Three-Year Trial Shows Mediterranean Diet Eases Food Addiction Symptoms but Only for Some

A landmark three-year randomized trial has delivered some of the strongest evidence yet that food addiction, a controversial but increasingly studied pattern of disordered eating, can be measurably improved through a structured dietary intervention. The findings, drawn from the PREDIMED-Plus trial and published in BMC Medicine, suggest that an intensive lifestyle programme built around an energy-reduced Mediterranean diet can reduce food addiction symptoms and improve cardiometabolic health in older adults with overweight or obesity and metabolic syndrome. But the study also carries a sobering caveat: people who entered the trial with food addiction symptoms were less able to stick to the programme and gained less from it over the long term, raising the prospect that food addiction may be a clinically meaningful marker for tailoring obesity treatment.

The research team, led by Lucía Camacho-Barcia and Fernando Fernández-Aranda of Bellvitge University Hospital and the Bellvitge Biomedical Research Institute in Barcelona, together with Jordi Salas-Salvadó of Universitat Rovira i Virgili, analysed data from 448 participants enrolled in the PREDIMED-Plus Cognition sub-study. Participants had a mean age of 65.3 years and were randomly assigned to one of two groups. The intensive intervention group received an energy-reduced Mediterranean diet combined with a physical activity programme and behavioural support, while the control group was advised to follow a Mediterranean diet ad libitum, meaning without any deliberate calorie restriction or structured support. The researchers tracked food addiction symptomatology using the Yale Food Addiction Scale 2.0, a validated instrument that applies criteria analogous to those used for substance-related disorders in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders, alongside anthropometric measurements, diet and physical activity adherence, and quality of life at baseline and at one and three years.

Food addiction is not a formal diagnosis in current psychiatric classification manuals, but the construct has gained traction among researchers because it captures a cluster of behaviours that many patients with obesity describe: loss of control over eating, continued consumption despite negative consequences, unsuccessful attempts to cut down, and withdrawal-like distress when palatable foods are unavailable. The Yale Food Addiction Scale 2.0 operationalises these symptoms and allows researchers to quantify both the presence of a food addiction phenotype and the severity of symptoms on a continuous scale. Critics have long argued that the label risks medicalising ordinary overeating, while proponents counter that a subset of patients genuinely exhibits an addictive-like relationship with food, particularly with ultra-processed products engineered for high palatability. The new trial is notable because it examines this phenotype not in a laboratory or a specialist eating disorders clinic, but within one of the largest and most rigorous nutrition intervention trials ever conducted.

The results on symptom change were nuanced. Food addiction severity declined in both study arms over the three-year follow-up, which the authors interpret as partly reflecting the natural fluctuation of symptoms and the effects of study participation itself. However, only the intensive intervention group sustained those reductions to the three-year mark, and at that final assessment the intensive group had significantly lower food addiction scores than the control group, a difference that reached statistical significance at p equals 0.011. In other words, simply being advised to eat a Mediterranean diet freely was not enough to produce lasting change in addictive-like eating, whereas a structured, energy-restricted, behaviourally supported programme was. This pattern mirrors what addiction science would predict: sustained change in compulsive behaviours typically requires active, repeated intervention rather than passive advice.

Where the trial becomes particularly striking is in the stratified analysis by baseline food addiction status. Among participants who did not meet criteria for food addiction at the start of the study, the intensive intervention dramatically increased the odds of achieving clinically meaningful outcomes. The likelihood of losing at least eight percent of body weight was nearly tripled, with an odds ratio of 2.87 and a 95 percent confidence interval of 1.99 to 4.14. The odds of achieving at least a five percent reduction in waist circumference, a proxy for visceral adiposity and cardiometabolic risk, were similarly elevated, with an odds ratio of 3.01 and a confidence interval of 2.20 to 4.11. Improvements in adherence to the Mediterranean diet were even more pronounced, with participants without food addiction in the intensive arm showing an odds ratio of 5.82 for meaningful adherence gains, significant at p less than 0.001.

For participants with food addiction at baseline, the picture was considerably less encouraging. The intensive intervention still produced some benefit, and food addiction symptoms did improve, but the magnitude of weight loss, waist circumference reduction, and dietary adherence gains was attenuated compared with their counterparts without the phenotype. The researchers used structural equation modelling, a statistical technique that allows researchers to test hypothesised networks of causal pathways in observational and trial data, to explore how these variables related over the three-year period. The models indicated that improvements in Mediterranean diet adherence mediated, or statistically explained, the long-term improvements in body weight, waist circumference, and quality of life, but only in individuals without food addiction. In the food addiction group, that mediating pathway was absent or substantially weakened, suggesting that the behavioural mechanism through which the diet programme usually works was disrupted.

These findings have immediate implications for how clinicians and researchers think about personalised obesity medicine. If food addiction marks a subgroup of patients who respond less well to standard lifestyle interventions, then screening for it at the outset of treatment could identify who needs additional or different support, perhaps drawing on techniques from addiction treatment such as cognitive behavioural therapy, motivational interviewing, or pharmacological approaches that target craving and reward circuitry. The study’s authors frame food addiction as a clinically relevant behavioural phenotype, and the trial provides the kind of longitudinal, randomised evidence that has been largely missing from a debate often fuelled by cross-sectional studies and self-selected samples. It is one thing to show that people with addictive-like eating differ cross-sectionally from others; it is quite another to show, in a randomised framework over three years, that the phenotype predicts differential adherence and differential benefit from a specific intervention.

The trial also speaks to the ongoing scientific controversy over ultra-processed foods and their role in addictive eating. Although the published report focuses on the intervention effects rather than on specific food categories, the food addiction construct is closely entwined with research on ultra-processed products, which are typically high in refined carbohydrates and fats and are thought to hijack reward pathways in ways that whole foods do not. The Mediterranean diet, rich in olive oil, nuts, legumes, fish, vegetables, and whole grains, represents almost the nutritional opposite of the ultra-processed food environment, and the marked adherence improvements in the intensive group suggest that replacing the food environment, with support, is feasible even in older adults. That feasibility matters, because much of the food addiction literature has been criticised for documenting the problem without testing solutions.

Several limitations deserve attention when weighing the findings. The participants were older adults with metabolic syndrome enrolled in a cardiovascular prevention trial, so generalisation to younger populations or to patients with severe obesity remains untested. Food addiction was measured by self-report questionnaire rather than clinical interview, and the construct itself remains scientifically contested. The structural equation models, while sophisticated, describe statistical associations consistent with mediation rather than proving causal chains. The trial was registered with the International Standard Randomized Controlled Trial Number registry as ISRCTN 89898870 in July 2014, and the analysis was embedded within the broader PREDIMED-Plus Cognition sub-study, meaning the food addiction assessment was added to a trial designed primarily for other endpoints. None of these caveats undermines the central result, but they do bound the conclusions.

What makes the study resonate beyond the specialist literature is its reframing of a familiar clinical frustration. Weight-loss programmes fail for many people, and clinicians have long observed that some patients seem unable to sustain adherence no matter how well-designed the programme. This trial provides prospective, randomised evidence that one measurable psychological phenotype, food addiction, predicts exactly that pattern of attenuated response, while simultaneously showing that the phenotype itself is not fixed: symptoms improved under intensive intervention. The dual message, that food addiction both limits standard treatment response and can itself be treated, points toward a future in which obesity care begins with a brief behavioural assessment and routes patients toward the intensity and type of support their profile requires. For a field searching for ways to move beyond one-size-fits-all dietary advice, that is a genuinely consequential step.

Subject of Research: The effect of an intensive energy-reduced Mediterranean diet lifestyle intervention on food addiction symptoms and weight-loss outcomes in older adults with overweight or obesity and metabolic syndrome

Article Title: Targeting food addiction through an intensive energy-reduced Mediterranean diet lifestyle intervention: evidence from the 3-year PREDIMED-plus trial

Article References: Camacho-Barcia, L., Jimenez-Murcia, S., Granero, R., de la Torre, R., Babio, N., Pintó, X., Corella, D., Gazzoli, S., Frances, F., Riera-Mestre, A., Fitó, M., Cuenca-Royo, A., Gómez-Martínez, C., Gearhardt, A. N., Forcano, L., Salas-Salvadó, J., & Fernández-Aranda, F. (2026). Targeting food addiction through an intensive energy-reduced Mediterranean diet lifestyle intervention: evidence from the 3-year PREDIMED-plus trial. BMC Medicine. https://doi.org/10.1186/s12916-026-05283-z

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05283-z

Keywords: food addiction, Mediterranean diet, PREDIMED-Plus, obesity, weight loss, metabolic syndrome, Yale Food Addiction Scale 2.0, lifestyle intervention, dietary adherence, quality of life, randomized controlled trial, personalised nutrition