Keto for Mood in 2025: What the New Evidence Says About Depression—and How to Apply It Safely Today

What changed this week?

• A November 5, 2025 JAMA Psychiatry meta-analysis pooling 50 studies (41,718 participants) reported a significant association between ketogenic diets and reduced depressive symptoms in randomized trials; effects were stronger in studies that monitored ketones and used very‑low‑carb prescriptions. No consistent benefit for anxiety in RCTs. ^[3]
• News coverage echoed the cautious optimism: benefits for depression appear modest and contingent on bona fide ketosis and program fidelity; anxiety data remain preliminary. ^[4]
• In the background literature this fall, small, uncontrolled pilot data in college students suggested large depression score drops on a well‑formulated keto diet—useful for hypothesis generation but not clinical proof. ^[5]

“Ketogenic diets show potential associations with reduced depressive symptoms… evidence for anxiety is inconclusive; larger, high‑quality trials are needed.” ^[6]

Science snapshot: what’s proven vs. promising

More solid (moderate evidence)

• Depression symptoms: Modest improvements in adults on ketogenic diets, especially when ketosis is biochemically verified. Quality and consistency of the intervention matter. ^[7]
• Not a cure‑all: Anxiety outcomes were not significantly improved in RCTs to date. ^[8]

Promising, needs confirmation

• Serious mental illness (bipolar, schizophrenia): Early trials, case series, and an editorial in 2025 suggest feasibility and possible symptom improvements, but robust RCTs remain limited. ^[9]
• Sleep/cognition: In healthy adults, a 3‑week crossover RCT found no meaningful differences in sleep or mood vs. high‑carb; in operational stress models (sleep deprivation), short keto phases sometimes improved vigilance—small samples, short duration. ^[10]
• Exogenous ketones for sleep: A February 2025 RCT of D‑β‑hydroxybutyrate improved some self‑reported sleep domains; another trial under hypoxia showed no sleep quality benefit—do not extrapolate to mood. ^[11]

Why keto might affect mood (mechanistic plausibility)

• Energy metabolism: Ketones provide an alternative brain fuel, potentially improving mitochondrial efficiency and reducing oxidative stress. ^[12]
• Neurotransmission and inflammation: Reviews in 2025 highlight potential effects on GABA/glutamate balance and inflammatory signaling—relevant to mood pathways. Evidence is mixed and largely preclinical/early clinical. ^[13]

Who might consider a “mood‑smart” ketogenic trial?

• Adults with depressive symptoms under clinician care who want a nutrition adjunct (not a replacement) to therapy/medication. Strongest effects were seen with very‑low‑carb protocols and ketone monitoring. ^[17]

Who should be cautious or avoid?

• People on SGLT2 inhibitors (for diabetes): Risk of euglycemic ketoacidosis rises with carbohydrate restriction—coordinate closely with your prescriber. ^[18]
• History of eating disorders, pregnancy/breastfeeding, or complex medical conditions: require individualized medical supervision.

How to do “mood‑smart keto” in practice

Macronutrient targets and ketosis

• Start: 20–30 g net carbs/day, protein ~1.2–1.6 g/kg reference body weight, remaining calories from fat. This typically yields nutritional ketosis (β‑HB ~0.5–1.5 mmol/L) for many adults; verify with a blood ketone meter and adjust food quality and carb tolerance accordingly. Effects in depression were stronger when ketosis was actually monitored. ^[19]

Electrolytes and supplements (keep it simple)

• Sodium 4–5 g/day from foods + broth as tolerated; potassium and magnesium from leafy greens, nuts, seeds; consider magnesium glycinate 200–400 mg if intake is low (discuss with your clinician).
• Omega‑3 (EPA/DHA) from fatty fish 2–3x/week; consider a supplement if intake is low—benefits mood generally, independent of keto (evidence base outside keto scope).
• Exogenous ketones: Not recommended for mood at this time; sleep data are mixed and do not establish an antidepressant effect. ^[20]

3-day mood‑smart keto sample (nutrient‑dense, high‑fiber, whole foods)

How to track what matters (and avoid over‑interpreting)

• Outcomes: PHQ‑9 weekly, sleep quality (bed/wake times), step count or light activity, and subjective energy/focus. Reassess at 4–8 weeks.
• Ketosis: Aim for nutritional ketosis (e.g., β‑HB ≥0.5 mmol/L) several days per week; consistent, verified ketosis was a moderator of benefit in depression studies. ^[22]
• Safety checks: If on diabetes meds—especially SGLT2 inhibitors—review a sick‑day plan and DKA precautions with your clinician before starting low‑carb. ^[23]

Common pitfalls (and fixes)

Where this could go next (and what we still need)

• Better RCTs that verify ketosis, standardize protocols, and track durability beyond 8–12 weeks. ^[25]
• Clarify subgroups (e.g., insulin‑resistant depression, inflammatory phenotypes) most likely to respond—an active question in metabolic psychiatry. ^[26]
• Sleep: Mixed findings to date; more trials should integrate objective sleep metrics alongside mood scales. ^[27]

Actionable summary

• If you and your clinician choose a keto trial for depression, commit 4–8 weeks, verify ketosis, and track PHQ‑9 weekly.
• Build meals around protein + non‑starchy veg + olive oil/avocado/nuts; use broth/electrolytes to blunt early side effects.
• Continue standard mental‑health care; revisit the plan if depression scores don’t improve by week 4–6.
• On SGLT2 inhibitors or complex regimens? Don’t start without medical supervision due to DKA risk. ^[28]

References

• Janssen‑Aguilar R, et al. Ketogenic Diets and Depression and Anxiety: A Systematic Review and Meta‑analysis. JAMA Psychiatry. Published online November 5, 2025. ^[29]
• AJMC coverage of the meta‑analysis (summary, context). November 5, 2025. ^[30]
• BJPsych Open editorial on ketogenic diet and neuropsychiatric disorders (May 2025). ^[31]
• J Clin evidence review on serious mental illness and ketogenic diet (2024–2025 synthesis). ^[32]
• Randomized crossover trial: 3 weeks of nutritional ketosis vs. high‑carb showed no differences in sleep/mood in healthy adults. ^[33]
• Randomized trial: D‑β‑hydroxybutyrate and sleep quality (healthy adults). February 6, 2025. ^[34]
• Trial: Overnight hypoxia with/without ketone ester—no improvement in sleep quality. ^[35]
• FDA safety communication on SGLT2 inhibitors and (eu)glycemic ketoacidosis—relevant when combining low‑carb with these agents. ^[36]