Alcohol consumption, drinking patterns, and ischemic heart disease: a narrative review of meta-analyses and a systematic review and meta-analysis of the impact of heavy drinking occasions on risk for moderate drinkers

Background Alcohol consumption is a major global risk factor for mortality and morbidity. Much discussion has revolved around the diverse findings on the complex relationship between alcohol consumption and the leading cause of death and disability, ischemic heart disease (IHD). Methods We conducted a systematic search of the literature up to August 2014 using Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines to identify meta-analyses and observational studies examining the relationship between alcohol drinking, drinking patterns, and IHD risk, in comparison to lifetime abstainers. In a narrative review we have summarized the many meta-analyses published in the last 10 years, discussing the role of confounding and experimental evidence. We also conducted meta-analyses examining episodic heavy drinking among on average moderate drinkers. Results The narrative review showed that the use of current abstainers as the reference group leads to systematic bias. With regard to average alcohol consumption in relation to lifetime abstainers, the relationship is clearly J-shaped, supported by short-term experimental evidence and similar associations within strata of potential confounders, except among smokers. Women experience slightly stronger beneficial associations and also a quicker upturn to a detrimental effect at lower levels of average alcohol consumption compared to men. There was no evidence that chronic or episodic heavy drinking confers a beneficial effect on IHD risk. People with alcohol use disorder have an elevated risk of IHD (1.5- to 2-fold). Results from our quantitative meta-analysis showed that drinkers with average intake of <30 g/day and no episodic heavy drinking had the lowest IHD risk (relative risk = 0.64, 95% confidence interval 0.53 to 0.71). Drinkers with episodic heavy drinking occasions had a risk similar to lifetime abstainers (relative risk = 1.12, 95% confidence interval 0.91 to 1.37). Conclusions Epidemiological evidence for a beneficial effect of low alcohol consumption without heavy drinking episodes is strong, corroborated by experimental evidence. However, episodic and chronic heavy drinking do not provide any beneficial effect on IHD. Thus, average alcohol consumption is not sufficient to describe the risk relation between alcohol consumption and IHD. Alcohol policy should try to reduce heavy drinking patterns. Electronic supplementary material The online version of this article (doi:10.1186/s12916-014-0182-6) contains supplementary material, which is available to authorized users.


Additional files
Supplement to: Roerecke M & Rehm J. Alcohol consumption, drinking patterns, and ischaemic heart disease: a narrative review of meta-analyses and a systematic review and meta-analysis of the impact of heavy drinking occasions on risk for moderate drinkers.

Content list
Text S1. Systematic review protocols for narrative and quantitative reviews  Table S1. Characteristics of 7 studies on IHD risk in drinkers of 1-30 g/day average alcohol consumption in comparison to lifetime abstainers Figure S3. IHD risk among non-heavy drinkers with total average intake of 1-30 g/day compared to lifetime abstainers Figure S4. IHD risk among episodic heavy drinkers with total average intake of 1-30 g/day compared to lifetime abstainers

Review question 1
What systematic evidence exists for the relationship between alcohol consumption and IHD risk?

Literature searches
Using PRISMA guidelines [1], we searched electronic databases from 1980 to second week of August 2014 for meta-analyses on total alcohol consumption and IHD risk ( Figure S1).

Types of studies to be included initially
Meta-analyses.

Strategy for data synthesis
The most comprehensive and recent meta-analysis reporting data by dimensions of alcohol consumption (lifetime drinking status, average consumption, drinking pattern) by sex and IHD endpoint (mortality vs. morbidity) was used in a narrative review.

Review question 2
What is the relative risk for IHD among heavy-and non-heavy alcohol drinkers?

Literature search
Using PRISMA guidelines [1], we searched electronic databases from 1980 to fourth week of March 2014 for original articles, excluding letters, editorials, conference abstracts, reviews, and comments for variations of search terms for the exposure (alcohol consumption), outcome (IHD), and study design from 1980 to fourth week of March 2014 ( Figure S2). Additionally, we hand searched references of identified papers and relevant reviews and metaanalyses.

Participants/population
Inclusion criteria: Adults (≥18 years) from population samples, IHD was analyzed as a separate outcome (ICD-9: 410-414, ICD-10: I20-25), case-control or prospective or historical cohort study design, exposure measurement had to cover a reference period of more than 2 weeks for average alcohol consumption at baseline, a drinking group that either specifically excluded or included episodic heavy drinking among current drinkers with an average alcohol consumption <30 g of pure alcohol per day, a measure of risk in comparison to lifetime abstainers and its corresponding measure of variability was reported (or sufficient data to calculate these), and English-, German-, or Spanish-language.
Exclusion criteria: Adolescents (<18 years), population samples from people with IHD-related conditions. We excluded self-reported IHD outcomes, as well as studies reporting estimates on cardiovascular outcomes combined rather than IHD separately and studies with precursors as outcome.

Intervention/exposure
Non-heavy drinking (1-2 drinks on average and usual consumption <30 g/occasion), episodic heavy drinking (1-2 drinks on average and 5+ drinks per occasion or intoxication) are the exposures of interest.

Comparators/controls
Measure of relative risk in comparison to lifetime abstainers in population studies.

Types of studies to be included initially
Observational studies on alcohol consumption and ischaemic heart disease. (alcohol drinking or alcoholic beverages or heavy drinking occasion* or heavy episodic drinking or binge drinking or alcoholic intoxication or problem drinking or hangover* or irregular drinking or drinking pattern or inebriation).mp. 5 exp drinking behavior/ or exp alcohol drinking/ or exp binge drinking/ 6 4 or 5 7 (myocardial ischemia or myocardial infarction or myocardial infarct$ or coronary disease or heart diseases or coronary artery disease or coronary heart disease or angina or cardiac death$ or ischaemic heart disease or ischemic heart disease or cardiac event$ or coronary event$).mp. 8 exp myocardial ischemia/ or exp coronary artery disease/ 9 7 or 8 10 exp Case-Control Studies/ 11 exp cohort studies/ or exp follow-up studies/ or exp longitudinal studies/ or exp prospective studies/ or exp retrospective studies/ 12 exp risk/ 13 10 or 11 or 12 14 3 and 6 and 9 and 13 15 limit 14 to yr="1980 -2014"

Data extraction
From all relevant articles we extracted authors' names, year of publication, country, calendar year(s) of baseline examination, follow-up period, setting, assessment of IHD and alcohol consumption, mean and range of age at baseline, sex, number of observed IHD cases or deaths among participants by drinking group, number of total participants by drinking group, adjustment for potential confounders, and relative risk and its standard error. We used the most adjusted relative risk reported. Information found in related papers from the same cohort was used where possible. The first author performed the literature search and abstracted the data. Full-text articles with uncertain eligibility were discussed by both authors until consensus was reached. Primary authors were not contacted in case there was not enough information presented in the article.

Risk of bias
Most quality scores are tailored for meta-analyses of randomized trials of interventions [2][3][4][5] and many criteria do not apply to epidemiological studies like the ones examined here. Also, their use in meta-analyses remains controversial [5,6]. Thus, quality assessment was incorporated differently by including quality components such as study design and alcohol measurement into the inclusion and exclusion criteria (please see Data abstraction and Table S1 for details). Quality checklists therefore would not have been able to distinguish the quality of selected studies in our analysis.

Strategy for data synthesis
Hazard ratios, odds ratios, and relative risks were treated as equivalent measures of risk. Analyses were stratified by sex where possible. If necessary, relative risks within studies were re-calculated based on the method described by Hamling et al. [7] and pooled across studies using inverse-variance weighted DerSimonian-Laird random-effect models to allow for between-study heterogeneity [8]. We quantified between-study heterogeneity using Cochran's Q [9] and the I 2 statistic [10]. I 2 can be interpreted as the proportion of the total variation other than chance that is due to heterogeneity between studies. We tested for potential publication bias using Egger's test [11]. Sensitivity analyses for the influence of single studies on the pooled relative risks were conducted omitting one study at a time and re-estimating the pooled relative risk. All meta-analytical procedures were conducted on the natural log scale in Stata statistical software, version 12.1 (Stata Corp, College Station, Texas), and p<0·05 (two-sided) was considered statistically significant.

Analysis of subgroups or subsets
Subgroup analyses were completed for different classification of alcohol exposure (episodic heavy drinking, nonheavy drinking. Articles retrieved in full-text (n=19) Not meta-analysis or not total alcohol consumption (n=5) Articles excluded because more comprehensive metaanalysis has been published (n=3) Most comprehensive articles on total alcohol consumption selected for narrative review (11)

Figure S2. Search results for population studies on non-heavy and heavy alcohol consumption and IHD risk
Articles identified: Web of Science (n=1258) Handsearch (n=2) Unique articles (n=2769) Excluded based on title or abstract with minimal uncertainty (n=2306) Articles retrieved in full-text (n=463) Articles excluded (n=457): Not case-control or cohort study design (n=21) No data for alcohol intake (n=198) IHD is not the outcome or self-reported (n=64) No data on non-heavy or heavy drinking (n=131) Estimates not age-adjusted (n=1) Reference group is not lifetime abstainers (n=19) No deaths recorded (n=1) Chronic heavy drinking (n=12) Articles for quantitative analysis: Heavy and non-heavy drinking up to 30 g/day on average (n=7) alcohol is less risky alcohol is more risky Figure S4. IHD risk among episodic heavy drinkers with total average intake of 1-30 g/day compared to lifetime abstainers alcohol is less risky alcohol is more risky