Critique 017: Moderate alcohol intake is associated with lower risk of total mortality than are either abstinence or heavy drinking

Critique 017                                                                                                                                   5 September 2010

Holahan CJ, Schutte KK, Brennan PL, Holahan CK, Moos BS, Moos RH.  Late-Life Alcohol Consumption and 20-Year Mortality.  Alcoholism: Clinical and Experimental Research 2010;34:in press, November 2010

Authors’ Abstract

Background: Growing epidemiological evidence indicates that moderate alcohol consumption is associated with reduced total mortality among middle-aged and older adults.  However, the salutary effect of moderate drinking may be overestimated owing to confounding factors.  Abstainers may include former problem drinkers with existing health problems and may be atypical compared to drinkers in terms of sociodemographic and social-behavioral factors.  The purpose of this study was to examine the association between alcohol consumption and all-cause mortality over 20 years among 1,824 older adults, controlling for a wide range of potential confounding factors associated with abstention.

Methods: The sample at baseline included 1,824 individuals between the ages of 55 and 65.  The database at baseline included information on daily alcohol consumption, sociodemographic factors, former problem drinking status, health factors, and social-behavioral factors. Abstention was defined as abstaining from alcohol at baseline. Death across a 20-year follow-up period was confirmed primarily by death certificate.

Results: Controlling only for age and gender, compared to moderate drinkers, abstainers had a more than 2 times increased mortality risk, heavy drinkers had 70% increased risk, and light drinkers had 23% increased risk. A model controlling for former problem drinking status, existing health problems, and key sociodemographic and social-behavioral factors, as well as for age and gender, substantially reduced the mortality effect for abstainers compared to moderate drinkers.  However, even after adjusting for all covariates, abstainers and heavy drinkers continued to show increased mortality risks of 51 and 45%, respectively, compared to moderate drinkers. 

Authors’ Conclusions: Findings are consistent with an interpretation that the survival effect for moderate drinking compared to abstention among older adults reflects 2 processes.  First, the effect of confounding factors associated with alcohol abstention is considerable.  However, even after taking account of traditional and nontraditional covariates, moderate alcohol consumption continued to show a beneficial effect in predicting mortality risk.

Forum Comments

Comments on the present study:  This study followed 1,824 adults, aged 55-65 at baseline, for 20 years.  It did not have data on potential changes in alcohol intake during follow up or on drinking patterns, and all analyses were based on baseline alcohol intake.  “Lifetime abstainers” were excluded from the study, hence all subjects presumably consumed alcohol at some time during their lives.  This allowed the authors to separate former drinkers reporting “former problem drinking” (based on responses to the Drinking Problems Index) from other former drinkers not reporting problem drinking.  However, numbers of subjects in these categories are not given. 

All 11 measures of lifestyle/behavior that were included as potential confounders showed the expected relations with mortality.  For example, mortality was greater among former problem drinkers, those with obesity, smokers, and those with depressive symptoms or poor avoidance coping skills.  Mortality was lower for subjects with a higher sociodemographic status, married individuals, physically active people, and those with more social support.  Subjects reporting “moderate” alcohol intake (up to 3 drinks/day) had the most favorable profile of lifestyle factors and abstainers generally had the worst profile.  In Cox analyses including all lifestyle factors and in comparison with moderate drinkers, abstainers had the highest mortality risk (HR = 1.49, CI = 1.20 – 1.84) followed by “heavy” drinkers (HR = 1.42, CI = 1.17 – 1.73).  There were similar results for “light” drinkers as for “moderate” drinkers.  Overall, the study supports much previous research on the subject; its main innovative contributions may be the exclusion of lifetime abstainers and substantial data on associated lifestyle factors of its subjects.

Because the definition of “moderate” alcohol consumption used in the main analyses in this study exceeded the 1 to 2 drinks per day used in some studies, the authors conducted an additional analysis to examine the potential mortality effect of contrasting levels of alcohol consumption within the moderate drinker group.  Specifically, they contrasted mortality risk for moderate drinkers who consumed 1 to 2 drinks per day (from 14 to 28 g of alcohol, n = 345) with that of moderate drinkers who consumed > 2 drinks per day to < 3 drinks per day (from more than 28 to less than 42 g of alcohol, n = 215).  Considering that the NIAAA guidelines suggest that older adults should limit alcohol intake to no more than one drink per day, the results of this study are particularly interesting: the 20-year mortality risk for moderate drinkers who consumed 1 to 2 drinks per day was nearly identical to that of moderate drinkers who consumed > 2 to < 3 drinks per day (OR = 0.97, p = 0.80, 95% CI = 0.74, 1.26), and both groups had lower risk than consumers of < 1 drink/day.

In their Discussion, the authors describe in detail potential weaknesses of their study (not able to judge effects of changes in intake over many years, unable to compare results with those of lifetime abstainers, etc.), but also some real strengths of their study.  The latter include long follow up and very good data on many of the aspects of socio-economic and lifestyle behaviors, all of which were related to mortality.  As stated, their fully adjusted model indicated 51% higher risk for abstainers and 45% higher risk for heavy drinkers when compared with moderate drinkers.  These are quite remarkable effects on the risk of mortality, and are larger than those associated with most drugs.  The paper provides information that may be especially important for American physicians, who tend to be very cautious when giving advice to patients regarding the consumption of alcohol.

Weaknesses of the present study include its relatively small size; it is likely underpowered to assess the relative merits of red wine, white wine, beer and spirits, or to determine whether the drink of choice was different between moderate and heavy drinkers.  The potential influence of diet was not assessed in the present study.  Further, in this and all observational research, one can never be sure to have included all confounders — given that moderate people tend to do moderate things, it is not possible to rule out residual confounding. 

Also, there may be selection bias in the subjects included in the study, making generalizations to the population difficult.  The sample consisted of subjects who had outpatient contact with a health care facility in the previous three years. Requiring contact with a health care facility means that this sample may have excluded the very healthy (who may have been less likely to see a physician within the preceding 3 years) and perhaps excluded those who could not afford health care.

Previous research on the topic:  Numerous studies have documented alcohol’s complex effects on health and mortality.  It was shown by Pearl in 1923 that alcohol consumption and mortality had a U-shaped relation, with intermediate alcohol consumption associated with lower mortality than complete abstention or heavy alcohol use1.  Many studies since have confirmed these results, with moderate drinkers having lower mortality than heavy drinkers or nondrinkers2-7.  In a meta-analysis of 34 studies in 2006, Di Castelnuovo et al8 found that low levels of alcohol intake (1-2 drinks per day for women and 2-4 drinks per day for men) were inversely associated with total mortality in both men and women.

More recent reports have confirmed lower mortality for moderate drinkers and attempted to judge the degree to which associated lifestyle factors may affect this association.  A study by Lee et al9 supports most previous studies suggesting that moderate drinkers tend to have higher socio-economic status and fewer functional limitations, which in themselves are associated with lower total mortality.  The question is to what degree are some of these other advantages the result of the moderate intake of alcohol, or is moderate drinking just a marker of a healthier lifestyle?  These investigators carried out very elaborate analyses to obtain a more precise measure of the effects that can be attributed to the alcohol drinking itself.  Their analyses estimated that moderate drinking, in comparison with no alcohol intake, was associated with 38% lower mortality risk (OR = 0.62; CI = 0.48 – 0.80)9

Costanzo et al10 found that similar associations between alcohol and mortality as reported for the general population were also present among patients with known cardiovascular disease.  They concluded, “Cardiologists should be aware that regular, moderate alcohol consumption, in the context of a healthy lifestyle (increased physical activity, no smoking), dietary habits (decreased dietary fat intake, high consumption of fruit and vegetables), and adequate drug therapy, would put their patients at a level of cardiovascular or mortality risk substantially lower than either abstainers or heavy or binge drinkers10.” 

An inverse association between moderate drinking and mortality is also present among the elderly, as recently reported by McCaul et al11.  From an analysis of more than 24,000 elderly community-living adults, these authors concluded: “In people over the age of 65 years, alcohol intake of four standard drinks per day for men and two standard drinks per day for women was associated with lower mortality risk.” 

Effects of alcohol on mortality are modified by drinking pattern.  Rehm et al12 reported on mortality among a representative sample (n = 5,072) of the US population, finding a significant influence of drinking alcohol on mortality with a J-shaped association for males and an insignificant relation of the same shape for females.  When the largest categories of equivalent average volume of consumption were divided into people with and without heavy drinking occasions, serving as an indicator of drinking pattern, this differentiation proved important in predicting mortality.  Light to moderate drinkers had higher mortality risks when they reported heavy drinking occasions (defined by either eight drinks per occasion or getting drunk at least monthly).  Rehm et al13 also found in a study from Canada that light to moderate drinkers had higher mortality risks when they reported heavy drinking occasions but not when the group excluded such binge drinkers.  Information on drinking pattern was not available for the present study.

The message to the public:  This article was used as the basis for an article in Time magazine entitled, “Why Do Heavy Drinkers Outlive Nondrinkers?”  The article was by John Cloud and released on 30 August 2010; it can be accessed at,8599,2014332,00.html?hpt=Sbin#ixzz0ybMyqUQg

Several Forum members commented that the fact that this article was chosen for a major report in Time magazine illustrates the dramatic distance between the world of science and the general views of the public.  For scientists the issue is quite clear, and none of the members of the International Scientific Forum on Alcohol Research were surprised by what Time considered to be “incredible” news.  Although the study could be more extended or more accurate in defining confounding variables and possible bias, Forum members agree on its conclusions, which are in line with much previous research. 

The issue of the net effects of moderate drinking is particularly relevant since, while scientists are convinced that moderation is far better than abstinence and excessive use, regulatory bodies often refe only to abuse when discussing alcohol use.  One Forum member describes the following to explain how something that at one “dose” may be toxic may have very different net effects at a different “dose.”

(1)     A food or drink (or, in general, a given behavior) that at a certain level may increase the risk of one or more diseases does not necessarily reflect a corresponding increase in risk of total morality.  This is particularly true of alcohol, as excessive intake is associated with increased risk of certain cancers, deaths from accidents, etc.

(2)     In toxicology the relationship between dose and effect is almost never linear.  The largely discussed J-shaped curve of risk from alcohol intake is not something peculiar of alcoholic beverages.  It is, instead, practically a rule for a large number of xenobiotics.  The role of hormesis (the term for generally-favorable biological responses to low exposures to toxins and other stressors), known for centuries, is today becoming more precisely focused as a crucial element for protecting health – particularly in the area of degenerative diseases and cancer.  The most typical example is the nuclear factor NRf2, that is activated by electrofiles (oxidants) and than primes the expression of a battery of antioxidant defenses and repair mechanisms.  Thus, it is not rare that something that may be “toxic” at some level leads, in moderation, to protection of health.

The stated goal of the International Scientific Forum on Alcohol Research is to critically review emerging science and to provide sound and balanced information on alcohol and health.  It should help the public understand concepts that “being good or being bad for you” cannot always be as simple as black and white.  We should attempt to translate science into understandable format.  Lay publications can play an important role in disseminating balanced information to the public14.

A key message for the public (and governmental agencies) is that we should not criminalize alcohol.  We should stress the fact that while in excess alcohol can be toxic, in moderation it may provide considerable health benefits.  Even key vitamins such as vitamins A and C, and elements such as iron, zinc, and manganese, can be considered “poisonous” if taken in excess.  It will be important that physicians who have focused only on the toxic aspects of alcohol drinking be cognizant of the potentially beneficial effects of moderate alcohol intake.

Summary of Forum Comments:  The present analyses were based on data from 1,824 predominantly Caucasian Americans from the Western part of the United States who were recruited into a longitudinal project that has examined late-life patterns of alcohol consumption and drinking problems.  Lifetime abstainers were not included in the study, which focused on stress and coping processes among problem and non-problem drinkers.  The sample at baseline included only subjects aged 55 to 65 years who had had outpatient contact with a health care facility in the previous 3 years. 

The database at baseline included information on daily alcohol consumption, sociodemographic factors, former problem drinking status, health factors, and social-behavioral factors.  Subjects who were not lifetime abstainers but did not report drinking at the time of the baseline examination were classified as “abstainers.”  Follow-up data on alcohol consumption during the course of the study were not collected.  Death across a 20-year follow-up period was confirmed primarily by death certificate. 

The key results of the paper are that even when adjusting for sociodemographic factors, former problem drinking status, health factors, and social-behavioral factors, moderate drinking was associated with considerably lower risk of all-cause mortality.  In comparison with “moderate drinkers” (subjects reporting up to 3 drinks/day), abstainers had 51 % higher mortality risk and heavy drinkers had 45% higher risk.  The study supports most previous scientific studies showing that moderate drinking, in comparison with both abstinence and heavy drinking, is associated with lower risk of total mortality.

References from Forum review

1.  Pearl R.  Alcohol and mortality.  In: Starling E, ed. The Action of Alcohol on Man.  London: Longmans, Green and Co, 1923, pp 213–286.

2. Poikolainen K. Alcohol and mortality: A review. J Clin Epidemiol 1995;48:455–465.

3. Thun MJ, Peto R, Lopez AD et al. Alcohol consumption and mortality among middle-aged and elderly U.S. adults. N Engl J Med 1997;337:1705–1714.

4. Gronbaek M, Deis A, Becker U et al. Alcohol and mortality: Is there a U-shaped relation in elderly people? Age Ageing 1998;27:739–744.

5. Gmel G, Gutjahr E, Rehm J. How stable is the risk curve between alcohol and all-cause mortality and what factors influence the shape? A precision-weighted hierarchical meta-analysis. Eur J Epidemiol 2003;18:631–642.

6. Room R, Babor T, Rehm J. Alcohol and public health. Lancet 2005;365:519–530.

7. Klatsky AL, Udaltsova N. Alcohol drinking and total mortality risk. Ann Epidemiol 2007;17:S63–S67.

8.  Di Castelnuovo A, Costanzo S,  Bagnard V, Donati MD, Iacoviello L, de Gaetano G.  Alcohol Dosing and Total Mortality in Men and Women: An Updated Meta-analysis of 34 Prospective Studies.  Arch Intern Med. 2006;166:2437-2445

9.  Lee SJ, Sudore RL, Williams BA, Lindquist K, Chen HL, Covinsky KE. Functional limitations, socioeconomic status, and all-cause mortality in moderate alcohol drinkers.  J Am Geriatr Soc 2009;57:955-962.

10.  Costanzo S, Di Castelnuovo A, Donati MB, Iacoviello L, de Gaetano G.  Alcohol consumption and mortality in patients with cardiovascular disease.  A meta-analysis.  J Am Coll Cardiol 2010;55:1339–1347

11.  McCaul KA, Almeida OP, Hankey GJ, Jamrozik K, Byles JE, Flicker L. Alcohol use and mortality in older men and women.  Addicition 2010.  On-line prior to publication:  doi:10.1111/j.1360-0443.2010.02972.x

12.  Rehm J, Greenfield TK, Rogers JD.  Average Volume of Alcohol Consumption, Patterns of Drinking, and All-Cause Mortality: Results from the US National Alcohol Survey.  Am J Epidemiol 2001;153:64-71.

13.  Rehm J, Patra J, Taylor B.  Harm, Benefits, and Net Effects on Mortality of Moderate Drinking of Alcohol Among Adults in Canada in 2002.  Ann Epidemiol 2007;17:S81–S86.

14.  Stuttaford T. The influence of print media on their readers’ understanding of the benefits of moderate drinking.  Ann Epidemiol 2007;17(Suppl):S108-S109.

                                                                             *                      *                      *

Comments on this paper provided by the following members of the International Scientific Forum on Alcohol Research:

Roger Corder, PhD, MRPharmS, William Harvey Research Institute, Queen Mary University of London, UK

R. Curtis Ellison, MD, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

Ross McCormick PhD, MSC, MBChB, Associate Dean, Faculty of Medical and Health Sciences, The University of Auckland, Auckland, New Zealand

Erik Skovenborg, MD, Scandinavian Medical Alcohol Board, Practitioner, Aarhus, Denmark

Gordon Troup, MSc, DSc, School of Physics, Monash University, Victoria, Australia

Fulvio Ursini, MD, Dept. of Biological Chemistry, University of Padova, Padova, Italy

David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa

Andrew L. Waterhouse, PhD, Marvin Sands Professor, Department of Viticulture and Enology, University of California, Davis