40 Grams of Fibre: The Number Nobody Recommended

“The ketogenic diet is not effective; consumption of whole foods with fibre intake above 40 g per day improves health markers.”

Where it circulates

The Diary Of A CEO

Fat Loss Scientist: It's Easy To Lose Weight, But Here's Why You WON'T Do It! | Dr Andy Galpin

Aug 13, 2026 2:12:19 1,293,301 views on Sep 7, 2026

@075ad1 2mo ago

A commenter setting out the claim in full: ketogenic diets are dismissed as ineffective, while eating whole foods with more than 40 g of fibre a day is said to improve health markers across the board, with carbohydrates defended as the body's main fuel and only refined ones worth avoiding.

9 likes 1 reply

Nothing in this episode supports the claim. The comment originated here, but the claim let go of its source – that is a finding, not a gap.

The adjudicated claim is the commenter's, not the host's.

Dr. Gabrielle Lyon

The Muscle Myth GLP-1 Users Have Been Lied to About | GLS #227

Aug 13, 2026 14:08 17,484 views on Sep 7, 2026

@f4f2da 2mo ago

A commenter carrying the same pairing into a discussion about GLP-1 drugs and muscle: the high-fibre whole-food route is presented as what actually improves markers, and carbohydrate-restricted eating as the approach that does not work.

7 likes 2 replies

Nothing in this episode supports the claim. The comment originated here, but the claim let go of its source – that is a finding, not a gap.

The adjudicated claim is the commenter's, not the host's.

Dr. Eric Berg DC

The #1 Healthiest Carb in the World

Aug 8, 2026 8:36 282,115 views on Sep 7, 2026

@70bcce 2mo ago

A commenter under an episode about carbohydrates and insulin resistance, restating the claim as settled: fibre above 40 g a day from whole foods improves health markers, and the ketogenic approach does not deliver.

4 likes 3 replies

Nothing in this episode supports the claim. The comment originated here, but the claim let go of its source – that is a finding, not a gap.

The adjudicated claim is the commenter's, not the host's.

The Proof (Simon Hill)

The Keto Study Everyone Got Wrong: What Science Says About LDL & Heart Disease | Dr Terry Simpson

Aug 4, 2026 1:56:18 21,512 views on Sep 7, 2026

@1366ca 2mo ago

A commenter reading an episode-length criticism of how one imaging study was interpreted as a general result: the ketogenic diet is taken to be ineffective, and high-fibre whole-food eating above 40 g a day is offered as the alternative that works.

3 likes 2 replies

Nothing in this episode supports the claim. The comment originated here, but the claim let go of its source – that is a finding, not a gap.

The adjudicated claim is the commenter's, not the host's.

Four podcast episodes. Between them, 1.6 million views. And a claim that landed in their comment sections four times in six days, always in the same shape: eat whole foods with more than 40 g of fibre a day and your health markers improve — and the ketogenic diet does not work.

We read all four transcripts looking for that sentence. It is in none of them.

That is not a technicality, and it is not a gotcha. It is the subject of this piece. Because the claim is not absurd — in its strongest form it is close to something true. Fibre is among the best-supported interventions in all of nutrition: raise it and blood pressure, glucose and lipids move, in trials, without changing anything else. And the worry about ketogenic diets is raised by cardiologists rather than by cranks, because those diets raise LDL cholesterol in some people, and sharply in some of them. Someone reading four episodes about exactly these things could assemble that sentence in good faith.

What happened between the episodes and the comment sections is worth watching in slow motion, because it happens constantly and it usually leaves no trace. Here it left one.

What everyone says

The claim reached us as 4 occurrences across 4 channels in a 6-day window, measured on 3 September 2026. One occurrence per channel: a two-hour interview about fat loss, a short episode on GLP-1 drugs and muscle, an eight-minute episode about carbohydrates, and a two-hour episode about how one keto study was read.

The four episodes are not variations on one theme. One is a fat-loss interview that passes through ketogenic diets and cholesterol, one is about a drug class and mentions fibre in passing, one is an eight-minute explainer on carbohydrates, and one is an episode-length criticism of how a single keto study was read. What they share is that the claim appears underneath them and in none of them.

Here is what each episode actually contains, read from the transcript.

The Diary Of A CEO discusses ketogenic diets and cholesterol between roughly 33 and 37 minutes, then fibre as a way of lowering cholesterol in the minute and a half that follows. On the point that matters, it says the opposite of the claim: that a rise in cholesterol on this diet does not mean ketogenic diets are bad or unsafe.

Dr Gabrielle Lyon talks about fibre maxing and GLP-1 drugs. The figure of 40 to 60 g is in that episode — as a description of an online trend, accompanied by caution about it. It is named as something happening, not as a target to hit.

The Proof spends an episode on how one study was read: the KETO-CTA plaque-progression cohort. It criticises an interpretation, not a diet. Fibre appears only in the context of colorectal cancer, with the observation that 92% of people fall short of adequate intake.

Dr Eric Berg discusses carbohydrates, glycaemic index and insulin resistance in eight minutes. It contains no assertion about either half of the claim.

In the material gathered for this dossier, the audience splits nearly evenly — 8 comments supporting the claim against 9 opposing it — while the literature runs 21 supporting studies against 12 opposing. That is the reverse of what we found in our second episode, and worth naming: on this question the comment sections are less settled than the evidence, not more. The disagreement is loud in both directions. What is missing from the argument is the four episodes it claims to be about.

How the claim was assembled

Four passages, four different things lost. None of these operations requires bad faith, and none of them has a culprit. They describe what happens to a sentence when it is carried, not what any person intended.

Read together they are one mechanism with four faces: each step drops the part of the sentence that limits it. A caveat, a negation, a scope, a subject. What is left is shorter, cleaner, easier to repeat, and no longer true.

It is worth being exact about what this does and does not say. It does not say the audience is careless. Reconstructing a claim from four long conversations is what listening to podcasts is for, and the reconstruction here is a reasonable one. It says that the four hosts did not make this claim, and that the parts of their episodes that would have blocked it are precisely the parts that did not travel.

What the evidence shows

55 primary sources. The claim carries two assertions inside one sentence, and the evidence adjudicates them in opposite directions. Three questions, answered separately.

Direction of effect

Partial. Evidence: high.

The fibre half holds. In adults with diabetes or prediabetes, meta-analysed controlled feeding and supplementation trials show that raising fibre intake alone — with no other change in lifestyle — improves HbA1c (−2.00 mmol/mol), fasting glucose (−0.56 mmol/L), HOMA-IR, LDL, triglycerides, body weight and CRP. In adults with hypertension or cardiovascular disease, pooled randomised trials show blood pressure falling by roughly 4/3 mmHg at high GRADE certainty, with the benefit holding irrespective of cardioprotective medication. This is not an association read off a questionnaire. It is fibre, given deliberately, moving markers.

The keto half does not hold. A systematic review and meta-regression of 53 randomised controlled trials finds ketogenic diets lowering triglycerides (−22.31 mg/dL) and raising HDL-C (+3.52 mg/dL) at moderate certainty, while raising LDL-C (+8.22 mg/dL) and total cholesterol (+8.06 mg/dL) at low certainty. Systematic review evidence in overweight and obese women with PCOS reports 7–13% weight loss over six to twenty-four weeks, with improvement in androgens and menstrual regularity, at low-to-moderate certainty. An exploratory randomised trial in polycystic kidney disease found reduced body fat and liver volume. A pilot randomised trial against a low-fat diet found equivalent weight loss and no between-group superiority, and a sports-nutrition position stand concludes that adherence and energy deficit — not carbohydrate restriction as such — drive changes in body composition.

So the honest description is a diet that moves some markers favourably and others unfavourably, comparably to other patterns people can stick to. That is a trade-off profile. “Not effective” is a different statement, and it is the wrong one. The concerns raised on the supporting side are real — LDL-C rising by 18 to 70 mg/dL in normal-weight adults, and fibre and micronutrient shortfalls in strictly animal-based patterns — but a diet with a cost is not a diet with no effect.

Thresholds

Fails. Evidence: high.

No primary source in this dossier reports 40 g/day as a boundary. Not one.

The canonical source on these numbers is the 2019 Lancet series of systematic reviews and dose-response meta-analyses: 185 prospective cohorts, roughly 135 million person-years, 58 randomised trials, with robustness tested through sensitivity analyses, meta-regression and dose-response modelling, and certainty graded with GRADE. It reports 15–30% lower all-cause and cardiovascular mortality for the highest versus the lowest intake, and it locates the greatest risk reduction between 25 and 29 g a day, with the curves suggesting — not demonstrating — further benefit above that range.

Nothing in it designates 40. The 40 appears to be a reading of an upward-sloping curve continued past where the data support it: the number is an extrapolation, not a finding.

Every independent quantification sits lower:

• 35 g/day (or an increase of 15 g/day) is the target in the diabetes fibre meta-analysis;
• ≥35 g/day is what a PROSPERO-registered meta-analysis of 35 randomised trials in type 1 diabetes actually tested, finding HbA1c −0.46% and hypoglycaemia −0.81 episodes per month;
• 7.6–8.3 g/day of soluble fibre is the modelled optimum in a dose-response analysis of 29 randomised trials in type 2 diabetes;
• 26.3 g/day is the inflection point of a U-shaped mortality curve in an NHANES spline analysis of 39,184 adults, above which risk begins to rise again.

And a dose-response meta-analysis of 15 cohorts finds a smooth linear relation — RR 0.91 per 10 g/day for cardiovascular disease. A straight line is the signature of a continuous variable. It is what a quantity looks like when it has no threshold anywhere in it, at 40 or at any other number.

A multi-society joint advisory and a major dietary policy review decline to set any numeric fibre cut-off at all, judging foods and patterns rather than nutrient boundaries — which is the professional way of saying the same thing.

The claim overshoots the canonical estimate by roughly 40%, and it converts a gradient into a gate.

Causal direction

Partial. Evidence: moderate.

For fibre, a substantial part of the association survives designs built to exclude reverse causation and confounding. Randomised increases in fibre lowered systolic and diastolic blood pressure (−4.3 and −3.1 mmHg) at high GRADE certainty. A blinded randomised crossover trial found that a high-fermentable-fibre diet raised faecal and plasma short-chain fatty acids (P = 0.02) and measurably shifted circulating B- and T-cell populations — fermentation altering immune markers, under randomisation. And two-sample Mendelian randomisation separates the food types: genetically predicted whole-grain-type cereal intake is protective for coronary heart disease, myocardial infarction, heart failure and stroke, while refined cereal products raise risk. That last result matters because it is the whole-food half of the claim tested by a design that does not depend on who chooses to eat what.

The counterweights are real and we do not discount them. A second Mendelian randomisation across 83 dietary habits produced a biologically implausible result — cooked vegetables as the strongest risk factor — which is a warning about the method’s reach. An acute crossover trial found that 15 g of added fibre raised GLP-1 and PYY without changing appetite or energy intake at all: a marker moved, a behaviour did not. And a methodological review documents healthy-user bias, collinearity and exposure misclassification running through nutritional cohorts generally.

On the ketogenic side causal attribution is weaker in both directions, which is the point people on each side tend to skip. Reviews concede that improvements in insulin sensitivity cannot be separated from the accompanying weight loss. No controlled trial supports keto-mediated beta-cell preservation in type 1 diabetes; the evidence there is six case reports. And the strongest datum offered against the lipid worry — the KETO-CTA imaging cohort, where in 100 lean, metabolically healthy people on a ketogenic diet with very high ApoB, Bayesian inference favoured no association between ApoB and one-year non-calcified plaque progression by a factor of 6 to 10 — is a single-arm, one-year, hypothesis-generating study in a selected phenotype, set against LDL-C’s established cumulative causal role in atherosclerosis. It is genuinely informative. It is not a resolution. A single case report of seven years at very high LDL-C without plaque is not one either.

The adjudication

Take the sentence apart and each piece gets a different answer.

The direction is right. Whole-food, fibre-rich eating improves health markers, in randomised trials, partly independently of everything else that tends to travel with it. That is as well established as nutrition gets, and the Mendelian randomisation on cereal type suggests it is not merely the healthy-user effect.

The number is not. It is the only part of the claim stated with complete precision, and it is the part with no source. The canonical dose-response analysis puts the greatest gain at 25 to 29 g. Every independent quantification lands lower than 40. The relation is a straight line. A gradient has been rewritten as a threshold, and the threshold has been placed about 40% above the best available estimate.

And the confident half is the false one. “The ketogenic diet is not effective” is contradicted by 53 randomised trials that show it moving markers in both directions — favourably for triglycerides and HDL, unfavourably for LDL and total cholesterol, with heterogeneity high throughout (I² 60–93%) and no evidence of publication bias (Egger’s p > 0.05 for every outcome). The defensible criticism of ketogenic diets is about an unresolved lipid signal, long-term adherence, and fibre and micronutrient adequacy. That criticism is available, it is serious, and the claim does not make it.

There is a pattern in which half of a claim fails, and it is not random. The part that survived the journey through four comment sections is the part that is short, numeric and absolute. The parts that did not survive are the ones that limit it: the caution attached to the number, the sentence saying a cholesterol rise does not make a diet unsafe, the fact that a criticism was aimed at one study. Precision travelled. Qualification did not.

Verdict: partial. The direction of effect is sound and partly causal. The 40 g threshold is not reported by any primary source and overshoots the canonical estimate. The dismissal of ketogenic diets is contradicted by randomised evidence. And none of it was said by the four people whose comment sections carried it.

Adjudicated against 55 primary sources. Circulation measured from our own collection on 3 September 2026. Sources, study designs and DOIs are listed below in order of evidential hierarchy.