Keto made easy

The PURE Salt Study: Is Low Sodium Bad for You?

Quick answer: The PURE study estimated sodium intake in 101,945 people across 17 countries and found that 3 to 6 g of sodium a day was associated with fewer deaths and cardiovascular events than eating more or less (O'Donnell et al., 2014). That "J-curve" is a real signal, but it came from a single morning urine sample per person, and it's an association, not a trial. Guideline bodies still set lower targets. If you're on keto, you probably need more sodium than you think, and if you have high blood pressure or kidney disease, that's a conversation for your doctor.

On this page

What PURE did

PURE is a large international cohort study. Its sodium paper followed 101,945 people in 17 countries. Instead of asking people what they ate, the researchers estimated each person's sodium from one morning spot urine sample, then watched who went on to die or have a heart attack or stroke (O'Donnell et al., 2014).

A sister paper from the same study looked at blood pressure. The link between sodium and blood pressure was steeper at higher intakes, in people with hypertension and in older people, and potassium went the other way: more potassium, lower pressure (Mente et al., 2014).

The J-curve

Plot risk against sodium and you'd expect a straight line: more salt, more risk. PURE drew a J instead. Risk was lowest in the 3 to 6 g a day range and rose both above it and below it.

The high end surprises nobody. The low end is the controversial part, because it suggests that the low targets many guidelines recommend might be too low for some people.

Why one urine sample matters

Here's the problem with the low end of that J. One morning sample is a snapshot, and estimating a whole day's usual intake from it adds error. That single-sample method is the caveat attached to the study.

The second problem is that it's observational. PURE watched people; it didn't assign anyone a salt level. People who eat very little salt may differ in other ways. Some of them may be eating little because they're already unwell. An association can't separate "low salt caused harm" from "sick people ate less".

So what you do with PURE is treat it as a reason to question very low sodium, not as proof that 5 g a day is the healthy number for you.

What the guidelines say instead

The bodies that set targets still come down lower than PURE's sweet spot:

Source What it says
National Academies, 2019 Sodium adequate intake 1,500 mg/day; reduce intakes if above 2,300 mg/day (NASEM, 2019)
WHO, 2012 Adults under 2,000 mg/day sodium (under 5 g of salt) (WHO)
American Heart Association No more than 2,300 mg/day, "optimal goal" 1,500 mg; notes 1,500 may not apply to people losing a lot of sodium in sweat (AHA)
PURE, 2014 3,000–6,000 mg/day associated with the lowest death and CV-event rates (O'Donnell et al., 2014)

Notice salt and sodium aren't the same number. WHO's 5 g of salt lines up with its 2,000 mg sodium figure, so a salt number is about two and a half times the sodium number.

Why keto changes the sodium question

Most of the sodium in a typical diet rides in on processed food, bread and snacks. Cut those and your intake can drop a long way without you touching the salt shaker.

At the same time, insulin tells your kidneys to hold onto sodium. In a classic experiment, infusing insulin while holding blood glucose steady cut the sodium healthy men excreted (DeFronzo et al., 1975). Lower your insulin by cutting carbs and the reasoning runs the other way: you let more sodium go. Nobody has run a big trial measuring this on a ketogenic diet, so this is mechanism from fasting and insulin studies, but it fits how a lot of people feel in week one.

So you're eating less sodium and keeping less of it. That's a different situation from the average person a low-sodium guideline was written for.

What the research says

The trials that actually changed people's sodium mostly point one way:

That's why the experts still argue. The trials show lowering sodium lowers blood pressure. PURE suggests going very low might not lower risk. Both can be partly true at once, depending on who's eating the salt.

KetoDick's take

In his electrolytes video, he argued the old sodium recommendation was aimed at people with salt-sensitive hypertension, which he described as a genetic subgroup (video -dHjt_Bpruk). He described a mortality study that, as he put it, found around 5,000 mg of sodium a day optimal, with more risk at 3,500 than at 6,500, and said he aims for about 5,000 mg himself. He didn't name the study; PURE is the best-known one with that shape, and his 5,000 mg sits inside its 3 to 6 g range.

Where he and the literature differ: the guideline bodies don't accept that the low targets are only for a subgroup. The trials above show blood pressure drops with less sodium in people without hypertension too, just by less. And PURE is a single-sample observational study, not the final word.

Where he's on firm ground is the keto-specific part. He's said that without processed food and starches, salt mostly disappears from the diet, and that in ketosis you retain less of it (video -dHjt_Bpruk). His fix is cheap: a chicken bouillon cube three times a day, each cup with a quarter teaspoon of potassium-chloride salt. He's also been specific that sodium and potassium imbalances "can end you up in the hospital", so talk to a doctor before taking supplements (video 9r2Y74AnjUQ), and that people with pre-diabetes or type 2 diabetes often have some kidney impairment and should bring this to their doctor (videos lbRPlGZ3Rvw, GojQvo_2GS0).

Key takeaways