Diet Headlines: What "Linked To" Really Means
"Linked to" means a study found two things moving together, not that one caused the other, and many diet headlines also skip whether the study was in people and whether it measured a real health outcome.
On this page
- Headline words and what they usually mean
- Association is not cause
- Mice and cells are not people
- Lab markers are not outcomes
- Five questions before you share a headline
- Related
Headline words and what they usually mean
| Headline says | What the study often found |
|---|---|
| "Linked to," "tied to," "associated with" | An observational study found two things go together |
| "Slashes risk by 30%" | A relative difference between two groups, not your personal odds |
| "Boosts," "protects," "fights" | Often a cell or animal result, or a change in a blood marker |
| "Scientists say" | One study, sometimes one author's comment |
| "New study proves" | Single studies rarely prove anything on their own |
Association is not cause
Large observational studies follow people over years and look for patterns. They are useful for spotting what to test next, but the people being compared choose their own diets and differ in many other ways.
Take fiber. Reynolds et al., 2019 compared the highest and lowest fiber eaters and found 15–30% lower rates of death, heart disease, stroke, type 2 diabetes and colorectal cancer in the high-fiber group. A fair headline is "high fiber intake linked to lower disease risk." An unfair one is "fiber cuts your cancer risk by 30%."
The 15–30% is also relative. It compares two groups; it does not say how many people in 100 that represents. See relative vs absolute risk and, for the difference between study types, Observational Studies vs Randomized Trials.
Mice and cells are not people
Several of the most-cited papers on ketones and gut bacteria were done in animals or in cells. They show possible mechanisms. They do not show what happens to a person on a diet.
| Study | What was studied | What it found |
|---|---|---|
| Youm et al., 2015 | Mouse models and human immune cells | The ketone BHB blocked an inflammation pathway (NLRP3), separately from its role as fuel |
| Shimazu et al., 2013 | Mice | BHB blocked certain enzymes (class I HDACs) and raised genes linked to resisting oxidative stress |
| Desai et al., 2016 | Germ-free mice given set bacteria | On a fiber-free diet, gut bacteria broke down the colon's mucus layer |
| Sonnenburg et al., 2016 | Mice with human gut bacteria | A low-fiber diet reduced bacterial diversity, with losses building over generations |
| Ang et al., 2020 | A human diet study plus mice | A ketogenic diet shifted the gut microbiome, with lower bifidobacteria |
How to read these: "Ketones fight inflammation" is a reach from Youm 2015, which did not test people eating a ketogenic diet. "Low-fiber diets starve your gut" is a reach from Desai and Sonnenburg, which were in mice. Ang 2020 included people, so it sits a step closer, but it measured gut bacteria, not health outcomes. See ketosis explained and the keto gut microbiome.
Lab markers are not outcomes
A surrogate outcome is a lab number used as a stand-in for a real event. Blood pressure stands in for stroke. LDL cholesterol stands in for heart attacks. HbA1c stands in for diabetes complications. Surrogates are useful, but a better number does not always mean fewer events.
| Study | Marker measured | What it found |
|---|---|---|
| Hallberg et al., 2018 (Virta) | HbA1c, weight, medication use | Non-randomized, 262 vs 87 usual care: HbA1c 7.6% → 6.3%, weight about −12%, less diabetes medication |
| Bueno et al., 2013 | Weight and blood lipids | 13 trials of 12 months or more: very-low-carb ketogenic diets −0.91 kg vs low-fat; lower triglycerides and diastolic blood pressure, higher HDL-C and higher LDL-C |
| Graudal et al., 2017 | Blood pressure and hormones | Sodium reduction, white adults with high blood pressure: systolic −5.51 mmHg; with normal blood pressure: systolic −1.09 mmHg; renin +55%, aldosterone +127% |
How to read these:
- Hallberg 2018 is a strong marker result, but it was not randomized, so the groups may have differed at the start.
- Bueno 2013 shows markers moving in both directions at once. Whether the net effect on heart disease is good or bad is not answered by the markers alone.
- Graudal 2017 shows a blood pressure drop alongside hormone changes. Which matters more for long-term health is a separate question.
KetoDick on this: he says cholesterol may spike early on keto and that the HDL/LDL ratio and particle size improve, with markers improving over 6 to 12 months; he attributes most cardiovascular improvement to weight loss (video UygKGoP_BtA). The literature is more mixed: Bueno 2013 found higher HDL-C and higher LDL-C, and the National Lipid Association statement reports a variable effect on LDL-C. Both views are here so you can take them to your doctor. See cholesterol on keto.
If you take blood pressure or diabetes medication, a better lab number on a new diet is a reason to see your doctor soon, since doses may need to change.
Five questions before you share a headline
- Was it in people? Look for "mice," "rats," "cells" or "in vitro" in the abstract.
- Was it randomized? If not, "linked to" is the most it can say.
- What was measured? A real event, or a lab marker?
- How big was the effect? Find the actual numbers, not just "lower risk."
- Does the story name the study? If not, you cannot check it. See How to Read a Nutrition Study Abstract.
KetoDick on this: in his rebuttal of a 2019 Men's Health keto article, he objects to its anecdote-led structure and checks its claims against what keto practice looks like, such as its statement that keto "strictly caps" vegetables (video 1JbK1_ClDmY). Read the full rebuttal in the Men's Health keto article.
Related
Guide: Chapter 12: Reading nutrition news and studies · How to read an abstract · Observational studies vs randomized trials · Red flags in diet advice · Ketosis explained · Fiber on keto · Cholesterol on keto
Blog: Relative vs absolute risk · The keto gut microbiome · Saturated fat on keto · The Men's Health keto article
Videos: Keto basics for beginners · Men's Health hack job on keto
Last verified: 2026-10-31