Circulation Notes · About this section

Food and Circulation: What Goes in This Section, and What Does Not

A recipe site does not need a section about blood vessels. This one has one because of what arrives in our inbox, and because the claims arriving there are checkable. Here is the rule we apply, what it lets in, what it keeps out, and where the paid pages sit.

Why this section exists

Most of what readers send us is a recipe question. A smaller pile is not: it is a screenshot of something a relative forwarded, usually a glass of something in the morning, and usually with a promise attached. Beets and ginger. Honey, garlic and cinnamon. A tea instead of the evening coffee.

We kept answering those one at a time until it was obvious they were the same question wearing different groceries: somebody says this food does something to my circulation, and I have no way to tell whether that is true. That question has an answer more often than you would expect, because these claims are specific enough to have been tested in people. So we started looking the trials up and writing down what they said.

That is the whole section. It is not a health section, we have no clinician on staff, and nothing here is a plan for anybody.

The rule, and it is one sentence

A claim gets a page here when it is about a food, specific enough to be wrong, and already measured in people — and the page then reports what the measurement said, including when it said nothing.

Each of those three does work. About a food keeps the section inside what this site knows; we have no business reviewing a drug. Specific enough to be wrong is what separates “beets lower your blood pressure by five points” from “beets support wellness”, and only the first can be checked. Measured in people is the line we refuse to move: a result in a rat is a reason to run a trial, not a reason to change a breakfast, and where we can only find animal data we label it as animal data and leave the claim open.

The last clause is the one that costs us readers. When the trials come out empty, the page says so. That has already happened: the claim readers send us most often about leafy greens has two trials built to test it, and both found nothing.

What goes in

• Claims about an ordinary grocery and a measurable outcome — a blood pressure number, an artery measurement, a blood lipid.
• The dose actually used in the trial, in grams or millilitres, because “eat more garlic” and “1.2 g a day of a standardised extract” are not the same instruction.
• The population actually tested. A result in untreated mild hypertension says nothing about a person whose pressure is normal, and the trials are explicit about that split.
• Interactions with medication where somebody has measured them in people, and a plain label when the only data is from animals.
• Every identifier, linked to the public record, so that a reader can check us.

What stays out

• Anything that reads as treatment. No page here says a food treats, prevents or cures a condition, and none ever will.
• Dosing advice. We report the dose a trial used. We do not tell anybody what to take.
• Stopping or changing medication. If a page could be read that way it gets rewritten. The most common dangerous claim in this whole area is “as good as the prescription”, and where it has been tested head to head it has lost.
• Before-and-after accounts, and personal results. They are real for the person and predict nothing for a reader.
• Hard outcomes. No trial we have read measured a heart attack, a stroke or a death. Until one does, this section does not discuss them.

How a page gets checked

We search the public index for human studies on the food and the outcome, read what we can reach, and write down the design, the number of participants, the dose, the result and the authors' own stated limits. Where a review and a single trial disagree, both go on the page. We do not pool results ourselves and we do not turn a relative change into an absolute one, which is the most common way a modest number becomes an impressive one.

The date of the last source check sits under every headline in the section and moves when we check again. If a trial lands that changes an answer, the answer changes and the old one is not quietly deleted. If you find a number here that does not match the record it links to, write to us: that is a correction, it gets made, and it gets dated.

Where to start

The section's first page takes the three groceries readers ask about most and checks six claims against the published trials, with every identifier linked: Claims Review: Arugula, Garlic and Hibiscus for Blood Flow. Two of the three have trials that found nothing, which is a strange way to open a section and the honest one.