Protocols by Charles Sieg

Diet

Keto

What ketosis is, what the trials show for weight, diabetes, epilepsy, and training, and the LDL question that comes with it.

Overall evidence Mixed22 sourcesReviewed November 5, 2026

One randomized trial supports keto for childhood epilepsy and non-randomized trials support it for type 2 diabetes; it is roughly even with other diets for weight loss; and its effect on LDL and heart risk is unsettled.

The short version

  • I follow a keto diet. It is a tool with solid evidence in a few places and real open questions in others, not a cure-all.
  • Nutritional ketosis means blood beta-hydroxybutyrate (BHB) of at least 0.5 mM, compared with about 0.1 mM on a balanced diet1.
  • For weight loss, it edges out low-fat diets in a 13-trial meta-analysis by under 1 kg2, and a large head-to-head trial found no significant difference3.
  • For type 2 diabetes, the Virta program cut HbA1c and medications in a non-randomized trial, with results held at two years4,5. For childhood epilepsy, a randomized trial found a large drop in seizures6.
  • It raises LDL in some people, sometimes a lot7. The one plaque study so far is reassuring but small and limited8. That is the honest caveat.

The protocol

This is what the research defines, not a prescription. My own specifics are below the table.

ParameterWhat the research uses
KetosisBlood BHB of at least 0.5 mM1
CarbohydrateVaries by trial; one randomized LDL trial used under 20 g a day7
MonitoringBlood ketones; the 3 mM line marks the start of ketoacidosis territory, which is a medical problem, not a goal1
DurationTrials run from 3 weeks7 to 2 years5; long-term (beyond 2 years) safety data are minimal9

Why it works

Cut carbohydrate enough and the liver makes ketone bodies, with BHB being the main one measured in blood1. Several effects are worth separating:

  • Appetite and medication needs. The National Lipid Association statement says low-carb and very-low-carb diets may help appetite control, triglycerides, and medication use in type 2 diabetes9.
  • Not a metabolic advantage for fat loss. In a meta-analysis of 32 controlled feeding studies, swapping carbohydrate for fat at equal calories slightly lowered energy expenditure (about 26 kcal a day) and fat loss (about 16 g a day)10. Whatever keto does for weight, it is probably not magic physics.

What the research says

Ketone levels

Strong Nutritional ketosis is generally defined as blood BHB of at least 0.5 mM. Ketoacidosis, the medical emergency, starts above 3 mM, and diabetic ketoacidosis runs from 3 to 20 mM1.

Weight loss versus low-fat diets

Mixed A meta-analysis of 13 randomized trials lasting 12 months or longer (1,415 patients) found very-low-carbohydrate ketogenic diets produced 0.91 kg more weight loss (95% CI 0.17 to 1.65) than low-fat diets. Triglycerides were lower (-0.18 mmol/L), HDL was higher (+0.09 mmol/L), and LDL was higher (+0.12 mmol/L)2. DIETFITS randomized 609 adults to healthy low-fat or healthy low-carb eating for 12 months: 5.3 kg lost versus 6.0 kg, a difference of 0.7 kg that was not statistically significant, and neither genotype nor insulin secretion predicted who did better on which diet3. The National Lipid Association concluded that these diets are not superior for weight loss9. Controlled feeding studies, which hold calories equal, gave a slight edge to lower-fat diets10. My read: the best diet is the one you can hold.

Type 2 diabetes

Moderate The Virta trial enrolled 349 adults with type 2 diabetes: 262 in a continuous remote care program with nutritional ketosis and 87 in usual care. At 1 year, 83% of the program group was still in, HbA1c fell from 7.6% to 6.3%, weight dropped 13.8 kg, and insulin was reduced or stopped in 94% of insulin users. Usual care showed no significant change4. At 2 years, glycemic medication other than metformin fell from 55.7% to 26.8% of users, insulin use fell 62%, and the authors reported diabetes reversal in 53.5% and remission in 17.6% of the program group5. Spine bone density was unchanged5. The catch: both papers were open-label and non-randomized, and the program bundled coaching, monitoring, and medication changes with the diet4. LDL cholesterol rose 10% at one year4.

Epilepsy

Moderate One good randomized trial supports it in children. In that trial of 145 children aged 2 to 16, those starting a ketogenic diet had 62.0% of their baseline seizures at 3 months versus 136.9% in the control group, and 38% had more than a 50% reduction versus 6%6. Side effects were constipation, vomiting, low energy, and hunger6. It was unblinded, and only 103 of the 145 were analyzed6. It tells us little about adults using keto for general health.

Training performance and muscle

Mixed In elite race walkers, 3 weeks of intensified training on a ketogenic diet (n=10) did not improve a 10 km walk (-1.6%, 90% CI -8.5 to 5.3), while the high-carbohydrate (+6.6%) and periodized-carbohydrate (+5.3%) groups improved; keto also worsened exercise economy11. For muscle and strength the picture is kinder. In 25 young men resistance training for 10 weeks, both ketogenic and Western diet groups gained lean mass and improved strength and power equally12. In 20 middle-aged trained men on a calorie-restricted diet for 8 weeks, ketogenic and non-ketogenic low-carb diets kept strength and lost similar lean and fat mass13. All three are small and short. My summary: for lifting, keto looks like it can hold its own; for high-intensity endurance racing, the one elite trial says it costs something.

The honest caveat: LDL and ApoB

Mixed In a randomized trial, 30 healthy, normal-weight young adults ate under 20 g of carbohydrate a day or their usual diet for 3 weeks. LDL-C rose from 2.2 to 3.1 mmol/L on keto while controls stayed at 2.5 (a 44% rise versus controls; individual changes ran from 5% to 107%), and ApoB went up significantly7. Pooled across the long-term trials in the meta-analysis, the average LDL rise was smaller, +0.12 mmol/L2, and the National Lipid Association called LDL effects mixed9.

Some lean, active people show big jumps. A survey analysis of 548 adults on carbohydrate-restricted diets found that lower BMI went with larger LDL rises, and described a subgroup of 100 "lean mass hyper-responders" with LDL-C of 200 mg/dL or higher, HDL-C of 80 or higher, and triglycerides of 70 or lower; in five patients, reintroducing carbohydrate markedly lowered LDL14. That is self-reported survey data with no outcomes14.

The KETO-CTA study (the "KETO Trial") looked at plaque. It matched 80 people on keto for a mean 4.7 years with LDL-C of 190 or higher (mean 272 mg/dL), high HDL, and low triglycerides to controls with mean LDL-C of 123. Coronary calcium was median 0 versus 1 (P=0.520), plaque scores did not differ (P=0.357), and LDL did not track with plaque in either group8. That is reassuring on its face, but it has hard limits: 80 pairs, selected for a specific lipid profile, one time point, plaque instead of heart attacks, and controls from a different cohort8. The causal evidence says LDL exposure adds up over years15, and a few years of follow-up cannot rule out a long-run effect. So I watch ApoB and LDL on actual labs; see the heart page for the tests16.

Rhonda Patrick on ketosis

Rhonda Patrick has covered ketosis on FoundMyFitness with several guests. These are conversations, not trials, and they do not mean she endorses this site.

  • Eric Verdin (Buck Institute) on ketogenic diets and aging: ketones as HDAC inhibitors, the role of PPAR-alpha, declining NAD+ with age, keto safety, and intermittent fasting as an alternative17.
  • Dominic D'Agostino, 2016: the therapeutic potential of ketosis, what to eat to stay in nutritional ketosis, ketones and the brain, and whether ketone supplements are a shortcut18.
  • Dominic D'Agostino, 2022: types of keto diets, exogenous ketones, exercise performance, brain health, neurodegeneration, cancer, and the carnivore diet19.

Cautions

  • SGLT2 inhibitors. These diabetes drugs have been linked to ketoacidosis even when blood glucose is not very high: a published case series described 13 episodes in nine people20. Pairing them with a diet that raises ketones is something to settle with a prescriber first.
  • Type 1 diabetes. Ketoacidosis begins above 3 mM BHB1, and the case series flagged type 1 diabetes as needing great caution with SGLT2 drugs20. Keto in type 1 is a conversation for an endocrinologist, not a self-experiment.
  • Pregnancy. I did not find a human trial. A mouse study found ketogenic feeding before and during pregnancy altered embryo organ growth21. That is an animal signal, not a human result.
  • Kidney stones. In 195 children on a medical ketogenic diet for epilepsy, 13 (6.7%) developed kidney stones, and potassium citrate was associated with fewer (3.2% versus 10.0%)22. Adult rates on a self-directed diet are not established.
  • High LDL or genetic lipid disorders. The National Lipid Association statement specifically reviewed safety for people with genetic lipid disorders and noted that LDL effects are mixed9.
  • Long-term safety. Data beyond 2 years are minimal9.

References

  1. Anderson JC, Mattar SG, Greenway FL, Lindquist RJ. Measuring ketone bodies for the monitoring of pathologic and therapeutic ketosis. Obesity Science & Practice. 2021;7(5):646-656.ReviewPubMed 34631141 · DOI

    What it found: Review of how ketones are measured: nutritional ketosis (keto-adaptation) corresponds to blood beta-hydroxybutyrate of at least 0.5 mM, a person eating a balanced diet sits near 0.1 mM, and ketoacidosis starts above 3 mM, with diabetic ketoacidosis ranging from 3 to 20 mM.

    Limits: A review of measurement methods; ketosis thresholds vary a little between sources and are conventions, not sharp biological lines.

  2. Bueno NB, de Melo IS, de Oliveira SL, da Rocha Ataide T. Very-low-carbohydrate ketogenic diet v. low-fat diet for long-term weight loss: a meta-analysis of randomised controlled trials. The British Journal of Nutrition. 2013;110(7):1178-1187.Meta-analysisPubMed 23651522 · DOI

    What it found: Meta-analysis of 13 randomized trials lasting 12 months or longer (1,415 patients for weight). Very-low-carbohydrate ketogenic diets produced 0.91 kg more weight loss (95% CI 0.17 to 1.65), lower triglycerides (-0.18 mmol/L), higher HDL (+0.09 mmol/L), and higher LDL-C (+0.12 mmol/L, 95% CI 0.04 to 0.20) than low-fat diets.

    Limits: The extra weight loss is under 1 kg on average, and a 2013 pooling of 13 trials cannot say how the result holds up for any one person.

  3. Gardner CD, Trepanowski JF, Del Gobbo LC, Hauser ME, Rigdon J, Ioannidis JPA, et al. Effect of Low-Fat vs Low-Carbohydrate Diet on 12-Month Weight Loss in Overweight Adults and the Association With Genotype Pattern or Insulin Secretion: The DIETFITS Randomized Clinical Trial. JAMA. 2018;319(7):667-679.Randomized trialPubMed 29466592 · DOI

    What it found: DIETFITS randomized 609 overweight adults to a healthy low-fat or healthy low-carbohydrate diet for 12 months; weight loss was 5.3 kg versus 6.0 kg, a difference of 0.7 kg (95% CI -0.2 to 1.6) that was not statistically significant. Neither genotype pattern nor baseline insulin secretion predicted who did better on which diet.

    Limits: A comparison of low-fat versus low-carbohydrate eating that does not report ketone levels, and 21% of participants (128 of 609) did not finish the 12 months.

  4. Hallberg SJ, McKenzie AL, Williams PT, Bhanpuri NH, Peters AL, Campbell WW, et al. Effectiveness and Safety of a Novel Care Model for the Management of Type 2 Diabetes at 1 Year: An Open-Label, Non-Randomized, Controlled Study. Diabetes Therapy. 2018;9(2):583-612.Cohort studyPubMed 29417495 · DOI

    What it found: Virta trial: 349 adults with type 2 diabetes (262 in a continuous remote care program with nutritional ketosis, 87 in usual care); 83% of the program group stayed to 1 year. HbA1c fell from 7.6% to 6.3%, weight fell 13.8 kg, insulin was reduced or stopped in 94% of users, and LDL cholesterol rose 10%; usual care showed no significant change.

    Limits: Open-label and non-randomized, and the program bundled coaching, monitoring, and medication changes with the diet, so the diet's own effect cannot be isolated.

  5. Athinarayanan SJ, Adams RN, Hallberg SJ, McKenzie AL, Bhanpuri NH, Campbell WW, et al. Long-Term Effects of a Novel Continuous Remote Care Intervention Including Nutritional Ketosis for the Management of Type 2 Diabetes: A 2-Year Non-randomized Clinical Trial. Frontiers in Endocrinology. 2019;10:348.Cohort studyPubMed 31231311 · DOI

    What it found: Two-year follow-up of the same Virta cohort (262 in the program, 87 in usual care): glycemic medication other than metformin fell from 55.7% to 26.8% of users, insulin use fell 62%, diabetes reversal was reported in 53.5% and remission in 17.6% of the program group (none in usual care), and spine bone density was unchanged.

    Limits: Non-randomized and open-label, and the abstract does not give numeric HbA1c, weight, or LDL changes for year 2.

  6. Neal EG, Chaffe H, Schwartz RH, Lawson MS, Edwards N, Fitzsimmons G, et al. The ketogenic diet for the treatment of childhood epilepsy: a randomised controlled trial. The Lancet Neurology. 2008;7(6):500-506.Randomized trialPubMed 18456557 · DOI

    What it found: Unblinded trial randomizing 145 children aged 2 to 16 to start a ketogenic diet now or after a 3-month delay (103 analyzed). At 3 months the diet group had 62.0% of baseline seizures versus 136.9% in controls, and 38% versus 6% had more than a 50% reduction; common side effects were constipation, vomiting, low energy, and hunger.

    Limits: Unblinded, only 3 months of comparison, many dropouts, and in children with epilepsy, so it says little about adults using keto for general health.

  7. Retterstøl K, Svendsen M, Narverud I, Holven KB. Effect of low carbohydrate high fat diet on LDL cholesterol and gene expression in normal-weight, young adults: A randomized controlled study. Atherosclerosis. 2018;279:52-61.Randomized trialPubMed 30408717 · DOI

    What it found: Randomized trial in which 30 healthy, normal-weight young adults (15 per group) ate under 20 g of carbohydrate a day or their usual diet for 3 weeks. LDL-C rose from 2.2 to 3.1 mmol/L on the diet while controls stayed at 2.5 (+44% versus controls, individual changes 5% to 107%), and ApoB rose significantly.

    Limits: Thirty completers (9 people dropped out) over 3 weeks, so it shows the direction of change, not what happens over years.

  8. Budoff M, Manubolu VS, Kinninger A, Norwitz NG, Feldman D, Wood TR, et al. Carbohydrate Restriction-Induced Elevations in LDL-Cholesterol and Atherosclerosis: The KETO Trial. JACC: Advances. 2024;3(8):101109.Cohort studyPubMed 39372369 · DOI

    What it found: Eighty people on a ketogenic diet for a mean 4.7 years with LDL-C of 190 mg/dL or higher (mean 272), HDL-C of 60 or higher, and triglycerides of 80 or lower were matched 1:1 to controls with mean LDL-C of 123 mg/dL. Coronary calcium was median 0 versus 1 (P=0.520), total plaque score was not different (P=0.357), and LDL-C did not track with plaque in either group.

    Limits: Eighty matched pairs selected for a high-HDL, low-triglyceride profile, one time point, controls drawn from a different cohort, and plaque is not the same as heart attacks, so it cannot show long-term safety or apply to people with other lipid profiles.

  9. Kirkpatrick CF, Bolick JP, Kris-Etherton PM, Sikand G, Aspry KE, Soffer DE, et al. Review of current evidence and clinical recommendations on the effects of low-carbohydrate and very-low-carbohydrate (including ketogenic) diets for the management of body weight and other cardiometabolic risk factors: A scientific statement from the National Lipid Association Nutrition and Lifestyle Task Force. Journal of Clinical Lipidology. 2019;13(5):689-711.e1.Position standPubMed 31611148 · DOI

    What it found: National Lipid Association scientific statement concluding that low-carb and very-low-carb diets are not superior to other diets for weight loss, may help appetite, triglycerides, and medication use in type 2 diabetes, have mixed effects on LDL cholesterol (some studies show an increase), and lack long-term (more than 2 years) safety data.

    Limits: A 2019 expert review, so it predates the KETO-CTA study and the longer Virta follow-up, and the abstract gives no pooled numbers.

  10. Hall KD, Guo J. Obesity Energetics: Body Weight Regulation and the Effects of Diet Composition. Gastroenterology. 2017;152(7):1718-1727.e3.ReviewPubMed 28193517 · DOI

    What it found: Review including a meta-analysis of 32 controlled feeding studies: at equal calories, swapping carbohydrate for fat produced slightly lower energy expenditure (about 26 kcal a day) and about 16 g a day less body fat loss than lower-fat diets.

    Limits: Controlled feeding studies are short and tightly supervised, so they test physiology rather than how well people stick to a diet in real life.

  11. Burke LM, Ross ML, Garvican-Lewis LA, Welvaert M, Heikura IA, Forbes SG, et al. Low carbohydrate, high fat diet impairs exercise economy and negates the performance benefit from intensified training in elite race walkers. The Journal of Physiology. 2017;595(9):2785-2807.Randomized trialPubMed 28012184 · DOI

    What it found: Three energy-matched diets for 3 weeks of intensified training in elite race walkers (high carbohydrate n=9, periodized carbohydrate n=10, ketogenic low-carb high-fat n=10). The 10 km walk improved 6.6% and 5.3% in the two carbohydrate groups but changed -1.6% (90% CI -8.5 to 5.3) on keto, which also worsened exercise economy despite peak fat oxidation of 1.57 g/min.

    Limits: Only 29 elite race walkers over 3 weeks, so it cannot show what a longer adaptation period or other sports would do.

  12. Wilson JM, Lowery RP, Roberts MD, Sharp MH, Joy JM, Shields KA, et al. Effects of Ketogenic Dieting on Body Composition, Strength, Power, and Hormonal Profiles in Resistance Training Men. Journal of Strength and Conditioning Research. 2020;34(12):3463-3474.Randomized trialPubMed 28399015 · DOI

    What it found: Randomized trial of 25 college-aged men on an isocaloric ketogenic or Western diet for 10 weeks of resistance training. Lean mass rose in both groups, fat mass fell in both (-2.2 kg keto, -1.5 kg Western), and strength and power improved equally; total testosterone rose on keto (118 ng/dL) and fell on the Western diet (-36 ng/dL).

    Limits: Twenty-five young men over 10 weeks, and the abstract does not report LDL or ApoB.

  13. Vidić V, Ilić V, Toskić L, Janković N, Ugarković D. Effects of calorie restricted low carbohydrate high fat ketogenic vs. non-ketogenic diet on strength, body-composition, hormonal and lipid profile in trained middle-aged men. Clinical Nutrition. 2021;40(4):1495-1502.Randomized trialPubMed 33743284 · DOI

    What it found: Randomized trial of 20 resistance-trained middle-aged men eating a calorie-restricted ketogenic or non-ketogenic low-carb high-fat diet for 8 weeks (after a 4-week run-in). Both groups lost similar lean and fat mass, kept their maximal upper and lower body strength, and showed no difference in lipid profile.

    Limits: Twenty men over 8 weeks on a calorie deficit, and the abstract gives F-statistics rather than group means.

  14. Norwitz NG, Feldman D, Soto-Mota A, Kalayjian T, Ludwig DS. Elevated LDL Cholesterol with a Carbohydrate-Restricted Diet: Evidence for a 'Lean Mass Hyper-Responder' Phenotype. Current Developments in Nutrition. 2022;6(1):nzab144.Cross-sectional studyPubMed 35106434 · DOI

    What it found: Analysis of web-survey data from 548 adults on carbohydrate-restricted diets plus a five-patient case series. Lower BMI was linked to larger LDL-C rises, and a subgroup of 100 with LDL-C of 200 mg/dL or higher, HDL-C of 80 or higher, and triglycerides of 70 or lower was labeled lean mass hyper-responders; carbohydrate reintroduction markedly lowered LDL-C in the five cases.

    Limits: Self-reported survey data with no control group and no outcomes, so it describes a pattern but cannot show whether high LDL-C in this group is harmless.

  15. Ference BA, Ginsberg HN, Graham I, Ray KK, Packard CJ, Bruckert E, et al. Low-density lipoproteins cause atherosclerotic cardiovascular disease. 1. Evidence from genetic, epidemiologic, and clinical studies. A consensus statement from the European Atherosclerosis Society Consensus Panel. European Heart Journal. 2017;38(32):2459-2472.Position standPubMed 28444290 · DOI

    What it found: Consensus review of genetic studies, more than 200 prospective cohorts, Mendelian randomization studies, and randomized trials covering more than 2 million participants, over 20 million person-years, and over 150,000 cardiovascular events. It found a consistent, dose-dependent, log-linear link between LDL exposure and cardiovascular risk that grows with the duration of exposure, and concluded that LDL causes atherosclerotic cardiovascular disease.

    Limits: A consensus review of other studies, not a new trial; it rests on the quality of the underlying genetic, cohort, and trial data.

  16. Sniderman AD, Thanassoulis G, Glavinovic T, Navar AM, Pencina M, Catapano A, et al. Apolipoprotein B Particles and Cardiovascular Disease: A Narrative Review. JAMA Cardiology. 2019;4(12):1287-1295.ReviewPubMed 31642874 · DOI

    What it found: Narrative review arguing that every atherogenic particle (VLDL, LDL, chylomicron remnant, Lp(a)) carries one apoB molecule, so plasma apoB counts the particles; cholesterol enters the artery wall only inside apoB particles, and the number entering depends mainly on how many are in the blood. The authors conclude that apoB is a better marker of atherogenic risk than LDL-C or non-HDL-C.

    Limits: A narrative review by authors who advocate apoB testing, not a systematic review or a trial.

  17. Patrick R. #040 Dr. Eric Verdin on Ketogenic Diet Longevity, Beta-Hydroxybutyrate & HDAC Inhibitors. FoundMyFitness, December 13, 2017.PodcastListen

    What it found: Interview with Eric Verdin, head of the Buck Institute, on how ketogenic diets might affect aging: ketones acting as HDAC inhibitors, PPAR-alpha's role, declining NAD+ with age, keto safety, and intermittent fasting as an alternative.

    Limits: A podcast conversation from 2017, not a trial; it discusses mechanisms and hypotheses.

  18. Patrick R. #023 Dominic D'Agostino, Ph.D. on Modified Atkins Diet, Ketosis, Supplemental Ketones and More. FoundMyFitness, March 23, 2016.PodcastListen

    What it found: Interview with Dominic D'Agostino, a University of South Florida researcher, on the therapeutic potential of ketosis, what to eat to stay in nutritional ketosis, ketones and the brain, and whether ketone supplements are a shortcut.

    Limits: A podcast conversation from 2016, not a trial.

  19. Patrick R. #074 Dr. Dominic D'Agostino on Developing a Well-Designed Ketogenic Diet and Harnessing Its Benefits. FoundMyFitness, April 28, 2022.PodcastListen

    What it found: Interview with Dominic D'Agostino covering types of keto diets, exogenous ketones, exercise performance, brain health, neurodegeneration, cancer, and the carnivore diet.

    Limits: A podcast conversation, not a trial.

  20. Peters AL, Buschur EO, Buse JB, Cohan P, Diner JC, Hirsch IB. Euglycemic Diabetic Ketoacidosis: A Potential Complication of Treatment With Sodium-Glucose Cotransporter 2 Inhibition. Diabetes Care, 2015.ArticlePubMed 26078479 · DOI

    What it found: Case series of 13 episodes of ketoacidosis or ketosis in nine people taking SGLT-2 inhibitors (seven with type 1 diabetes, two with type 2); blood glucose was not markedly high, so recognition was slow. The authors advise testing ketones in anyone on these drugs who is nauseated, vomiting, or unwell, and extreme caution in type 1 diabetes.

    Limits: A small case series identified incidentally, so it cannot give how often this happens and does not test a keto diet itself.

  21. Sussman D, van Eede M, Wong MD, Adamson SL, Henkelman M. Effects of a ketogenic diet during pregnancy on embryonic growth in the mouse. BMC Pregnancy and Childbirth. 2013;13:109.Lab or mechanistic studyPubMed 23656724 · DOI

    What it found: Mice fed a ketogenic or standard diet for 30 days before and through pregnancy; embryo imaging showed altered organ growth in the ketogenic group (for example, relatively smaller brain and larger heart at day 13.5, then smaller heart and thymus at day 17.5).

    Limits: A mouse study, so it raises a question about human pregnancy without answering it.

  22. Sampath A, Kossoff EH, Furth SL, Pyzik PL, Vining EPG. Kidney stones and the ketogenic diet: risk factors and prevention. Journal of Child Neurology. 2007;22(4):375-378.Cohort studyPubMed 17621514 · DOI

    What it found: Cohort of 195 children started on the ketogenic diet for intractable epilepsy: 13 (6.7%) developed kidney stones, and oral potassium citrate was associated with fewer stones (3.2% versus 10.0%, P=0.049).

    Limits: Children on a medical keto diet for epilepsy at one center, observational, so adult rates on a self-directed diet are unknown.