The part of a plant you can't digest. What it does depends on the kind: some fiber forms a gel, some feeds your gut bacteria, and some just adds bulk.
Mechanisms, studies and evidence grades for each claim above. This page began from a thread on X that splits fiber into four types; its claims are checked below.
The usual split is soluble versus insoluble. It is the least useful way to sort fiber, because it says nothing about the two properties that predict effects in people:
The two properties don't go together. Psyllium is soluble and makes a strong gel, but bacteria barely ferment it.6 Inulin is soluble and ferments quickly, but it forms no gel, so it doesn't help LDL or constipation much and it does make gas. "Soluble fiber feeds good bacteria and lowers cholesterol" packs two different fibers into one sentence.
| Fiber | Gel? | Fermented? | What it's good for | Where you find it |
|---|---|---|---|---|
| Psyllium | Strong | Barely | LDL, blood sugar, constipation, IBS | Husk supplement |
| Beta-glucan | Yes | Yes | LDL, blood sugar | Oats, barley |
| Pectin | Yes | Yes | LDL | Apples, citrus, berries |
| Inulin / fructans | No | Quickly | Butyrate (and gas) | Onion, garlic, asparagus, Jerusalem artichoke, chicory |
| Resistant starch | No | Slowly | Butyrate, maybe insulin sensitivity | Cooked-then-chilled rice and potatoes, green banana, raw potato starch |
| Wheat bran | No | Partly | Stool bulk | Bran, whole wheat |
| Cellulose | No | Barely | Stool bulk | Vegetable skins, stalks |
A gel-forming fiber binds bile acids in the gut, and they leave in the stool instead of being reabsorbed. The liver replaces them from cholesterol, which means pulling more LDL from the blood.
A 1999 meta-analysis of 67 trials found that each gram of soluble fiber from oats, psyllium or pectin lowered LDL by about 2 mg/dL, and the three sources worked about equally well.2 The authors called the effect "small within the practical range of intake": three bowls of oatmeal lowers LDL by about 5 mg/dL. Psyllium is where the dose becomes useful. A 2018 meta-analysis of 28 trials (1,924 people) found that about 10 g a day lowered LDL by 13 mg/dL and also lowered non-HDL cholesterol and apoB. The apoB result was graded high quality.3 (Strong)
This is one of the few dietary levers that moves apoB itself. The Statins & LDL paper argues that apoB counts, even though LDL-C is a flawed way to make decisions.
↑ back to summaryAcross seven trials in chronic constipation, 77% of people who took fiber responded, against 44% on placebo. Stools were more frequent and softer, and there was more gas. The authors rated the evidence as low quality.4 Psyllium's gel holds water, so it softens hard stool and firms loose stool. Coarse bran works differently: it irritates the colon lining into secreting water and mucus.
A viscous gel slows how quickly the stomach empties and how quickly glucose reaches the gut wall. A meta-analysis of 35 psyllium trials found the effect grows with how poor glucose control already is. It did nothing measurable in people with normal blood sugar and a little in prediabetes. In treated type 2 diabetes, fasting glucose fell by 37 mg/dL and HbA1c by about 1 point when psyllium was taken before meals.6 (Moderate)
In a 12-week Dutch primary-care trial, 275 people with IBS took 10 g a day of psyllium, bran, or rice-flour placebo. Psyllium beat placebo in months one and two (57% vs 35% responders). Bran only helped in month three, and that result didn't survive the worst-case analysis. People in the bran group dropped out earliest, mostly because their IBS got worse.7 A meta-analysis of 14 trials agreed: soluble fiber helped (for every 7 people treated, 1 more got relief), and bran did not.8
What makes IBS worse is fermentable fiber. The low-FODMAP diet removes fructans (onion, garlic, wheat), galacto-oligosaccharides (beans) and some sugars, and it reduced IBS symptoms in a controlled crossover trial.9 It cuts out specific fermentable carbohydrates, not fiber overall.
↑ back to summaryFermentable fiber and resistant starch reach the colon undigested, and bacteria there turn them into short-chain fatty acids: acetate, propionate and butyrate. Butyrate is the main fuel for colon cells. That much is textbook.
What's thinner is the jump from "makes butyrate" to "lowers inflammation" in people. In a 17-week Stanford trial, a high-fiber diet increased the bacteria's fiber-digesting enzymes, but microbial diversity didn't change and the main immune score didn't move. The fermented-food arm (yogurt, kefir, kimchi, kombucha) was the one that raised diversity and lowered inflammatory markers.10 On the high-fiber diet, what happened depended on the bacteria people started with, and those with the least diverse microbiomes did worst. If your gut has lost the bacteria that ferment a fiber, eating more of that fiber doesn't bring them back. (Moderate for butyrate production; Early for downstream health effects)
↑ back to summaryCooling works, but it adds little. When cooked starch cools, some of it recrystallizes (retrogradation) into a form that digestive enzymes can't break down. In an Indonesian study, freshly cooked white rice had 0.64 g resistant starch per 100 g. After 10 hours at room temperature it had 1.30 g, and after 24 hours in the fridge then reheated it had 1.65 g. The chilled-and-reheated rice produced a glucose curve about 18% lower in 15 healthy adults.11 Reheating did not undo it. A cup of rice cooked the day before gives you about 2–3 g of resistant starch, roughly 2 g more than fresh rice.
The insulin trials used ten times that. Two small trials found better insulin sensitivity with 15–30 g a day of high-amylose maize starch, a supplement. An Oxford study of 10 healthy people used 30 g a day for four weeks and measured with the gold-standard clamp.12 In a crossover of 33 people with large waists, men improved and women didn't. That trial was run by a contract research firm funded by National Starch, which makes the product, and two authors were National Starch employees.13 Leftover rice won't get you there. Unheated raw potato starch comes closer, at roughly 8 g of resistant starch per tablespoon stirred into something cold. Heating turns it back into ordinary starch.
↑ back to summaryA 2019 Lancet set of meta-analyses (185 cohort studies, about 135 million person-years) found 15–30% lower all-cause and cardiovascular mortality, coronary disease, stroke, type 2 diabetes and colorectal cancer in the people who ate the most fiber compared with those who ate the least. The benefit was clearest at 25–29 g a day.1 The authors graded the evidence moderate. Those numbers come from cohort studies: people who eat 30 g of fiber also eat fewer processed foods, and adjusting for that is never perfect.
When fiber was tested head-to-head against colon cancer's precursor, it failed. Two NEJM trials published together in 2000 found no effect on adenoma recurrence. One gave 13.5 g a day of wheat-bran fiber to 1,429 people for about three years.14 The other used a low-fat, high-fiber, high-produce diet in 2,079 people for four years, and 39.7% versus 39.5% had a recurrent adenoma.15 Adenoma recurrence over three or four years isn't cancer over thirty, but those were the tests available, and fiber didn't pass them. (Moderate for the association; Contested as cause)
↑ back to summaryA 12-post thread from 30 September 2026 by Cooking with Chris (@coookwithchris, 1.4M followers, "I fix gut issues, skin problems & stubborn weight"). It ends by selling gut-healing protocols built around the buyer's stool test. The first post's image is a stock photo of fiber foods. Each claim was checked against the primary papers.
| Claim | What the literature shows | Does it hold up? |
|---|---|---|
| "There are 4 different types of fiber": soluble, insoluble, fermentable, resistant starch | These aren't four separate groups. Soluble/insoluble is one property and fermentable is another, and resistant starch is itself a fermentable fiber. Psyllium is soluble but barely fermented; inulin is soluble and highly fermented.6 | A muddled map. Gel-forming versus fermentable predicts effects better (table above). |
| Soluble fiber "slows digestion, softens stool, feeds good bacteria, and helps stabilize blood sugar"; best for "high cholesterol" | The LDL and blood-sugar effects are real and come from gel-formers: psyllium, oat beta-glucan, pectin.2,3,6 The best of them, psyllium, is the one bacteria barely eat. | Mostly holds. "Raw carrots" as a lead source is odd; you'd need a lot of them to reach the gram doses in the LDL trials. |
| Insoluble fiber is "best for constipation"; sources include "wheat bran, nuts, seeds" | For IBS, bran lost to psyllium and made some people worse.7,8 The constipation trials that worked mostly used psyllium.4 | Backwards on the details. Soluble gel-formers have the better constipation record. (This site doesn't recommend nuts anyway: see Linoleic Acid.) |
| Fermentable fiber makes short-chain fatty acids "that reduce inflammation, support immunity, and protect the gut lining"; best sources garlic, onions, asparagus, Jerusalem artichokes, green bananas | Fermentation into butyrate is real. In the one careful 17-week human trial, a high-fiber diet didn't lower inflammatory markers or raise diversity on average; fermented foods did.10 Four of the five sources are fructans, which are the high-FODMAP foods cut in IBS.9 | The mechanism is right; the human outcomes are thin. And the thread lists these foods without saying they're the classic IBS triggers. |
| Resistant starch "forms when you cook starchy foods and then let them cool. Potatoes, rice, and oats" | True. Rice went from 0.64 to 1.65 g per 100 g after a night in the fridge, and stayed there when reheated.11 | Holds, but the amount is small: a gram or two per serving. |
| Resistant starch is best for "rebuilding the microbiome after antibiotics" | I found no human trial of resistant starch or any other fiber after antibiotics. The closest human trial tested probiotics, which slowed recovery.20 | No evidence. |
| Resistant starch is best for "improving insulin sensitivity" | Two small trials, at 15–30 g a day of maize starch. In one the effect was in men only, and that trial was funded by the manufacturer.12,13 | Early, and at roughly ten times the dose that leftovers provide. |
| "Even GI doctors mess up": they add fiber when the gut is inflamed; with IBS or IBD "too much fiber too fast will make things worse" | "Too fast" is fair: more fiber means more gas at first.4 But for IBS, soluble fiber is the evidence-based first step,8 and in Crohn's without a stricture a normal-fiber diet did no worse than low-residue over 29 months.16 | Half right. Adding fiber slowly is good advice. Treating IBS or IBD as a reason to avoid fiber is not. |
| "Compromised gut (IBS, IBD, post-infection): drop below 10g/day" | No guideline or trial I found supports a blanket cut to 10 g. The evidence points the other way for IBS (psyllium helps),7,8 and a high-fiber diet was well tolerated in UC in remission.17 The low-FODMAP diet targets specific carbohydrates, not total fiber.9 | Not supported. The exception is a Crohn's stricture, which the thread doesn't mention. |
| "Healthy gut: aim for 25–38g/day" | Matches the US adequate intake (14 g per 1,000 kcal) and the Lancet sweet spot of 25–29 g.1 | Holds. |
| With SIBO, fermentable fibers "feed the wrong bacteria… This is why you test before you supplement" | Plausible, and low-fermentation diets are used in SIBO. SIBO is diagnosed with a breath test or a small-bowel sample.18,19 | Reasonable, but the "test" he sells can't diagnose SIBO (next row). |
| "I build protocols around your actual stool test results. This removes any guesswork" | A stool sample reflects the colon, not the small intestine, so it can't find SIBO. A 2025 international expert consensus in Lancet Gastroenterology & Hepatology says direct-to-consumer microbiome tests are sold "without any consensus on their regulation or any proven value in clinical practice."21 | The sales pitch doesn't hold. No validated method turns a consumer stool test into a fiber prescription. This is the product the thread exists to sell. |
Verdict on the thread: the practical advice for a healthy person (eat 25–38 g, add it slowly, chill your starches) is fine. The sorting into four types is confused. The advice for sick guts contradicts the IBS trials, and it builds toward a stool test that can't answer the questions the thread raises.
↑ back to summary