Critique #281 – Why do only some cohort studies find health benefits from low-volume alcohol use? A systematic review and meta-analysis of study characteristics that may bias mortality risk estimates

Authors

Stockwell T; Zhao J; Clay J; Levesque C; Sanger N; Sherk A; Naimi T

Citation

Journal of Studies on Alcohol and Drugs 2024 85(4): 441–452. doi: 10.15288/jsad.23-00283

Author’s abstract

Objectives Assumptions about alcohol’s health benefits profoundly influence global disease burden estimates and drinking guidelines. Using theory and evidence, we identify and test study characteristics that may bias estimates of all-cause mortality risk associated with low-volume drinking.

Methods We identified 107 longitudinal studies by systematic review with 724 estimates of the association between alcohol consumption and all-cause mortality for 4,838,825 participants with 425,564 recorded deaths. “Higher-quality” studies had a mean cohort age of 55 years or younger, followed up beyond 55 years, and excluded former and occasional drinkers from abstainer reference groups. “Low-volume” alcohol use was defined as between one drink per week (>1.30 g ethanol/day) and two drinks per day (<25 g ethanol/ day). Mixed linear regression was used to model relative risks (RRs) of mortality for subgroups of higher- versus lower-quality studies.

Results As predicted, studies with younger cohorts and separating former and occasional drinkers from abstainers estimated similar mortality risk for low-volume drinkers (RR = 0.98, 95% CI [0.87, 1.11]) as abstainers. Studies not meeting these quality criteria estimated significantly lower risk for low-volume drinkers (RR = 0.84, [0.79, 0.89]). In exploratory analyses, studies controlling for smoking and/or socioeconomic status had significantly reduced mortality risks for low-volume drinkers. However, mean RR estimates for low-volume drinkers in nonsmoking cohorts were above 1.0 (RR = 1.16, [0.91, 1.41]).

Conclusions Studies with life-time selection biases may create misleading positive health associations. These biases pervade the field of alcohol epidemiology and can confuse communications about health risks. Future research should investigate whether smoking status mediates, moderates, or confounds alcohol-mortality risk relationships.

Summary

The recent study by Stockwell et al. (2024) claims that there is no safe level of alcohol consumption, which disregards decades of robust evidence. It has been consistently shown that those who choose to drink alcohol moderately and regularly tend to have longer and healthier lives than those who choose to abstain and those who drink more heavily.  The authors base their claim on a highly selective number of incorrectly referenced studies misinterpreting the original data and misrepresenting the original conclusions. Their selected 5 out of 107 studies are evaluated according to inconsistent and unclear quality criteria. Furthermore, the authors fail to recognize the methodological rigor of contemporary studies. Future research should continue to refine our understanding of alcohol’s health/harm effects. This paper, however, does not contribute to that issue, nor does it convincingly refute the existing evidence of the potential benefits of low-volume drinking.

Forum Comments

Background

The opinion that low to moderate alcohol consumption protects against all-cause mortality in general populations continues to be controversial. Observational studies show that moderate drinkers have longer life expectancy mainly because they are less likely to die from cardiovascular diseases than abstainers (Di Castelnuovo et al., 2022).

In 2023, Zhao et al. (2023) concluded: “Daily low or moderate alcohol intake was not significantly associated with all-cause mortality risk, while increased risk was evident at higher consumption levels, starting at lower levels for women than men.” In Critique #264, ISFAR asserted that the consistent finding of lower cardiovascular disease risk among moderate drinkers in all well-done cohort studies is strongly supported by robust animal and human experimental evidence of the biological mechanisms.

ISFAR was subsequently invited to author a commentary (Stockley et al. 2024) in the Journal of Studies on Alcohol, which describes the mechanisms by which moderate alcohol consumption has been shown to decrease essentially all risk factors for cardiovascular disease. These include low HDL-cholesterol, elevated LDL-cholesterol, endothelial dysfunction, coagulation pathologies, inflammation, abnormal glucose metabolism, as well as many others (Brien et al., 2011, Ronksley et al., 2011, Hendriks, 2020). Stockwell et al. were then asked to respond to the commentary. Stockwell et al. (2024a) responded.

In early 2024, Shield and Rehm (2024) reviewed both commentary and response in Judgment-Based Errors in Epidemiological Studies: A Commentary on Stockley et al. (2024) and Stockwell et al. (2024). They suggested that “given the susceptibility of Zhao and colleagues’ (2023) decisions to both systematic and random error, the ISFAR critique (2024) highlights, although not explicitly, a fundamental, universally overlooked limitation of tools like the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), which Zhao and colleagues used to design their systematic review—namely, the lack of standardized guidelines or rules for formulating judgments. This absence applies not only to systematic reviews but also to all epidemiological studies.”  

Critique

The most recent paper by Stockwell et al. (2024b) seeks to address Shield and Rehm (2024) by identifying and testing “study characteristics that may bias estimates of all-cause mortality risk associated with low-volume drinking,” where they conclude that “studies with lifetime selection biases may create misleading positive health associations. These biases pervade the field of alcohol epidemiology and can confuse communications about health risks. Future research should investigate whether smoking status mediates, moderates, or confounds alcohol-mortality risk relationships.”

These conclusions ignore the fact that since approximately 1995, the majority of past research has indeed investigated potential confounders to show that the overwhelming body of observational scientific data, as well as an immense number of experimental studies, support the contention that, for most middle-aged and older men and women who choose to do so, the regular consumption of small amounts of an alcoholic beverage can be considered as one of four or five additive components of a “healthy lifestyle.” Such a habit has been repeatedly shown to be associated with a lower risk of CVD and total mortality (Spencer et al., 2005, Holahan et al., 2010, McCaul et al., 2010, Ford et al., 2011, Simons et al., 2011, Larsson et al., 2017, Barbarsko et al., 2018, Li et al., 2018, Ding et al., 2022). Ding et al. (2022) examining modifiable risk factors and longevity in 85 346 participants from the Nurses’ Health Study and the Health Professionals Follow-up Study, concludes that there is “important evidence” that maintaining healthy behaviours should be recommended to individuals through mid-life to late adulthood as well as at young ages.

Further, the 2014 meta-analysis by Roerecke and Rehm concluded that “For drinkers having one to two drinks per drinking day without episodic heavy drinking, there is substantial and consistent evidence from epidemiological and short-term experimental studies for a beneficial association with ischaemic heart disease (IHD) risk when compared to lifetime abstainers. The alcohol-IHD relationship fulfills all criteria for a causal association proposed by Hill.”

This is not a new study by Stockwell et al. (2024) nor new conclusions, but a revision of Zhao et al. (2023) with additional filters on papers to be included in their meta-analysis. Twenty-one ’fit for purpose’ papers are further reduced to six and then to five, such that on the balance of these five papers, low to moderate drinkers are at a higher risk of all-cause mortality than abstainers. Stockwell et al. (2024) do not make a compelling case for abstinence, as they continue to disregard the supporting biological mechanisms for a J-shaped relationship between alcohol and health.

Furthermore, Stockwell et al. (2024) omit the references to these five papers in their paper, and the references provided in supplementary data do not correspond to the author and publication dates of published papers in PubMed. Such errors should have been found and corrected by the JSAD editors. The PubMed-published papers from those five authors that Stockwell et al. (2024) may refer to Figure 2 in Stockwell et al. (2016), which have mixed results regarding a j-shaped curve. Two support and two do not support its existence and one is undecided hardly a clear conclusion (Rehm et al. 2001, Sempos et al. 2003, Nakaya et al. 2004, Mäkelä et al. 2005, Zaridze et al. 2014). Indeed, their interpretation of the data is inconsistent and inaccurate, given some of the papers that they claim to support their conclusions do not. For example, Mäkelä et al.  (presumed to be 2005, not 2012 as shown) conclude that “moderate drinking is associated with a lower risk of IHD, whereas drinking in a heavy episodic manner (often referred to as “binge drinking”) is not”.  In JSAD supplemental material Figure 4, Mäkelä et al. show an all-cause mortality risk slightly higher than 1.0.  This data cannot have come from Mäkelä et al. 2005, as all drinking categories show a risk <1 except the highest consumption category.   Indeed, collectively these results do not suggest or support the shared conclusions of Stockwell et al. (2024) and Zhao et al. (2023) that “estimates of mortality risk associated with low volume use become smaller and insignificant when adjustment is made for key study characteristics”.

In addition, there is considerable confounding in this study of Zaridze et al. (presumed to be 2014, not 2019 as shown), where the invited commentary of Rehm et al. (2014) stated: “On its own, the overall volume of alcohol consumed in Russia, albeit high, cannot explain the high alcohol-attributable mortality; it is the combination of high overall volume with the specific pattern of episodic binges that is necessary to explain the high level and fluctuating trends of total and alcohol-attributed mortality in Russia…Since the prevalence of smoking in men was so high, varying from 68·9% in the lowest alcohol intake group to 89·4% in the highest alcohol intake group, Zaridze and colleagues limited their main analyses to the effects of drinking among smokers. So perhaps, particularly at older ages, some of the alcohol effect reported could be attributable to an interaction between the effects of smoking and of alcohol consumption.”

While Stockwell et al. (2024) conclude that “future research should investigate whether smoking status mediates, moderates, or confounds alcohol-mortality risk relationships, one of the key five papers, Rehm et al. (presumed to be 2001, not 2009 as they claim), interpret their data as “the volume of drinking and mortality showed a curvilinear relation for both sexes….neither choosing lifetime abstainers as a reference category nor adjusting for social class, smoking, or marital status altered the basic J-shaped relation.”

Lastly, confounders for alcohol-mortality risk relationships are already well-documented (Klatsky 2001). As studies of disease outcomes in alcohol epidemiology are likely to be indefinitely limited to observational studies, potential confounding will remain an issue concerning both beneficial and adverse effects (Klatsky and Udaltsova 2013).

The following comments are fresh perspectives from ISFAR forum members on Stockwell et al. (2024) which add to those already published by ISFAR in Critique #264 on Zhao et al. (2023).  

Comments by specific Forum Members and colleagues

Forum member Giovanni de Gaetano and colleagues Augusto Di Castelnuovo and Simona Costanzo state: “A group of researchers authored the paper committed to the hypothesis that no level of alcohol consumption is safe and presents a critical view of the health benefits associated with low-volume alcohol consumption. While the authors aim to challenge the prevailing consensus that moderate drinking can be part of a healthy lifestyle, their arguments and methodology exhibit several weaknesses that undermine the validity and impact of their conclusions as follows:

1. Study Selection and Quality Assessment

The authors conducted a systematic review of 107 longitudinal studies, emphasizing the importance of study quality by setting criteria such as cohort age, follow-up duration, and exclusion of former and occasional drinkers from the abstainer reference group. While these criteria are valid for ensuring robust study designs, the paper overlooks that most recent, large-scale epidemiological studies have already implemented such controls. The insistence on these criteria suggests a lack of acknowledgment of the advancements in study designs and the rigorous adjustments made in contemporary research to address these biases.

2. The Sick Quitter Hypothesis

It is well-established that including former drinkers, who may have stopped drinking due to health issues, in the control group can bias results. However, the authors seem to overstate the prevalence of this issue. Modern studies have recognized and addressed the “sick quitter” hypothesis extensively. By continuing to emphasize this point, the authors ignore the progress made in recent alcohol epidemiology, which has adjusted for this bias to present more accurate associations between alcohol consumption and health outcomes.

3. Smoking as a Confounder

The paper’s focus on smoking as a major confounder in alcohol-mortality studies is valid. It is indeed widely recognized that smoking can confound the relationship between alcohol consumption and health outcomes. However, the authors fail to acknowledge that most recent studies rigorously controlled for smoking status. Their assertion that studies controlling for smoking still show increased risks for low-volume drinkers lacks substantive evidence and seems to disregard the understanding of how multiple lifestyle factors interact. Moreover, the authors’ exploratory analyses, which suggest that non-smoking cohorts show no J-shaped curve, do not convincingly support their claim and require further robust analysis.

4. Health Benefits of Low-Volume Drinking

The conclusion that the few studies meeting “minimal quality criteria” do not show significant health benefits for low-volume drinkers is overly simplistic. A substantial body of literature supports the protective effects of moderate alcohol consumption, particularly in terms of cardiovascular health. By dismissing these findings, the authors ignore the multifaceted nature of alcohol’s impact on health and the necessity of considering a broader range of evidence.

5. Alcohol and Socio-Economic Status (SES)

The authors’ claim that “none of the 107 studies presented stratified results by SES” demonstrates a significant gap in their review. They fail to acknowledge the comprehensive discussion on the relationship between alcohol and health across different SES levels, as thoroughly addressed in recent literature, including the following paper (Di Castelnouvo et al. 2023). This omission reveals a rather superficial analysis of the different studies and undermines the credibility of their critique.

In summary, while the authors make some valid points regarding potential biases in alcohol epidemiology, their critique lacks a balanced consideration of the advancements in study designs and the comprehensive body of evidence supporting moderate alcohol consumption as part of a healthy lifestyle. Their insistence on outdated issues and failure to recognize the methodological rigor of contemporary studies weaken their argument. Future research should indeed continue to refine our understanding of alcohol’s health/harm effects. This paper, however, does not convincingly refute the existing evidence of the potential benefits of low-volume drinking.

The paper under critique was published in The Journal of Studies on Alcohol and Drugs, with a relatively low impact factor of 2.4. While some media outlets may pick up the story, the paper is unlikely to have a lasting impact and will probably fade into obscurity like other similar papers. We also attract our attention to their Conflict-of-Interest Statement.

When analysing a study that reports the beneficial effects of alcohol in moderation, it is essential to verify whether the authors have material, cultural, or ideological conflicts of interest that could influence their conclusions. Similarly, it is equally important to examine whether the authors have opposing conflicts of interest, such as receiving compensation or support from organizations or entities that are openly against moderate alcohol consumption. Transparency in this area is fundamental to maintaining the trust of the public and the scientific community in published research.”

Tim Stockwell, Adam Sherk, and Tim Naimi declare previous receipt of funding from government alcohol monopolies in Finland, Sweden, and Canada to conduct public health-oriented research on alcohol. Tim Stockwell and Tim Naimi have also received personal fees from these bodies in the past for the same purposes. Tim Stockwell and Tim Naimi have also received travel support from IOGT-NTO, a not-for-profit nongovernmental organization that advocates against alcohol use, to prepare reports on alcohol and health.

Forum member Erik Skovenborg writes when “concentrating on cohort age, ‘the usual reference group of abstainers fills up among older cohorts with people who have stopped or cut down on their drinking for health reasons’. This issue is more complicated than Stockwell et al. (2024b) have us believe.

Alcohol consumption changes with age. A study reporting patterns of alcohol consumption in a large, Californian health plan population (adults ages 65 to 90) found that those who quit drinking were more likely than current drinkers to report having medical problems, including diabetes, heart problems, and worse self-reported health (Satre et al. 2007). However, reasons for decreasing consumption with age include an ageing effect (i.e. that people drink less because of physiological, biological, and social changes that occur during chronological aging), cohort effects (i.e. that groups of people born at different times use alcohol differentially, with older cohorts typically consuming less alcohol), and the mortality hypothesis (i.e. that heavier drinkers die younger) (Stall et al. 1987).

A cross-sectional National Health Interview Survey in the United States of 40,556 Americans aged 60 described a trend of decreasing alcohol intake between aged ≥ 60 years and older and found that the proportions of men and women consuming higher quantities of alcohol (≥2 drinks/day) decreased while the proportions drinking most frequently (260-365 days per year) increased = a healthier drinking pattern (Breslow et al. 2004).

A UK study examined alcohol consumption trajectories, reporting both average weekly volume and frequency, using data from nine UK-based prospective cohorts with at least three repeated alcohol consumption measures on individuals (combined sample size of 59,397 with 174,666 alcohol observations), with data spanning from adolescence to very old age (90 years plus). The results showed that, for men, mean consumption rose sharply during adolescence, peaked at around 25 years at 20 units per week, and then declined and plateaued during mid-life, before declining from around 60 years. A similar trajectory was seen for women but with lower overall consumption (peak of around 7 to 8 units per week). Frequent drinking (daily or most days of the week) became more common during mid to older age, most notably among men, reaching above 50% of men (Britton et al. 2015).

A cohort study of British civil servants totalling 6838 men and 3372 women aged 34–55 years at baseline, followed for a mean 19.1 years found that consumption within baseline drinking categories is generally stable during the life course, except among heavier baseline drinkers, for whom intakes decline with increasing age. This shift does not appear to be driven by transitions to non-drinking. Cohorts of older people may be at particular risk of misclassifying former heavy drinkers as moderate consumers of alcohol (Knott et al. 2018).

Finally, an analysis of 21 peer-reviewed publications and six reports, including data from 17 national surveys and 10 general practice and community samples published since 2000 found that older adults in developed countries report different rates of abstention and alcohol consumption.  This places obvious limitations on extrapolating results from specific research findings (Gell et al. 2015).”

Forum member Richard Harding comments that “last week a study from Dr Tim Stockwell at the University of Victoria, Canada claimed that there is no safe level of alcohol intake, which flew in the face of decades of remarkably robust evidence. We have consistently been told that those who choose to drink alcohol moderately but regularly tend to have longer and healthier lives than those who choose to abstain and those who drink more heavily.

How can this be? Epidemiology is the study of populations to determine the risk factors for disease. It seeks to identify correlations between variables. Some correlations turn out to be causal, some purely coincidental. For example, the correlation between smoking and lung cancer we now know is causal. The correlation between the birth rate in Germany and changes in the local stork population probably isn’t. Consequently, it is wrong to assume that correlations are automatically causal, but that is exactly what Dr Stockwell has done.

He has assumed that in his selected six out of 107 studies on which his conclusions are based it is alcohol consumption per se that is responsible for his observed reduction in lifespan, without any other supporting evidence. The plain fact is that it is impossible to establish causality by epidemiology alone.

What evidence for causality is there? There is plenty, from human and animal experiments on the biological mechanisms. We know that moderate alcohol consumption leads to favourable changes in several cardiovascular biomarkers. These effects are not minor. The authors of a 2011 review found that moderate alcohol consumption increased high-density lipoprotein — “good” cholesterol — greater than any available pharmacological therapy, a finding corroborated by others. Moderate alcohol consumption also normalises insulin and glucose levels, and increases insulin sensitivity, so little wonder that moderate drinkers experience a significant lowering of diabetes risk. These are some important reasons why we know that the curve is J-shaped.

If Dr Stockwell thinks that one alcoholic drink per day is responsible for shortening our lives, he should produce some evidence to explain why. Otherwise, we are clearly in stork territory. What evidence there is indicates the complete opposite.”

[This comment was also published in The Times on Wednesday, July 31, 2024]

Forum member Curtis Ellison considers that “it is annoying that this same set of authors continues to generate papers that, without sound evidence, attack the results of extremely well-founded and well-established science. One of the most consistent findings of solid cohort studies for decades is that regular, moderate alcohol consumption is associated with a lower risk of total mortality.  In other words, moderate drinkers live longer than abstainers or abusers of alcohol. This is undoubtedly related to favourable effects on the physiologic and biologic mechanisms associated with cardiovascular disease and diabetes among older individuals when most deaths occur.

The authors of this latest re-hash of the same biased arguments again dared to invent criteria for what is a ‘higher quality’ study, then use their unvalidated criteria to exclude all but 5 or 6 out of 107 cohort studies to announce that they had finally found the ‘truth’ about alcohol and mortality. Is it surprising that these happened to fail to show the benefits of moderate drinking on mortality? They show graphs of these few studies and state that they increase the risk of mortality, even though none showed statistically significant effects.

No one is urging everyone to drink, but the public should have well-founded, accurate data on what unbiased research shows for the effects of consuming a beverage containing alcohol. Current research continues to show that drinking that is excessive, in binges, and when consumed without food, may have net adverse effects on health. Current research on moderate drinking, especially during meals, is also consistently shown to be associated with fewer of the diseases of ageing and with a significantly lower risk of dying prematurely.”  

Forum member Fulvio Mattivi suggests that “reality is complex and diverse. Consequently, a simplistic message toward absolute conclusions may not be the right interpretation. Especially if we do not overcome some intrinsic problems common to the studies we rely upon today. In addition to pointing (rightly) the finger at some known critical aspects of the old studies, which have already been implemented for years in the most robust studies, in my opinion, the time has come to systematically include biomarkers of ethanol consumption in studies on diet, lifestyle, and health. Another common problem in epidemiological studies is indeed the lack of adoption of clinical biomarkers that can verify the correspondence between alcohol consumption and the declared amount. A topical but still underestimated issue, revised in Trius-Soler et al. (2023).

The reliability of the tools for assessing alcohol intake in the diet is debated and it is assumed that a large percentage of subjects report incorrectly, especially those who declare abstinence or low alcohol consumption. For various reasons, man is a liar, and tends to deny consumption or, at least, to underestimate it. Using biomarkers allows us to identify the most striking deviations, almost always upwards.

Reliable biomarkers to discriminate between abstinence and low or moderate intake, both in the long and short term could, therefore, have a considerable impact on future research in alcohol studies. Furthermore, with the assistance of objective markers, the attention of exposure questionnaires can be focused on the type of beverage consumed and the pattern of consumption. Since no one – fortunately – drinks “ethanol”, knowing the real matrix most consumed and the pattern of intake is essential for the quality of the results, for real and not just formal progress.”

Forum member Andrew Waterhouse remarks that “when I studied Rehm et al. (presumably 2001) which is included in the list of five ‘higher quality’ papers, looking at the lowest regular consumption, 0-2 (seems odd to go to 0 since there is another category for non-drinkers), the average of the male and female RR’s is 1.  Unfortunately, Stockwell et al. (2024) overlook the RR <1 for the higher moderate drinking categories.  Further, the 0-2 band includes very occasional drinkers who should be in a separate category.  The lowest daily drinking category should be 1-2.  Thus, the best characterization of moderate daily consumption is 2-4, and that RR is clearly <1. 

In addition, Zaridze et al. (presumably 2014), also included in the ‘higher quality’ list, are looking only at the high consumption of vodka in Russia.  Thus, this paper cannot be considered a valid comparison to the other four papers. Also, non-drinkers are not clearly defined, but appear in the results tables as a “bonus.”  Consequently, I question the “quality” screening criteria used by Stockwell et al. (2024).”

Conclusions

As concluded by Shield and Rehm (2024), the results of Zhao et al. (2023) and correspondingly that of the related Stockwell et al. (2024) “are susceptible to errors introduced during the judgment-making process. Accordingly, to improve the validity of such studies…there is a need for standardized guidelines on how judgments in observational research should be formulated, along with standardized tools enabling assessment of the risk of bias introduced by the judgment-making process.”

A fitting conclusion to this critique comes from the late Forum member Arthur Klatsky: “One obvious lesson is that we need better instruments to measure alcohol intake histories. Another is that complete description of methodology greatly aids evaluation and comparison of reports. Alcohol researchers can perform an important service by presenting verbatim expositions showing exactly how all information was obtained and how alcohol consumption was categorized. Such specificity, accompanied by frank discussion of implications and limitations, would move us in the right direction” (Klatsky 2008).

References

Barbaresko, J., Rienks, J., and Nöthlings, U. (2018) Lifestyle indices and cardiovascular disease. Risk: a meta-analysis. American Journal of Preventative Medicine, 55(4):555-564.

Breslow, R. A. and Smothers, B. (2004) Drinking patterns of older Americans: National Health Interview Surveys, 1997–2001. Journal of Studies on Alcohol, 65:232-40.

Brien, S.E., Ronksley, P.E., Turner, B.J., et al. (2011). Effect of alcohol consumption on biological markers associated with risk of coronary heart disease: systematic review and meta-analysis of interventional studies. British Medical Journal, 342: d636. https://doi.org/10.1136/bmj.d636.

Britton, A., Ben-Shlomo, Y., Benzeval, M. et al. (2015) Life course trajectories of alcohol consumption in the United Kingdom using longitudinal data from nine cohort studies. BMC Medicine, 13:47.

Connor J. (2017) Alcohol consumption as a cause of cancer. Addiction, 112: 222–228.

Di Castelnuovo, A., Costanzo, S., Bonaccio, M., et al. (2022). Alcohol intake and total mortality in 142 960 individuals from the MORGAM Project: a population-based study. Addiction (Abingdon, England), 117(2), 312–325. https://doi.org/10.1111/ADD.15593.

Di Castelnuovo A, Bonaccio M, Costanzo S, … et al. MORGAM Study Investigators; de Gaetano G. (2023) Drinking alcohol in moderation is associated with lower rate of all-cause mortality in individuals with higher rather than lower educational level: findings from the MORGAM project. European Journal of Epidemiology, 38(8):869-881. https://doi.org/10.1007/s10654-023-01022-3.

Ding, M., Fitzmaurice, G.M., Arvizu, M, et al. (2022) Associations between patterns of modifiable risk factors in mid-life to late life and longevity: 36-year prospective cohort study. British Medical Journal, 1(1):e000098. https://doi.org/10.1136/bmjmed-2021-000098.

Ford, E.S., Zhao, G., Tsai, K., and Li, C. (2011). Low-risk lifestyle behaviors and all-cause mortality: findings from the National Health and Nutrition Examination Survey III Mortality Study. American Journal of Public Health, 101:1922–1929.

Gell, L., Meier, P.S., and Goyder, E. (2015) Alcohol consumption among the over 50s: International comparisons. Alcohol and Alcoholism, 50:1–10.

Hendriks, H.F.J. (2020). Alcohol and human health: What is the evidence? Annual Review of Food Science and Technology, 11(1):1–21. https://doi.org/10.1146/annurev-food-032519-051827.

Holahan, C.J., Schutte, K.K., Brennan, P.L., et al. (2010). Late-life alcohol consumption and 20-year mortality. Alcohol: Clinical and Experimental Research, 34:1961–1971.

Klatsky, A.L. (2001) Diet, alcohol, and health: a story of connections, confounders, and cofactors. American Journal of Clinical Nutrition, 74: 279–280.

Klatsky, A.L. (2008) Invited commentary: never, or hardly ever? It could make a difference. American Journal of Epidemiology, 168(8):872-5; discussion 876-7. https://doi.org/10.1093/aje/kwn192.

Klatsky, A.L., and Udaltsova, N. (2013) Abounding confounding: sick quitters and healthy drinkers. Addiction, 108(9):1549-52. https://doi.org/10.1111/add.12157. PMID: 23947729.

Knott, C.S., Bell, S., and Britton, A. (2018) The stability of baseline-defined categories of alcohol consumption during the adult life-course: a 28-year prospective cohort study. Addiction, 113:34-43.

Larsson, S.C., Drca, N., and Wolk, A. (2014). Alcohol consumption and risk of atrial fibrillation: a prospective study and dose-response meta-analysis. Journal of the American College of Cardiology, 64(3):281-289.

Li, Y., Pan, A., Wang, D.D., et al. (2018) Impact of healthy lifestyle factors on life expectancies in the US population.  Circulation. 137:00–00. (Pre-publication).  https://doi.org/10.1161/CIRCULATIONAHA.117.032047.

Mäkelä, P., Paljärvi, T., and Poikolainen K. (2005) Heavy and nonheavy drinking occasions, all-cause and cardiovascular mortality and hospitalizations: A follow-up study in a population with a low consumption level. Journal of Studies on Alcohol, 66:722–728. https://doi.org/10.15288/jsa.2005.66.72.

McCaul, K.A., Almeida, O.P., Hankey, G.J., et al. (2010). Alcohol use and mortality in older men and women. Addiction. 105(8):1391-1400.

Nakaya, N., Kurashima, K., Yamaguchi, J., et al.  (2004) Alcohol consumption and mortality in Japan: The Miyagi Cohort Study. Journal of Epidemiology, 14:S18-S25. https://doi.org/10.2188/jea.14.S18.

J. Rehm et al. (2001) Alcohol and all-cause mortality: A pooled analysis. Contemporary Drug Problems.

Rehm, J. (2014) Russia: lessons for alcohol epidemiology and alcohol policy. The Lancet, 383(9927):1440-1442. https://doi.org/10.1016/S0140-6736(14)60038-6.

Rehm, J., Greenfield, T.K., and Rogers, J.D. (2001) Average volume of alcohol consumption, patterns of drinking, and all-cause mortality: results from the US national alcohol survey. American Journal of Epidemiology, 153(1):64-71. https://doi.org/10.1093/aje/153.1.64.

Roerecke, M., and Rehm, J. (2014) 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. BMC Medicine, 12:182.

Ronksley, P. E., Brien, S. E., Turner, B. J., et al. (2011). Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. British Medical Journal, 342:d671. https://doi.org/10.1136/bmj.d671.

Satre, D.D., Gordon, N.P., and Weisner, C. (2007) Alcohol consumption, medical conditions, and health behavior in older adults. American Journal of Health Behaviours, 31: 238–248.

Sempos, C.T., Rehm, A., Wu T.J., Crespo, C.J., and Trevisan M. (2003) Average volume of alcohol consumption and all-cause mortality in African Americans: the NHEFS cohort. Alcohol: Clinical and Experimental Research, 27(1):88-92. https://doi.org/10.1111/j.1530-0277.2003.tb02726.x.

Shield, K., and Rehm, J. (2024) Judgment-based errors in epidemiological studies: A commentary on Stockley et al. (2024) and Stockwell et al. (2024). Journal of Studies on Alcohol and Drugs, 85:1, 141-142.

Simons, L.A., Simons, J., Friedlander, Y., and McCallum, J. (2011) Predictors of long-term mortality in the elderly: the Dubbo Study. Internal Medicine Journal, 41(7):555-60. https://doi.org/10.1111/j.1445-5994.2009.02106.x.

Spencer, C.A., Jamrozik, K., Norman, P.E., Lawrence-Brown, M. (2005). A simple lifestyle score predicts survival in healthy elderly men. Preventative Medicine, 40:712–717.

Stall R. (1987) Research issues concerning alcohol consumption among aging populations. Drug and Alcohol Dependence, 19:195–213.

Stockley, C.S., Hendriks, H. F. J., and Ellison, R.C. (2024). ISFAR reiterates its defense of moderate alcohol consumption’s health benefits. Journal of Studies on Alcohol and Drugs, 85(1), 136–140. https://doi.org/10.15288/jsad.23-00293.

Stockwell, T., Zhao, J., and Naimi, T. (2024a). Apologizing for the alcohol industry? A response to ISFAR’s defense of alcohol’s purported health benefits. Journal of Studies on Alcohol and Drugs, 85(1), 133–135. https://doi.org/10.15288/jsad.23-00193.

Stockwell T; Zhao J; Clay J; Levesque C; Sanger N; Sherk A; Naimi T. (2024b). Why do only some cohort studies find health benefits from low-volume alcohol use? A systematic review and meta-analysis of study characteristics that may bias mortality risk estimates. Journal of Studies on Alcohol and Drugs 2024 85(4): 441–452. https://doi.org/10.15288/jsad.23-00283.

Stockwell, T., Zhao, J., Panwar, S., et al. (2016). Do “moderate” drinkers have reduced mortality risk? A systematic review and meta-analysis of alcohol consumption and all-cause mortality. Journal of Studies on Alcohol and Drugs, 77(2), 185-198.

Trius-Soler, M., Praticò, G., Gürdeniz, G., et al. (2023) Biomarkers of moderate alcohol intake and alcoholic beverages: a systematic literature review. Genes Nutrition, 18(1):7. https://doi.org/10.1186/s12263-023-00726-1. Erratum in: Genes Nutr. 2023 May 15;18(1):9. doi: 10.1186/s12263-023-00728-z.

Zaridze, D., Lewington, S., Boroda, A., et al. (2014) Alcohol and mortality in Russia: prospective observational study of 151 000 adults. The Lancet, 383(9927):1465-1473. https://doi.org/10.1016/S0140-6736(13)62247-3.

Zhao, J., Stockwell, T., Naimi, T., Churchill, S., Clay, J., and Sherk, A. (2023). Association between daily alcohol intake and risk of all-cause mortality: A systematic review and meta-analyses. JAMA Network Open, 6(3):e236185. https://doi.org/10.1001/jamanetworkopen.2023.6185.

ISFAR signatories

Creina Stockley, PhD, MBA, Independent Consultant and Adjunct Senior Lecturer in the School of Agriculture, Food and Wine at the University of Adelaide, Australia

Henk Hendriks, PhD, Independent Consultant, The Netherlands

Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy

Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark

Richard Harding, PhD, Formerly Head of Consumer Choice, Food Standards and Special Projects Division, Food Standards Agency, UK

Fulvio Mattivi, MSc, Scientific Advisor, Research and Innovation Centre, Fondazione Edmund Mach, in San Michele all’Adige, Italy

Andrew Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA

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

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