- Email: [email protected]
Alcohol consumption and risk of coronary artery disease: A dose-response meta-analysis of prospective studies
Accepted Manuscript Alcohol consumption and risk of coronary artery disease: a dose–response metaanalysis of prospective studies Yang Yang, Dong-Chen Liu, Qi-Ming Wang, Qing-Qing Long, Shan Zhao, Zhen Zhang, Yao Ma, Ze-Mu Wang, Lei-Lei Chen, Lian-Sheng Wang PII:
S0899-9007(15)00474-8
DOI:
10.1016/j.nut.2015.11.013
Reference:
NUT 9660
To appear in:
Nutrition
Received Date: 8 July 2015 Revised Date:
29 November 2015
Accepted Date: 29 November 2015
Please cite this article as: Yang Y, Liu D-C, Wang Q-M, Long Q-Q, Zhao S, Zhang Z, Ma Y, Wang Z-M, Chen L-L, Wang L-S, Alcohol consumption and risk of coronary artery disease: a dose–response metaanalysis of prospective studies, Nutrition (2016), doi: 10.1016/j.nut.2015.11.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Alcohol consumption and risk of coronary artery disease: a dose–response meta-analysis of prospective studies Yang Yang†; Dong-Chen Liu†; Qi-Ming Wang; Qing-Qing Long; Shan Zhao; Zhen Zhang; Yao Ma; Ze-Mu Wang; Lei-Lei Chen*; and Lian-Sheng Wang, *
RI PT
Department of Cardiology, the First Affiliated Hospital of Nanjing Medical
Yang Yang and Dong-Chen Liu contributed equally to this work.
M AN U
†
SC
University, Nanjing, China.
*Corresponding authors: Lian-Sheng Wang, MD, PhD, Department of Cardiology, the First Affiliated Hospital of Nanjing Medical University, 300 Guangzhou Road, Nanjing, 210029, Jiangsu Province, China. E-mail: [email protected], Tel./Fax:
TE D
0086-25-83724440; Lei-Lei Chen, MD, PhD, Department of Cardiology, the First Affiliated Hospital of Nanjing Medical University, 300 Guangzhou Road, Nanjing,
EP
210029, Jiangsu Province, China. E-mail address: [email protected],
AC C
Tel./Fax: 0086-25-83724440.
ACCEPTED MANUSCRIPT Abstract Objective: To investigate and quantify the potential dose-response association between alcohol consumption and risk of coronary artery disease (CAD). Methods and Results: We searched the PubMed database from inception to March 2015 and reviewed the reference list of relevant articles to identify prospective studies
RI PT
assessing the association between alcohol consumption and risk of CAD. Study-specific relative risk (RR) estimates were combined using a random-effects model. Publication bias was estimated using Begg’ funnel plot and Egger’s regression asymmetry test. The meta-analysis included 18 prospective studies, with a total of
SC
214340 participants and 7756 CAD cases. The pooled adjusted RRs were 0.62 [95% confidence interval (CI) 0.56–0.68] for highest alcohol consumption amount versus
M AN U
lowest amount. Begg and Egger regression tests provided no evidence of substantial publication bias (P=0.762 for begg’s test and 0.172 for egger’s test).In a dose-response analysis, We observed a non-linear association between alcohol consumption and risk of CAD(P for non-linearity
0.00 ). Compared with
non-drinkers, the RRs (95% CI) of CAD across levels of alcohol consumption were
TE D
0.75 (0.70-0.80) for 12g/day, 0.70 (0.66-0.75) for 24g/day, 0.69 (0.64-0.75) for 36g/day, 0.70 (0.64-0.77) for 60g/day, 0.74 (0.67-0.83) for 90g /day, and 0.83 (0.67-1.04) for 135g/day.
EP
Conclusion: Alcohol consumption in moderation is associated with a reduced risk of
AC C
CAD with 36 grams/day of alcohol conferring a lower risk than other levels.
Keywords: coronary artery disease; alcohol consumption; meta-analysis; prospective studies
ACCEPTED MANUSCRIPT
Introduction Coronary artery disease (CAD) is now a severe public health problem, affecting
RI PT
millions of people in both developed and developing countries. CAD is the major cause of death in developed countries [1], accounting for up to 40% of all fatal events [2]. In low and middle income countries, the incidence of cardiovascular disease has significantly increased. By 2020, the disease is expected to be the leading cause of
SC
morbidity and mortality in most developing countries[3]. Increased CAD risk is associated with increased atherosclerosis, which is more frequent in patients with
M AN U
CKD, especially the dialysis ones, than in the general population[4]. CAD is by far the most common initiating cause of HF (in ∼ 70% of patients), things even worse[5].
which makes
Alcohol is one of the most widely consumed beverages in the world and is the fifth leading reason for death and disability accounting for 4% of life years lost due to
TE D
disease[6]. Disease burden and disability are closely associated with the volume of alcohol consumption[7]. Consumption of alcohol is believed to have potential health benefits, such as anti-inflammatory effect on atherosclerosis[8], increases in
EP
high-density lipoprotein cholesterol[9], inhibitory effects on platelet aggregation[9], influence on antioxidant capacity and insulin sensitivity, thus restraining
AC C
atherosclerosis[10-11]. Nevertheless, it can also have harmful effects. Alcohol use is associated with a worse hematological picture of total cholesterol, LDL cholesterol (LDL-C) levels, and systolic blood pressure [12]. The association between alcohol consumption and the risk of CAD is of great
importance because of the popularity of alcohol and the huge burden caused by CAD. The role of alcohol in the risk of CAD is well established, and has been assessed in a large number of observational epidemiological studies or meta-analyses[12-16]. Though all demonstrated a U-shaped or J-shaped association between alcohol intake and CAD such that both non-drinkers and heavy drinkers appear to have higher morbidity rates than light and moderate drinkers, few study has designed to
ACCEPTED MANUSCRIPT quantitatively assess the amount of alcohol, referring rather than qualitative levels of alcohol consumption(e.g., “heavy” versus “low”). Thus, we aimed to better quantify the association between alcohol consumption and the risk of CAD through a comprehensive systematic literature review and dose–response meta-analysis that can intuitively reflect the relationship between alcohol consumption and risk of CAD.
RI PT
Methods Literature search and selection:
We identified potentially relevant articles regardless of language by a computerized search of the PubMed database from inception to April 2015.The
SC
following search terms were used: coronary artery disease, coronary heart disease, cardiovascular disease, myocardial infarction, ischemic heart disease, CAD, IHD
M AN U
combined with alcohol consumption, drink, drinking, ethanol. The reference lists of pertinent articles were reviewed to identify additional studies. Two investigators independently reviewed all identified studies. Eligibility criteria for inclusion in the present meta-analysis were: (1) the study was prospective design; (2) the exposure was alcohol consumption; (3) the outcome was total CAD incidence (including MI,
TE D
CAD, non-stroke cardiovascular disease, and other coronary events); (4) the population was free from CAD at baseline; and (5) relative risks (RRs) with 95% confidence intervals (CIs) , adjusted for at least age, were reported. If a study reported
EP
more than one set of RR estimates, we used the maximally adjusted RRs. We considered gender-specific estimates in cases where a study reported data separately
AC C
for men and women. Data Extraction
From each study, we extracted the first author’s last name, country in which the
study was performed, duration of follow-up, sex and age, method of alcohol consumption assessment, total numbers of cases and participants, number of cases and person-years in every exposure category, the most fully adjusted RRs with corresponding 95% CIs for each category of alcohol consumption, and covariates adjusted for in the analysis. If there was disagreement between the two investigators about eligibility of the data, it was resolved by consensus with a third reviewer. Statistical methods
ACCEPTED MANUSCRIPT We standardized alcohol consumption across studies using a common scale, i.e, alcoholic g/day in order to pool the study-specific RRs. When a study reported alcohol consumption in drinks/week, we converted the intake into g/day assuming that one drink contains 12 g of alcohol. For each study, we assigned the median or mean alcohol consumption for the category to each corresponding RR. When the median or
RI PT
mean consumption was not reported, we assigned the midpoint of the upper and lower boundaries in each category as the median consumption. If the upper boundary for the highest category was not provided, the midpoint of the category was set at 1.5 times the lower boundary. When the lowest category was open-ended, we set the lower
SC
boundary to zero. If the number of cases and person-years were not available, we used
obtain a summary estimate.
M AN U
the relative risks comparing the highest versus lowest categories of alcohol intake to
The heterogeneity among studies was estimated by the Cochran Q test and I2 statistic. Heterogeneity was confirmed with a significance level of P≤0.10. The I2 statistic describes the percentage of total variation in point estimates that can be attributed to heterogeneity. For the I2 metric, we considered low, moderate, and high
TE D
I2 values to be 25%, 50%, and 75%, respectively. We used a fixed effect model (Mantel-Haenszel method) when heterogeneity was negligible, and a random effect model (DerSimonian and Laird method) when heterogeneity was significant.
EP
Publication bias was evaluated using Begg’ funnel plot [17] and Egger regression asymmetry test [18].
AC C
To evaluate the dose–response association between alcohol consumption and risk of CAD, a 2-stage random-effects dose–response meta-analysis taking into account the between-study heterogeneity was performed proposed by Orsini et al to compute the trend from the correlated logRR estimates across categories of alcohol consumption[19]. Firstly, a restricted cubic spline model, with 4 knots at the 5th, 35th, 65th, and 95th percentiles of alcohol consumption levels, was estimated using generalized least square regression taking into account the correlation within each set of published RRs[20-21]. Secondly, the restricted maximum likelihood method in a multivariate random-effects meta-analysis was used to combine the study-specific estimates[22]. A P value for nonlinearity and overall significance of the curve was
ACCEPTED MANUSCRIPT calculated using the method proposed by Greenland and Longnecker [23]. All statistical analyses were performed using Stata (version 12.0, StataCorp, College Station, TX, USA).All reported probabilities (P values) were two-sided, with P ≤0.05 considered statistically significant. Results
RI PT
Search Results and Study Characteristics The detailed steps of our literature search are shown in Fig. 1. Finally, 13 articles including 18 studies involving 214340 participants and 7756 CAD cases were used in this meta-analysis. Among the 18 prospective studies, 9 studies reported the result
SC
from United States, 5 studies from Japan, 2 studies from China, and 2 studies from
High Versus Low Analyses
M AN U
Spain. The characteristics of these studies are presented in Table 1.
Data from 13 articles including 18 prospective studies were used in this meta-analysis [24-36]. Pooled results suggested that highest alcohol consumption amount versus lowest amount was 0.62(95% CI 0.56-0.68; I-squared 28.5%,p=0.126). (Fig.2).Additionally, Begg and Egger regression tests provided no evidence of
TE D
substantial publication bias (P=0.762 for begg’s test and 0.172 for egger’s test). Fig. 3 presents the funnel plot of log RR versus variance of log RR for all the studies included in the meta-analysis.
EP
In subgroup analyses for country, inverse associations of alcohol consumption and CAD risk were found both in USA[summary RR = 0.61, 95%CI = 0.54-0.69] and
AC C
in Asia [summary RR = 0.66, 95%CI = 0.55-0.78]. When we conducted the subgroup analysis by follow-up duration ( >10 years and ≤10years), significant associations were found both in >10 years and ≤10years follow-up. When stratified analysis for study outcome (CAD and MI), sex, number of cases and diabetes mellitus(Yes or No),similar results were also found. The detailed results are summarized in Table 2. Dose–response Analysis For dose–response analysis, data from 13 articles including 18 prospective studies involving 7756 CAD cases were used for the dose-response relation between alcohol and CAD risk. We observed a non-linear association between alcohol consumption and risk of CAD(P for non-linearity 0.00 )(Fig. 4). Compared with
ACCEPTED MANUSCRIPT non-drinkers, the RRs (95% CI) of CAD across levels of alcohol consumption were 0.75 (0.70-0.80) for 12g/day, 0.70(0.66-0.75) for 24g/day, 0.69 (0.64-0.75) for 36g/day, 0.70 (0.64-0.77) for 60g/day, 0.74 (0.67-0.83) for 90g /day, and 0.83 (0.67-1.04) for 135g/day. Discussion
RI PT
Results from this meta-analysis showed that alcohol consumption in moderation is associated with a reduced risk of CAD. The dose–response meta-analysis indicated that consumption of 36 grams per day is associated with a 31% lower risk of CAD. Our dose–response meta-analysis suggested a non-linear association of alcohol
SC
consumption with risk of CAD, which is consistent with the observed U-shaped relationship between alcohol consumption and CAD [16-17]. A possible explanation
M AN U
for the non-linear association is that consumption of large amounts of alcohol may increase blood pressure, low-density lipoprotein (LDL) cholesterol, and triglycerides, which are risk factors for CAD[7].
There are several possible mechanisms for protection against CAD by alcohol. The best established one is related to the levels of plasma lipoproteins, including
TE D
increases in high-density lipoprotein (HDL) cholesterol and Apo lipoprotein A-I, and, to a lesser degree, a decrease in LDL-cholesterol [37-38]. The central roles of lipids in CAD development include a positive relation with LDL-cholesterol and an inverse
EP
relation with HDL-cholesterol. Inverse relations of HDL to CAD risk primarily operate via the removal of lipid deposits in large blood vessels, but HDL also binds
AC C
with cholesterol in the tissues and may result in preventing tissue oxidation of LDL-cholesterol[39]. Several studies examining the hypothesis that higher HDL-cholesterol in drinkers mediates CAD protection suggested that higher HDL levels mediated approximately half of the lower CAD risk [40–42]. There is data suggesting that alcohol might increase HDL but it also might change it functionality into dysfunctional HDL [43]. Besides, light to moderate alcohol consumption may decrease platelet aggregation, decrease fibrinogen, increase tissue plasminogen activator, and improve insulin sensitivity[44-45]. Antithrombotic actions of alcohol could partially account for the lower CAD risk at very light drinking level. Different beverages may have different effects upon CAD. The French Paradox
ACCEPTED MANUSCRIPT suggests that red wine confers a greater cardiovascular benefit than other alcoholic beverages[46]. Red wine contains high amount of polyphenolic compounds like resveratrol (RES) (trans-3,4,5-trihydroxystilbene), epicatechin, catechin, gallic acid, quercetin. Primarily RES is thought to be responsible for the cardioprotective effect of red wine[47]. RES, one of the main polyphenol extracts in grape skin/seed, red wine,
RI PT
and the root of Polygonum cuspidatum, is found to have protective potentials to multi-targets related to CAD [48-49]. The protective effect of RES has been shown to involve various mechanisms. RES improves endothelial function, and has numerous beneficial effects on vascular tone and vessels in human and animal models[50]. It has
SC
also been shown that RES improves the release of nitric oxide (NO) and prostacyclin (PGI) which plays a prominent role in the maintenance of endothelial function [51].
M AN U
Besides, several studies supported its antioxidant activity, its ability to decrease LDL oxidation, and function as a direct free radical scavenger [52]. The effect of RES on CAD is controversial. A meta-analysis of 10 randomized controlled trials did not indicate any benefit of RES supplementation on plasma CRP. Furthermore, RES supplementation was not found to alter plasma levels of TC, LDL-C, TG and glucose,
TE D
and additionally slightly decreased HDL-C concentrations[53]. A reason for these effects could be that serum levels of RES in lower dose may not be adequate for a beneficial effect[53].
EP
Besides the type, the drinking pattern also plays an important role in the effect of alcohol. Drinking-pattern aspects with medical implications for CAD are the
AC C
frequency of drinking (number of days per week), variability over time and whether the alcohol is taken with food [39]. Data exist that support that alcohol drinking at mealtimes is more favorable for protecting against CAD and hypertension than drinking apart from meals and that frequency of intake may be a stronger factor in benefit than amount [54-56]. The main strength of this meta-analysis is its dose-response design that provides better quantification of the associations between specified amounts of alcohol and risk of CAD. A dose-response meta-analysis should be the first choice when performing a systematic review, rather than conducting a meta-analysis based only on the comparison of the extreme categories of consumption (highest versus lowest), which
ACCEPTED MANUSCRIPT could vary considerably among studies. Another strength of this meta-analysis is that only prospective studies were included. The prospective design minimized the potential influence of recall and selection bias, which could be a problem in retrospective case–control studies. Nevertheless, our meta-analysis also had several limitations. First, although most
RI PT
studies controlled for major risk factors for CAD, we cannot rule out the possibility that the results were affected by confounding completely, because the studies were observational. Another limitation is the misclassification of alcohol consumption, which was inevitable considering that alcohol consumption was self-reported.
SC
Misclassification of alcohol consumption because of measurement error or changes in consumption during follow-up would most likely lead to inaccurate estimation of the
M AN U
true association between alcohol consumption and CAD risk[57]. Third, our meta-analysis focus our attention only on the amount of total alcohol, the type of beverage (wine, beer or spirits), the amount that produces the putative benefit of alcoholic beverages and the drinking pattern are ignored.
The association we observed between alcohol consumption and risk of CAD is
TE D
consistent with the U-shaped curve which describes a relationship where the relative risk decreases at low-to-moderate alcohol intake, then rises as alcohol intake increases, but does not rise above the level of the reference population of abstainers[58]. Though
EP
the RR (95% CI) of CAD was 0.83 (0.67-1.04) for 135g/day compared with non-drinkers, we absolutely do not support heavy drink as heavy drink is associated cardiomyopathy, systemic hypertension,
AC C
with increased cardiovascular risks
supraventricular arrhythmias, hemorrhagic stroke and heart failure that is not associated with CAD. No final conclusion has yet been reached regarding the recommendation of alcohol consumption to improve cardiovascular health. Advice regarding the health effects of alcohol drinking needs to be individualized according to the specific medical history and risks of any concerned person. However, a few common-sense rules are suggested: (1) The overall health risk of a heavy drinker is likely to be reduced by intake reduction or abstinence. (2) Owing to the unknown risk of progression to heavier drinking, abstainers should not indiscriminately be advised to drink for CAD health benefit. (3) Established light-to-moderate drinkers,
ACCEPTED MANUSCRIPT accounting for the majority of drinkers in the USA and Western Europe, do not need to change their drinking habits, except in special circumstances [39]. In summary, this meta-analysis provides evidence that alcohol consumption in moderation is associated with a reduced risk of CAD with 36 grams/day of alcohol
Acknowledgements
TE D
M AN U
SC
RI PT
conferring a lower risk than other levels.
This work was supported by grants from the National Natural Science Foundation of
Group
AC C
Innovation
EP
China (No 81270255), the Project Funded by the Science and Technological of
Jiangsu
Higher
Project”(JX2161015030). Conflict of interest
The authors state no conflict of interest.
Education
Institution
“Qing-Lan
ACCEPTED MANUSCRIPT
References
[1] Chen L, Wang X, Bao J, Geng C, Xia Y, Wang J. Direct comparison of cardiovascular magnetic resonance and single-photon emission computed tomography for detection of coronary artery disease: a meta-analysis. PLoS One 2014; 9:e88402. [2] Franchini M1, Peyvandi F, Mannucci PM. The genetic basis of coronary artery disease: from
RI PT
candidate genes to whole genome analysis. Trends Cardiovasc Med 2008; 18:157–62 . [3] Celermajer DS, Chow CK, Marijon E, Anstey NM, Woo KS. Cardiovascular disease in the developing world: prevalences, patterns, and the potential of early disease detection.J am CollCardiol 2012;60:1207-16.
SC
[4] Bielecka-Dabrowa A1, von Haehling S, Aronow WS, Ahmed MI, Rysz J, Banach M. Heart failure biomarkers in patients with dilated cardiomyopathy. Int J Cardiol. 2013; 168(3): 2404-10.
M AN U
[5] Gluba-Brzózka A1, Michalska-Kasiczak M, Franczyk-Skóra B, Nocuń M, Banach M, Rysz J. Markers of increased cardiovascular risk in patients with chronic kidney disease. Lipids Health Dis 2014; 13:135.
[6] Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in
2012; 380:2224-60.
TE D
21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study . Lancet
[7] Rehm J, Mathers C, Popova S, Thavorncharoensap M, Teerawattananon Y, Patra J. Global
EP
burden of disease and injury and economic cost attributable to alcohol use and alcohol-us e disorders. Lancet 2009; 373(9682):2223 – 33.
AC C
[8] Jerrard-Dunne P1, Sitzer M, Risley P, Steckel DA, Buehler A, von Kegler S, Markus HS. Carotid Atherosclerosis Progression Study. Interleukin-6 promoter polymorphism modulates the effects of heavy alcohol consumption on early carotid artery atherosclerosis: the Carotid Atherosclerosis Progression Study (CAPS).Stroke 2003; 34:402-7. [9] Rimm EB1, 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; 319:1523-8. [10] Booyse FM1, Pan W, Grenett HE, Parks DA, Darley-Usmar VM, Bradley KM, Tabengwa EM. Mechanism by which alcohol and wine polyphenols affect coronary heart disease risk. Ann Epidemiol 2007; 17(5 Suppl):S24-31.
ACCEPTED MANUSCRIPT [11] Vasdev S1, Gill V, Singal PK. Beneficial effect of low ethanol intake on the cardiovascular system: possible biochemical mechanisms. Vasc Health Risk Manag 2006; 2:263-76. [12] Perissinotto E, Buja A, Maggi S, Enzi G, Manzato E et al. Alcohol consumption and cardiovascular risk factors in older lifelong wine drinkers: the Italian Longitudinal Study on Aging. Nutr Metab Cardiovasc Dis 2010; 20(9):647 – 55.
RI PT
[13] Corrao G, Rubbiati L, Bagnardi V, Zambon A, Poikolainen K. Alcohol and coronary heart disease: a meta-analysis. Addiction 2000; 95:1505-23.
[14] Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and
SC
meta-analysis. BMJ 2011; 342:d671.
[15] Tolstrup J, Jensen MK, Tjønneland A, Overvad K, Mukamal KJ, Grønbaek M. Prospective
M AN U
study of alcohol drinking patterns and coronary heart disease in women and men. BMJ 2006; 27;332(7552): 1244-8
[16] Rehm J1, Sempos CT, Trevisan M. Alcohol and cardiovascular disease - more than one paradox to consider. Average volume of alcohol consumption, patterns of drinking and risk of coronary heart disease-a review. J Cardiovasc Risk 2003; 10(1):15-20.
TE D
[17] Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994; 50:1088–101.
[18] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple,
EP
graphical test. BMJ 1997; 315:629–34.
[19] Orsini N1, Li R, Wolk A, Khudyakov P, Spiegelman D. Meta-analysis for
linear and
AC C
nonlinear dose–response relations: examples, an evaluation of approximations, and software. Am J Epidemiol 2012; 175:66–73. [20] Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med 1989; 8:551–61.
[21] Orsini N, Bellocco RSG. Generalized least squares for trend estimation of summarized dose–response data. Stat J 2006; 6:40–57. [22]Jackson D, White IR, Thompson SG. Extending DerSimonian and Laird’s methodology to perform multivariate random effects meta-analyses. Stat Med 2010; 29:1282–97. [23] Greenland S1, Longnecker MP. Methods for trend estimation from summarized dose–response data, with applications to meta-analysis. Am J Epidemiol 1992; 135:1301–9.
ACCEPTED MANUSCRIPT [24] Ikehara S, Iso H, Yamagishi K, Yamamoto S, Inoue M, Tsugane S; JPHC Study Group. Alcohol consumption, social support,and risk of stroke and coronary heart disease among Japanes e men: the JPHC Study. Alcohol Clin Exp Res 2009; 33(6):1025-32. [25] Bazzano LA1, Gu D, Reynolds K, Chen J et a1. Alcohol consumption and risk of coronary heart disease among Chinese men. Int J Cardiol 2009; 135:78-85.
RI PT
[26] Kitamura A1, Iso H, Sankai T, Naito Y, Sato S et a1. Alcohol intake and premature coronary heart disease in urban Japanese men. Am J Epidemiol 1998; 147:59-65.
[27] Iso H1, Kitamura A, Shimamoto T, Sankai T et a1. Alcohol intake and the risk of
SC
cardiovascular disease in middle-aged Japanese men. Stroke 1995; 26767-773.
[28]Caren G. Solomon,Frank B. Hu, Meir J. Stampfer, Graham A. Colditz,et al. Moderate Alcohol
Circulation 2000; 102:494-499.
M AN U
Consumption and Risk of Coronary Heart Disease Among Women With Type 2 Diabetes Mellitus.
[29] Britton KA1, Gaziano JM, Sesso HD, Djoussé L. Relation of Alcohol Consumption and Coronary Heart Disease in Hypertensive Male Physicians (from the Physicians’ Health Study) .Am J Cardiol 2009; 104:932–935.
[30] Beulens JW1, Rimm EB, Ascherio A, Spiegelman D, Hendriks HF, Mukamal KJ. Alcohol
Med 2007; 146:10-19.
TE D
Consumption and Risk for Coronary Heart Disease among Men with Hypertension. Ann Intern
[31] Arriola L1, Martinez-Camblor P, Larrañaga N, Basterretxea M et al. Alcohol intake and the
EP
risk of coronary heart disease in the Spanish EPIC cohort study. Heart 2010; 96:124–130. [32] Mukamal KJ1, Chiuve SE, Rimm EB. Alcohol Consumption and Risk for Coronary Heart
AC C
Disease in Men With Healthy Lifestyles. Arch Intern Med 2006; 166:2145-2150. [33] Djoussé L1, Lee IM, Buring JE, Gaziano JM. Alcohol Consumption and Risk of Cardiovascular Disease and Mortality in Women: Potential Mediating Mechanisms. Circulation 2009; 120(3): 237–244.
[34] Tanasescu M1, Hu FB, Willett WC, Stampfer MJ, Rimm EB. Alcohol Consumption and Risk of Coronary Heart Disease Among Men With Type 2 Diabetes Mellitus. J Am Coll Cardiol 2001; 38:1836 – 42. [35]Umed A. Ajani, . Michael Gaziano, Paulo A. Lotufo, Simin Liu, MD et al. Alcohol Consumption and Risk of Coronary Heart Disease by Diabetes Status. Circulation 2000; 102:500-505.
ACCEPTED MANUSCRIPT [36] Rajpathak SN1, Freiberg MS, Wang C, Wylie-Rosett J, Wildman RP. Alcohol consumption and the risk of coronary heart disease in postmenopausal women with diabetes: Women’s Health Initiative Observational Study. Eur J Nutr 2010; 49:211–218. [37] Gaziano, J. M. et al. Moderate alcohol intake, increased levels of high-density lipoprotein and its subfractions, and decreased risk of myocardial infarction. N. Engl. J. Med 1993; 329,
RI PT
1829–1834. [38] 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.
SC
[39] Klatsky AL. Alcohol and cardiovascular diseases. Expert Rev Cardiovasc Ther. 2009 May;7(5):499-506.
M AN U
[40] 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(3), 910–915.
[41] Criqui MH, Golomb BA. Epidemiologic aspects of lipid abnormalities. Am. J. Med. 1998; 105(1A), S48–S57.
TE D
[42] Rimm EB, Williams P, Fosher K et al. Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ 1999; 31 9 :1523–1528.
EP
[43] Otocka-Kmiecik A1, Mikhailidis DP, Nicholls SJ, Davidson M, Rysz J, Banach M. Dysfunctional HDL: a novel important diagnostic and therapeutic target in cardiovascular disease?
AC C
Prog Lipid Res. 2012 Oct;51(4):314-24. [44] Rimm EB, Williams P, Fosher K et al. Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ 1999; 31 9 :1523–1528.
[45] Krenz M1, Korthuis RJ. Moderate ethanol ingestion and cardiovascular protection: from epidemiologic associations to cellular mechanisms Mol Cell Cardiol 2012; 52:93–104. [46] Renaud, S. & Gueguen, R. The French paradox and wine drinking. Novartis Found. Symp. 216, 208–217 (1998). [47]Magyar
K1, Halmosi
R, Palfi
A, Feher
G
et
al.
Cardioprotection by resveratrol:
A human clinical trial in patients with stable coronary artery disease.
Clin
Hemorheol
ACCEPTED MANUSCRIPT Microcirc. 2012; 50(3):179-87. [48] Baur JA, Sinclair DA. Therapeutic potential of resvera-trol: the in vivo evidence. Nat Rev Drug Discov 2006; 5:493–506. [49] Opie LH, Lecour S. The red wine hypothesis: from concepts to protective signalling molecules. Eur Heart J 2007; 28:1683–1693.
RI PT
[50] Riccioni G, Gammone MA, Tettamanti G, et al. Resveratrol and anti-atherogenic effects. Int J Food Sci Nutr. 2015 ;66(6):603-10.
[51] O. Rakici, U. Kiziltepe, B. Coskun, S. Aslamaci and F. Akar. Effects of resveratrol on vascular tone and endothelial function of human saphenous vein and internal mammary artery, Int
SC
J Cardiol 2005; 105: 209–215.
[52] D.S. Celermajer, K.E. Sorensen, V.M. Gooch, D.J. Spiegelhalter, O.I. Miller, I.D. Sullivan, et
M AN U
al. Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis, Lancet 1992; 340: 1111–1115.
[53] Sahebkar A1, Serban C2, Ursoniu S3, Wong ND4, Muntner P5, Graham IM6, Mikhailidis DP7, Rizzo M8, Rysz J9, Sperling LS10, Lip GY11, Banach M12; Lipid and Blood Pressure Meta-analysis Collaboration Group. Lack of efficacy of resveratrol on C-reactive protein and
TE D
selected cardiovascular risk factors — Results from a systematic review and meta-analysis of randomized controlled trials. Int J Cardiol. 2015 Jun 15;189:47-55. [54] Della Valle E, Stranges S, Trevisan M et al. Drinking habits and health in Northern Italian and
EP
American men. Nutr. Metab. Cardiovasc. Dis 2009; 19(2), 115–122. [55] Trevisan M, Schisterman E, Mennotti A, Farchi G, Conti S. Risk Factor and Life Expectancy
AC C
Research Group. Drinking pattern and mortality: the Italian Risk Factor and Life Expectancy pooling project. Ann. Epidemiol. 2001; 11(5), 312–319. [56] Stranges S, Wu T, Dorn JM et al. Relationship of alcohol drinking pattern to risk of hypertension: a population-based study. Hypertension 2004; 44(6), 813–819. [57] Larsson SC1, Orsini N. Coffee Consumption and Risk of Stroke: A Dose-Response Meta-Analysis of Prospective Studies. Am J Epidemiol 2011; 174(9):993–1001. [58] Fernández-Solà J. Cardiovascular risks and benefits of moderate and heavy alcohol consumption. Nat Rev Cardiol. 2015; 12(10):576-87.
ACCEPTED MANUSCRIPT Figure legends Fig. 1 Flow chart of study selection. . Fig. 2 The forest plot between highest versus lowest amount of alcohol consumption and risk of CAD.CI, confidence interval; CHD, coronary heart disease; MI, myocardial infarction; UA, unstable angina.
RI PT
Fig. 3 Funnel plot for studies included in the meta-analysis of alcohol consumption and the risk of CAD.
Fig. 4 Association between gram per day of alcohol and risk of CAD. Alcohol
SC
consumption was modeled with restricted cubic splines in a multivariate random-effects dose–response model. The lowest value of zero (non-drinkers) was
M AN U
used to estimate all relative risks. The solid line and the long dash line represent the estimated relative risk and its 95% confidence interval. Short dash line represents
AC C
EP
TE D
the linear relationship
ACCEPTED MANUSCRIPT Table 1 Characteristics of prospective studies of alcohol consumption and risk of CAD study population author(year)
country
Follow-up
Size of
No. of
period
cohort(sex)
cases(sex)
(year)
alcohol consumption Lowest
Relative risk
Highest
Adjusted for
RI PT
age, smoking status, body mass index, history of hypertension,
Satoyo Ikehaa et
Japan
al.(2009)
19356
CAD
(M)
207(M)
history of diabetes, sports at leisure
9.9
Never
450g/week
1.09
time, levels of mental stress,
presence of flushing, presence of
Kathryn A et al.(2009)
USA
26399
CVD
(F)
1039(F)
5164
MI
(M+F )
623(M+F)
64597(M) Lydia A et al.(2009)
China
12.2
17.9
0
CAD
<1drinks/w
>8drinks/wee
eek
k
None
976(M)
None
EP
MI 725(M)
AC C
al.(2010)
USA
3198(F)
15630(M) L Arriola et al.(2009)
35drinks/wee
1.02
CAD
336(M)
128(F)
vegetable intake , menopausal status , and family history of
age, body mass index, smoking, 0.57
exercise, diabetes, multivitamin use, vegetable intake, breakfast cereal intake, and cholesterol
0.58
age, body mass index, average systolic blood pressure, physical
k
activity, cigarette smoking, diabetes education, urban or rural residence, 35drinks/wee
0.46
and living in North China age, race/ethnicity, body mass index, smoking, family history of MI, hypertension, high cholesterol,
7
None
4g/day
0.63
hormone use, regular aspirin use, quintiles of physical activity, duration of diabetes, intake of saturated fat, PUFA, fiber model adjusted for centre, smoking
None
481(M)
CAD
physical activity , fruit and
(yes/no)
>90g/day
0.5
status, height and educational level stratified by age, physical activity
10 25808(F)
age, body mass index , smoking,
k
CAD
Spain
and area.
premature myocardial infarction
8
64597(M)
Swapnil N et
30g/day
M AN U
et al.(2009)
USA
TE D
Luc Djoussé
SC
job, marital status, medical checkups
index, waist/hip ratio, vitamin E, None
30-90g/day
0.21
antithrombotic and anti-hemorrhagic drugs and energy intake
CAD
Diabetics 510(M+F)
Ajani et
USA No diabetes
al.(2000)
21214(M+ F)
120(Diabeti
Never
cs M+F) CAD
daily(
drink per day)
1827(No
mass index, parental history of MI, Never
diabetes
daily(
drink per day)
CAD
None
83(M)
69g/day
0.59
Japan
8476(M)
MI 54(M)
8.8
cigarette smoking, body mass index,
None
69g/day
0.65
None
MI 106(M)
11711(M)
MI 653(M)
16
16
0
0
30g/day
50g/day
0.41
0.7
2792(F)
CAD
Iso et al.(1995)
Japan
M 2890
34(M)
hypercholesterolemia, individual
physical activity, and smoking
0.41
age and smoking
age (5-year categories), time period (7 periods), body mass index (5 categories), cigarette smoking, parental history of MI before age 60 y, hypertension,
14
0
5g/day
0.45
295(F)
CAD
history of hypertension or
levels of body mass index, diet,
Diabetics
EP
USA
Diabetics
age, parental history of myocardial infarction, regular aspirin use,
TE D
USA
8867(M)
69g/day
M AN U
USA
SC
pectoris
hypercholesterolemia, menopausal status/postmenopausal hormone use aspirin use, multivitamin use,
AC C
al.(2000)
left ventricular hypertrophy, and history of diabetes mellitus
32(M)
Solomon et
hypertension, high cholesterol, and
age, serum total cholesterol,
8476(M)
al.(2007)
0.67
diabetes mellitus
angina
Joline W.J et
personal history of angina,
1
al.(1998)
al.(2006)
assignment (aspirin, b-carotene), smoking, physical activity, body
5.5
Akihiko
Kenneth J et
0.66
M+F)
8476(M)
Krtamura et
age, randomized treatment
1
RI PT
Umed A.
ACCEPTED MANUSCRIPT
vitamin E supplement use, and physical activity level (vigorous exercise at least once per week).
age, hypertension, serum total 10.5
0
70g/day
0.8
cholesterol level, smoking and diabetes mellitus
RI PT
ACCEPTED MANUSCRIPT
No cases
No studies
All studies Sex Male Female Country USA Asia Outcome CHD MI No of cases >300 ≤300 Duration of follow-up >10years ≤10years Diabetes mellitus Yes No
7756
18
RR(95%CI)
0.62(0.56-0.68)
Heterogeneity test I-squared(%) P-value 28.5 0.126
12 3
0.63(0.56-0.70) 0.54(0.37-0.71)
28.9 72.4
0.162 0.027
9 7
0.61(0.54-0.69) 0.66(0.55-0.78)
25.7 43.3
0.215 0.102
4487 2232
10 6
0.63(0.56-0.70) 0.53(0.40-0.66)
44.8 0
0.061 0.924
6660 1096
8 10
0.62(0.55-0.69) 0.61(0.49-0.74)
15 42
0.312 0.077
EP
M AN U
Subgroups
SC
Table 2 Stratified summary RRs for the highest versus the lowest category of alcohol intake.
2750 5006
6 12
0.54(0.41-0.67) 0.64(0.57-0.71)
35.4 21.7
0.171 0.23
486 7270
3 15
0.61(0.52-0.70) 0.63(0.54-0.71)
54.4 27.6
0.112 0.153
3758 1462
AC C
TE D
5070 2111
ACCEPTED MANUSCRIPT from
Articles identified from reference lists (n = 2)
Excluded articles (n = 4867) Title and/or abstract were not relevant to the inclusion and exclusion criteria
SC
Full-text articles assessed for eligibility(n=46)
RI PT
Identified articles PubMed (n=4911)
M AN U
Excluded articles (n = 33): Did not evaluate this association(n=20) Case-control studies (n = 6) Reported CAD mortality(n=3) Did not report RR and/OR 95% CI(n=3) Conducted on the same study populations as other included studies (n = 1)
TE D
Articles included in the final analysis(n=13)
AC C
EP
Fig. 1 Flow chart of study selection. .
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig.2 The forest plot between highest versus lowest amount of alcohol consumption and risk of CAD.CI, confidence interval; CHD, coronary heart disease;MI, myocardial infarction;UA, unstable angina.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 3. Funnel plot for studies included in the meta-analysis of alcohol consumption and the risk of CAD.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 4 Association between gram per day of alcohol and risk of CAD.Alcohol consumption was modeled with restricted cubic splines in a multivariate random-effects dose–response model. The lowest value of zero (non-drinkers) was used to estimate all relative risks.The solid line and the long dash line represent the estimated relative risk and its 95% confidence interval.Short dash line represents the linear relationship
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
1. We examined the relationship between alcohol consumption and coronary arter disease. 2. We observed a non-linear association between alcohol consumption and risk of coronary artery disease. 3. 36 grams/day of alcohol conferred a lower risk than other levels.
S0899-9007(15)00474-8
DOI:
10.1016/j.nut.2015.11.013
Reference:
NUT 9660
To appear in:
Nutrition
Received Date: 8 July 2015 Revised Date:
29 November 2015
Accepted Date: 29 November 2015
Please cite this article as: Yang Y, Liu D-C, Wang Q-M, Long Q-Q, Zhao S, Zhang Z, Ma Y, Wang Z-M, Chen L-L, Wang L-S, Alcohol consumption and risk of coronary artery disease: a dose–response metaanalysis of prospective studies, Nutrition (2016), doi: 10.1016/j.nut.2015.11.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Alcohol consumption and risk of coronary artery disease: a dose–response meta-analysis of prospective studies Yang Yang†; Dong-Chen Liu†; Qi-Ming Wang; Qing-Qing Long; Shan Zhao; Zhen Zhang; Yao Ma; Ze-Mu Wang; Lei-Lei Chen*; and Lian-Sheng Wang, *
RI PT
Department of Cardiology, the First Affiliated Hospital of Nanjing Medical
Yang Yang and Dong-Chen Liu contributed equally to this work.
M AN U
†
SC
University, Nanjing, China.
*Corresponding authors: Lian-Sheng Wang, MD, PhD, Department of Cardiology, the First Affiliated Hospital of Nanjing Medical University, 300 Guangzhou Road, Nanjing, 210029, Jiangsu Province, China. E-mail: [email protected], Tel./Fax:
TE D
0086-25-83724440; Lei-Lei Chen, MD, PhD, Department of Cardiology, the First Affiliated Hospital of Nanjing Medical University, 300 Guangzhou Road, Nanjing,
EP
210029, Jiangsu Province, China. E-mail address: [email protected],
AC C
Tel./Fax: 0086-25-83724440.
ACCEPTED MANUSCRIPT Abstract Objective: To investigate and quantify the potential dose-response association between alcohol consumption and risk of coronary artery disease (CAD). Methods and Results: We searched the PubMed database from inception to March 2015 and reviewed the reference list of relevant articles to identify prospective studies
RI PT
assessing the association between alcohol consumption and risk of CAD. Study-specific relative risk (RR) estimates were combined using a random-effects model. Publication bias was estimated using Begg’ funnel plot and Egger’s regression asymmetry test. The meta-analysis included 18 prospective studies, with a total of
SC
214340 participants and 7756 CAD cases. The pooled adjusted RRs were 0.62 [95% confidence interval (CI) 0.56–0.68] for highest alcohol consumption amount versus
M AN U
lowest amount. Begg and Egger regression tests provided no evidence of substantial publication bias (P=0.762 for begg’s test and 0.172 for egger’s test).In a dose-response analysis, We observed a non-linear association between alcohol consumption and risk of CAD(P for non-linearity
0.00 ). Compared with
non-drinkers, the RRs (95% CI) of CAD across levels of alcohol consumption were
TE D
0.75 (0.70-0.80) for 12g/day, 0.70 (0.66-0.75) for 24g/day, 0.69 (0.64-0.75) for 36g/day, 0.70 (0.64-0.77) for 60g/day, 0.74 (0.67-0.83) for 90g /day, and 0.83 (0.67-1.04) for 135g/day.
EP
Conclusion: Alcohol consumption in moderation is associated with a reduced risk of
AC C
CAD with 36 grams/day of alcohol conferring a lower risk than other levels.
Keywords: coronary artery disease; alcohol consumption; meta-analysis; prospective studies
ACCEPTED MANUSCRIPT
Introduction Coronary artery disease (CAD) is now a severe public health problem, affecting
RI PT
millions of people in both developed and developing countries. CAD is the major cause of death in developed countries [1], accounting for up to 40% of all fatal events [2]. In low and middle income countries, the incidence of cardiovascular disease has significantly increased. By 2020, the disease is expected to be the leading cause of
SC
morbidity and mortality in most developing countries[3]. Increased CAD risk is associated with increased atherosclerosis, which is more frequent in patients with
M AN U
CKD, especially the dialysis ones, than in the general population[4]. CAD is by far the most common initiating cause of HF (in ∼ 70% of patients), things even worse[5].
which makes
Alcohol is one of the most widely consumed beverages in the world and is the fifth leading reason for death and disability accounting for 4% of life years lost due to
TE D
disease[6]. Disease burden and disability are closely associated with the volume of alcohol consumption[7]. Consumption of alcohol is believed to have potential health benefits, such as anti-inflammatory effect on atherosclerosis[8], increases in
EP
high-density lipoprotein cholesterol[9], inhibitory effects on platelet aggregation[9], influence on antioxidant capacity and insulin sensitivity, thus restraining
AC C
atherosclerosis[10-11]. Nevertheless, it can also have harmful effects. Alcohol use is associated with a worse hematological picture of total cholesterol, LDL cholesterol (LDL-C) levels, and systolic blood pressure [12]. The association between alcohol consumption and the risk of CAD is of great
importance because of the popularity of alcohol and the huge burden caused by CAD. The role of alcohol in the risk of CAD is well established, and has been assessed in a large number of observational epidemiological studies or meta-analyses[12-16]. Though all demonstrated a U-shaped or J-shaped association between alcohol intake and CAD such that both non-drinkers and heavy drinkers appear to have higher morbidity rates than light and moderate drinkers, few study has designed to
ACCEPTED MANUSCRIPT quantitatively assess the amount of alcohol, referring rather than qualitative levels of alcohol consumption(e.g., “heavy” versus “low”). Thus, we aimed to better quantify the association between alcohol consumption and the risk of CAD through a comprehensive systematic literature review and dose–response meta-analysis that can intuitively reflect the relationship between alcohol consumption and risk of CAD.
RI PT
Methods Literature search and selection:
We identified potentially relevant articles regardless of language by a computerized search of the PubMed database from inception to April 2015.The
SC
following search terms were used: coronary artery disease, coronary heart disease, cardiovascular disease, myocardial infarction, ischemic heart disease, CAD, IHD
M AN U
combined with alcohol consumption, drink, drinking, ethanol. The reference lists of pertinent articles were reviewed to identify additional studies. Two investigators independently reviewed all identified studies. Eligibility criteria for inclusion in the present meta-analysis were: (1) the study was prospective design; (2) the exposure was alcohol consumption; (3) the outcome was total CAD incidence (including MI,
TE D
CAD, non-stroke cardiovascular disease, and other coronary events); (4) the population was free from CAD at baseline; and (5) relative risks (RRs) with 95% confidence intervals (CIs) , adjusted for at least age, were reported. If a study reported
EP
more than one set of RR estimates, we used the maximally adjusted RRs. We considered gender-specific estimates in cases where a study reported data separately
AC C
for men and women. Data Extraction
From each study, we extracted the first author’s last name, country in which the
study was performed, duration of follow-up, sex and age, method of alcohol consumption assessment, total numbers of cases and participants, number of cases and person-years in every exposure category, the most fully adjusted RRs with corresponding 95% CIs for each category of alcohol consumption, and covariates adjusted for in the analysis. If there was disagreement between the two investigators about eligibility of the data, it was resolved by consensus with a third reviewer. Statistical methods
ACCEPTED MANUSCRIPT We standardized alcohol consumption across studies using a common scale, i.e, alcoholic g/day in order to pool the study-specific RRs. When a study reported alcohol consumption in drinks/week, we converted the intake into g/day assuming that one drink contains 12 g of alcohol. For each study, we assigned the median or mean alcohol consumption for the category to each corresponding RR. When the median or
RI PT
mean consumption was not reported, we assigned the midpoint of the upper and lower boundaries in each category as the median consumption. If the upper boundary for the highest category was not provided, the midpoint of the category was set at 1.5 times the lower boundary. When the lowest category was open-ended, we set the lower
SC
boundary to zero. If the number of cases and person-years were not available, we used
obtain a summary estimate.
M AN U
the relative risks comparing the highest versus lowest categories of alcohol intake to
The heterogeneity among studies was estimated by the Cochran Q test and I2 statistic. Heterogeneity was confirmed with a significance level of P≤0.10. The I2 statistic describes the percentage of total variation in point estimates that can be attributed to heterogeneity. For the I2 metric, we considered low, moderate, and high
TE D
I2 values to be 25%, 50%, and 75%, respectively. We used a fixed effect model (Mantel-Haenszel method) when heterogeneity was negligible, and a random effect model (DerSimonian and Laird method) when heterogeneity was significant.
EP
Publication bias was evaluated using Begg’ funnel plot [17] and Egger regression asymmetry test [18].
AC C
To evaluate the dose–response association between alcohol consumption and risk of CAD, a 2-stage random-effects dose–response meta-analysis taking into account the between-study heterogeneity was performed proposed by Orsini et al to compute the trend from the correlated logRR estimates across categories of alcohol consumption[19]. Firstly, a restricted cubic spline model, with 4 knots at the 5th, 35th, 65th, and 95th percentiles of alcohol consumption levels, was estimated using generalized least square regression taking into account the correlation within each set of published RRs[20-21]. Secondly, the restricted maximum likelihood method in a multivariate random-effects meta-analysis was used to combine the study-specific estimates[22]. A P value for nonlinearity and overall significance of the curve was
ACCEPTED MANUSCRIPT calculated using the method proposed by Greenland and Longnecker [23]. All statistical analyses were performed using Stata (version 12.0, StataCorp, College Station, TX, USA).All reported probabilities (P values) were two-sided, with P ≤0.05 considered statistically significant. Results
RI PT
Search Results and Study Characteristics The detailed steps of our literature search are shown in Fig. 1. Finally, 13 articles including 18 studies involving 214340 participants and 7756 CAD cases were used in this meta-analysis. Among the 18 prospective studies, 9 studies reported the result
SC
from United States, 5 studies from Japan, 2 studies from China, and 2 studies from
High Versus Low Analyses
M AN U
Spain. The characteristics of these studies are presented in Table 1.
Data from 13 articles including 18 prospective studies were used in this meta-analysis [24-36]. Pooled results suggested that highest alcohol consumption amount versus lowest amount was 0.62(95% CI 0.56-0.68; I-squared 28.5%,p=0.126). (Fig.2).Additionally, Begg and Egger regression tests provided no evidence of
TE D
substantial publication bias (P=0.762 for begg’s test and 0.172 for egger’s test). Fig. 3 presents the funnel plot of log RR versus variance of log RR for all the studies included in the meta-analysis.
EP
In subgroup analyses for country, inverse associations of alcohol consumption and CAD risk were found both in USA[summary RR = 0.61, 95%CI = 0.54-0.69] and
AC C
in Asia [summary RR = 0.66, 95%CI = 0.55-0.78]. When we conducted the subgroup analysis by follow-up duration ( >10 years and ≤10years), significant associations were found both in >10 years and ≤10years follow-up. When stratified analysis for study outcome (CAD and MI), sex, number of cases and diabetes mellitus(Yes or No),similar results were also found. The detailed results are summarized in Table 2. Dose–response Analysis For dose–response analysis, data from 13 articles including 18 prospective studies involving 7756 CAD cases were used for the dose-response relation between alcohol and CAD risk. We observed a non-linear association between alcohol consumption and risk of CAD(P for non-linearity 0.00 )(Fig. 4). Compared with
ACCEPTED MANUSCRIPT non-drinkers, the RRs (95% CI) of CAD across levels of alcohol consumption were 0.75 (0.70-0.80) for 12g/day, 0.70(0.66-0.75) for 24g/day, 0.69 (0.64-0.75) for 36g/day, 0.70 (0.64-0.77) for 60g/day, 0.74 (0.67-0.83) for 90g /day, and 0.83 (0.67-1.04) for 135g/day. Discussion
RI PT
Results from this meta-analysis showed that alcohol consumption in moderation is associated with a reduced risk of CAD. The dose–response meta-analysis indicated that consumption of 36 grams per day is associated with a 31% lower risk of CAD. Our dose–response meta-analysis suggested a non-linear association of alcohol
SC
consumption with risk of CAD, which is consistent with the observed U-shaped relationship between alcohol consumption and CAD [16-17]. A possible explanation
M AN U
for the non-linear association is that consumption of large amounts of alcohol may increase blood pressure, low-density lipoprotein (LDL) cholesterol, and triglycerides, which are risk factors for CAD[7].
There are several possible mechanisms for protection against CAD by alcohol. The best established one is related to the levels of plasma lipoproteins, including
TE D
increases in high-density lipoprotein (HDL) cholesterol and Apo lipoprotein A-I, and, to a lesser degree, a decrease in LDL-cholesterol [37-38]. The central roles of lipids in CAD development include a positive relation with LDL-cholesterol and an inverse
EP
relation with HDL-cholesterol. Inverse relations of HDL to CAD risk primarily operate via the removal of lipid deposits in large blood vessels, but HDL also binds
AC C
with cholesterol in the tissues and may result in preventing tissue oxidation of LDL-cholesterol[39]. Several studies examining the hypothesis that higher HDL-cholesterol in drinkers mediates CAD protection suggested that higher HDL levels mediated approximately half of the lower CAD risk [40–42]. There is data suggesting that alcohol might increase HDL but it also might change it functionality into dysfunctional HDL [43]. Besides, light to moderate alcohol consumption may decrease platelet aggregation, decrease fibrinogen, increase tissue plasminogen activator, and improve insulin sensitivity[44-45]. Antithrombotic actions of alcohol could partially account for the lower CAD risk at very light drinking level. Different beverages may have different effects upon CAD. The French Paradox
ACCEPTED MANUSCRIPT suggests that red wine confers a greater cardiovascular benefit than other alcoholic beverages[46]. Red wine contains high amount of polyphenolic compounds like resveratrol (RES) (trans-3,4,5-trihydroxystilbene), epicatechin, catechin, gallic acid, quercetin. Primarily RES is thought to be responsible for the cardioprotective effect of red wine[47]. RES, one of the main polyphenol extracts in grape skin/seed, red wine,
RI PT
and the root of Polygonum cuspidatum, is found to have protective potentials to multi-targets related to CAD [48-49]. The protective effect of RES has been shown to involve various mechanisms. RES improves endothelial function, and has numerous beneficial effects on vascular tone and vessels in human and animal models[50]. It has
SC
also been shown that RES improves the release of nitric oxide (NO) and prostacyclin (PGI) which plays a prominent role in the maintenance of endothelial function [51].
M AN U
Besides, several studies supported its antioxidant activity, its ability to decrease LDL oxidation, and function as a direct free radical scavenger [52]. The effect of RES on CAD is controversial. A meta-analysis of 10 randomized controlled trials did not indicate any benefit of RES supplementation on plasma CRP. Furthermore, RES supplementation was not found to alter plasma levels of TC, LDL-C, TG and glucose,
TE D
and additionally slightly decreased HDL-C concentrations[53]. A reason for these effects could be that serum levels of RES in lower dose may not be adequate for a beneficial effect[53].
EP
Besides the type, the drinking pattern also plays an important role in the effect of alcohol. Drinking-pattern aspects with medical implications for CAD are the
AC C
frequency of drinking (number of days per week), variability over time and whether the alcohol is taken with food [39]. Data exist that support that alcohol drinking at mealtimes is more favorable for protecting against CAD and hypertension than drinking apart from meals and that frequency of intake may be a stronger factor in benefit than amount [54-56]. The main strength of this meta-analysis is its dose-response design that provides better quantification of the associations between specified amounts of alcohol and risk of CAD. A dose-response meta-analysis should be the first choice when performing a systematic review, rather than conducting a meta-analysis based only on the comparison of the extreme categories of consumption (highest versus lowest), which
ACCEPTED MANUSCRIPT could vary considerably among studies. Another strength of this meta-analysis is that only prospective studies were included. The prospective design minimized the potential influence of recall and selection bias, which could be a problem in retrospective case–control studies. Nevertheless, our meta-analysis also had several limitations. First, although most
RI PT
studies controlled for major risk factors for CAD, we cannot rule out the possibility that the results were affected by confounding completely, because the studies were observational. Another limitation is the misclassification of alcohol consumption, which was inevitable considering that alcohol consumption was self-reported.
SC
Misclassification of alcohol consumption because of measurement error or changes in consumption during follow-up would most likely lead to inaccurate estimation of the
M AN U
true association between alcohol consumption and CAD risk[57]. Third, our meta-analysis focus our attention only on the amount of total alcohol, the type of beverage (wine, beer or spirits), the amount that produces the putative benefit of alcoholic beverages and the drinking pattern are ignored.
The association we observed between alcohol consumption and risk of CAD is
TE D
consistent with the U-shaped curve which describes a relationship where the relative risk decreases at low-to-moderate alcohol intake, then rises as alcohol intake increases, but does not rise above the level of the reference population of abstainers[58]. Though
EP
the RR (95% CI) of CAD was 0.83 (0.67-1.04) for 135g/day compared with non-drinkers, we absolutely do not support heavy drink as heavy drink is associated cardiomyopathy, systemic hypertension,
AC C
with increased cardiovascular risks
supraventricular arrhythmias, hemorrhagic stroke and heart failure that is not associated with CAD. No final conclusion has yet been reached regarding the recommendation of alcohol consumption to improve cardiovascular health. Advice regarding the health effects of alcohol drinking needs to be individualized according to the specific medical history and risks of any concerned person. However, a few common-sense rules are suggested: (1) The overall health risk of a heavy drinker is likely to be reduced by intake reduction or abstinence. (2) Owing to the unknown risk of progression to heavier drinking, abstainers should not indiscriminately be advised to drink for CAD health benefit. (3) Established light-to-moderate drinkers,
ACCEPTED MANUSCRIPT accounting for the majority of drinkers in the USA and Western Europe, do not need to change their drinking habits, except in special circumstances [39]. In summary, this meta-analysis provides evidence that alcohol consumption in moderation is associated with a reduced risk of CAD with 36 grams/day of alcohol
Acknowledgements
TE D
M AN U
SC
RI PT
conferring a lower risk than other levels.
This work was supported by grants from the National Natural Science Foundation of
Group
AC C
Innovation
EP
China (No 81270255), the Project Funded by the Science and Technological of
Jiangsu
Higher
Project”(JX2161015030). Conflict of interest
The authors state no conflict of interest.
Education
Institution
“Qing-Lan
ACCEPTED MANUSCRIPT
References
[1] Chen L, Wang X, Bao J, Geng C, Xia Y, Wang J. Direct comparison of cardiovascular magnetic resonance and single-photon emission computed tomography for detection of coronary artery disease: a meta-analysis. PLoS One 2014; 9:e88402. [2] Franchini M1, Peyvandi F, Mannucci PM. The genetic basis of coronary artery disease: from
RI PT
candidate genes to whole genome analysis. Trends Cardiovasc Med 2008; 18:157–62 . [3] Celermajer DS, Chow CK, Marijon E, Anstey NM, Woo KS. Cardiovascular disease in the developing world: prevalences, patterns, and the potential of early disease detection.J am CollCardiol 2012;60:1207-16.
SC
[4] Bielecka-Dabrowa A1, von Haehling S, Aronow WS, Ahmed MI, Rysz J, Banach M. Heart failure biomarkers in patients with dilated cardiomyopathy. Int J Cardiol. 2013; 168(3): 2404-10.
M AN U
[5] Gluba-Brzózka A1, Michalska-Kasiczak M, Franczyk-Skóra B, Nocuń M, Banach M, Rysz J. Markers of increased cardiovascular risk in patients with chronic kidney disease. Lipids Health Dis 2014; 13:135.
[6] Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in
2012; 380:2224-60.
TE D
21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study . Lancet
[7] Rehm J, Mathers C, Popova S, Thavorncharoensap M, Teerawattananon Y, Patra J. Global
EP
burden of disease and injury and economic cost attributable to alcohol use and alcohol-us e disorders. Lancet 2009; 373(9682):2223 – 33.
AC C
[8] Jerrard-Dunne P1, Sitzer M, Risley P, Steckel DA, Buehler A, von Kegler S, Markus HS. Carotid Atherosclerosis Progression Study. Interleukin-6 promoter polymorphism modulates the effects of heavy alcohol consumption on early carotid artery atherosclerosis: the Carotid Atherosclerosis Progression Study (CAPS).Stroke 2003; 34:402-7. [9] Rimm EB1, 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; 319:1523-8. [10] Booyse FM1, Pan W, Grenett HE, Parks DA, Darley-Usmar VM, Bradley KM, Tabengwa EM. Mechanism by which alcohol and wine polyphenols affect coronary heart disease risk. Ann Epidemiol 2007; 17(5 Suppl):S24-31.
ACCEPTED MANUSCRIPT [11] Vasdev S1, Gill V, Singal PK. Beneficial effect of low ethanol intake on the cardiovascular system: possible biochemical mechanisms. Vasc Health Risk Manag 2006; 2:263-76. [12] Perissinotto E, Buja A, Maggi S, Enzi G, Manzato E et al. Alcohol consumption and cardiovascular risk factors in older lifelong wine drinkers: the Italian Longitudinal Study on Aging. Nutr Metab Cardiovasc Dis 2010; 20(9):647 – 55.
RI PT
[13] Corrao G, Rubbiati L, Bagnardi V, Zambon A, Poikolainen K. Alcohol and coronary heart disease: a meta-analysis. Addiction 2000; 95:1505-23.
[14] Ronksley PE, Brien SE, Turner BJ, Mukamal KJ, Ghali WA. Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and
SC
meta-analysis. BMJ 2011; 342:d671.
[15] Tolstrup J, Jensen MK, Tjønneland A, Overvad K, Mukamal KJ, Grønbaek M. Prospective
M AN U
study of alcohol drinking patterns and coronary heart disease in women and men. BMJ 2006; 27;332(7552): 1244-8
[16] Rehm J1, Sempos CT, Trevisan M. Alcohol and cardiovascular disease - more than one paradox to consider. Average volume of alcohol consumption, patterns of drinking and risk of coronary heart disease-a review. J Cardiovasc Risk 2003; 10(1):15-20.
TE D
[17] Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994; 50:1088–101.
[18] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple,
EP
graphical test. BMJ 1997; 315:629–34.
[19] Orsini N1, Li R, Wolk A, Khudyakov P, Spiegelman D. Meta-analysis for
linear and
AC C
nonlinear dose–response relations: examples, an evaluation of approximations, and software. Am J Epidemiol 2012; 175:66–73. [20] Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med 1989; 8:551–61.
[21] Orsini N, Bellocco RSG. Generalized least squares for trend estimation of summarized dose–response data. Stat J 2006; 6:40–57. [22]Jackson D, White IR, Thompson SG. Extending DerSimonian and Laird’s methodology to perform multivariate random effects meta-analyses. Stat Med 2010; 29:1282–97. [23] Greenland S1, Longnecker MP. Methods for trend estimation from summarized dose–response data, with applications to meta-analysis. Am J Epidemiol 1992; 135:1301–9.
ACCEPTED MANUSCRIPT [24] Ikehara S, Iso H, Yamagishi K, Yamamoto S, Inoue M, Tsugane S; JPHC Study Group. Alcohol consumption, social support,and risk of stroke and coronary heart disease among Japanes e men: the JPHC Study. Alcohol Clin Exp Res 2009; 33(6):1025-32. [25] Bazzano LA1, Gu D, Reynolds K, Chen J et a1. Alcohol consumption and risk of coronary heart disease among Chinese men. Int J Cardiol 2009; 135:78-85.
RI PT
[26] Kitamura A1, Iso H, Sankai T, Naito Y, Sato S et a1. Alcohol intake and premature coronary heart disease in urban Japanese men. Am J Epidemiol 1998; 147:59-65.
[27] Iso H1, Kitamura A, Shimamoto T, Sankai T et a1. Alcohol intake and the risk of
SC
cardiovascular disease in middle-aged Japanese men. Stroke 1995; 26767-773.
[28]Caren G. Solomon,Frank B. Hu, Meir J. Stampfer, Graham A. Colditz,et al. Moderate Alcohol
Circulation 2000; 102:494-499.
M AN U
Consumption and Risk of Coronary Heart Disease Among Women With Type 2 Diabetes Mellitus.
[29] Britton KA1, Gaziano JM, Sesso HD, Djoussé L. Relation of Alcohol Consumption and Coronary Heart Disease in Hypertensive Male Physicians (from the Physicians’ Health Study) .Am J Cardiol 2009; 104:932–935.
[30] Beulens JW1, Rimm EB, Ascherio A, Spiegelman D, Hendriks HF, Mukamal KJ. Alcohol
Med 2007; 146:10-19.
TE D
Consumption and Risk for Coronary Heart Disease among Men with Hypertension. Ann Intern
[31] Arriola L1, Martinez-Camblor P, Larrañaga N, Basterretxea M et al. Alcohol intake and the
EP
risk of coronary heart disease in the Spanish EPIC cohort study. Heart 2010; 96:124–130. [32] Mukamal KJ1, Chiuve SE, Rimm EB. Alcohol Consumption and Risk for Coronary Heart
AC C
Disease in Men With Healthy Lifestyles. Arch Intern Med 2006; 166:2145-2150. [33] Djoussé L1, Lee IM, Buring JE, Gaziano JM. Alcohol Consumption and Risk of Cardiovascular Disease and Mortality in Women: Potential Mediating Mechanisms. Circulation 2009; 120(3): 237–244.
[34] Tanasescu M1, Hu FB, Willett WC, Stampfer MJ, Rimm EB. Alcohol Consumption and Risk of Coronary Heart Disease Among Men With Type 2 Diabetes Mellitus. J Am Coll Cardiol 2001; 38:1836 – 42. [35]Umed A. Ajani, . Michael Gaziano, Paulo A. Lotufo, Simin Liu, MD et al. Alcohol Consumption and Risk of Coronary Heart Disease by Diabetes Status. Circulation 2000; 102:500-505.
ACCEPTED MANUSCRIPT [36] Rajpathak SN1, Freiberg MS, Wang C, Wylie-Rosett J, Wildman RP. Alcohol consumption and the risk of coronary heart disease in postmenopausal women with diabetes: Women’s Health Initiative Observational Study. Eur J Nutr 2010; 49:211–218. [37] Gaziano, J. M. et al. Moderate alcohol intake, increased levels of high-density lipoprotein and its subfractions, and decreased risk of myocardial infarction. N. Engl. J. Med 1993; 329,
RI PT
1829–1834. [38] 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.
SC
[39] Klatsky AL. Alcohol and cardiovascular diseases. Expert Rev Cardiovasc Ther. 2009 May;7(5):499-506.
M AN U
[40] 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(3), 910–915.
[41] Criqui MH, Golomb BA. Epidemiologic aspects of lipid abnormalities. Am. J. Med. 1998; 105(1A), S48–S57.
TE D
[42] Rimm EB, Williams P, Fosher K et al. Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ 1999; 31 9 :1523–1528.
EP
[43] Otocka-Kmiecik A1, Mikhailidis DP, Nicholls SJ, Davidson M, Rysz J, Banach M. Dysfunctional HDL: a novel important diagnostic and therapeutic target in cardiovascular disease?
AC C
Prog Lipid Res. 2012 Oct;51(4):314-24. [44] Rimm EB, Williams P, Fosher K et al. Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ 1999; 31 9 :1523–1528.
[45] Krenz M1, Korthuis RJ. Moderate ethanol ingestion and cardiovascular protection: from epidemiologic associations to cellular mechanisms Mol Cell Cardiol 2012; 52:93–104. [46] Renaud, S. & Gueguen, R. The French paradox and wine drinking. Novartis Found. Symp. 216, 208–217 (1998). [47]Magyar
K1, Halmosi
R, Palfi
A, Feher
G
et
al.
Cardioprotection by resveratrol:
A human clinical trial in patients with stable coronary artery disease.
Clin
Hemorheol
ACCEPTED MANUSCRIPT Microcirc. 2012; 50(3):179-87. [48] Baur JA, Sinclair DA. Therapeutic potential of resvera-trol: the in vivo evidence. Nat Rev Drug Discov 2006; 5:493–506. [49] Opie LH, Lecour S. The red wine hypothesis: from concepts to protective signalling molecules. Eur Heart J 2007; 28:1683–1693.
RI PT
[50] Riccioni G, Gammone MA, Tettamanti G, et al. Resveratrol and anti-atherogenic effects. Int J Food Sci Nutr. 2015 ;66(6):603-10.
[51] O. Rakici, U. Kiziltepe, B. Coskun, S. Aslamaci and F. Akar. Effects of resveratrol on vascular tone and endothelial function of human saphenous vein and internal mammary artery, Int
SC
J Cardiol 2005; 105: 209–215.
[52] D.S. Celermajer, K.E. Sorensen, V.M. Gooch, D.J. Spiegelhalter, O.I. Miller, I.D. Sullivan, et
M AN U
al. Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis, Lancet 1992; 340: 1111–1115.
[53] Sahebkar A1, Serban C2, Ursoniu S3, Wong ND4, Muntner P5, Graham IM6, Mikhailidis DP7, Rizzo M8, Rysz J9, Sperling LS10, Lip GY11, Banach M12; Lipid and Blood Pressure Meta-analysis Collaboration Group. Lack of efficacy of resveratrol on C-reactive protein and
TE D
selected cardiovascular risk factors — Results from a systematic review and meta-analysis of randomized controlled trials. Int J Cardiol. 2015 Jun 15;189:47-55. [54] Della Valle E, Stranges S, Trevisan M et al. Drinking habits and health in Northern Italian and
EP
American men. Nutr. Metab. Cardiovasc. Dis 2009; 19(2), 115–122. [55] Trevisan M, Schisterman E, Mennotti A, Farchi G, Conti S. Risk Factor and Life Expectancy
AC C
Research Group. Drinking pattern and mortality: the Italian Risk Factor and Life Expectancy pooling project. Ann. Epidemiol. 2001; 11(5), 312–319. [56] Stranges S, Wu T, Dorn JM et al. Relationship of alcohol drinking pattern to risk of hypertension: a population-based study. Hypertension 2004; 44(6), 813–819. [57] Larsson SC1, Orsini N. Coffee Consumption and Risk of Stroke: A Dose-Response Meta-Analysis of Prospective Studies. Am J Epidemiol 2011; 174(9):993–1001. [58] Fernández-Solà J. Cardiovascular risks and benefits of moderate and heavy alcohol consumption. Nat Rev Cardiol. 2015; 12(10):576-87.
ACCEPTED MANUSCRIPT Figure legends Fig. 1 Flow chart of study selection. . Fig. 2 The forest plot between highest versus lowest amount of alcohol consumption and risk of CAD.CI, confidence interval; CHD, coronary heart disease; MI, myocardial infarction; UA, unstable angina.
RI PT
Fig. 3 Funnel plot for studies included in the meta-analysis of alcohol consumption and the risk of CAD.
Fig. 4 Association between gram per day of alcohol and risk of CAD. Alcohol
SC
consumption was modeled with restricted cubic splines in a multivariate random-effects dose–response model. The lowest value of zero (non-drinkers) was
M AN U
used to estimate all relative risks. The solid line and the long dash line represent the estimated relative risk and its 95% confidence interval. Short dash line represents
AC C
EP
TE D
the linear relationship
ACCEPTED MANUSCRIPT Table 1 Characteristics of prospective studies of alcohol consumption and risk of CAD study population author(year)
country
Follow-up
Size of
No. of
period
cohort(sex)
cases(sex)
(year)
alcohol consumption Lowest
Relative risk
Highest
Adjusted for
RI PT
age, smoking status, body mass index, history of hypertension,
Satoyo Ikehaa et
Japan
al.(2009)
19356
CAD
(M)
207(M)
history of diabetes, sports at leisure
9.9
Never
450g/week
1.09
time, levels of mental stress,
presence of flushing, presence of
Kathryn A et al.(2009)
USA
26399
CVD
(F)
1039(F)
5164
MI
(M+F )
623(M+F)
64597(M) Lydia A et al.(2009)
China
12.2
17.9
0
CAD
<1drinks/w
>8drinks/wee
eek
k
None
976(M)
None
EP
MI 725(M)
AC C
al.(2010)
USA
3198(F)
15630(M) L Arriola et al.(2009)
35drinks/wee
1.02
CAD
336(M)
128(F)
vegetable intake , menopausal status , and family history of
age, body mass index, smoking, 0.57
exercise, diabetes, multivitamin use, vegetable intake, breakfast cereal intake, and cholesterol
0.58
age, body mass index, average systolic blood pressure, physical
k
activity, cigarette smoking, diabetes education, urban or rural residence, 35drinks/wee
0.46
and living in North China age, race/ethnicity, body mass index, smoking, family history of MI, hypertension, high cholesterol,
7
None
4g/day
0.63
hormone use, regular aspirin use, quintiles of physical activity, duration of diabetes, intake of saturated fat, PUFA, fiber model adjusted for centre, smoking
None
481(M)
CAD
physical activity , fruit and
(yes/no)
>90g/day
0.5
status, height and educational level stratified by age, physical activity
10 25808(F)
age, body mass index , smoking,
k
CAD
Spain
and area.
premature myocardial infarction
8
64597(M)
Swapnil N et
30g/day
M AN U
et al.(2009)
USA
TE D
Luc Djoussé
SC
job, marital status, medical checkups
index, waist/hip ratio, vitamin E, None
30-90g/day
0.21
antithrombotic and anti-hemorrhagic drugs and energy intake
CAD
Diabetics 510(M+F)
Ajani et
USA No diabetes
al.(2000)
21214(M+ F)
120(Diabeti
Never
cs M+F) CAD
daily(
drink per day)
1827(No
mass index, parental history of MI, Never
diabetes
daily(
drink per day)
CAD
None
83(M)
69g/day
0.59
Japan
8476(M)
MI 54(M)
8.8
cigarette smoking, body mass index,
None
69g/day
0.65
None
MI 106(M)
11711(M)
MI 653(M)
16
16
0
0
30g/day
50g/day
0.41
0.7
2792(F)
CAD
Iso et al.(1995)
Japan
M 2890
34(M)
hypercholesterolemia, individual
physical activity, and smoking
0.41
age and smoking
age (5-year categories), time period (7 periods), body mass index (5 categories), cigarette smoking, parental history of MI before age 60 y, hypertension,
14
0
5g/day
0.45
295(F)
CAD
history of hypertension or
levels of body mass index, diet,
Diabetics
EP
USA
Diabetics
age, parental history of myocardial infarction, regular aspirin use,
TE D
USA
8867(M)
69g/day
M AN U
USA
SC
pectoris
hypercholesterolemia, menopausal status/postmenopausal hormone use aspirin use, multivitamin use,
AC C
al.(2000)
left ventricular hypertrophy, and history of diabetes mellitus
32(M)
Solomon et
hypertension, high cholesterol, and
age, serum total cholesterol,
8476(M)
al.(2007)
0.67
diabetes mellitus
angina
Joline W.J et
personal history of angina,
1
al.(1998)
al.(2006)
assignment (aspirin, b-carotene), smoking, physical activity, body
5.5
Akihiko
Kenneth J et
0.66
M+F)
8476(M)
Krtamura et
age, randomized treatment
1
RI PT
Umed A.
ACCEPTED MANUSCRIPT
vitamin E supplement use, and physical activity level (vigorous exercise at least once per week).
age, hypertension, serum total 10.5
0
70g/day
0.8
cholesterol level, smoking and diabetes mellitus
RI PT
ACCEPTED MANUSCRIPT
No cases
No studies
All studies Sex Male Female Country USA Asia Outcome CHD MI No of cases >300 ≤300 Duration of follow-up >10years ≤10years Diabetes mellitus Yes No
7756
18
RR(95%CI)
0.62(0.56-0.68)
Heterogeneity test I-squared(%) P-value 28.5 0.126
12 3
0.63(0.56-0.70) 0.54(0.37-0.71)
28.9 72.4
0.162 0.027
9 7
0.61(0.54-0.69) 0.66(0.55-0.78)
25.7 43.3
0.215 0.102
4487 2232
10 6
0.63(0.56-0.70) 0.53(0.40-0.66)
44.8 0
0.061 0.924
6660 1096
8 10
0.62(0.55-0.69) 0.61(0.49-0.74)
15 42
0.312 0.077
EP
M AN U
Subgroups
SC
Table 2 Stratified summary RRs for the highest versus the lowest category of alcohol intake.
2750 5006
6 12
0.54(0.41-0.67) 0.64(0.57-0.71)
35.4 21.7
0.171 0.23
486 7270
3 15
0.61(0.52-0.70) 0.63(0.54-0.71)
54.4 27.6
0.112 0.153
3758 1462
AC C
TE D
5070 2111
ACCEPTED MANUSCRIPT from
Articles identified from reference lists (n = 2)
Excluded articles (n = 4867) Title and/or abstract were not relevant to the inclusion and exclusion criteria
SC
Full-text articles assessed for eligibility(n=46)
RI PT
Identified articles PubMed (n=4911)
M AN U
Excluded articles (n = 33): Did not evaluate this association(n=20) Case-control studies (n = 6) Reported CAD mortality(n=3) Did not report RR and/OR 95% CI(n=3) Conducted on the same study populations as other included studies (n = 1)
TE D
Articles included in the final analysis(n=13)
AC C
EP
Fig. 1 Flow chart of study selection. .
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig.2 The forest plot between highest versus lowest amount of alcohol consumption and risk of CAD.CI, confidence interval; CHD, coronary heart disease;MI, myocardial infarction;UA, unstable angina.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 3. Funnel plot for studies included in the meta-analysis of alcohol consumption and the risk of CAD.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 4 Association between gram per day of alcohol and risk of CAD.Alcohol consumption was modeled with restricted cubic splines in a multivariate random-effects dose–response model. The lowest value of zero (non-drinkers) was used to estimate all relative risks.The solid line and the long dash line represent the estimated relative risk and its 95% confidence interval.Short dash line represents the linear relationship
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
1. We examined the relationship between alcohol consumption and coronary arter disease. 2. We observed a non-linear association between alcohol consumption and risk of coronary artery disease. 3. 36 grams/day of alcohol conferred a lower risk than other levels.