The konjac fibre sold to fill you up was tested in 11 trials of people with diabetes or excess weight. Nobody lost weight, and the one good result, on fasting blood sugar, vanished when the maths changed

Konjac is a plant whose underground stem has been made into a firm, almost tasteless jelly in Japan and China for centuries. The useful part is a fibre called glucomannan. Stir a gram of it into water and it soaks up the liquid and turns thick and gluey, far more so than most fibres. That is the whole sales pitch: a gel that slows down how fast food leaves the stomach, so you feel full sooner and sugar from a meal reaches the blood more slowly. It is sold as capsules, powders and gummies, and the European Union allows it to carry two health claims, one on cholesterol and one on weight loss within a calorie-cut diet.

A team from Mashhad University of Medical Sciences in Iran and Saint Louis University in the United States set out to test those promises in the people they are mostly aimed at: adults with type 2 diabetes, insulin resistance, overweight or obesity. Their review was published on 8 October 2026 in Food Science & Nutrition. They searched three databases up to March 2025 and found 11 controlled trials that gave glucomannan on its own and compared it with a placebo. The trials were small, between 11 and 80 people each, about 350 people in all by our count from the paper's table. Most gave 3 g a day in capsules for 3 to 12 weeks. Two baked the fibre into biscuits and one into a gummy sweet. The oldest trial dates from 1984 and the newest two from 2023.

Start with weight, because that is what most people buy it for. Nine trials measured it, and the pooled result was that people taking glucomannan ended up 0.45 kg heavier than those on placebo, a figure so uncertain that it could just as well be 0.7 kg lighter or 1.6 kg heavier. In plain terms: no effect. The trials were also fairly consistent with each other on this, and the authors rate the evidence for weight as moderate certainty, the best grade anything in the review gets. BMI and body fat did not change either.

Cholesterol was the next hope, since an earlier review in 2008 and some later ones reported that glucomannan lowered LDL. Here LDL cholesterol came out 7.4 mg/dL (about 0.19 mmol/L) lower on average, but the range of likely values ran from a drop of 17.9 to a rise of 3.1, so the result was not statistically clear. The ten trials disagreed with each other a great deal. Triglycerides, HDL and both blood pressure numbers did not change overall either. The authors grade the evidence on LDL, total cholesterol and blood pressure as very low certainty.

That leaves fasting blood sugar, the one result the authors highlight. Across seven trials it was 7 mg/dL (about 0.39 mmol/L) lower with glucomannan, and it just crossed the line for statistical significance. Then the authors tested how solid it was. When they left out a single small trial from Taiwan, the drop shrank to 1.8 mg/dL and stopped being significant. When they re-ran the numbers with a different, more cautious assumption about the five crossover trials, where the same people took both the fibre and the placebo, it shrank to 3.5 mg/dL and again was not significant. They rate this evidence as low certainty and say in their conclusion that it should be read with caution. The same re-run, by the way, nudged diastolic blood pressure to a small rise that did reach significance, which nobody would take as a real harm on evidence this thin either.

A few more things limit what this review can tell us. Only one of the 11 trials was judged at low risk of bias; three were at high risk. The groups were mixed: people with diabetes, people with obesity and one trial in insulin resistance, with different doses and lengths. Some subgroups looked better, such as higher doses or trials of four weeks or less, but the authors themselves call these exploratory, and each rests on a handful of tiny studies. We also noticed that the text says doses ran from 3,000 to 10,000 mg a day, while the paper's own table lists one trial at 1,500 mg and the gummy trial at 500 mg a day for two weeks.

Our reading: if you have diabetes or are trying to lose weight, the controlled trials so far don't support glucomannan as a way to do either, and they are small and old. A gel-forming fibre can still make a meal more filling, and that may help some people eat less, which is roughly what the EU's weight claim says: it works, if at all, alongside a diet with fewer calories. What does have strong evidence for steadier blood sugar and weight is fibre from food, such as beans, lentils, oats, whole grains, vegetables and fruit. If you do try glucomannan, the EU requires a choking warning on it for a reason: take it with plenty of water, and not at all if you have trouble swallowing.