The study, led by Julia-Sophia Scheuermann of Friedrich-Alexander-Universität Erlangen-Nürnberg together with colleagues in Germany and Austria, is a secondary data analysis of the BrainFit-Nutrition randomized controlled trial. That trial used a two-by-two factorial design, combining a computerized cognitive training intervention with a nutritional counseling intervention, allowing researchers to disentangle the effects of each component. The newly published analysis focused specifically on what happened to alcohol consumption over twelve months among participants who drank at the start of the study and who received one of two distinct nutritional counseling programs.

The technical details of the trial design matter for interpreting the findings. Of the 270 trial participants, 189 reported consuming alcohol at baseline and began one of the two nutritional counseling arms, forming the intention-to-treat sample for this analysis. Data were collected at a twelve-month follow-up, with values imputed for fifteen participants whose data were incomplete. Participants were randomly assigned either to whole-food plant-based nutritional group counseling or to nutritional group counseling based on the recommendations of the German Nutrition Society, the country’s leading dietary authority. Crucially, neither program focused solely on alcohol; the counseling addressed diet broadly, making any reduction in drinking an indirect, and arguably more naturalistic, outcome.

Alcohol consumption was quantified in a standardized way that allows cross-study comparison: standard drinks per day, with one standard drink defined as ten grams of pure alcohol. This metric sidesteps the ambiguity of self-reported glasses or bottles, which vary enormously in actual alcohol content. The researchers then applied multiple linear regression models to identify which baseline characteristics predicted who would reduce their drinking the most, with p-values corrected to guard against false-positive findings.

The headline result is striking in its simplicity. Both nutritional counseling programs led to a statistically significant reduction in alcohol consumption among people with mild cognitive impairment after twelve months. The effect size, expressed as Cohen’s d, was 0.30 for the whole-food plant-based program and 0.39 for the counseling based on German Nutrition Society recommendations. In behavioral intervention research, effect sizes in this range are considered small to moderate, but for a lifestyle outcome as entrenched as drinking habits in adults averaging 70.6 years of age, they represent a clinically meaningful shift. The sample was nearly gender-balanced, with 52 percent women, which strengthens the generalizability of the findings across sexes.

Perhaps the most intriguing finding is that the two dietary approaches performed almost identically. There was no significant difference between the counseling groups in how much alcohol consumption changed over time, with the comparison yielding a p-value of 0.173. This suggests that the active ingredient was not the specific dietary philosophy, whether plant-based or conventional, but the structure and support of group counseling itself. Regular contact with trained counselors and peers, attention to everyday eating and drinking routines, and the accountability of a program appear to have nudged participants away from alcohol without any explicit abstinence message.

Sex, however, turned out to be a decisive moderator. Women substantially reduced their alcohol consumption over the twelve months only in the German Nutrition Society counseling group, with an effect size of 0.33, while showing no comparable reduction in the plant-based arm. The authors interpret this pattern through the lens of the gender health gap, arguing that nutritional counseling programs may need to be designed differently for men and women to achieve equitable benefit. Their conclusion is direct: a gender-specific nutritional counseling program is advised in light of these differences.

To understand who benefited most, the researchers isolated the successful alcohol reducers, a subgroup of 84 participants, and modeled the predictors of greater reduction. Three baseline characteristics emerged as significant positive predictors. Higher baseline alcohol consumption was by far the strongest, with a standardized beta coefficient of 0.670 and a p-value below 0.001, meaning that heavier drinkers had the most room to cut back and did so most dramatically. Older age predicted greater reduction as well, with a beta of 0.158 and a p-value of 0.035, and stricter adherence within the counseling group also predicted larger reductions, with a beta of 0.184 and a p-value of 0.017. The overall regression model was highly significant, with F(7,76) equal to 18.539 and a p-value of 0.003.

One predictor ran in the opposite direction and deserves close attention. Higher cognitive functioning was a significant negative predictor of reduction, with a beta of minus 0.199 and a p-value of 0.009. In other words, participants whose cognition was relatively better preserved were less likely to reduce their drinking, while those with more impaired cognition cut back more. The authors draw a practical implication from this counterintuitive pattern: prevention efforts should focus on younger people living with mild cognitive impairment as well as higher educated people, especially men, to optimize the potential of modifiable risk factors for dementia. The logic is that those who are younger, better educated, and cognitively stronger may perceive less immediate threat from their drinking and therefore need more targeted motivational approaches.

Taken together, the findings carry implications well beyond a single trial. Dementia prevention research has increasingly emphasized modifiable lifestyle risk factors, and alcohol sits alongside physical inactivity, hypertension, and social isolation on that list. This study demonstrates that a broad, non-judgmental dietary intervention, delivered remotely to older adults with cognitive impairment, can produce measurable reductions in alcohol intake sustained over a full year. It also demonstrates that one-size-fits-all prevention may leave women behind, and that the people who seem least at risk on the surface may be the hardest to reach. As populations age and dementia prevalence climbs, the humble question of water versus wine, asked in a supportive online group setting, may prove to be one of the more scalable tools in the prevention toolkit. The trial was registered as ISRCTN10560738, and the study was funded by the Karl and Veronica Carstens Foundation, which had no role in the design, analysis, or publication of the research.

Subject of Research: Effects of online nutritional counseling on alcohol consumption in older adults with mild cognitive impairment

Article Title: Should I drink water or wine? Impacts of online nutritional counseling on the alcohol consumption of elderly individuals with mild cognitive impairment

Article References: Scheuermann, J.-S., Graessel, E., Scheerbaum, P., Hanslian, E., Schiele, J. K., Kessler, C. S., & Donath, C. (2026). Should I drink water or wine? Impacts of online nutritional counseling on the alcohol consumption of elderly individuals with mild cognitive impairment. BMC Public Health, 26(1), Article 2761. https://doi.org/10.1186/s12889-026-29728-5

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29728-5

Keywords: mild cognitive impairment, alcohol consumption, nutritional counseling, dementia prevention, randomized controlled trial, older adults, whole-food plant-based diet, German Nutrition Society, gender differences, lifestyle intervention, telehealth, public health