Ketogenic Diet Trial in Treatment-Resistant Depression

Key Takeaways

• The ketogenic diet was associated with a modestly greater reduction in depressive symptoms at 6 weeks, but there was no significant between-group difference at 12 weeks.
• Most secondary outcomes were similar between groups, although anxiety scores at week 12 favored the ketogenic diet.
• No serious adverse events were reported, and adherence declined after the intensive support phase ended.

The 95% CI was -4.33 to -0.03, with P=.05. In the UK trial, 88 adults with treatment-resistant depression were randomized to 6 weeks of a ketogenic diet or a matched phytochemical diet and followed through 12 weeks. The between-group separation in depressive symptoms was not maintained at 12 weeks.

Eligible participants were adults aged 18 to 65 years with diagnosed depression, PHQ-9 scores of 15 or greater, and 2 adequate antidepressant trials in the current episode. The UK study used a single-blind, 1:1 randomized design; baseline characteristics included a mean age of 42.1 years, and 69% were women. Participants in the ketogenic group received prepared meals, urine ketone strips, weekly individual dietetic support or coaching, and less than 30 g of carbohydrates daily, with individualized energy targets to avoid weight loss. The phytochemical group added a daily fruit or vegetable serving, replaced saturated fats with unsaturated fats, and followed the same weight-maintenance plan. Support intensity was matched across both groups throughout the intervention.

The prespecified primary outcome was the between-group difference in change in PHQ-9 score from baseline to week 6. Mean PHQ-9 change at week 6 was -10.5 in the ketogenic group and -8.3 in the phytochemical group. At 12 weeks, the between-group PHQ-9 difference was -1.85 (95% CI, -4.04 to 0.33; P=.10). Remission occurred in 11 ketogenic and 4 phytochemical participants at 6 weeks, and in 8 versus 4 at 12 weeks after 3 ketogenic relapses. The per-protocol analysis of 61 participants showed a similar direction, but the early separation in symptoms did not persist at follow-up.

Secondary and exploratory analyses assessed anxiety, anhedonia, cognitive impairment, quality of life, and functioning, with no overall between-group differences in most outcomes. One exception emerged for anxiety at week 12, with a GAD-7 difference of -2.02 (95% CI, -3.95 to -0.10; P=.04). Exploratory subgroup analyses suggested a larger response among participants with severe baseline depression than among those with moderately severe depression. Exploratory analyses did not show a significant association between ketone concentration and PHQ-9 change. Broader symptom and functional advantages were not shown overall.

No serious adverse events occurred, and no serious events were related to research procedures. Two participants in the ketogenic group stopped early, and 1 participant in the phytochemical group discontinued the diet. At least 84% of ketogenic participants measured urine ketones at least twice weekly, and at least 50% measured them daily. Twenty-eight participants maintained ketone concentrations of 1.5 mmol/L or greater at least 60% of the time. Adherence fell after the intensive support phase ended, with only 9% continuing the ketogenic diet nearly every day by week 12.