Keto Electrolytes: How Much Sodium, Potassium and Magnesium Do You Actually Need?

Cutting carbohydrate really does make you shed sodium and water — settled physiology, and the reason week one of keto feels like flu. What isn't settled is the number. Nearly every keto page quotes 3,000–5,000 mg of sodium a day as though it were a dietary reference value. It isn't. It is clinical practice guidance built on a mechanism, and the 2025 scoping review that went looking for the trials behind it concluded that quantified recommendations for salt supplementation during keto induction "do not appear to exist."¹ Here's what each mineral actually needs, where the famous figures come from, and the magnesium limit most keto guides break without noticing.

Why keto makes you lose electrolytes

Two things happen when you cut carbohydrate, and only one of them is the one people talk about.

The first is glycogen. Liver glycogen is close to depleted after 24 hours on a low-carbohydrate diet, and muscle glycogen falls about 20% in the first week.¹ Glycogen is stored with water, so the water goes with it.

The second actually drives the scale, and it is a kidney story. Insulin increases sodium reabsorption in the kidney.¹ The landmark demonstration infused insulin into six healthy people while holding blood glucose steady: urinary sodium excretion fell by roughly half, from 401 to 213 μEq/min, with no change in glomerular filtration rate — direct tubular reabsorption.² Run that in reverse. Carbohydrate drops, insulin drops, the brake comes off, and the kidney lets sodium go, with water following and some potassium behind that. The same review notes early keto weight loss is predominantly water lost to this natriuresis, not glycogen alone¹ — the physiology described in fasting half a century ago as "the natriuresis of fasting." For what these minerals do outside the keto context, start at our electrolytes hub.

"Keto flu": what the evidence actually says

The symptom cluster is real and well described. Across 89 studies, the scoping review put adult occurrence rates at 8–25% for headache, 18–25% for fatigue, 15–21% for dizziness, 8–16% for nausea and 3–37% for muscle cramps, with symptoms typically appearing within 2–3 days and mostly resolving inside 2–4 weeks.¹ The best-known "keto flu" paper is descriptive in a different way: it analysed 448 posts across 43 online forums, where the 101 users describing their own experience reported headache (24.8%), fatigue (17.8%), nausea (15.8%), dizziness (14.9%) and brain fog (10.9%).⁶ A useful record of what people report — not a trial, and its authors say so.

Here is the gap nobody names. The natriuresis is well documented. The symptoms are well documented. Nothing has connected the two, and nothing has shown that taking electrolytes resolves them — in the review's phrasing, "no studies have tied these states directly to symptoms of keto-induction."¹

Hold that alongside the fact that an abrupt calorie drop, a caffeine change and plain dehydration produce the same symptom list, and all usually happen in week one too. So a C is the honest grade: strong mechanism, no trial evidence. Salt is cheap enough that trying it is sensible; "electrolytes fix keto flu" is a claim nobody has earned.

Sodium on keto: where the 3,000–5,000 mg figure comes from

It comes from low-carbohydrate clinical practice. The peer-reviewed version, documented in the scoping review, is Volek and Phinney's advice to drink 1–2 cups of broth or bouillon a day while low-carb — which the review notes provides 1–2 grams of extra sodium on top of what you already eat, enough to blunt the weakness, fatigue and light-headedness of hypovolaemia.¹ Average US sodium intake is about 3,400 mg a day.⁷ Add 1–2 g and you land at 4,400–5,400 mg — which is where the round "3,000–5,000 mg" keto target comes from. It is an increment, restated as a total.

That reframing matters, because the increment is defensible and the total looks alarming next to population guidance:

• The Adequate Intake for sodium in adults is 1,500 mg/day.³
• The Chronic Disease Risk Reduction intake advises adults to reduce intakes if above 2,300 mg/day — about one teaspoon of table salt.³⁷
• The keto figure is therefore more than double the level at which national guidance asks people to cut back, and no randomised trial has tested it.¹

Whether that matters for a healthy person is genuinely contested, and the disagreement is really about study design. DASH-Sodium randomised 412 people through 30 days at each of three sodium levels under controlled feeding: high to intermediate lowered systolic blood pressure 2.1 mmHg, intermediate to low a further 4.6 mmHg.¹⁰ PURE followed 101,945 people in 17 countries, estimated sodium from a single morning urine sample, and found the lowest rate of death and cardiovascular events between 3 and 6 g/day.¹¹ One randomises the exposure and measures blood pressure over a month; the other observes populations, infers intake from a spot urine, and counts events over 3.7 years. Different questions — and we won't pretend one settles the other.

For everyone else, the practical version is cheap: salt your food, and drink a cup of broth or bouillon if week one feels rough. Sodium is the least expensive nutrient on earth. Nothing in this section requires a purchase.

Potassium on keto: the one to be careful with

Start with the number. The Adequate Intake for potassium, updated by a NASEM committee in 2019, is 3,400 mg/day for men and 2,600 mg/day for women.³⁴ The 4,700 mg quoted across keto sites is the Daily Value used on Nutrition Facts labels, not the intake recommendation.⁴ Different instruments, different jobs — and quoting the label figure inflates the target by about 40%.

Keto does make potassium harder, because the foods carrying most dietary potassium are the ones the diet removes: a baked potato holds about 610 mg, a cup of cooked lentils 731 mg, a banana 422 mg.⁴ The keto-compatible replacements are real but smaller — two cups of raw spinach at 334 mg, three ounces of Atlantic salmon at 326 mg, an ounce of cashews at 187 mg.⁴ The clinical guidance routes potassium through food too: roughly 4 g/day, "mainly through bone broth or vegetables like avocados, nuts and seeds."¹ Food-first is not our hedge. It is what the practitioners actually said.

There is no Tolerable Upper Intake Level for potassium, which is easy to misread as a safety clearance. It isn't: the committee found case reports that very large supplement doses can cause heart abnormalities and death, but judged them insufficient to set a UL.⁴

Magnesium on keto — and the upper limit most keto guides miss

This is the factual error copied across the whole category.

The RDA for magnesium is 310–420 mg a day, counting food and supplements together.⁵ Separately, the Tolerable Upper Intake Level for supplemental magnesium is 350 mg for adults — pills and magnesium-containing medications only, not the magnesium in food.⁵ So when a keto page prints "take 300–500 mg of magnesium daily" as a supplement target, the top of that range sits above the supplemental ceiling. Plenty of them do exactly that.

That ceiling is not a toxicity line. It is set where supplemental magnesium reliably causes diarrhoea, nausea and abdominal cramping.⁵ Which is exactly why it matters more here than anywhere else: your gut is already renegotiating fibre, fat and fluid in week one.

The honest version: aim for the RDA in total, keep the supplemental share at or under 350 mg of elemental magnesium, split it in two, and let loose stools be your ceiling signal. The word doing the work is elemental — a "400 mg magnesium glycinate" capsule contains about 56 mg of magnesium, because glycinate is roughly 14% magnesium by weight. Our guide to how much magnesium per day works that math out form by form.

Form matters on keto specifically. Citrate is osmotically active by design — useful for constipation, unhelpful when your digestion is already unsettled; glycinate is the gentlest common form. The form-by-form breakdown grades all of them, and the glycinate vs. citrate comparison settles the everyday choice.

We may earn a commission if you buy through this link, at no extra cost to you. It doesn't change the guidance above — 350 mg of supplemental elemental magnesium is the ceiling whatever brand you choose.

Magnesium is the one keto shortfall where a supplement is genuinely the sensible route rather than a convenience, which is why it is the only product recommendation on this page. The magnesium hub grades what the mineral is and isn't good for.

How long do you actually need to do this?

The section nobody writes, and the most useful one.

The losses are front-loaded. The increased natriuresis and kaliuresis of keto induction subside after about 14 days, the same pattern seen in fasting.¹ Symptoms follow the same curve: onset within 2–3 days, most resolved within 2–4 weeks.¹ Exercise capacity takes longer — impairment appears to last 3–4 weeks before most adults return to normal.¹

So the aggressive sodium numbers describe an adaptation window, not a permanent requirement. Expect to need less salt at month three than in week one. It varies between people and moves with heat, sweat and training load — but writing 5,000 mg/day as a lifelong keto tax, which is what the category does, isn't supported by the physiology it cites.

One caution: liberal fluid intake is widely recommended during keto induction, and the review flags that no experimental evidence supports it.¹ Drinking a great deal of plain water while your kidneys are actively dumping sodium is the one way to make week one measurably worse.¹² Salt first, then drink to thirst.

Do you need a keto electrolyte powder?

Three minerals, three different answers.

Sodium: no. Salt is free, broth works, and the increment the clinical guidance calls for is 1–2 g a day on top of normal eating.¹ A powder is a delivery mechanism for something already in your kitchen.

Potassium: no, and this is where a supplement carries real risk — the 99 mg cap and the medication interactions above are the reason.⁴ Food is the route the practitioners themselves recommend.¹

Magnesium: reasonably, yes. A common shortfall, a well-understood supplemental route, a ceiling clearly marked at 350 mg.⁵

If you still want a powder — for convenience, or because a fixed dose helps you stay consistent — the sugar-free formulations are the relevant category, since a sugary hydration mix defeats the point of the diet. Our broader look at whether electrolyte supplements are worth it covers that decision for everyone, not just keto dieters.

Who shouldn't be running this experiment

Everything above assumes a healthy adult choosing a diet. Two things it doesn't cover.

Ketogenic diets have a real clinical history — drug-resistant epilepsy is the one guideline-supported indication, everything else remains investigational — and a real contraindication list.⁸ Rare fat-oxidation disorders and porphyria are absolute; acute pancreatitis, acute liver failure and advanced chronic kidney disease (stages G3b–G5) are relative.⁹ Pregnancy and a history of disordered eating are also reasons not to start unsupervised.⁸

Medication changes the picture entirely. Carbohydrate restriction lowers blood pressure and blood glucose enough that antihypertensive and glucose-lowering doses often need adjusting by the prescriber, and combining a ketogenic diet with an SGLT2 inhibitor carries a ketoacidosis risk.⁹ The supervised version of this diet monitors sodium, potassium, magnesium and bicarbonate⁸ — it measures what this article can only estimate. None of this is medical advice.

The bottom line

Carbohydrate restriction genuinely makes you dump sodium and water, and most of the fix is a salt shaker. The 3,000–5,000 mg figure quoted everywhere is clinical practice guidance built on a real mechanism, not a trial result, and it's more than double the 2,300 mg level national guidance asks adults to stay under — so if you have high blood pressure, kidney disease or heart failure, that's a clinician's call and this page gives you no number. Get potassium from food. Keep supplemental magnesium at or under 350 mg elemental. And expect to need less of all of it after the first couple of weeks.

Frequently asked questions

Do you need electrolytes on keto?

You need more sodium than usual, at least early on, because falling insulin makes the kidneys shed sodium and the water that follows it.² Most of that is covered by salting your food, not by buying a powder.

How much sodium do you need on keto?

Keto sources commonly say 3,000–5,000 mg a day. That's clinical practice guidance built on the mechanism rather than a trial result,¹ and it's more than double the 2,300 mg level at which national guidance asks adults to cut back³ — a reason to be careful if you have high blood pressure, kidney disease or heart failure.

What is keto flu and how long does it last?

A cluster of self-reported symptoms — headache, fatigue, brain fog, nausea, dizziness — usually starting within 2–3 days of cutting carbohydrate and mostly resolving within 2–4 weeks.¹ Much of the evidence base is people's own accounts rather than clinical trials.⁶

How much magnesium should you take on keto?

Aim for the RDA of 310–420 mg from all sources, and keep supplemental magnesium at or under 350 mg of elemental magnesium a day — that's the tolerable upper intake level for supplements, set because higher supplemental doses cause diarrhoea.⁵

How do you get potassium on keto?

From food: leafy greens, salmon, avocado, mushrooms, nuts and seeds. Two cups of raw spinach carry about 334 mg and three ounces of salmon about 326 mg.⁴ Potassium supplements are the one place to be cautious, especially with kidney problems or blood-pressure medication.⁴

Why do you lose so much water at the start of keto?

Glycogen is stored with water, and liver glycogen is nearly depleted within a day of carbohydrate restriction.¹ The larger driver is that falling insulin releases the kidney's brake on sodium excretion, and water follows the sodium out.²

Do electrolytes fix keto flu?

The mechanism makes it plausible and many people report it helps, but no study has tied the electrolyte losses directly to the symptoms, and quantified salt-supplementation recommendations for keto don't exist in the literature.¹ Treat it as a reasonable, cheap thing to try — not a proven fix.