Ketogenic diet · mild cognitive impairment

Ketogenic Diet, Mild Cognitive Impairment and the Screening Workup

The ketogenic diet produces modest, reproducible cognitive gains in mild cognitive impairment, MMSE up 1.25 points across ten pooled trials with daily function unchanged, but those trials excluded the insulin-resistant patient most panels carry. The screening workup is a metabolic and cognitive baseline before any supervised trial.

Patients with memory complaints now arrive asking about ketones. The evidence is modest, specific and built on exclusions, and it names the baseline numbers a supervised trial needs.

Questions that pair the ketogenic diet, mild cognitive impairment and memory now reach primary care before physicians raise the subject, and most of what patients have read is either promotion or dismissal. The evidence reviewed in The Starved Brain by Dr. Gurpreet Singh Padda, MD, MBA, MHP, and in its chapter video, One Correction, Many Diseases, supports a narrower position: guideline-level therapy in drug-resistant epilepsy, modest reproducible cognitive effects in MCI and early Alzheimer’s disease, and an unresolved signal in Parkinson’s disease. For the screening physician the useful question is not whether to endorse a diet. It is who warrants a metabolic and cognitive baseline, and what that baseline changes. Measura [Cardiometabolic and Autonomic Health Analysis] provides that measurement layer and does not treat.

Does a ketogenic diet help mild cognitive impairment? Effect sizes at their true size

• BENEFIC randomized 122 adults aged 55 and older with MCI to 30 g a day of ketogenic MCT emulsion or calorie-matched placebo for six months; 83 completed. Word recall, verbal fluency, naming and Trail-Making errors improved after adjustment, and the change in blood ketones correlated with cognitive change (r = +0.325 for fluency). Discontinuation was 32%, gastrointestinal adverse events occurred in 74% of drink completers against 40% on placebo, and no Alzheimer’s biomarker moved.
• Rong 2024 pooled ten randomized trials, 691 participants, 12 to 60 weeks: MMSE +1.25 (95% CI 0.46 to 2.04), ADAS-Cog −3.43, activities of daily living unchanged (P = 0.95). Eight trials used MCT, and the funnel plot suggested publication bias for MMSE.
• The whole-diet crossover (Phillips 2021) in 26 people with probable Alzheimer’s: daily function +3.13 against a prespecified meaningful difference of 2, quality of life +3.37, cognition unchanged (P = 0.24).

For context, lecanemab slowed decline by 0.45 points on an 18-point scale over 18 months with imaging edema in 12.6%, and donanemab by 0.70 points over 76 weeks with edema in 24.0%. Scales, durations and constructs differ, so these are not head-to-head comparisons, and improvement is a different quantity from slowed decline. They do set the backdrop for a modest nutritional signal, following a registry record in which 99.6% of Alzheimer’s drug trials registered between 2002 and 2012 failed.

Who did the ketogenic diet trials exclude?

BENEFIC excluded uncontrolled diabetes, cardiac, hepatic and renal disease, uncontrolled hypertension and dyslipidemia, and B12 deficiency. The Alzheimer’s diet trial excluded moderate or severe depression, substantial cerebrovascular disease, and abnormal B12, folate or thyroid values. The PET studies establishing preserved ketone uptake selected participants for a similar metabolic phenotype. The insulin-resistant older adult with hepatic steatosis who dominates a primary care panel was, by design, absent.

Two consequences follow. First, the nutrient hypothesis, that a whole-food ketogenic diet repletes B12, choline and DHA as well as correcting fuel supply, was screened out rather than tested; a trial that excludes deficiency cannot detect a repletion effect. Second, response in the metabolically complex patient is uncertain in both directions. The BENEFIC authors anticipated a reduced response in prediabetic older adults, while the 2026 Washington University trial in prediabetes with fatty liver, holding weight loss at 10% in every arm, found hepatic insulin sensitivity improved two to three times more on the ketogenic diet than on the Mediterranean diet (p < 0.001), with no between-group difference in LDL cholesterol or apoB. That trial is so far reported in abstract. Uncertainty of that shape argues for baseline measurement, not for a blanket recommendation in either direction.

Does APOE4 status predict response to a ketogenic diet?

Carrier status is the objection patients raise most, and the data do not run one way. In a Stockholm cohort of 2,157 adults aged 60 and over without dementia at entry, 569 of them carriers, the highest meat-intake quintile carried a dementia subhazard of 0.45 (95% CI 0.21 to 0.95) among carriers, with a non-significant genotype interaction for dementia and a clearer interaction on global cognition. Against that, pooled MCT trials favored non-carriers (SMD 1.87), and an umbrella review found dietary interventions other than the Mediterranean pattern generally ineffective in older carriers, while small exploratory BENEFIC subgroups pointed the other way. The defensible reading is heterogeneity: carriers appear more sensitive to metabolic inputs, and the direction in an individual has to be observed rather than assumed.

The baseline that makes a supervised trial interpretable

For a patient with MCI, subjective decline or a strong family history who is considering supervised ketogenic therapy, a defensible pre-intervention set looks like this:

• Cognitive assessment, because every positive trial reported change on a scored instrument, and an individual needs the same anchor. Pairing it with fall-risk work is covered in cognitive assessment and fall prevention.
• Laboratory panels with fasting insulin reported as a number, lipids including apoB or particle measures, and a plan for beta-hydroxybutyrate monitoring through the supervising clinician. Across 62 randomized trials in adults, triglycerides fell 19.96 mg/dL while LDL cholesterol rose 8.49 mg/dL; the Parkinson’s pilot saw LDL rise 0.70 mmol/L in eight weeks. The diet trials that separated ran beta-hydroxybutyrate between 0.95 and 1.3 mmol/L, and the 2026 Parkinson’s crossover defined nutritional ketosis as above 0.5 mM.
• Bioimpedance body composition, since weight fell 2.62 kg in the Alzheimer’s diet trial, and an older adult losing lean mass is a different outcome from one losing visceral fat.

Dr. Padda’s threshold for concern is a fasting insulin above 10 µIU/mL, stated as his clinic cutoff; the published neighbor is a level above 9.0 that identified 80% of prediabetes in one practice. In 683 older Manhattan adults, hyperinsulinemia carried HR 2.1 for Alzheimer’s disease.

What safety signals come before ordering a ketogenic diet?

The Parkinson’s pilot reported transient worsening of tremor or rigidity in 12 of 24 ketogenic patients during weeks one to four. A 2026 case series described nine cases of hypomania or mania within two months of starting a ketogenic diet, seven without prior bipolar history, with no denominator. A bipolar pilot recorded euglycemic ketoacidosis in a participant taking an SGLT2 inhibitor. Diabetes, antihypertensive and cholinesterase-inhibitor regimens were excluded from the MCI trial rather than managed within it. Those are prescriber conversations that precede any dietary counseling, and measurement ordered under standing orders for screening makes the sequence reproducible.

Documentation and workflow

A cognitive baseline and metabolic panel fit structures practices already run. The annual wellness visit is a natural home for a documented cognitive score that becomes the comparison point for the next visit, and findings can be recorded against the quality concepts described in MIPS and quality reporting. The system failure is familiar: nutrition is delegated to a handout while visit time goes to prescriptions, and a diet nobody measures cannot be told apart from a diet that failed. Adherence is the other constraint; a modified Atkins pilot found 27 willing patients in two and a half years. A repeat panel at three months tells physician and patient whether the correction is happening at all. Primary studies, methods, sponsorship and limits are laid out in the Chapter 9 companion evidence guide, and the pre-diet safety workup and re-measure schedule continue in low-carb diet safety for physicians.

Frequently asked questions

References

• Fortier, M., Castellano, C.-A., St-Pierre, V., Myette-Côté, É., Langlois, F., Roy, M., Morin, M.-C., Bocti, C., Fulop, T., Godin, J.-P., Delannoy, C., Cuenoud, B., & Cunnane, S. C. (2021). A ketogenic drink improves cognition in mild cognitive impairment: Results of a 6-month RCT. Alzheimer’s & Dementia, 17(3), 543–552. https://doi.org/10.1002/alz.12206
• Rong, L., Peng, Y., Shen, Q., Chen, K., Fang, B., & Li, W. (2024). Effects of ketogenic diet on cognitive function of patients with Alzheimer’s disease: A systematic review and meta-analysis. The Journal of Nutrition, Health and Aging, 28(8), 100306. https://doi.org/10.1016/j.jnha.2024.100306
• Phillips, M. C. L., Deprez, L. M., Mortimer, G. M. N., Murtagh, D. K. J., McCoy, S., Mylchreest, R., Gilbertson, L. J., Clark, K. M., Simpson, P. V., McManus, E. J., Oh, J.-E., Yadavaraj, S., King, V. M., Pillai, A., Sequeira, B., Rodgers, B., Wiltshire, C., Tanner, S., & Tisch, S. (2021). Randomized crossover trial of a modified ketogenic diet in Alzheimer’s disease. Alzheimer’s Research & Therapy, 13(1), 51. https://doi.org/10.1186/s13195-021-00783-x
• van Dyck, C. H., Swanson, C. J., Aisen, P., Bateman, R. J., Chen, C., Gee, M., Kanekiyo, M., Li, D., Reyderman, L., Cohen, S., Froelich, L., Katayama, S., Sabbagh, M., Vellas, B., Watson, D., Dhadda, S., Irizarry, M., Kramer, L. D., & Iwatsubo, T. (2023). Lecanemab in early Alzheimer’s disease. The New England Journal of Medicine, 388(1), 9–21. https://doi.org/10.1056/NEJMoa2212948
• Sims, J. R., Zimmer, J. A., Evans, C. D., Lu, M., Ardayfio, P., Sparks, J., Wessels, A. M., Shcherbinin, S., Wang, H., Monkul Nery, E. S., Collins, E. C., Solomon, P., Salloway, S., Apostolova, L. G., Hansson, O., Ritchie, C., Brooks, D. A., Mintun, M., Skovronsky, D. M., & TRAILBLAZER-ALZ 2 Investigators. (2023). Donanemab in early symptomatic Alzheimer disease: The TRAILBLAZER-ALZ 2 randomized clinical trial. JAMA, 330(6), 512–527. https://doi.org/10.1001/jama.2023.13239
• Cummings, J. L., Morstorf, T., & Zhong, K. (2014). Alzheimer’s disease drug-development pipeline: Few candidates, frequent failures. Alzheimer’s Research & Therapy, 6(4), 37. https://doi.org/10.1186/alzrt269
• Norgren, J., Carballo-Casla, A., Grande, G., Börjesson-Hanson, A., Xu, H., Eriksdotter, M., Laukka, E. J., & Garcia-Ptacek, S. (2026). Meat consumption and cognitive health by APOE genotype. JAMA Network Open, 9(3), Article e266489. https://doi.org/10.1001/jamanetworkopen.2026.6489
• Sun, L., Ye, K. X., Wong, H. L. K., Wang, L., Lim, S. L., Chao, Y. X., Zhang, C., Yap, K. Z., & Feng, L. (2023). The effects of medium chain triglyceride for Alzheimer’s disease related cognitive impairment: A systematic review and meta-analysis. Journal of Alzheimer’s Disease, 94(2), 441–456. https://doi.org/10.3233/JAD-230406
• Petersen, M. C., Smith, G. I., Farabi, S. S., Palacios, H. H., Shankaran, M., Hellerstein, M. K., Patterson, B. W., & Klein, S. (2026). Effect of diet macronutrient content on the cardiometabolic response to weight loss: A randomized clinical trial. Cell Metabolism. Advance online publication. https://doi.org/10.1016/j.cmet.2026.07.020
• Chang, C., Liu, Y., Rohani, P., Khodadadi, N., Prabahar, K., & Sohouli, M. H. (2026). The impact of the ketogenic diet on the lipid profile in adults: A comprehensive review and meta-regression analysis of randomized controlled trials. Endocrine Practice, 32(5), 819–828. https://doi.org/10.1016/j.eprac.2026.01.009

Related reading

• Periodontal Disease and Dementia: A Screening Workflow
• Low-Carb Diet Safety: A Physician’s Baseline and Re-Measure Plan
• Cognitive Assessment

Medically reviewed by Dr. Gurpreet Singh Padda, MD, MBA, MHP, medical director of Measura. Last reviewed September 12, 2026.