Why some supplements show 'Studies not yet linked'

What the label means

Each supplement and condition page is built from a table of findings. A finding is one measured result for one supplement, such as "blood pressure" for magnesium. It can show a grade, a direction (such as "Improves"), the size of the effect, and the studies behind it.

Some findings have none of the things that would back a direction: no grade, no written description of what the studies found, and no study linked to them. Until now, many of those still showed "Improves" with an arrow, read from a direction stored in the database. Nothing on the page said which study found it, in whom, or how big the change was.

Those findings now say Studies not yet linked instead, with no arrow and no size word. Its tooltip says that no study is linked to the finding yet, so it shows no direction. Our How we grade page puts the rule in one line: a benefit is shown only once a study is linked to the finding.

The change went live on 9 November 2026 on all but a small number of pages, which keep the old reading for now.

Where the old arrow came from

The database behind the site was first compiled from published evidence summaries. Many findings came with a stored direction: the measure went up or down. That record is a starting point, not evidence on its own.

"Up" and "down" are not the same as better and worse. Lower blood pressure is usually good; lower muscle strength is not. Turning a stored direction into "Improves" takes a guess about which way is better, and a trust that the original summary read its studies correctly. Without a study on the page, you can't check either.

Why a gap is more honest than an arrow

An arrow and the word "Improves" read as a finding. People see them and assume there is a trial behind them. When there isn't one on the page, the arrow claims more than we can show.

A blank tells you exactly where things stand: we haven't yet linked a study we've checked for this result. That fits the rest of our rules. We don't show a study count for a citation that doesn't link to a paper, and we don't call an effect helpful or harmful when the data doesn't say which it is (our editorial policy lists these). A finding with nothing behind it now follows the same rule.

How a study gets linked to a finding

Linking is a check, not a lookup. For each finding, we start from candidate studies: PubMed searches, and the study lists of the evidence summaries the database was first built from. A candidate is only a lead. We read its record on PubMed, and the full paper when the record leaves a detail unclear, such as how the supplement was taken.

A study is linked only when all of these hold:

• It tested this supplement on its own. A blend where this supplement's effect can't be separated doesn't count. A trial that adds the supplement to a medicine or a training plan counts only when the comparison group gets the same medicine or training.
• It was taken the way the page is about. For a supplement you swallow, a skin cream, an injection or a nasal spray doesn't count.
• It had a comparison group. People taking the supplement are compared with people taking a placebo, no treatment or usual care, and the study reports the difference between the groups, not only how one group changed over time.
• It studied the right people and the right result. The people match the finding's condition, and the study measured this outcome or a direct measure of it.
• It was done in people. Studies in animals or cells can stay listed on a page, but they don't earn a grade.
• It passes our integrity checks. It isn't retracted and isn't by an author on our internal integrity list. For a review that pools trials, we check whether its result for this outcome leans on a retracted or flagged trial (our post on how we handle retracted research explains how).

Then we compare the study with what the finding showed. If the study supports it, the study is linked and listed under the finding. If the study found no difference, or found a harm, the finding is corrected to match the study rather than left as it was.

What changes when a study is linked

Each fix is listed in our corrections log: what the finding said, what it says now, and why, with how many studies were added. Some findings go back to "Improves", now with the trials behind it listed. Others change: a finding that showed "Improves" with no study can become "No effect" once the only trial is linked and found nothing.

Linking a study doesn't give a finding a grade by itself. A grade follows the rules on How we grade: how many good human trials there are, and whether they agree.

What the label doesn't mean

"Studies not yet linked" is not a verdict on the supplement. It doesn't mean the supplement doesn't work, and it doesn't mean it is safe or unsafe. It means we haven't yet checked and linked a study for that one result. The same supplement can have other findings on the same page with studies, grades and directions.

It also doesn't cover every finding without a linked study. A finding that has a grade or a written description keeps its reading while it waits for its studies. As How we grade says, a finding without a "Reviewed" label hasn't been re-checked yet, so read its letter as a starting point.

How you can help

If you know a good trial for a finding that says "Studies not yet linked", or you think a linked study doesn't fit its finding, email hello@winnowleaf.com with the page address and the study. Our corrections policy says what happens next, and the fix will appear in the corrections log.