Small antidepressant effect with keto diet seen in treatment-resistant depression

New research suggests that ketogenic diet, when used as an adjunct to pharmacotherapy, may confer a small antidepressant benefit in adults with treatment-resistant depression (TRD).

After 6 weeks of intervention, depression severity decreased substantially more for participants who received ketogenic diet (keto group) than for those who received a control phytochemical diet (phyto group). The mean change in 9-item Patient Health Questionnaire (PHQ-9) score was –10.5 vs –8.3, respectively, corresponding to a treatment effect of −2.18 (95 percent confidence interval [CI], −4.33 to −0.03; p=0.05). [JAMA Psychiatry 2026;doi:10.1001/jamapsychiatry.2025.4431]

Further analyses indicated that the response to the ketogenic diet was greater in participants with severe depression (PHQ-9 score, 20-27) than in those with moderate depression (PHQ-9, 15-19). The 6-week difference in PHQ-9 score between the keto and the phyto groups was −4.73 (95 percent CI, −8.16 to −1.30) for participants with severe depression and 0.16 (95% CI, −2.30 to 2.63) for those with moderately severe depression (p=0.02 for the difference in treatment effects).

Remission from depression (PHQ-9 score ≤4) at 6 weeks occurred in 25 percent of participants in the keto group and in 9 percent of those in the phyto group. At the 12-week follow-up, 7 percent in the keto group experienced relapse, while 18 percent had sustained remission.

The keto group showed no significant improvements in secondary outcomes such as anxiety, cognition, and functional measures compared with the phyto group.

There were no incidents of serious adverse events.

Uncertain relevance

“This randomized clinical trial provides preliminary evidence that adherence to a ketogenic diet may have small antidepressant benefits in people with TRD,” according to lead study investigator Dr Min Gao from the University of Oxford in Oxford, UK.

Gao described the clinical relevance of ketogenic diet’s effect in TRD to be “uncertain,” as the observed PHQ-9 score difference of −2.2 points between the keto and the phyto groups was lower than the prespecified minimal clinically important difference of 5 points.

“The per-protocol analysis did not suggest a larger effect in those who were adherent. Evidence of improvement persisted in different analyses but with wide 95 percent CIs, and ketone concentrations were not associated with depression improvement,” Gao said.

Furthermore, adherence to the ketogenic diet proved challenging, and only 9 percent of participants reported diet continuation after the provision of keto meals and support ended, he added.

Results of semistructured interviews conducted in a subset of participants in the keto group after the completion of the study brought to light the demanding nature of ketogenic diets and associated logistical burdens (eg, meal planning, social restrictions) and highlighted the importance of external support. Many participants expressed the value and convenience of having prepared meals supplied, which simplified diet compliance by eliminating the time and cost of extensive meal planning and reducing the risk of consuming noncompliant foods.

“Notably, participants did not experience relapse to previous levels of depression,” despite the fact that only a few continued with the ketogenic diet after all support stopped, Gao pointed out. “[This suggests] either that a short period of carefully supported dietary intervention may have enduring benefits or that the mechanism of any improvement was unrelated to ketosis and perhaps linked to study participation.”

Study details

A total of 88 adult participants (mean age 42.1 years, 69 percent female, 88 percent White, mean BMI 32.3 kg/m²) with TRD were randomly assigned to either the keto group (n=44) or the phyto group (n=44). The median duration of the current depressive episode was 16 months, and 93 percent were receiving antidepressant monotherapy. The mean PHQ-9 score at baseline was 19.5.

Participants in the keto group received three prepared ketogenic diet meals per day, snacks, and urine ketone strips for free for 6 weeks. They were instructed to test first-morning urine for ketones at least twice weekly, with results recorded and reported during weekly 30-min counselling sessions with a dietitian or trained coach. Individualized energy targets were set and snack portions adjusted if weight changed by 0.5 kg or more on 2 consecutive days, with the aim of avoiding the confounding effect of weight loss.

Meanwhile, participants in the phyto group followed a diet aimed at increasing phytochemical intake by adding one differently coloured fruit or vegetable each day and replacing saturated animal fats with unsaturated plant oils. They received food vouchers to support the purchase of recommended foods, as well as dietary support of similar frequency and duration as the keto group.