovascular diseases in men from the general population Shinji Makita a, *, Toshiyuki Onoda b, Masaki Ohsawa b, Fumitaka Tanaka a, Toshie Segawa a, Tomohiro Takahashi a, Kenyu Satoh a, Kazuyoshi Itai b, Kozo Tanno b, Kiyomi Sakata b, Shinichi Omama c, Yuki Yoshida c, Yasuhiro Ishibashi d, Tomiko Koyama e, Toru Kuribayashi f, Kuniaki Ogasawara c, Akira Ogawa c, Akira Okayama g, Motoyuki Nakamura a a
Division of Cardiology, Department of Internal Medicine, Iwate Medical University, 19-1 Uchimaru, Morioka 020-8505, Japan Department of Hygiene and Preventive Medicine, Iwate Medical University, Morioka, Japan c Department of Neurosurgery, Iwate Medical University, Morioka, Japan d Department of Neurology, Iwate Medical University, Morioka, Japan e Iwate Health Service Association, Morioka, Japan f Department of Health and Physical Education, Faculty of Education, Iwate University, Morioka, Japan g First Institute of Health Service, Japan Anti-Tuberculosis Association, Tokyo, Japan b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 January 2012 Received in revised form 29 May 2012 Accepted 5 July 2012 Available online 20 July 2012
Background and methods: There is controversy about the association between mild-to-moderate alcohol consumption and a reduced risk of cardiovascular diseases. The relationships between daily alcohol consumption and the incidence of acute myocardial infarction (MI) or ischemic stroke (IS) were examined in men in a community-based, prospective cohort study (n ¼ 8014, age 40e80 years, mean age ¼ 64.1 years). Alcohol consumption was categorized into 3 groups (A1, none or occasional; A2, 25 g/ day; A3, >25 g/day as ethanol) at baseline. Results: During the mean follow-up of 5.5 years, 53 MIs and 186 ISs occurred. On Cox regression analysis adjusted for age, hypertension, diabetes, dyslipidemia, smoking index, and body mass index (BMI), the hazard ratio (HR) for incident MI was significantly lower in the A2 group than in the A1 group (HR ¼ 0.49, p ¼ 0.043). The HR for incident MI in the A3 group tended to be lower than in the A1 group (HR ¼ 0.53, p ¼ 0.10). In obese subjects, while a significantly lower HR for incident MI in the A2 group was retained (HR ¼ 0.29, p ¼ 0.049), no significant difference in the HR of the A3 group compared with the A1 group was found. No significant differences were found in the IS-free curve among the 3 groups of alcohol consumption. Conclusions: Alcohol consumption may have a protective effect on the onset of MI but not on IS in the general population. A U-shaped relation between alcohol consumption and incident MI was found in obese subjects. An appropriate limit for daily alcohol consumption, depending on the risk of ischemic heart disease, may need to be established. Ó 2012 Elsevier Ireland Ltd. All rights reserved.
Keywords: Epidemiology Alcohol Acute myocardial infarction Stroke
1. Introduction Moderate drinkers are considered to have less risk of cardiovascular diseases and death than nondrinkers. Furthermore, it has been shown that a U-shaped or J-shaped relationship exists between daily alcohol consumption and incident cardiovascular diseases, such as ischemic stroke [1e3] and coronary heart disease
* Corresponding author. Tel.: þ81 19 651 5111x2324; fax: þ81 19 651 7072. E-mail address: [email protected] (S. Makita). 0021-9150/$ e see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2012.07.004
[4e9]. According to the most recent and largest meta-analysis, moderate alcohol consumption was shown to have beneficial effects for cardiovascular disease mortality, coronary heart disease incidence and mortality, stroke incidence and mortality, and allcause mortality [10]. However, only a few validation studies investigating the effects of alcohol consumption on both incident coronary heart disease and stroke events have been done in a single cohort, and the difference in the effects of alcohol consumption on ischemic heart disease and ischemic stroke has been unclear. Although Iso et al. showed protective effects of light-to-moderate drinking on both
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
ischemic stroke and coronary heart disease in a validation study of the Japanese general population, significant differences were not seen [11]. Moreover, enrollment of this cohort was started from 1975 and the social background and intensities of various risk factor management strategies might have differed considerably from those of today. The present study investigated the relationships between daily alcohol consumption and incident myocardial infarction and ischemic stroke in a large, single cohort of a Japanese rural population. 2. Methods
223
(HDL-C) levels <40 mg/dl or in subjects with a history of cholesterol-lowering therapy. Diabetes mellitus was diagnosed in patients with a blood glucose level 200 mg/dl, HbA1c level (NGSP) 6.5%, or a history of antidiabetic therapy. Obese subjects consisted of overweight or obese individuals (defined as a body mass index (BMI) 25.0 kg/m2). The smoking index was defined as the product of packs per day and years of smoking (pack-years). Although an electrocardiogram was not done in 225 males (2.8%), these participants were considered to have no atrial fibrillation (Af) if they had no history of Af, and they were included in the present analysis. Daily alcohol consumption was categorized into three groups (A1, none or occasional; A2, 25 g/day, A3, >25 g/day as ethanol averaged from the amount in a week) at baseline.
2.1. Study subjects 2.4. Outcome measures The study subjects were recruited from the community-dwelling population living in districts of Iwate in northern Japan (the IwateKenpoku Cohort study). This study was conducted as part of a government-sponsored, multiphasic health check-up program aimed at the general population. Between April 2002 and January 2005, invitations to participate in this health check-up program were issued by government offices in 17 rural municipalities located in these districts; 26,469 individuals (9161 males) took part in the program and agreed to join the present study. After subjects aged over 80 years (280 males) and those under 40 years (300 males) or subjects without information about alcohol consumption (23 males), as well as those with a history of cardiovascular disease or stroke (526 males), were excluded, the data of 8059 males (mean age, 64.1 9.7 years) were included in the present study. The study was approved by our institutional ethics committee, and all of the participants provided their written, informed consent. 2.2. Blood sampling and baseline assessment Blood samples were collected from an antecubital vein. The samples were collected into vacuum tubes containing EDTA or a serum separator gel. After sampling, the tubes were stored immediately in an icebox and centrifuged at 1500 g for 10 min within 8 h of collection. Aliquots of serum were stored at 20 C, and routine hematology and biochemistry tests were done within a few days after blood sampling. The estimated glomerular filtration rate (eGFR) was calculated using the modified equation of the Modification of Diet in Renal Disease (MDRD) study [12]. Glycosylated hemoglobin (HbA1c) was measured by the HPLC method as a Japan Diabetes Society (JDS) value, and the result was presented as a National Glycohemoglobin Standardization Program (NGSP) equivalent value transformed by adding 0.4% uniformly. Baseline clinical examinations included a standard 12-lead electrocardiogram, and a self-reported questionnaire was administered to document subjects’ medical history and lifestyle. Systolic blood pressure (sBP) and diastolic blood pressure (dBP) were measured by trained nurses using an automatic digital sphygmomanometer placed on the upper arm at the height of the heart with the patient in a sitting position after having rested for a few minutes. The average of two measurements was used for the present analysis. 2.3. Risk factor definition Hypertension was defined as at least one of: systolic blood pressure 140 mmHg; a diastolic blood pressure 90 mmHg; or current antihypertensive therapy. Dyslipidemia was diagnosed in subjects with serum low-density lipoprotein cholesterol (LDL-C) levels 140 mg/dl and/or high-density lipoprotein cholesterol
In this cohort study, the outcome event was all-cause death, as well as any non-fatal cardiovascular events, such as myocardial infarction, cerebral infarction, or other strokes. The dates of death and moving away from the region were confirmed by the investigators reviewing population-register sheets at each local government office. Persons who were known to be alive at the end of follow-up and those who had moved away from the study area were treated as censored cases. Stroke events were identified by accessing the Iwate prefecture stroke registration program, which included the entire area where the subjects lived; details of this registry have been described previously [13]. Since 1991, the stroke registration program has been coordinated by the Iwate prefecture government and the Iwate Medical Association; the medical records of all medical facilities within the survey area are verified to ensure complete capture of all data. Incidents of acute myocardial infarction were identified by accessing data from the Northern Iwate Heart Disease Registry Consortium, which has been collecting data since 2002. The registration of acute myocardial infarction and sudden death was based on the criteria of the MONICA study [14]. To verify the accuracy of the data, a physician or trained research nurse visited and checked the medical records of the referral hospitals. 2.5. Statistical analysis Comparisons of baseline demographic data among the three categories of alcohol consumption levels were performed using analysis of variance (ANOVA) with Scheffe’s post hoc test. The values in the tables are expressed as means standard deviation. The cumulative survival curves (free of acute myocardial infarction or ischemic stroke) by the three categories of alcohol consumption and the relative risk of group A2 and A3 compared to the reference A1 group were determined using the multivariate Cox regression hazards model adjusted by age, presence of hypertension, diabetes, dyslipidemia, BMI, and smoking index. In the analysis concerning ischemic stroke, further adjustment by the presence of Af was performed. The results are expressed as hazard ratios (HRs) and the corresponding 95% confidence interval (CI). The data of adjusted factors such as height, weight, blood pressures, and smoking status were missing as random phenomena in 47 subjects, and thus, 8014 subjects having a complete data set were analyzed in the Cox model. The analyses were performed using the SPSS statistical package, version 19.0 (Chicago, IL, USA). 3. Results Overall baseline demographic data for the 3 categories of daily alcohol consumption are shown in Table 1. The A3 group was younger and had higher blood pressures, higher BMI, higher eGFR,
224
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
Table 1 Baseline demographic data for three categories of alcohol consumption (overall analysis). Alcohol consumption category （average alcohol consumption)
A1 (0.1 g/day)
A2 (17.4 g/day)
A3 (49.7 g/day)
p-value
Number of subjects Age, years Height, cm Weight, kg Body mass index, kg/m2 Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Serum creatinine, mg/dl eGFR, ml/min/1.73 m2 Uric acid, mg/dl Hemoglobin A1c, % Triglycerides, mg/dl LDL-cholesterol, mg/dl HDL-cholesterol, mg/dl Smoking index, pack-years High sensitivity CRP, mg/l Hypertension, n(%) Diabetes mellitus, n(%) Dyslipidemia, n(%) Atrial fibrillation, n(%)
3246 65.7 9.4 161.1 6.3 62.2 9.7 23.9 3.1 128 19
2441 64.9 9.4a 161.3 6.3 62.1 9.0 23.8 2.8 131 20a
2372 60.9 9.6a,b 162.8 6.4a,b 63.8 9.4a,b 24.0 2.9b 134 19a,b
e <0.001 <0.001 <0.001 <0.05 <0.001
76 11
79 11a
82 11a,b
<0.001
0.85 0.27 73.6 15.0 5.6 0.8 5.6 0.8 124 76 119 29 51 13 18 22
0.81 0.15a 76.2 14.7a 5.7 1.3a 5.5 0.7a 120 79 113 28a 57 15a 16 19a
0.79 0.14a,b 80.2 15.4a,b 6.0 1.4a,b 5.5 0.7a 133 97a,b 107 29a,b 62 17a,b 24 21a,b
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
1.47 4.90
1.42 5.15
1.25 3.83
1295 (39.9%) 328 (10.1%) 952 (29.3%) 77 (2.4%)
1151 (47.2%) 227 (9.3%) 440 (18.0%) 75 (3.1%)
1259 (53.1%) 194 (8.2%) 357 (15.0%) 69 (2.9%)
0.22 <0.001 <0.001 <0.001 0.23
eGFR; estimated glomerular filtration rate. a p < 0.05 vs A1. b p < 0.05 vs A2.
4. Discussion
higher uric acid, higher triglyceride, lower LDL-cholesterol, higher HDL-cholesterol, and a higher smoking index than the A2 group. The prevalence of hypertension was significantly higher with larger amounts of alcohol consumption. The prevalence of diabetes and dyslipidemia was significantly lower with larger amounts of alcohol consumption. A similar trend in the baseline data was obtained with respect to obese subjects, except for the prevalence of diabetes (Table 2). Table 2 Baseline demographic data for three categories of alcohol consumption (obese subjects). Alcohol consumption category （average alcohol consumption)
A1 (0.1 g/day)
A2 A3 (17.0 g/day) (48.9 g/day)
p-value
Number of subjects Age, years Height, cm Weight, kg Body mass index, kg/m2 Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Serum creatinine, mg/dl eGFR, ml/min/1.73 m2 Uric acid, mg/dl Hemoglobin A1c, % Triglycerides, mg/dl LDL-cholesterol, mg/dl HDL-cholesterol, mg/dl Smoking index, pack-years High sensitivity CRP, mg/l Hypertension, n(%) Diabetes mellitus, n(%) Dyslipidemia, n(%) Atrial fibrillation, n(%)
1155 64.3 9.5 161.5 6.2 71.0 9.7 27.2 1.9 132 19
789 64.3 9.5 161.8 6.0 70.7 7.2 27.0 1.8a 135 19a
824 60.9 9.2a,b 162.8 6.4a,b 72.1 7.8a,b 27.1 1.9 137 18a,b
e <0.001 <0.001 <0.001 0.05 <0.001
79 11
81 10a
84 10a,b
<0.001
0.86 0.25 72.8 14.6 5.9 1.3 5.7 0.8 151 93 124 29 49 11 18 23 1.68 5.96 574 (49.7%) 147 (12.7%) 467 (40.4%) 31 (2.7%)
0.83 0.16a 74.8 14.5a 6.0 1.3 5.7 0.8 141 84 119 26a 52 12a 16 19 1.71 5.96 434 (55.0%) 92 (11.7%) 187 (23.7%) 23 (2.9%)
0.81 0.14a 77.6 15.0a,b 6.2 1.5a,b 5.6 0.7a 157 113a,b 112 29a,b 57 15a,b 22 22a,b 1.36 4.28 523 (63.5%) 83 (10.1%) 171 (20.8%) 34 (4.1%)
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.30 <0.001 0.19 <0.001 0.17
eGFR; estimated glomerular filtration rate. a p < 0.05 vs A1. b p < 0.05 vs A2.
The mean follow-up period was 5.5 years. During follow-up, 53 subjects (cumulative incidence rate 0.66%) had a first, non-fatal myocardial infarction (MI) and 186 subjects (cumulative incidence rate 2.32%) had a first ischemic stroke (IS). MI-free curves are shown in Fig. 1. The HR for incident MI was significantly lower in the A2 group than in the A1 group when adjusted for age, hypertension, diabetes, dyslipidemia, smoking index, and BMI (HR ¼ 0.49, 95%CI 0.24e0.98, p ¼ 0.043). The HR for incident MI tended to be lower in the A3 group than in the A1 group when adjusted for the same factors, but it was not significant (HR ¼ 0.53, 95%CI 0.25e1.12, p ¼ 0.10). On the other hand, within the obese subjects, while a significantly lower HR for incident MI was retained in the A2 group (HR ¼ 0.29, 95%CI 0.08e0.99, p ¼ 0.049), no significant difference in the HR between the A3 group and the A1 group was found (HR ¼ 0.61, 95%CI 0.22e1.64, p ¼ 0.32). The IS-free curves are shown in Fig. 2. No significant differences were found in the IS-free curves among the 3 categories of alcohol consumption when adjusted for age, hypertension, diabetes, dyslipidemia, smoking index, Af, and BMI. Similar results were observed in obese patients. The shapes of the relationships between alcohol consumption category and incident MI or IS in overall subjects and obese subjects are shown in Fig. 3. For incident MI, while an L-shaped relationship with alcohol consumption was found in the analysis of overall subjects, a nearly U-shaped relationship was observed in obese subjects.
The present cohort study indicated that the risk for incident MI was lower in subjects who drank alcohol than in those who did not. This appeared to be particularly true for the subjects with mild-tomoderate alcohol consumption. In obese subjects, while there was a reduced risk for incident MI in the mild-to-moderate drinkers, there was no significant difference in the risk for incident MI between heavy drinkers and those who did not drink. There was no effect of alcohol consumption on incident IS, regardless of the amount of alcohol consumption. Given the baseline data in the 3 alcohol categories, the favorable effects of alcohol intake appeared to be linked with elevated eGFR, decreased HbA1c, decreased LDL-C, and increased HDL-C level in the present study. The prevalence rates of dyslipidemia and diabetes were also lower in the subjects who drank alcohol. Although these factors seemed to contribute to the protective effect on MI, the causality between alcohol consumption and these risk factors was obscure because there were many opportunities for instructional intervention about alcohol avoidance prior to the baseline investigation in the subjects with renal dysfunction, diabetes, and dyslipidemia. Regarding HDL-C levels, an acute increase effect of alcohol intake was shown in a controlled study [15]. However, when adjustment for HDL-C was performed in the present model, the relationship between alcohol consumption and the risk of incident MI was essentially unaltered, which suggests that other cardioprotective mechanisms may be involved. In fact, although HDL-C is a crucial mediator of the atheroprotective properties of alcohol, this protective effect has been established with light-tomoderate drinking only. No previous evidence that the behavior of HDL-C level contributes to the U-shaped relationships between alcohol intake and cardiovascular events has been presented in the literature. Previous reports suggested that risk reduction in ischemic heart disease was attributed to the beneficial impact of alcohol on hemostatic factors related to a thrombolytic profile [16], insulin sensitivity [17], inflammation reflected by the C-reactive protein level [18], and fibrinogen concentrations [19].
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
225
Fig. 1. Associations between incident acute myocardial infarction and daily alcohol consumption. The data are expressed as hazard ratios (HRs) (95% confidence interval). The data in the overall analysis are adjusted by age, hypertension, diabetes, dyslipidemia, smoking index, and BMI. In obese subjects, BMI is excluded from these adjusted factors.
The present study showed an L-shaped relationship between alcohol consumption and incident MI in the analysis of overall subjects, because there was no increase in MI risk associated with higher alcohol consumption. This finding is consistent with some results from a meta-analysis performed by Maclure et al. [20] and Ronksley et al. [10]. On the other hand, Corrao et al. reported a Jshaped relationship between alcohol intake and coronary heart disease in a meta-analysis [21]. Gigleux et al. showed that light-tomoderate consumption of alcohol (>15 g/d) had significant cardioprotective effects, which were more apparent in men with metabolic syndrome [22]. However, since information about heavy drinkers was lacking, the U- or J-shaped relationship between alcohol consumption and IHD risk was not demonstrated in their study results. In the present study, a U-shaped relationship was observed in obese subjects, indicating that a harmful effect of a large amount of alcohol consumption may appear in subjects with a high risk for IHD. With respect to the mechanism, other detrimental effects observed only in the A3 group of obese subject should be considered. However, within the baseline characteristics that were investigated in the present study, the inter-group differences in obese subjects were very similar to those of overall subjects in all clinical indices. Therefore, the decreased beneficial effect of heavy drinking in obese subjects cannot be explained by increased blood pressure, increased smoking index, or decreased lipid profile improvement. Hyperuricemia has been shown to increase all-cause mortality and incident cardiovascular disease [23]. Since uric acid (UA) increase was found in the A3 group of obese subjects, this effect can be hypothesized to increase MI only in heavy-drinking obese subjects. Moreover, in overall analysis, the increase in UA level for the A2 group was negligible. However, no consensus has yet been reached regarding whether UA is
a protective anti-oxidant factor or a risk factor. Other detrimental factors that show a U-shaped configuration by amount of alcohol consumption only in obese subjects need to be explored. Imhof et al. showed that light-to-moderate daily alcohol intake (up to 20 or 40 g/day) was associated with lower CRP and white blood counts compared to non-drinking and heavy-drinking, even after adjustment for various potential confounders [18]. Although such anti-inflammatory effects may contribute to the U-shaped relationship between alcohol consumption and ischemic heart disease, no significant difference in inflammation represented by the high sensitivity CRP level was found among categories of alcohol consumption in the obese subjects of the present study. A risk reduction in IS, as well as MI, should be also found if an antithrombotic effect [24] is the main contributor to the cardioprotective effect of alcohol consumption. However, favorable effects of alcohol consumption on IS were not documented in the present study. Increased blood pressure caused by alcohol consumption may be partly involved in this mechanism. In fact, the influence of increased systolic blood pressure on incident IS was larger compared with that on incident MI in the overall analysis of the present subjects (risk for MI: HR ¼ 1.008, p ¼ 0.22; risk for IS: HR ¼ 1.015, p < 0.001, by a 1-mmHg increase in sBP in men when adjusted by age). Alternatively, the high prevalence of hypertension at baseline in the present cohort might have had a sufficiently strong effect to eliminate the much weaker effect of alcohol consumption. Some limitations of the present study warrant consideration. Given the many previous investigations, the results or the shape of the relationships between alcohol consumption and cardiovascular risk can be altered by the choice of cutoff level for ‘light’ or ‘moderate’ alcohol consumption. Since the present study did not
Fig. 2. Associations between incident ischemic stroke and daily alcohol consumption. The data are expressed as hazard ratios (HRs) (95% confidence interval). The data in the overall analysis are adjusted by age, hypertension, diabetes, dyslipidemia, smoking index, atrial fibrillation, and BMI. In obese subjects, BMI is excluded from these adjusted factors.
226
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
Obese subjects
2.50
2.50
2.00
2.00
1.50
1 .50
Hazard ratio
Hazard ratio
Overall analysis
1.00
1.00
0 .50
0.50
0.10
0.10 A1
A2
A3
Category of alcohol consumption
A1
A2
A3
Category of alcohol consumption
Fig. 3. The shapes of the relationship between alcohol consumption and incident myocardial infarction or ischemic stroke. The graphs show the hazard ratios (circles) and 95% confidence interval (error bars) for incident myocardial infarction (solid lines) or ischemic stroke (dashed lines) in overall subjects and obese subjects. The vertical axis is the logscale hazard ratio.
have enough subjects or clinical incidents to examine the detailed effect of range setting in the alcohol consumption categories, the suitability of the threshold (25 g/day) in the categorization of alcohol consumption is unclear. However, Brien et al. showed in a meta-analysis that ‘moderate’ consumption of alcohol (up to two drinks or 30 g alcohol a day for men) has beneficial effects on a variety of biomarkers linked to the risk of coronary heart disease [25], and this level of alcohol consumption may be similar to the cutoff level used in the present study. Second, alcohol consumption was evaluated only once at baseline, and intake variation during follow-up [26] was not investigated. Drinking patterns were not taken into consideration. An inaccurate estimate daily alcohol consumption cannot be excluded due to self-reporting of the consumption in the present study. In addition, the present study did not investigate the kinds of beverages. However, beer, wine, and spirits are generally known to have similar effects on high-density lipoprotein cholesterol [16]. A possibility that physical activity may have confounded the results in the present analysis cannot be ruled out. Finally, since female subjects were excluded from the analysis due to their very small low alcohol consumption (group A1, 14,291; group A2, 1290; group A3, 153 of 15,734 total females), sex differences in the risk were unclear. In conclusion, alcohol consumption may have a protective effect on the onset of MI, but not on IS, in the general population. While an L-shaped relationship between alcohol consumption and incident MI was found in overall subjects, a U-shaped relationship was observed in obese subjects. An appropriate limit or a recommended amount of daily alcohol consumption, depending on the risk of ischemic heart disease, may need to be established. Disclosure The authors report no conflicts of interest. References [1] Iso H, Baba S, Mannami T, et al. JPHC study group. alcohol consumption and risk of stroke among middle-age men: the JPHC study cohort I. Stroke 2004; 35:1124e9. [2] Reynolds K, Lewis B, Nolen JD, Kinney GL, Sathya B, He J. Alcohol consumption and risk of stroke: a meta-analysis. J Am Med Assoc 2003;289:579e88.
[3] Berger K, Ajani UA, Kase CS, et al. Light-to-moderate alcohol consumption and risk of stroke among U.S. male physicians. N Engl J Med 1999;341: 1557e64. [4] Yano K, Rhoads GG, Kagan A. Coffee, alcohol and risk of coronary heart disease among Japanese men living in Hawaii. N Engl J Med 1977;297: 405e9. [5] Langer RD, Criqui MH, Reed DM. Lipoproteins and blood pressure as biological pathways for effect of moderate alcohol consumption on coronary heart disease. Circulation 1992;85:910e5. [6] Shaper AG, Wannamethee G, Walker M. Alcohol and coronary heart disease: a perspective from the British regional heart study. Int J Epidemiol 1994;23: 482e94. [7] Gordon T, Kannel WB. Drinking habits and cardiovascular disease: the Framingham study. Am Heart J 1985;110:331e4. [8] Miller GJ, Beckles GL, Maude GH, Carson DC. Alcohol consumption: protection against coronary heart disease and risks to health. Int J Epidemiol 1990;19: 923e30. [9] Kitamura A, Iso H, Sankai T, et al. Alcohol intake and premature coronary heart disease in urban Japanese men. Am J Epidemiol 1998;147:59e65. [10] Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ 2011;342:d671. [11] Iso H, Kitamura A, Shimamoto T, et al. Alcohol intake and the risk of cardiovascular disease in middle-aged Japanese men. Stroke 1995;26:767e73. [12] Imai E, Horio M, Nitta K, et al. Modification of the modification of diet in renal disease (MDRD) study equation for Japan. Am J Kidney Dis 2007;50: 927e37. [13] Omama S, Yoshida Y, Ogawa A, Onoda T, Okayama A. Differences in circadian variation of cerebral infarction, intracerebral haemorrhage and subarachnoid haemorrhage by situation at onset. J Neurol Neurosurg Psychiatr 2006;77: 1345e9. [14] Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, Arveiler D, Rajakangas AM, Pajak A. Myocardial infarction and coronary deaths in the world health organization MONICA project: registration procedures, event rates, and case fatality rates in 38 populations from 21 countries in four continents. Circulation 1994;90:583e612. [15] Sierksma A, Patel H, Ouchi N, et al. Effect of moderate alcohol consumption on adiponectin, tumor necrosis factor-alpha, and insulin sensitivity. Diabetes Care 2004;27:184e9. [16] 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;319:1523e8. [17] Kiechl S, Willeit J, Poewe W, et al. Insulin sensitivity and regular alcohol consumption: large, prospective, cross sectional population study (Bruneck study). BMJ 1996;313:1040e4. [18] Imhof A, Woodward M, Doering A, et al. Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille). Eur Heart J 2004;25: 2092e100. [19] Sierksma A, van der Gaag MS, Kluft C, Hendriks HF. Moderate alcohol consumption reduces plasma C-reactive protein and fibrinogen levels; a randomized, diet-controlled intervention study. Eur J Clin Nutr 2002;56: 1130e6.
S. Makita et al. / Atherosclerosis 224 (2012) 222e227 [20] Maclure M. Demonstration of deductive meta-analysis: ethanol intake and risk of myocardial infarction. Epidemiol Rev 1993;15:328e51. [21] Corrao G, Rubbiati L, Bagnardi V, Zambon A, Poikolainen K. Alcohol and coronary heart disease: a meta-analysis. Addiction 2000;95:1505e23. [22] Gigleux I, Gagnon J, St-Pierre A, et al. Moderate alcohol consumption is more cardioprotective in men with the metabolic syndrome. J Nutr 2006;136: 3027e32. [23] Verdecchia P, Schillaci G, Reboldi G, Santeusanio F, Porcellati C, Brunetti P. Relation between serum uric acid and risk of cardiovascular disease in essential hypertension. The PIUMA study. Hypertension 2000;36:1072e8.
227
[24] Rubin R, Rand ML. Alcohol and platelet function. Alcohol Clin Exp Res 1994; 18:105e10. [25] Brien SE, Ronksley PE, Turner BJ, Mukamal KJ, Ghali WA. Effect of alcohol consumption on biological markers associated with risk of coronary heart disease: systematic review and meta-analysis of interventional studies. BMJ 2011;342:d636. [26] Emberson JR, Shaper AG, Wannamethee SG, Morris RW, Whincup PH. Alcohol intake in middle age and risk of cardiovascular disease and mortality: accounting for intake variation over time. Am J Epidemiol 2005; 161:856e63.
Contents lists available at SciVerse ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Influence of mild-to-moderate alcohol consumption on cardiovascular diseases in men from the general population Shinji Makita a, *, Toshiyuki Onoda b, Masaki Ohsawa b, Fumitaka Tanaka a, Toshie Segawa a, Tomohiro Takahashi a, Kenyu Satoh a, Kazuyoshi Itai b, Kozo Tanno b, Kiyomi Sakata b, Shinichi Omama c, Yuki Yoshida c, Yasuhiro Ishibashi d, Tomiko Koyama e, Toru Kuribayashi f, Kuniaki Ogasawara c, Akira Ogawa c, Akira Okayama g, Motoyuki Nakamura a a
Division of Cardiology, Department of Internal Medicine, Iwate Medical University, 19-1 Uchimaru, Morioka 020-8505, Japan Department of Hygiene and Preventive Medicine, Iwate Medical University, Morioka, Japan c Department of Neurosurgery, Iwate Medical University, Morioka, Japan d Department of Neurology, Iwate Medical University, Morioka, Japan e Iwate Health Service Association, Morioka, Japan f Department of Health and Physical Education, Faculty of Education, Iwate University, Morioka, Japan g First Institute of Health Service, Japan Anti-Tuberculosis Association, Tokyo, Japan b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 January 2012 Received in revised form 29 May 2012 Accepted 5 July 2012 Available online 20 July 2012
Background and methods: There is controversy about the association between mild-to-moderate alcohol consumption and a reduced risk of cardiovascular diseases. The relationships between daily alcohol consumption and the incidence of acute myocardial infarction (MI) or ischemic stroke (IS) were examined in men in a community-based, prospective cohort study (n ¼ 8014, age 40e80 years, mean age ¼ 64.1 years). Alcohol consumption was categorized into 3 groups (A1, none or occasional; A2, 25 g/ day; A3, >25 g/day as ethanol) at baseline. Results: During the mean follow-up of 5.5 years, 53 MIs and 186 ISs occurred. On Cox regression analysis adjusted for age, hypertension, diabetes, dyslipidemia, smoking index, and body mass index (BMI), the hazard ratio (HR) for incident MI was significantly lower in the A2 group than in the A1 group (HR ¼ 0.49, p ¼ 0.043). The HR for incident MI in the A3 group tended to be lower than in the A1 group (HR ¼ 0.53, p ¼ 0.10). In obese subjects, while a significantly lower HR for incident MI in the A2 group was retained (HR ¼ 0.29, p ¼ 0.049), no significant difference in the HR of the A3 group compared with the A1 group was found. No significant differences were found in the IS-free curve among the 3 groups of alcohol consumption. Conclusions: Alcohol consumption may have a protective effect on the onset of MI but not on IS in the general population. A U-shaped relation between alcohol consumption and incident MI was found in obese subjects. An appropriate limit for daily alcohol consumption, depending on the risk of ischemic heart disease, may need to be established. Ó 2012 Elsevier Ireland Ltd. All rights reserved.
Keywords: Epidemiology Alcohol Acute myocardial infarction Stroke
1. Introduction Moderate drinkers are considered to have less risk of cardiovascular diseases and death than nondrinkers. Furthermore, it has been shown that a U-shaped or J-shaped relationship exists between daily alcohol consumption and incident cardiovascular diseases, such as ischemic stroke [1e3] and coronary heart disease
* Corresponding author. Tel.: þ81 19 651 5111x2324; fax: þ81 19 651 7072. E-mail address: [email protected] (S. Makita). 0021-9150/$ e see front matter Ó 2012 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2012.07.004
[4e9]. According to the most recent and largest meta-analysis, moderate alcohol consumption was shown to have beneficial effects for cardiovascular disease mortality, coronary heart disease incidence and mortality, stroke incidence and mortality, and allcause mortality [10]. However, only a few validation studies investigating the effects of alcohol consumption on both incident coronary heart disease and stroke events have been done in a single cohort, and the difference in the effects of alcohol consumption on ischemic heart disease and ischemic stroke has been unclear. Although Iso et al. showed protective effects of light-to-moderate drinking on both
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
ischemic stroke and coronary heart disease in a validation study of the Japanese general population, significant differences were not seen [11]. Moreover, enrollment of this cohort was started from 1975 and the social background and intensities of various risk factor management strategies might have differed considerably from those of today. The present study investigated the relationships between daily alcohol consumption and incident myocardial infarction and ischemic stroke in a large, single cohort of a Japanese rural population. 2. Methods
223
(HDL-C) levels <40 mg/dl or in subjects with a history of cholesterol-lowering therapy. Diabetes mellitus was diagnosed in patients with a blood glucose level 200 mg/dl, HbA1c level (NGSP) 6.5%, or a history of antidiabetic therapy. Obese subjects consisted of overweight or obese individuals (defined as a body mass index (BMI) 25.0 kg/m2). The smoking index was defined as the product of packs per day and years of smoking (pack-years). Although an electrocardiogram was not done in 225 males (2.8%), these participants were considered to have no atrial fibrillation (Af) if they had no history of Af, and they were included in the present analysis. Daily alcohol consumption was categorized into three groups (A1, none or occasional; A2, 25 g/day, A3, >25 g/day as ethanol averaged from the amount in a week) at baseline.
2.1. Study subjects 2.4. Outcome measures The study subjects were recruited from the community-dwelling population living in districts of Iwate in northern Japan (the IwateKenpoku Cohort study). This study was conducted as part of a government-sponsored, multiphasic health check-up program aimed at the general population. Between April 2002 and January 2005, invitations to participate in this health check-up program were issued by government offices in 17 rural municipalities located in these districts; 26,469 individuals (9161 males) took part in the program and agreed to join the present study. After subjects aged over 80 years (280 males) and those under 40 years (300 males) or subjects without information about alcohol consumption (23 males), as well as those with a history of cardiovascular disease or stroke (526 males), were excluded, the data of 8059 males (mean age, 64.1 9.7 years) were included in the present study. The study was approved by our institutional ethics committee, and all of the participants provided their written, informed consent. 2.2. Blood sampling and baseline assessment Blood samples were collected from an antecubital vein. The samples were collected into vacuum tubes containing EDTA or a serum separator gel. After sampling, the tubes were stored immediately in an icebox and centrifuged at 1500 g for 10 min within 8 h of collection. Aliquots of serum were stored at 20 C, and routine hematology and biochemistry tests were done within a few days after blood sampling. The estimated glomerular filtration rate (eGFR) was calculated using the modified equation of the Modification of Diet in Renal Disease (MDRD) study [12]. Glycosylated hemoglobin (HbA1c) was measured by the HPLC method as a Japan Diabetes Society (JDS) value, and the result was presented as a National Glycohemoglobin Standardization Program (NGSP) equivalent value transformed by adding 0.4% uniformly. Baseline clinical examinations included a standard 12-lead electrocardiogram, and a self-reported questionnaire was administered to document subjects’ medical history and lifestyle. Systolic blood pressure (sBP) and diastolic blood pressure (dBP) were measured by trained nurses using an automatic digital sphygmomanometer placed on the upper arm at the height of the heart with the patient in a sitting position after having rested for a few minutes. The average of two measurements was used for the present analysis. 2.3. Risk factor definition Hypertension was defined as at least one of: systolic blood pressure 140 mmHg; a diastolic blood pressure 90 mmHg; or current antihypertensive therapy. Dyslipidemia was diagnosed in subjects with serum low-density lipoprotein cholesterol (LDL-C) levels 140 mg/dl and/or high-density lipoprotein cholesterol
In this cohort study, the outcome event was all-cause death, as well as any non-fatal cardiovascular events, such as myocardial infarction, cerebral infarction, or other strokes. The dates of death and moving away from the region were confirmed by the investigators reviewing population-register sheets at each local government office. Persons who were known to be alive at the end of follow-up and those who had moved away from the study area were treated as censored cases. Stroke events were identified by accessing the Iwate prefecture stroke registration program, which included the entire area where the subjects lived; details of this registry have been described previously [13]. Since 1991, the stroke registration program has been coordinated by the Iwate prefecture government and the Iwate Medical Association; the medical records of all medical facilities within the survey area are verified to ensure complete capture of all data. Incidents of acute myocardial infarction were identified by accessing data from the Northern Iwate Heart Disease Registry Consortium, which has been collecting data since 2002. The registration of acute myocardial infarction and sudden death was based on the criteria of the MONICA study [14]. To verify the accuracy of the data, a physician or trained research nurse visited and checked the medical records of the referral hospitals. 2.5. Statistical analysis Comparisons of baseline demographic data among the three categories of alcohol consumption levels were performed using analysis of variance (ANOVA) with Scheffe’s post hoc test. The values in the tables are expressed as means standard deviation. The cumulative survival curves (free of acute myocardial infarction or ischemic stroke) by the three categories of alcohol consumption and the relative risk of group A2 and A3 compared to the reference A1 group were determined using the multivariate Cox regression hazards model adjusted by age, presence of hypertension, diabetes, dyslipidemia, BMI, and smoking index. In the analysis concerning ischemic stroke, further adjustment by the presence of Af was performed. The results are expressed as hazard ratios (HRs) and the corresponding 95% confidence interval (CI). The data of adjusted factors such as height, weight, blood pressures, and smoking status were missing as random phenomena in 47 subjects, and thus, 8014 subjects having a complete data set were analyzed in the Cox model. The analyses were performed using the SPSS statistical package, version 19.0 (Chicago, IL, USA). 3. Results Overall baseline demographic data for the 3 categories of daily alcohol consumption are shown in Table 1. The A3 group was younger and had higher blood pressures, higher BMI, higher eGFR,
224
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
Table 1 Baseline demographic data for three categories of alcohol consumption (overall analysis). Alcohol consumption category （average alcohol consumption)
A1 (0.1 g/day)
A2 (17.4 g/day)
A3 (49.7 g/day)
p-value
Number of subjects Age, years Height, cm Weight, kg Body mass index, kg/m2 Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Serum creatinine, mg/dl eGFR, ml/min/1.73 m2 Uric acid, mg/dl Hemoglobin A1c, % Triglycerides, mg/dl LDL-cholesterol, mg/dl HDL-cholesterol, mg/dl Smoking index, pack-years High sensitivity CRP, mg/l Hypertension, n(%) Diabetes mellitus, n(%) Dyslipidemia, n(%) Atrial fibrillation, n(%)
3246 65.7 9.4 161.1 6.3 62.2 9.7 23.9 3.1 128 19
2441 64.9 9.4a 161.3 6.3 62.1 9.0 23.8 2.8 131 20a
2372 60.9 9.6a,b 162.8 6.4a,b 63.8 9.4a,b 24.0 2.9b 134 19a,b
e <0.001 <0.001 <0.001 <0.05 <0.001
76 11
79 11a
82 11a,b
<0.001
0.85 0.27 73.6 15.0 5.6 0.8 5.6 0.8 124 76 119 29 51 13 18 22
0.81 0.15a 76.2 14.7a 5.7 1.3a 5.5 0.7a 120 79 113 28a 57 15a 16 19a
0.79 0.14a,b 80.2 15.4a,b 6.0 1.4a,b 5.5 0.7a 133 97a,b 107 29a,b 62 17a,b 24 21a,b
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
1.47 4.90
1.42 5.15
1.25 3.83
1295 (39.9%) 328 (10.1%) 952 (29.3%) 77 (2.4%)
1151 (47.2%) 227 (9.3%) 440 (18.0%) 75 (3.1%)
1259 (53.1%) 194 (8.2%) 357 (15.0%) 69 (2.9%)
0.22 <0.001 <0.001 <0.001 0.23
eGFR; estimated glomerular filtration rate. a p < 0.05 vs A1. b p < 0.05 vs A2.
4. Discussion
higher uric acid, higher triglyceride, lower LDL-cholesterol, higher HDL-cholesterol, and a higher smoking index than the A2 group. The prevalence of hypertension was significantly higher with larger amounts of alcohol consumption. The prevalence of diabetes and dyslipidemia was significantly lower with larger amounts of alcohol consumption. A similar trend in the baseline data was obtained with respect to obese subjects, except for the prevalence of diabetes (Table 2). Table 2 Baseline demographic data for three categories of alcohol consumption (obese subjects). Alcohol consumption category （average alcohol consumption)
A1 (0.1 g/day)
A2 A3 (17.0 g/day) (48.9 g/day)
p-value
Number of subjects Age, years Height, cm Weight, kg Body mass index, kg/m2 Systolic blood pressure, mmHg Diastolic blood pressure, mmHg Serum creatinine, mg/dl eGFR, ml/min/1.73 m2 Uric acid, mg/dl Hemoglobin A1c, % Triglycerides, mg/dl LDL-cholesterol, mg/dl HDL-cholesterol, mg/dl Smoking index, pack-years High sensitivity CRP, mg/l Hypertension, n(%) Diabetes mellitus, n(%) Dyslipidemia, n(%) Atrial fibrillation, n(%)
1155 64.3 9.5 161.5 6.2 71.0 9.7 27.2 1.9 132 19
789 64.3 9.5 161.8 6.0 70.7 7.2 27.0 1.8a 135 19a
824 60.9 9.2a,b 162.8 6.4a,b 72.1 7.8a,b 27.1 1.9 137 18a,b
e <0.001 <0.001 <0.001 0.05 <0.001
79 11
81 10a
84 10a,b
<0.001
0.86 0.25 72.8 14.6 5.9 1.3 5.7 0.8 151 93 124 29 49 11 18 23 1.68 5.96 574 (49.7%) 147 (12.7%) 467 (40.4%) 31 (2.7%)
0.83 0.16a 74.8 14.5a 6.0 1.3 5.7 0.8 141 84 119 26a 52 12a 16 19 1.71 5.96 434 (55.0%) 92 (11.7%) 187 (23.7%) 23 (2.9%)
0.81 0.14a 77.6 15.0a,b 6.2 1.5a,b 5.6 0.7a 157 113a,b 112 29a,b 57 15a,b 22 22a,b 1.36 4.28 523 (63.5%) 83 (10.1%) 171 (20.8%) 34 (4.1%)
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.30 <0.001 0.19 <0.001 0.17
eGFR; estimated glomerular filtration rate. a p < 0.05 vs A1. b p < 0.05 vs A2.
The mean follow-up period was 5.5 years. During follow-up, 53 subjects (cumulative incidence rate 0.66%) had a first, non-fatal myocardial infarction (MI) and 186 subjects (cumulative incidence rate 2.32%) had a first ischemic stroke (IS). MI-free curves are shown in Fig. 1. The HR for incident MI was significantly lower in the A2 group than in the A1 group when adjusted for age, hypertension, diabetes, dyslipidemia, smoking index, and BMI (HR ¼ 0.49, 95%CI 0.24e0.98, p ¼ 0.043). The HR for incident MI tended to be lower in the A3 group than in the A1 group when adjusted for the same factors, but it was not significant (HR ¼ 0.53, 95%CI 0.25e1.12, p ¼ 0.10). On the other hand, within the obese subjects, while a significantly lower HR for incident MI was retained in the A2 group (HR ¼ 0.29, 95%CI 0.08e0.99, p ¼ 0.049), no significant difference in the HR between the A3 group and the A1 group was found (HR ¼ 0.61, 95%CI 0.22e1.64, p ¼ 0.32). The IS-free curves are shown in Fig. 2. No significant differences were found in the IS-free curves among the 3 categories of alcohol consumption when adjusted for age, hypertension, diabetes, dyslipidemia, smoking index, Af, and BMI. Similar results were observed in obese patients. The shapes of the relationships between alcohol consumption category and incident MI or IS in overall subjects and obese subjects are shown in Fig. 3. For incident MI, while an L-shaped relationship with alcohol consumption was found in the analysis of overall subjects, a nearly U-shaped relationship was observed in obese subjects.
The present cohort study indicated that the risk for incident MI was lower in subjects who drank alcohol than in those who did not. This appeared to be particularly true for the subjects with mild-tomoderate alcohol consumption. In obese subjects, while there was a reduced risk for incident MI in the mild-to-moderate drinkers, there was no significant difference in the risk for incident MI between heavy drinkers and those who did not drink. There was no effect of alcohol consumption on incident IS, regardless of the amount of alcohol consumption. Given the baseline data in the 3 alcohol categories, the favorable effects of alcohol intake appeared to be linked with elevated eGFR, decreased HbA1c, decreased LDL-C, and increased HDL-C level in the present study. The prevalence rates of dyslipidemia and diabetes were also lower in the subjects who drank alcohol. Although these factors seemed to contribute to the protective effect on MI, the causality between alcohol consumption and these risk factors was obscure because there were many opportunities for instructional intervention about alcohol avoidance prior to the baseline investigation in the subjects with renal dysfunction, diabetes, and dyslipidemia. Regarding HDL-C levels, an acute increase effect of alcohol intake was shown in a controlled study [15]. However, when adjustment for HDL-C was performed in the present model, the relationship between alcohol consumption and the risk of incident MI was essentially unaltered, which suggests that other cardioprotective mechanisms may be involved. In fact, although HDL-C is a crucial mediator of the atheroprotective properties of alcohol, this protective effect has been established with light-tomoderate drinking only. No previous evidence that the behavior of HDL-C level contributes to the U-shaped relationships between alcohol intake and cardiovascular events has been presented in the literature. Previous reports suggested that risk reduction in ischemic heart disease was attributed to the beneficial impact of alcohol on hemostatic factors related to a thrombolytic profile [16], insulin sensitivity [17], inflammation reflected by the C-reactive protein level [18], and fibrinogen concentrations [19].
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
225
Fig. 1. Associations between incident acute myocardial infarction and daily alcohol consumption. The data are expressed as hazard ratios (HRs) (95% confidence interval). The data in the overall analysis are adjusted by age, hypertension, diabetes, dyslipidemia, smoking index, and BMI. In obese subjects, BMI is excluded from these adjusted factors.
The present study showed an L-shaped relationship between alcohol consumption and incident MI in the analysis of overall subjects, because there was no increase in MI risk associated with higher alcohol consumption. This finding is consistent with some results from a meta-analysis performed by Maclure et al. [20] and Ronksley et al. [10]. On the other hand, Corrao et al. reported a Jshaped relationship between alcohol intake and coronary heart disease in a meta-analysis [21]. Gigleux et al. showed that light-tomoderate consumption of alcohol (>15 g/d) had significant cardioprotective effects, which were more apparent in men with metabolic syndrome [22]. However, since information about heavy drinkers was lacking, the U- or J-shaped relationship between alcohol consumption and IHD risk was not demonstrated in their study results. In the present study, a U-shaped relationship was observed in obese subjects, indicating that a harmful effect of a large amount of alcohol consumption may appear in subjects with a high risk for IHD. With respect to the mechanism, other detrimental effects observed only in the A3 group of obese subject should be considered. However, within the baseline characteristics that were investigated in the present study, the inter-group differences in obese subjects were very similar to those of overall subjects in all clinical indices. Therefore, the decreased beneficial effect of heavy drinking in obese subjects cannot be explained by increased blood pressure, increased smoking index, or decreased lipid profile improvement. Hyperuricemia has been shown to increase all-cause mortality and incident cardiovascular disease [23]. Since uric acid (UA) increase was found in the A3 group of obese subjects, this effect can be hypothesized to increase MI only in heavy-drinking obese subjects. Moreover, in overall analysis, the increase in UA level for the A2 group was negligible. However, no consensus has yet been reached regarding whether UA is
a protective anti-oxidant factor or a risk factor. Other detrimental factors that show a U-shaped configuration by amount of alcohol consumption only in obese subjects need to be explored. Imhof et al. showed that light-to-moderate daily alcohol intake (up to 20 or 40 g/day) was associated with lower CRP and white blood counts compared to non-drinking and heavy-drinking, even after adjustment for various potential confounders [18]. Although such anti-inflammatory effects may contribute to the U-shaped relationship between alcohol consumption and ischemic heart disease, no significant difference in inflammation represented by the high sensitivity CRP level was found among categories of alcohol consumption in the obese subjects of the present study. A risk reduction in IS, as well as MI, should be also found if an antithrombotic effect [24] is the main contributor to the cardioprotective effect of alcohol consumption. However, favorable effects of alcohol consumption on IS were not documented in the present study. Increased blood pressure caused by alcohol consumption may be partly involved in this mechanism. In fact, the influence of increased systolic blood pressure on incident IS was larger compared with that on incident MI in the overall analysis of the present subjects (risk for MI: HR ¼ 1.008, p ¼ 0.22; risk for IS: HR ¼ 1.015, p < 0.001, by a 1-mmHg increase in sBP in men when adjusted by age). Alternatively, the high prevalence of hypertension at baseline in the present cohort might have had a sufficiently strong effect to eliminate the much weaker effect of alcohol consumption. Some limitations of the present study warrant consideration. Given the many previous investigations, the results or the shape of the relationships between alcohol consumption and cardiovascular risk can be altered by the choice of cutoff level for ‘light’ or ‘moderate’ alcohol consumption. Since the present study did not
Fig. 2. Associations between incident ischemic stroke and daily alcohol consumption. The data are expressed as hazard ratios (HRs) (95% confidence interval). The data in the overall analysis are adjusted by age, hypertension, diabetes, dyslipidemia, smoking index, atrial fibrillation, and BMI. In obese subjects, BMI is excluded from these adjusted factors.
226
S. Makita et al. / Atherosclerosis 224 (2012) 222e227
Obese subjects
2.50
2.50
2.00
2.00
1.50
1 .50
Hazard ratio
Hazard ratio
Overall analysis
1.00
1.00
0 .50
0.50
0.10
0.10 A1
A2
A3
Category of alcohol consumption
A1
A2
A3
Category of alcohol consumption
Fig. 3. The shapes of the relationship between alcohol consumption and incident myocardial infarction or ischemic stroke. The graphs show the hazard ratios (circles) and 95% confidence interval (error bars) for incident myocardial infarction (solid lines) or ischemic stroke (dashed lines) in overall subjects and obese subjects. The vertical axis is the logscale hazard ratio.
have enough subjects or clinical incidents to examine the detailed effect of range setting in the alcohol consumption categories, the suitability of the threshold (25 g/day) in the categorization of alcohol consumption is unclear. However, Brien et al. showed in a meta-analysis that ‘moderate’ consumption of alcohol (up to two drinks or 30 g alcohol a day for men) has beneficial effects on a variety of biomarkers linked to the risk of coronary heart disease [25], and this level of alcohol consumption may be similar to the cutoff level used in the present study. Second, alcohol consumption was evaluated only once at baseline, and intake variation during follow-up [26] was not investigated. Drinking patterns were not taken into consideration. An inaccurate estimate daily alcohol consumption cannot be excluded due to self-reporting of the consumption in the present study. In addition, the present study did not investigate the kinds of beverages. However, beer, wine, and spirits are generally known to have similar effects on high-density lipoprotein cholesterol [16]. A possibility that physical activity may have confounded the results in the present analysis cannot be ruled out. Finally, since female subjects were excluded from the analysis due to their very small low alcohol consumption (group A1, 14,291; group A2, 1290; group A3, 153 of 15,734 total females), sex differences in the risk were unclear. In conclusion, alcohol consumption may have a protective effect on the onset of MI, but not on IS, in the general population. While an L-shaped relationship between alcohol consumption and incident MI was found in overall subjects, a U-shaped relationship was observed in obese subjects. An appropriate limit or a recommended amount of daily alcohol consumption, depending on the risk of ischemic heart disease, may need to be established. Disclosure The authors report no conflicts of interest. References [1] Iso H, Baba S, Mannami T, et al. JPHC study group. alcohol consumption and risk of stroke among middle-age men: the JPHC study cohort I. Stroke 2004; 35:1124e9. [2] Reynolds K, Lewis B, Nolen JD, Kinney GL, Sathya B, He J. Alcohol consumption and risk of stroke: a meta-analysis. J Am Med Assoc 2003;289:579e88.
[3] Berger K, Ajani UA, Kase CS, et al. Light-to-moderate alcohol consumption and risk of stroke among U.S. male physicians. N Engl J Med 1999;341: 1557e64. [4] Yano K, Rhoads GG, Kagan A. Coffee, alcohol and risk of coronary heart disease among Japanese men living in Hawaii. N Engl J Med 1977;297: 405e9. [5] Langer RD, Criqui MH, Reed DM. Lipoproteins and blood pressure as biological pathways for effect of moderate alcohol consumption on coronary heart disease. Circulation 1992;85:910e5. [6] Shaper AG, Wannamethee G, Walker M. Alcohol and coronary heart disease: a perspective from the British regional heart study. Int J Epidemiol 1994;23: 482e94. [7] Gordon T, Kannel WB. Drinking habits and cardiovascular disease: the Framingham study. Am Heart J 1985;110:331e4. [8] Miller GJ, Beckles GL, Maude GH, Carson DC. Alcohol consumption: protection against coronary heart disease and risks to health. Int J Epidemiol 1990;19: 923e30. [9] Kitamura A, Iso H, Sankai T, et al. Alcohol intake and premature coronary heart disease in urban Japanese men. Am J Epidemiol 1998;147:59e65. [10] Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ 2011;342:d671. [11] Iso H, Kitamura A, Shimamoto T, et al. Alcohol intake and the risk of cardiovascular disease in middle-aged Japanese men. Stroke 1995;26:767e73. [12] Imai E, Horio M, Nitta K, et al. Modification of the modification of diet in renal disease (MDRD) study equation for Japan. Am J Kidney Dis 2007;50: 927e37. [13] Omama S, Yoshida Y, Ogawa A, Onoda T, Okayama A. Differences in circadian variation of cerebral infarction, intracerebral haemorrhage and subarachnoid haemorrhage by situation at onset. J Neurol Neurosurg Psychiatr 2006;77: 1345e9. [14] Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, Arveiler D, Rajakangas AM, Pajak A. Myocardial infarction and coronary deaths in the world health organization MONICA project: registration procedures, event rates, and case fatality rates in 38 populations from 21 countries in four continents. Circulation 1994;90:583e612. [15] Sierksma A, Patel H, Ouchi N, et al. Effect of moderate alcohol consumption on adiponectin, tumor necrosis factor-alpha, and insulin sensitivity. Diabetes Care 2004;27:184e9. [16] 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;319:1523e8. [17] Kiechl S, Willeit J, Poewe W, et al. Insulin sensitivity and regular alcohol consumption: large, prospective, cross sectional population study (Bruneck study). BMJ 1996;313:1040e4. [18] Imhof A, Woodward M, Doering A, et al. Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille). Eur Heart J 2004;25: 2092e100. [19] Sierksma A, van der Gaag MS, Kluft C, Hendriks HF. Moderate alcohol consumption reduces plasma C-reactive protein and fibrinogen levels; a randomized, diet-controlled intervention study. Eur J Clin Nutr 2002;56: 1130e6.
S. Makita et al. / Atherosclerosis 224 (2012) 222e227 [20] Maclure M. Demonstration of deductive meta-analysis: ethanol intake and risk of myocardial infarction. Epidemiol Rev 1993;15:328e51. [21] Corrao G, Rubbiati L, Bagnardi V, Zambon A, Poikolainen K. Alcohol and coronary heart disease: a meta-analysis. Addiction 2000;95:1505e23. [22] Gigleux I, Gagnon J, St-Pierre A, et al. Moderate alcohol consumption is more cardioprotective in men with the metabolic syndrome. J Nutr 2006;136: 3027e32. [23] Verdecchia P, Schillaci G, Reboldi G, Santeusanio F, Porcellati C, Brunetti P. Relation between serum uric acid and risk of cardiovascular disease in essential hypertension. The PIUMA study. Hypertension 2000;36:1072e8.
227
[24] Rubin R, Rand ML. Alcohol and platelet function. Alcohol Clin Exp Res 1994; 18:105e10. [25] Brien SE, Ronksley PE, Turner BJ, Mukamal KJ, Ghali WA. Effect of alcohol consumption on biological markers associated with risk of coronary heart disease: systematic review and meta-analysis of interventional studies. BMJ 2011;342:d636. [26] Emberson JR, Shaper AG, Wannamethee SG, Morris RW, Whincup PH. Alcohol intake in middle age and risk of cardiovascular disease and mortality: accounting for intake variation over time. Am J Epidemiol 2005; 161:856e63.