Critique 305 – The interaction of age and total lifetime drinks is associated with regional cortical perfusion and thickness in healthy adults with low-level alcohol consumption

Durazzo, T.C., Joseff, B.D.P., McNerney, M.W., Humphreys K., Meyerhoff D.J.

Alcohol (2026) 133:38–47 https://doi.org/10.1016/j.alcohol.2026.03.006

Abstract

Background: Low-level alcohol consumption at or below current guidelines (≤1 standard drink equivalent/day for females, ≤2 standard drink equivalents/day for males) has been proposed to carry minimal systemic health risks. The neurobiological correlates of low-level drinking in healthy adults are not well characterized. To date, no study has concurrently assessed the associations of low-level alcohol consumption, modelled as a continuous rather than categorical variable, with regional brain perfusion (blood flow) and morphometrics in healthy, non-smoking adults.

Methods: This study examined the associations between alcohol consumption and magnetic resonance measures of regional brain perfusion (n = 27) and cortical volumes and thickness (n = 45) in healthy non-smoking adults (22-70 years of age) with no history of alcohol use disorder.

Results: All participants consumed ≤60 standard drink equivalents/month over the 1-year preceding study. Average number of drinks/month over lifetime was 21 ± 11. We hypothesized that a greater multiplicative product of age and total lifetime drinks consumed predicts lower regional brain perfusion, volumes, and thickness. Greater age by total lifetime drinks were related to lower perfusion in numerous bilateral regions, primarily in the frontal, parietal, and occipital lobes, as well as lower bilateral average cortical perfusion. Greater age by total lifetime drinks were related to thinner cortex, primarily in the bilateral frontal and parietal lobes, as well as lower bilateral average cortical thickness.

Conclusions: Findings indicate alcohol consumption considered “low risk” may have consequences for the integrity of cortical tissue, particularly with advancing age. These results may have implications for current harm reduction strategies and alcohol consumption public health guidelines.

ISFAR Summary

In this study by Durazzo et al. (2026), a non-invasive technique called functional Magnetic Resonance Imaging (MRI) is used to measure perfusion in specific brain regions, and structural MRI data are combined to study the association between lifetime alcohol consumption and perfusion in those regions. The authors conclude that “low risk” alcohol consumption may have consequences for the integrity of cortical tissue, particularly with advancing age, based on their finding that greater age by total lifetime drinks was related to lower perfusion in some brain regions and to thinner thickness in some brain regions.

The Forum questions the validity of the reported findings and the conclusions drawn from them. The Forum’s critique concerns aspects of the study design: a small and highly variable group of participants was used, and their assessments of average monthly drinks over the lifetime and total lifetime drinks were considered prone to misclassification. Also, the findings were insufficiently corrected for various confounders, such as drinking pattern and diet. In addition, the statistical power was limited, which increases the risk of both type I and type II errors. Finally, the relevance of the measures performed was considered uncertain for brain health and brain function.

Background

Alcohol consumption affects health in various ways, largely because the association between alcohol consumption and disease outcomes is not always linear. For ischaemic heart disease, ischaemic stroke, and type 2 diabetes, the relationship is J-shaped, with lower risk at low-to-moderate consumption levels (Carr et al., 2026). Long-term and acute excessive alcohol consumption has deleterious effects on health. These two counteracting trends result in a J-shaped association between alcohol consumption and overall mortality (Stone KB & Calonge BN, 2025).

In addition, it should be recognised that alcoholic beverages are not consumed for their health effects or nutritional value, but rather for well-being. Feelings of relaxation, calmness, pleasure, and reduced stress arise when alcohol reaches the brain and temporarily affects signal transduction. Generally speaking, alcohol consumption appears to slow down brain activity. After moderate alcohol consumption, important short-term changes occur in various neurotransmitters, resulting in relaxation and pleasurable effects (Hendriks, 2020).

Excessive long-term alcohol consumption results in alcohol use disorder, with major effects on the brain, including addiction, Wernicke-Korsakoff syndrome, alcohol-related dementia, and foetal alcohol syndrome disorder. Conversely, moderate long-term alcohol consumption has been associated with a reduced risk of dementia (Zhang et al., 2025) and Parkinson’s disease (Hemeda et al., 2026) in meta-analytical reviews.

The relationship between alcohol consumption and brain-related diseases is difficult to study. Functional analyses may include behavioural changes, such as those seen in addiction, and disease outcomes, such as dementia. How alcohol consumption affects brain processes that may contribute to disease development or long-term behavioural changes is difficult to study, since the human brain is not easily accessible. In this study by Durazzo et al. (2026), a non-invasive technique called functional Magnetic Resonance Imaging (MRI) is used to measure perfusion in specific brain regions, using continuous arterial spin labelling (CASL). CASL data are combined with structural MRI data to study the association between lifetime alcohol consumption and perfusion in specific brain regions.

Critique

Whereas structural changes in alcohol use disorder have been extensively reported to date (Bernabeu et al., 2026), Durazzo et al. (2026) aim to add to that knowledge by “studying regional cortical perfusion of the brains of healthy adults with low-level alcohol consumption”. To that purpose, the researchers used a relatively small group of participants who varied considerably in age, included a low percentage of women, and included some obese (class I) subjects. In principle, participants were generally light-to-moderate alcohol consumers.

The drinking pattern of these participants was not analysed, but participants were excluded if they consumed more than 60 standard drinks (14 g alcohol) per month for men (30 for women) without a history of binge drinking in the year prior to the study. This criterion may have excluded some recent problem drinkers but may still have included participants who had been drinking more during earlier periods of their lives. Participants also varied considerably in their recent alcohol consumption; recent alcohol consumption ranged from 12 hours to 14 years. These sources of variation may have affected the results, yet a significant reduction in perfusion, volume, and thickness was still observed in specific brain regions.

Life-time drinking history (LDH) has been used to estimate the total number of drinks consumed over the lifetime of all participants. Indices of LDH validity range from 0.6 (Friesema et al., 2004) to 0.4–0.5 (Koenig et al., 2009) when compared with other pre- and post-prospective assessments of alcohol consumption. However, these instruments vary in structure, lack standardisation, and may fail to capture binge or episodic drinking, highlighting important gaps for refinement (Palmer et al., 2025).

Age and intracranial volume (like BMI) served as covariates in all regression models due to their well-established relationships with morphometric measures. These data were obtained from previous studies conducted on smokers (Durazzo et al., 2018; Durazzo et al., 2017). These studies showed that smoking negatively affected other parts of the brain and that the ageing process was an important factor in the reduction in thickness and volume of various brain regions. Age is also intrinsically linked to the authors’ ‘age by lifetime total drinks’ variable, as older individuals will generally have accumulated more drinks over a longer lifetime. Although age was included as a covariate in the regression models, it remains unclear whether this adjustment adequately accounts for age-related reductions in cortical thickness and volume. Consequently, it is difficult to determine the extent to which the reported associations reflect the independent effects of alcohol consumption rather than the cumulative effects of ageing itself, which may explain why the authors considered the lifetime average number of drinks per month, which could indicate that greater alcohol consumption is associated with greater reductions in cortical thickness and volume. However, age remains a potential confounder, as estimates of long-term alcohol consumption are influenced by both drinking duration and recall accuracy.

Average perfusion reductions associated with ‘one-year average drinks per month’ and ‘lifetime average drinks per month’ appear substantial, with reported reductions of approximately 40-60% in some brain regions. Given the magnitude of these associations, it would have been interesting for the authors to discuss the potential clinical significance of these findings and to compare them with perfusion changes observed in other neurological or vascular conditions. Furthermore, as acknowledged by the authors, other important aspects of brain perfusion, such as blood flow velocities in the cerebral arteries, were not evaluated. Consequently, it remains unclear whether the observed perfusion differences reflect altered neuronal activity, vascular function, or other underlying mechanisms.

The results would have been more convincing if participants had been selected using stricter inclusion and exclusion criteria, with comparisons between distinct drinking groups within defined age groups. This important limitation of the cross-sectional observational design precludes any inference of causality. While the study demonstrates associations between self-reported alcohol exposure and neuroimaging measures, it cannot determine whether alcohol consumption caused the observed differences in cortical perfusion and thickness. Alternative explanations remain plausible, including reverse causation, whereby pre-existing brain characteristics influence drinking behaviours, or residual confounding by unmeasured factors such as physical activity, diet, cardiovascular fitness, socioeconomic status, and other lifestyle variables.

Another important limitation of this study is its small sample size, particularly for the perfusion analyses (n = 27), which substantially reduces statistical power and increases the risk of unstable effect estimates. This concern is compounded by the large number of regional brain measures examined and the multiple alcohol exposure variables analysed. Although the authors applied false discovery rate correction to the primary analyses, many of the headline findings regarding the interaction between age and lifetime alcohol consumption were not corrected for multiple comparisons and were acknowledged by the authors as preliminary. Consequently, the possibility of false-positive findings cannot be excluded, and the reported associations require replication in substantially larger cohorts before firm conclusions can be drawn.

Errors in the text were also noted, including the standard deviation for total lifetime drinks and total lifetime years of drinking being identical, whereas the values differ considerably in Table 1. Beyond the conceptual difficulty of separating age from cumulative alcohol exposure, there is ambiguity surrounding the age × lifetime drinks “interaction” variable, which is a more significant methodological issue than the typographical mistake. Indeed, the authors state that the age × lifetime drinks findings were not FDR-corrected because they anticipated significance would disappear after correction. Yet these same findings become central to the discussion and conclusion. This is not a numerical error but an internal inconsistency between the statistical treatment of the results and the emphasis placed on them in the interpretation.

Altogether, the meaning of these observations is unclear, let alone their physiological relevance. It is therefore exaggerated to state in the conclusions that these results may have implications for current harm reduction strategies and public health guidelines on alcohol consumption. Consequently, the study’s findings should be viewed as hypothesis-generating rather than as evidence that low-risk alcohol consumption causes cortical injury or warrants changes to drinking guidelines.

Specific Comments

Forum member Ellison had one major concern about this publication. “While the very small total number of subjects, the wide range of ages, and the inclusion of a very small number of women pose problems, my key concern is using the number of self-reported drinks over varying periods of time as the sole estimate of alcohol use. We now appreciate that it is much better to determine how the drinks are consumed: the “pattern of drinking”. The pattern includes the type of beverage, how frequently (or infrequently) it is consumed, whether with or without food, and the rate of consumption. Underreporting of intake must also be taken into account when estimating consumption.

Most well-designed previous studies strongly suggest that regular, moderate wine consumption with meals is associated with more favourable health effects than other beverages, especially when those beverages are not consumed regularly and with food. Thus, the drinking pattern is a much better indicator of the health effects of “alcohol use” than the total number of drinks reported. The authors provide extensive data on measurements in the brain but offer no details on whether the “exposure variable” (number of drinks) they used in their analyses accurately reflects alcohol use by the subjects.

As for potential mechanisms, it would be useful to know how the peak blood alcohol level associated with alcohol consumption (which would be higher with irregular, rapid, or food-free drinking) may be an important factor in the effects on the brain. “Regular moderate consumption” cannot be judged just on the number of drinks consumed in a month, a year, or over a lifetime. There are real problems with assuming that simply using the monthly or yearly reported intake of alcohol relates to peak blood alcohol levels on most days. 

Other problems in these analyses include using linear regression for an exposure that clearly has different effects by dose and pattern.  Further, the 1-year average drinks/month ranged from 0 to 60, and total lifetime drinks ranged from 69 to 27,072; the number of days since last alcohol consumption prior to the study ranged from 12 hours to 14 years.  The subjects in this analysis were probably very different from one another in many ways.  Thus, it is difficult to determine how other subject characteristics may have influenced the results.  In addition, since age itself is a key factor in adverse changes in the brain, mixing very young and elderly subjects can cause difficulties in interpretation when combining all subjects into a single analysis.     

There must surely be serious brain damage caused by heavy drinking. However, until exposure to alcohol can be more clearly defined, policy concerns cannot be addressed by studies such as the present one. Large future studies, preferably of similar-aged subjects with more data on their pattern of consumption, will be essential for understanding alcohol’s effects on the brain and for developing appropriate guidelines on alcohol use for the public.

Forum member Romano considers that this study presents intriguing findings on the potential association between low-level alcohol consumption and changes in cortical perfusion and morphology. However, several methodological limitations warrant cautious interpretation of the conclusions. The most important limitation is the relatively small sample size, particularly for the cerebral perfusion analyses (n = 27), whereas structural analyses were conducted in 45 participants. Given the large number of cortical regions and alcohol-consumption variables examined, statistical power is limited, increasing the risk of both type I and type II errors. Although the authors employed bootstrap procedures and false discovery rate corrections for several analyses, they acknowledge that the associations involving the age-by-lifetime alcohol exposure interaction did not consistently withstand stringent correction for multiple comparisons and should therefore be regarded as preliminary.

A second critical issue concerns the cohort’s broad age range (22–70 years). Normal ageing is independently associated with reductions in cerebral perfusion, cortical thickness, and brain volume. Although age was included as a covariate and subsequently incorporated into an interaction term with cumulative alcohol exposure, the marked age heterogeneity may have introduced residual confounding that is difficult to fully control in a modest sample. Furthermore, older participants inevitably accumulated greater lifetime alcohol exposure, creating potential collinearity between age and alcohol burden that complicates the interpretation of the reported interaction effects.

The perfusion findings are the most consistent outcome of the study, with significant inverse associations between lifetime alcohol consumption and cortical blood flow across multiple frontal, parietal, and temporal regions. Nevertheless, the physiological interpretation of these findings remains limited. Perfusion measurements were obtained using arterial spin labelling with a single post-labelling delay and without complementary assessments of arterial flow velocity, cerebrovascular reactivity, or absolute cerebral blood flow. Consequently, it is not possible to determine whether the observed reductions reflect neuronal dysfunction, vascular alterations, age-related haemodynamic variability, or a combination of these mechanisms.

The authors propose cumulative oxidative stress as a biological mechanism linking alcohol exposure to reduced cortical perfusion and thinning. While this explanation is biologically plausible, it remains speculative because no direct biomarkers of oxidative stress, inflammation, endothelial dysfunction, or vascular injury were measured. In addition, the cross-sectional design precludes causal inference and does not rule out the possibility that pre-existing structural or functional brain differences contributed to the observed drinking patterns.

Overall, the study provides preliminary evidence that alcohol consumption traditionally regarded as “low risk” may be associated with age-related reductions in cortical perfusion and cortical thickness. However, the limited sample size, substantial age variability, and methodological constraints of the perfusion measurements mean the findings should be interpreted as exploratory. Confirmation through larger, longitudinal investigations incorporating comprehensive vascular and biological assessments will be necessary before firm conclusions regarding the neurobiological consequences of low-level alcohol consumption can be drawn.

Forum member McIntosh shares other members’ concerns about the small sample sizes. “I also think the paper is misleading. The authors produce graphs of certain brain characteristics plotted against alcohol use. At the top of each figure is a measure of goodness of fit, r, and its P value. One might be tempted to believe that, because the r values are significant, there is a significant negative relation between the brain characteristic and alcohol use. This is not what the figure shows. The significance of r is due to the inclusion of other variables, such as age and total intracranial volume, VPI. The true measure of the slopes of these lines is the regression coefficient for alcohol use. These are not reported, so we have no idea whether the relationship is significant. Topiwala et al. (2022, Figure 7) report these values, which are significant. Hence, their results are probably correct, but they provide no convincing evidence for this. As a result, the paper adds nothing to the literature on brain function and alcohol use.”

Forum member Harding’s comments on this paper concern the robustness of the data on alcohol consumption on which the conclusions are based. “I agree with the concerns expressed about the small sample size, wide age range, and lack of consideration of the pattern of drinking. The information on alcohol consumption was drawn from standardised questionnaires assessing lifetime alcohol consumption (Lifetime Drinking History, LDH). I have looked at the references cited to support this approach (Skinner and Sheu, 1982, Sobell et al., 1988, Abé et al., 2013). Skinner et al. (1982) says that the Lifetime Drinking History is a structured interview designed to provide quantitative data on patterns of alcohol consumption from the onset of regular drinking. It is based on the Michigan Alcoholism Screening Test (MAST) (Selzer, 1971). Questionnaires designed to identify alcohol misuse and addiction in a population cannot be an effective way of identifying those who are drinking moderately (less than 60 drinks/month), because they are designed for a completely different purpose. The Abé et al. (2013) paper does not seem to be relevant at all in this regard.

In addition, the Skinner et al. (1982)and Sobell et al. (1988) papers address the reliability of this self-reporting, but what this means in practice is the extent to which the answers to the same questions are consistent when subjects are asked them on different occasions, rather than how reliable they are in assessing actual intake, which is not known. It is about the consistency of the answers, not the extent to which the answers are correct.

This data should be treated with the same caution as the outcomes of food frequency questionnaires, which are usually based on current dietary intake and are notoriously unreliable. This study is based not on current intake, but on recall of intake over the previous three years and, indeed, over a lifetime. Furthermore, as Forum member Ellison points out, the literature is awash with assertions that self-reported alcohol consumption is routinely grossly under-estimated. This is determined by comparing what individuals say they drink with the amount of alcohol actually consumed (based on the amount sold or taxed). For example, Stockwell (2014) estimates this to be in the range 40%-65%.

Yet in Durazzo et al. (2026), none of these factors is explored in any detail, and the intake data is treated as reliably solid when compared with brain measurements. For all these reasons, I cannot see how any of the conclusions drawn are reasonable. At best, they should serve as a basis for designing more robust experimental protocols and cannot be regarded as, to quote the paper, ‘having implications for current harm reduction strategies and alcohol consumption public health guidelines’.”

References

Abé, C., Mon, A., Hoefer, M.E., Durazzo, T.C., Pennington, D.L., Schmidt, T.P., & Meyerhoff, D.J. (2013) Metabolic abnormalities in lobar and subcortical brain regions of abstinent polysubstance users: magnetic resonance spectroscopic imaging. Alcohol and Alcoholism, 48(5), 543-51.

Bernabeu, T., Vollstädt-Klein, S., Kiefer, F., & Rolland, B. (2026). Structural brain effects of alcohol use patterns in alcohol use disorder: A systematic review and Meta-analysis. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 146. https://doi.org/10.1016/j.pnpbp.2026.111684

Carr, S., Espinosa Dice, A. L., Gmel, G. E., Hassan, A. S., Shield, K. D., & Rehm, J. (2026). A review of the relationship between dimensions of alcohol consumption and the burden of disease: 2026 update including Mendelian randomisation studies. Addiction (Abingdon, England). https://doi.org/10.1111/ADD.70435

Durazzo, T. C., Joseff, B. D. P., McNerney, M. W., Humphreys, K., & Meyerhoff, D. J. (2026). The interaction of age and total lifetime drinks is associated with regional cortical perfusion and thickness in healthy adults with low-level alcohol consumption. Alcohol, 133, 38–47. https://doi.org/10.1016/j.alcohol.2026.03.006

Durazzo, T. C., Meyerhoff, D. J., & Yoder, K. K. (2018). Cigarette smoking is associated with cortical thinning in anterior frontal regions, insula and regions showing atrophy in early Alzheimer’s Disease. Drug and Alcohol Dependence, 192, 277–284. https://doi.org/10.1016/j.drugalcdep.2018.08.009

Durazzo, T. C., Meyerhoff, D. J., Yoder, K. K., & Murray, D. E. (2017). Cigarette smoking is associated with amplified age-related volume loss in subcortical brain regions. Drug and Alcohol Dependence, 177, 228–236. https://doi.org/10.1016/j.drugalcdep.2017.04.012

Friesema, I. H. M., Veenstra, M. Y., Zwietering, P. J., Knottnerus, J. A., Garretsen, H. F. L., & Lemmens, P. H. H. M. (2004). Measurement of lifetime alcohol intake: utility of a self-administered questionnaire. American Journal of Epidemiology, 159(8), 809–817. https://doi.org/10.1093/AJE/KWH102

Hemeda, S., Elmadani, M., Merzah, M., Odeh, A., Pintér, E., Balás, I., Janszky, J., Pintér, D., & Kovács, N. (2026). Impact of Alcohol Intake on Parkinson’s Disease Risk and Progression: A Systematic Review and Dose-Response Meta-Analysis of Prospective Studies. Movement Disorders: Official Journal of the Movement Disorder Society, 41(4). https://doi.org/10.1002/MDS.70200

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

Koenig, L. B., Jacob, T., & Haber, J. R. (2009). Validity of the lifetime drinking history: a comparison of retrospective and prospective quantity-frequency measures. Journal of Studies on Alcohol and Drugs, 70(2), 296–303. https://doi.org/10.15288/JSAD.2009.70.296

Palmer, A. J., Connor, J. P., Holtmann, G., Saunders, J. B., Rice, K., Yeo, J., Huang, A., & Clark, P. J. (2025). Measuring lifetime alcohol exposure: A scoping review and implications for translational research and alcohol-related harm. Addiction, 121(3), 472–487. https://doi.org/10.1111/ADD.70208;PAGE:STRING:ARTICLE/CHAPTER

Selzer, M.L. (1971) The Michigan Alcoholism Screening Test, the quest for a new diagnostic instrument. American Journal of Psychiatry, 127, 89-94.

Skinner, H.A. & Sheu, W. J. (1982). Reliability of alcohol use indices. The Lifetime Drinking History and the MAST. Journal of Studies on Alcohol, 43(11), 1157–1170.

Sobell, L.C., Sobell, M.B., Riley, D.M., Schuller, R., Pavan, D.S., Cancilla, A., Klajner, F., & Leo, G.I. (1988) The reliability of alcohol abusers’ self-reports of drinking and life events that occurred in the distant past. Journal of Studies on Alcohol, 49(3), 225-32. https://doi.org/10.15288/jsa.1988.49.225.

Stockwell, T., Zhao, J., & Macdonald, S. (2014) Who under-reports their alcohol consumption in telephone surveys and by how much? An application of the ‘yesterday method’ in a national Canadian substance use survey. Addiction, 109(10), 1657-66.

Stone KB, & Calonge BN. (2025). Review of Evidence on Alcohol and Health. Review of Evidence on Alcohol and Health. https://doi.org/10.17226/28582

Topiwala, A., Ebmeier, K.P., Maullin-Sapey, T., & Nichols, T.E. (2022) Alcohol consumption and MRI markers of brain structure and function: Cohort study of 25,378 UK Biobank participants. Neuroimage Clinical, 35, 103066.

Zhang, R., Li, B., & Miao, Y. (2025). Alcohol consumption and risk of dementia: a systematic review and meta-analysis. Internal Medicine Journal. https://doi.org/10.1111/IMJ.70288

Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:

Henk Hendriks, PhD, Independent consultant and partner of the Nutrition Consultants Cooperative, Netherlands

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

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

Raquel Romano, PhD, Independent consultant and Professor of Applied Technology at the University of Aconcagua, Argentina

James McIntosh, PhD, formerly Professor of Economics, Concordia University, Montreal, Canada

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