A ketogenic diet cuts carbohydrate low enough that the body stops running on glucose and runs on fat and ketones instead. Its one use with strong trial evidence is childhood epilepsy that has failed its medicines, where a supervised diet cut seizures by more than half in about four in ten children. Its newer use is type 2 diabetes: supervised programs lowered long-term blood sugar and let people reduce or stop medication, insulin included.
LDL cholesterol climbs in some people, the advantage over other diets narrows by twelve months as the diet gets hard to keep, and anyone on glucose-lowering medicine needs a clinician before starting. This page is about what you eat; time-restricted eating is about when, and the two are different levers.
Findings & Outcomes
What It Is
A ketogenic diet keeps carbohydrate very low: below roughly 50 grams a day, sometimes as low as 20. Protein stays high enough to keep muscle, and fat supplies the rest of the energy. At that level the body runs short of the glucose it usually burns, and the liver starts making ketones from fat, a state called nutritional ketosis. It began as medicine. A supervised ketogenic diet has treated drug-resistant epilepsy for about 100 years, and that remains its best-evidenced use. The version most people now run themselves, for weight and blood sugar, is newer and far less settled.
The evidence splits into three tiers of very different strength, and each card below grades one finding on its own terms. Most of these numbers are stand-in measures: blood sugar, weight and cholesterol read over months. No one has yet followed a ketogenic diet the decades it would take to weigh its effect on heart attacks, strokes or lifespan. Read each finding as what it measured, and no more.
The Research & Studies
Everything here is based on the research we have collected and checked, sorted into groups and ordered with the strongest evidence first. Click any claim to open the studies behind it.
Seizure Control
Cut seizures more than half in 38% of children with drug-resistant epilepsy
For children whose epilepsy did not respond to two or more medicines, a strictly supervised ketogenic diet cut seizures by more than half in about 4 in 10, against about 1 in 20 who carried on with usual care. This is the one use of the diet with strong randomized evidence.
In 145 children aged 2 to 16 with drug-resistant epilepsy, those randomized to a ketogenic diet had a mean seizure frequency of 62% of baseline at three months, versus 137% in controls (p<0.0001). 38% of the diet group achieved greater than 50% seizure reduction versus 6% of controls (p<0.0001), and 7% achieved greater than 90% reduction versus none. The diet is calorie- and fluid-prescribed and clinically supervised; common side effects were constipation, vomiting, low energy and hunger.
Who this may not transfer to:Measured in children; the strong seizure evidence does not transfer automatically to adults, in whom the trial base is thinner.
The study · 1
Neal et al., the ketogenic diet for the treatment of childhood epilepsy: a randomised controlled trial · Lancet Neurol 2008;7(6):500-506
Blood Sugar
HbA1c fell from 7.6% to 6.3% over a supervised year
In a closely supported program built around a ketogenic diet, people with type 2 diabetes lowered their long-term blood sugar (HbA1c) from 7.6% to 6.3% in a year and lost about 30 lb (13.8 kg). It was not a randomized trial, and a lot of coaching came bundled with the diet.
In a supervised continuous-care program built on nutritional ketosis, average HbA1c fell from 7.6% to 6.3% at one year while a usual-care group changed little, and average weight fell 30 lb (13.8 kg) (about 12% of body weight). Measured in: 262 adults with type 2 diabetes in the intervention arm, 87 in usual care; not randomized, before-and-after design; 83% remained enrolled at one year. Allocation was not randomized, and intensive coaching, biometric feedback and app support are bundled with the diet, so the diet's own contribution cannot be separated from the program around it. The comparison group was a separate clinic, not a matched control.
Who this may not transfer to:Measured in adults with type 2 diabetes in a supervised program; results away from that support are likely smaller.
The study · 1
Hallberg et al., effectiveness and safety of a novel care model for the management of type 2 diabetes at 1 year · Diabetes Ther 2018;9(2):583-612
Counts once: this finding and 1 other here come from the same source, so they are one body of evidence, not separate confirmations.
Insulin cut or stopped in 94%, every sulfonylurea withdrawn
Because the diet lowered blood sugar, the medical team was able to cut back medicines: insulin was reduced or stopped in 94% of insulin users, and sulfonylureas were dropped entirely. The clinicians lowered the doses as the numbers fell, which is what made it safe.
Over the same year, prescriptions for glucose-lowering drugs other than metformin fell from 57% to 30% of patients, insulin was reduced or stopped in 94% of those taking it, and every sulfonylurea prescription was withdrawn. Measured in: 262 adults with type 2 diabetes in a supervised program; before-and-after design. Deprescribing was done deliberately by the clinical team as blood sugar fell, which is exactly why it was safe here. The same glucose drop at home on unchanged doses is how hypoglycemia happens, so this result is inseparable from the supervision that produced it.
Who this may not transfer to:Deprescribing was clinician-led as glucose fell; the same is required for anyone on these drugs.
If you take insulin or a sulfonylurea, a ketogenic diet can lower your glucose enough that your usual dose becomes too much. This is a change to make with the person who prescribes for you, before you start, not after a low.
The study · 1
Hallberg et al., effectiveness and safety of a novel care model for the management of type 2 diabetes at 1 year · Diabetes Ther 2018;9(2):583-612
Counts once: this finding and 1 other here come from the same source, so they are one body of evidence, not separate confirmations.
Diabetes reversed in 53.5% and in remission in 17.6% at two years
Two years into the same supported program, about half the participants had brought their diabetes back under the diagnostic line on little or no medication. It was not randomized, and the people who leave hard diets are usually the ones it is not working for.
At two years in the same program, diabetes was reversed (HbA1c under 6.5% on metformin only or no drugs) in 53.5% and met a remission definition in 17.6%. Glucose-lowering drugs other than metformin fell from 55.7% to 26.8% of patients, and insulin use fell 62%. Measured in: 262 adults with type 2 diabetes enrolled; 194 (74%) completed two years; non-randomized. The same non-randomized, bundled-program limits apply, and attrition by two years tends to leave the people doing best, which flatters the averages.
Who this may not transfer to:Measured in adults with type 2 diabetes in a supervised program; the two-year figures reflect those who stayed enrolled.
The study · 1
Athinarayanan et al., long-term effects of a continuous care intervention including nutritional ketosis for type 2 diabetes: a 2-year non-randomized trial · Front Endocrinol (Lausanne) 2019;10:348
HbA1c reached 6.1% vs 6.7% on a moderate-carb diet, randomized
In a small year-long randomized trial, the ketogenic group lowered HbA1c more (to 6.1%) and lost more weight (about 18 lb (8 kg)) than a moderate-carb group, and more of them stopped a diabetes drug. Only 34 people, so it is a signal, not the final word.
In a 12-month randomized trial, a very-low-carb ketogenic diet lowered HbA1c from 6.6% to 6.1% versus 6.9% to 6.7% on a moderate-carb, calorie-reduced diet (between-group p=0.007), with 17 lb (7.9 kg) lost versus 3.7 lb (1.7 kg). 6 of 10 people on a sulfonylurea or DPP-4 inhibitor stopped it, versus 0 of 6 in the comparison arm. Measured in: 34 overweight adults with type 2 diabetes or prediabetes, randomized (16 ketogenic, 18 moderate-carb). Small (34 people), and both arms received online support, so this shows the ketogenic approach beating a specific moderate-carb comparator in a small sample, not a settled population effect.
Who this may not transfer to:Small adult sample with type 2 diabetes or prediabetes; a signal to confirm in larger trials.
The study · 1
Saslow et al., twelve-month outcomes of a randomized trial of a moderate-carbohydrate versus very low-carbohydrate diet in adults with type 2 diabetes or prediabetes · Nutr Diabetes 2017;7(12):304
Remission at six months in 57% vs 31% on control diets
Across 23 randomized trials, more people reached diabetes remission at six months on low-carb diets (57%) than on control diets (31%). The edge faded by a year and was much smaller under the strictest definition of remission.
Pooling 23 randomized trials (1,357 people), low-carbohydrate diets produced diabetes remission at six months in 57% versus 31% on control diets (risk difference 0.32, 95% CI 0.17 to 0.47). The effect shrank sharply when remission required an HbA1c under 6.5% with no medication, and in trials that included insulin users. Measured in: 1,357 adults with type 2 diabetes across 23 randomized controlled trials. The remission advantage was largest at six months and diminished by twelve, and studies varied in how strictly they defined remission. Under the most stringent definition the effect was small and not statistically significant.
Who this may not transfer to:Pooled adults with type 2 diabetes; the six-month advantage narrows by a year.
The study · 1
Goldenberg et al., efficacy and safety of low and very low carbohydrate diets for type 2 diabetes remission: systematic review and meta-analysis · BMJ 2021;372:m4743
Counts once: this finding and 2 others here come from the same source, so they are one body of evidence, not separate confirmations.
The blood-sugar edge over other diets is uncertain by twelve months
By twelve months the ketogenic diet's blood-sugar advantage over other diets is uncertain: one way of pooling the trials shows a clear HbA1c drop, another shows none. The most likely reason it fades is that few people keep carbohydrate that low for a year.
Across 8 randomized trials (606 people), very-low-carb ketogenic diets lowered HbA1c by 0.65 percentage points at twelve months in the two trials that reported absolute HbA1c (95% CI -0.99 to -0.31, p<0.001), but the four trials that reported change from baseline showed essentially no difference (0.01%, 95% CI -0.22 to 0.25, p=0.91). Triglycerides fell by 25 mg/dL (0.28 mmol/L). Measured in: 606 adults with prediabetes or type 2 diabetes across 8 randomized controlled trials. The review concluded the advantage over other dietary strategies is limited, and the split between the two analyzes tracks how the studies reported their data as much as any true difference. Keeping carbohydrate this low for a year is hard, and effects tend to converge with comparison diets as adherence slips.
Who this may not transfer to:Pooled adults with prediabetes or type 2 diabetes; the twelve-month effect depends on how trials reported their data.
The study · 1
Parry-Strong et al., very low carbohydrate (ketogenic) diets in type 2 diabetes: a systematic review and meta-analysis of randomized controlled trials · Diabetes Obes Metab 2022;24(12):2431-2442
Weight And Fat Loss
About 2 lb (0.9 kg) more weight loss than a low-fat diet at a year
After a year or more, the ketogenic diet edged out a low-fat diet for weight loss by under a kilogram. That gap is small compared with what either diet loses, so which diet you can stick to matters more than which one you choose.
Pooling 13 trials (1,415 people) at 12 months or longer, a very-low-carb ketogenic diet produced about 2 lb (0.9 kg) more weight loss than a low-fat diet (weighted mean difference -0.91 kg, 95% CI -1.65 to -0.17). Measured in: 1,415 adults across 13 randomized controlled trials, followed at least a year. Under a kilogram of difference after a year is small next to the several kilograms both diets lose, so at long follow-up which diet you can keep to matters more than which one you pick.
Who this may not transfer to:Adults across weight-loss trials; the sub-kilogram gap is small next to total loss.
The study · 1
Bueno et al., very-low-carbohydrate ketogenic diet v. low-fat diet for long-term weight loss: a meta-analysis of randomised controlled trials · Br J Nutr 2013;110(7):1178-1187
Counts once: this finding and 1 other here come from the same source, so they are one body of evidence, not separate confirmations.
Same weight loss as a healthy low-fat diet, 13 lb (6.0 kg) vs 12 lb (5.3 kg)
In a good year-long randomized trial, a healthy low-carb diet and a healthy low-fat diet lost about the same weight (roughly 11 to 13 lb (5 to 6 kg)), and nothing about a person's genes or insulin predicted which would suit them better.
In a 12-month randomized trial of 609 adults, a healthy low-carb diet and a healthy low-fat diet produced statistically indistinguishable weight loss (-13 lb (-6.0 kg) vs -12 lb (-5.3 kg); difference 0.7 kg, 95% CI -0.2 to 1.6). Neither baseline insulin secretion nor a genotype pattern predicted which diet worked better for a given person. Measured in: 609 adults with overweight or obesity, without diabetes, randomized. Both groups were coached to eat whole, minimally processed food and to cut added sugar and refined grain, which is not the same as an unguided keto diet. The trial tested diet quality with a carbohydrate difference on top, and at that quality the macronutrient split made no difference to the outcome.
Who this may not transfer to:Adults with overweight or obesity, without diabetes; both arms ate whole-food diets.
The study · 1
Gardner et al., effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults (DIETFITS randomized clinical trial) · JAMA 2018;319(7):667-679
Counts once: this finding and 1 other here come from the same source, so they are one body of evidence, not separate confirmations.
Cholesterol And Lipids
LDL cholesterol rises about 5 mg/dL on average, steeply in some
On average the ketogenic diet raises LDL ('bad') cholesterol a little while lowering triglycerides. Some people, often lean and otherwise healthy, see LDL rise sharply. That is why anyone doing this should have their cholesterol checked.
In pooled long-term trials the ketogenic diet raised LDL cholesterol by roughly 5 mg/dL (0.12 mmol/L) on average (95% CI 0.04 to 0.20, 1,255 people), while lowering triglycerides by 16 mg/dL (0.18 mmol/L). A subset of people, often lean and metabolically healthy, see much larger LDL rises. A separate 23-trial diabetes review found LDL worsening by twelve months. Measured in: 1,255 adults in the LDL analysis; separate 23-trial review for the 12-month signal. The average rise is small and comes alongside lower triglycerides and higher HDL, so the overall lipid picture is mixed. The concern is the individual outliers whose LDL climbs steeply, so a lipid check before and during the diet catches them.
Who this may not transfer to:Adults across long-term trials; the average is small but individual rises vary widely.
Have a lipid panel before you start and again a few months in. If LDL climbs sharply, that is information for you and your clinician, and it can often be softened by swapping saturated fat (butter, fatty meat) for unsaturated fat (olive oil, nuts, fish) without leaving ketosis.
The studies · 2
Bueno et al., very-low-carbohydrate ketogenic diet v. low-fat diet for long-term weight loss: a meta-analysis of randomised controlled trials · Br J Nutr 2013;110(7):1178-1187
Goldenberg et al., efficacy and safety of low and very low carbohydrate diets for type 2 diabetes remission: systematic review and meta-analysis · BMJ 2021;372:m4743
Counts once: this finding and 2 others here come from the same source, so they are one body of evidence, not separate confirmations.
Behavior Change
The early edge fades by twelve months as adherence slips
The ketogenic diet's early advantage tends to fade over a year, mostly because keeping carbohydrate that low is hard and people drift back. If you can sustain it, the benefit lasts longer; if you cannot, you end up about where other diets do.
Across the randomized evidence the metabolic gains from carbohydrate restriction are largest at six months and shrink by twelve, tracking a decline in adherence. In the 609-person DIETFITS trial both diet groups drifted back toward the middle in carbohydrate intake over the year. Measured in: Pooled randomized trial evidence in type 2 diabetes plus a 609-person weight-loss trial. Adherence, not physiology, is the usual reason the ketogenic edge fades, and that depends on the person: someone who can hold to it keeps the early benefit longer, and someone who cannot ends up about where the comparison diets land.
Who this may not transfer to:Adults across diabetes and weight-loss trials; sustaining the diet is the limiting factor.
The studies · 2
Goldenberg et al., efficacy and safety of low and very low carbohydrate diets for type 2 diabetes remission: systematic review and meta-analysis · BMJ 2021;372:m4743
Gardner et al., effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults (DIETFITS randomized clinical trial) · JAMA 2018;319(7):667-679
Counts once: this finding and 2 others here come from the same source, so they are one body of evidence, not separate confirmations.
How It Works
Carbohydrate raises blood glucose far more than protein or fat. Remove nearly all of it and the body has less glucose to clear and needs less insulin to do it. Blood insulin falls within a day or two, because insulin is largely a response to carbohydrate. This direct link is why the type 2 diabetes numbers hold up, and it runs on the same biology as the Insulin and glucose page.
Why the diet calms epilepsy is less settled. Three explanations compete: ketones, a shift in the brain's energy metabolism, and changes in the signaling chemical GABA. The diet helps many children before anyone can say exactly why.
Anatomy of the Practice
1Carbohydrate goes very low
Drop carbohydrate under roughly 20 to 50 grams a day and blood insulin falls, because insulin is largely a response to carbohydrate. The body draws down its stored glucose over a day or two.
2The liver makes ketones
With glucose scarce and insulin low, the liver turns fat into ketone bodies, which the brain, heart and muscle can burn in place of glucose. Blood ketones climb into nutritional ketosis over several days.
3Blood sugar steadies
Running on fat and ketones, with little carbohydrate coming in, day-to-day blood sugar flattens out and its swings shrink. Appetite often falls too, which is part of why people eat less without counting.
If you take insulin or a sulfonylurea, arrange your dose changes with a clinician before you start a ketogenic diet. It can lower blood sugar enough to make your usual dose dangerous.
Ways to Do It
There is nothing to buy here, and the packaged 'keto' products add nothing. The levers are how low you take carbohydrate, what you build the fat from, how gently you start, and who you do it with. If you take any glucose-lowering drug, the first rung is not optional.
For a healthy person wanting to try it, this is a food decision. On insulin or a sulfonylurea, or with heart, kidney or liver disease, it starts with your clinician, because the diet lowers blood sugar fast and doses have to move down with it.
Nutritional ketosis usually means carbohydrate under about 20 to 50 grams a day, protein kept high enough to hold muscle, and fat filling the rest. Below that carbohydrate level the liver makes ketones; above it most people are simply low-carb, which is a reasonable place to stop if full keto is more than you want.
Eggs, fish, meat, olive oil, nuts, avocado, above-ground vegetables and some dairy do the job without the ultra-processed bars and sweeteners sold under the keto label. Lean on unsaturated fats (olive oil, nuts, fish) over saturated ones (butter, fatty meat), which tends to keep LDL lower.
Lower carbohydrate over a week or two, not overnight. A gradual drop softens the keto flu, the tired, headachy, foggy patch many people hit in the first days as the body shifts fuel.
As insulin falls the kidneys shed water and electrolytes, and that loss drives most of the early fatigue, headaches and cramps. A bit more salt, enough water, and potassium and magnesium from food make the first weeks far easier.
Keep protein up and keep doing some resistance work, so the weight you lose is fat and your muscle holds. The protein and muscle page below has the detail.
A lipid check before you start gives your baseline. A second a few months in tells you which kind of responder you are, since LDL cholesterol climbs steeply in a minority. If your diabetes numbers improve, that is information for whoever prescribes for you, and any dose change is theirs to make.
Go Deeper
- Type 2 diabetes: where the supervised blood-sugar results come from.
- Insulin and glucose: the biology that makes cutting carbohydrate lower blood sugar.
- Protein and muscle: how to lose fat and hold muscle, the detail behind the sixth step.
- Carnivore diet: the all-meat version, taken further.
- Time-restricted eating: the timing lever, a different question from what you eat.
The Chinese Medicine View
In Chinese medicine, digestion belongs to two organs, the Spleen and the Stomach. The Stomach receives and ripens food; the Spleen transforms it and lifts the useful part up and out to nourish the body. The tradition has a name for very rich, fatty, sweet and greasy food: fei gan hou wei, the rich and greasy flavors. A steady load of them overtaxes the Spleen's transforming work and breeds dampness and phlegm. A diet built on fat and meat is a heavy dose of exactly those flavors. Four signs mark the strain: a greasy or thick tongue coat, heaviness after eating, loose stools and sluggish digestion.
The tradition favors warm cooked food at regular meals and is wary of cold and raw. A home-cooked keto plate of cooked meat, fish and vegetables sits closer to that than cold cereal and juice does. There is no blanket verdict here; it turns on the person's constitution. Strong digestion, a robust Stomach, handles rich food better. Someone with Spleen Qi deficiency reads differently, four familiar signs mark it: poor appetite, loose stools, fatigue that is worse after eating, and a pale tongue with tooth marks. That person is already unequal to their food, and piling on the richest flavors adds to a load already too much. The Spleen patterns and Stomach patterns pages carry the full differentiation. The classical advice and the modern trial are describing different things, and neither certifies the other.
Cautions For This Practice
Extra restraintEverything to be aware of is here, in one place. This practice suits most healthy people; a few situations call for real care.
Dangerous lows if an insulin or sulfonylurea dose is not cut
Because carbohydrate restriction lowers blood glucose quickly, glucose-lowering medication set for a higher-carb diet can cause hypoglycemia. In the supervised continuous-care programs, insulin was reduced or discontinued in 94% of users and sulfonylureas were withdrawn entirely, with clinicians adjusting doses as glucose fell. That proactive deprescribing is what kept it safe; the same glucose drop at home on unchanged insulin or sulfonylurea doses is how a serious low happens.Hallberg et al., effectiveness and safety of a novel care model for the management of type 2 diabetes at 1 yearAthinarayanan et al., long-term effects of a continuous care intervention including nutritional ketosis for type 2 diabetes: a 2-year non-randomized trial
Euglycemic ketoacidosis risk on an SGLT2-inhibitor drug
SGLT2 inhibitors promote ketone production, and a ketogenic diet does the same, so combining them can raise ketones and blood acid into diabetic ketoacidosis while glucose stays near normal (euglycemic DKA). Published case reports describe this in people with type 2 diabetes who started or continued an SGLT2 inhibitor on a very-low-carbohydrate diet. Case reports show the combination can cause it, not the rate; because glucose is normal, a home meter does not flag it, and the presentation is nausea, vomiting, breathlessness and malaise. Many clinicians stop an SGLT2 inhibitor before a ketogenic diet.Earle et al., euglycemic diabetic ketoacidosis in concurrent very low-carbohydrate diet and sodium-glucose transporter-2 inhibitor use: a case report
The keto flu peaks in the first week, settles within about four weeks
An analysis of 101 personal accounts drawn from 43 online forums found that starting a ketogenic diet is commonly followed by fatigue, headache, nausea, dizziness, lightheadedness, decreased energy and irritability, peaking in the first seven days and mostly resolving within four weeks. This is self-reported, self-selected data, not a controlled trial, and a published commentary questioned the method, so it maps the reported pattern, not establishing incidence. The symptoms are largely attributed to sodium and fluid loss as insulin falls, which is why extra salt, water and a gradual reduction in carbohydrate reduce them.Bostock et al., consumer reports of 'keto flu' associated with the ketogenic diet
Who should not do this without specialist supervision
Type 1 diabetes carries a serious risk of ketoacidosis and a ketogenic diet there is a specialist decision, not a self-start. Pregnancy and breastfeeding raise energy and nutrient needs and the diet has not been tested in either. A few inherited metabolic disorders, including fatty-acid oxidation defects, pyruvate carboxylase deficiency and porphyria, make a ketogenic diet unsafe. A history of pancreatitis or gallbladder disease, or advanced kidney or liver disease, also calls for medical advice before starting.
Have your cholesterol checked
On average the diet raises LDL cholesterol a little while lowering triglycerides, so the lipid picture is mixed. The reason to check is the minority, often lean and otherwise healthy, whose LDL climbs steeply. A lipid panel before you start and a few months in tells you which you are, and a sharp rise can often be softened by trading saturated fat for unsaturated without leaving ketosis.
When to talk to a doctor first
Speak to a doctor or pharmacist before starting if you take any glucose-lowering medicine, if you have heart, kidney or liver disease, if you are pregnant or breastfeeding, or if you have a history of an eating disorder or a very restrictive relationship with food, for which a highly rule-bound way of eating can be its own hazard.
Start slow, be smart, read the research, and consult a professional if you have any concerns. This is here to inform your choice, not make it for you.
Common Questions
How strong is the evidence in childhood epilepsy?
Strongest of all, and run as a medical treatment. In a randomized trial of 145 children, all with epilepsy that had failed two or more medicines, a supervised diet more than halved seizures in 38%. Only 6% did that well on usual treatment. It is a prescribed therapy, specific to that one condition. It does not carry over as an endorsement of the diet for anyone else.
How solid is the type 2 diabetes result?
Solid but qualified. In supervised programs, average HbA1c fell from 7.6% to 6.3% over a year, and insulin was reduced or stopped in 94% of users. Three things temper it. The strongest programs were not randomized, so their intensive coaching cannot be cleanly separated from the diet. Pooled randomized trials show more remission at six months, then the edge over other diets fades. And the deprescribing that made it safe was a clinical decision, taken as the numbers fell.
Is a ketogenic diet better than other diets for losing weight?
Barely, and only against the weakest comparison. Pooled year-long trials put a very-low-carb ketogenic diet under a kilogram ahead of a low-fat diet. In a randomized trial of 609 people, a healthy low-carb diet and a healthy low-fat diet lost about the same. Which diet you can keep to matters more than the carbohydrate level itself.
What is the keto flu, and how long does it last?
A cluster of temporary symptoms in the first days: fatigue, headache, nausea, dizziness and irritability. It is reported to peak in the first week. Much of it comes from losing salt and water as insulin drops. The salt, water and gradual ramp in the steps above blunt most of it, and the symptoms are mostly gone within about four weeks.
Explore Related
Other pages this one connects to, by the evidence they share, the outcomes they touch, and the ground they cover.
All 17 sources on this page independently checked and cross-referenced.
Thomas Dehli, Founder & Editor, Sacred Lotus
Sacred Lotus has published Chinese medicine reference material since 2001. Integrative pages are held to the same standard as the herb and formula library: cite the source, grade the claim at its real strength, and say where the research has not looked. This page is educational and it is not medical advice. Last reviewed and updated August 9, 2026.
Evidence strength
How confidently the research supports a claim. Strength describes the evidence, not our endorsement.