A 'pro-inflammatory' diet went with bowel leakage in the US, England and China. Then the researchers looked at people who had no leakage yet, and the diet score did not predict who went on to get it

Faecal incontinence, the accidental leakage of mucus or stool, is one of the health problems people are least likely to mention to a doctor. Diet often comes up when it is discussed, and a score called the Dietary Inflammatory Index has been at the centre of several studies. The index adds up how much of 25 or so nutrients a person eats, from fibre, magnesium and vitamin C to saturated fat and total calories, and weights each by how it has moved blood markers of inflammation in published research. A higher score means a diet that leans 'pro-inflammatory'. Earlier American studies found that people with higher scores were more likely to report leakage.

A paper published on 9 October 2026 in Food Science & Nutrition, by researchers at a hospital in Huai'an, China, set out to test that link more thoroughly, in three countries and over time. It used three large surveys: the US National Health and Nutrition Examination Survey for 2005 to 2010 (13,385 adults), the English Longitudinal Study of Ageing (4,141 people, mostly over 50) and the Chinese Longitudinal Healthy Longevity Survey (about 16,000 people aged 65 and over in 2008).

When diet and symptoms were measured at the same time, the link showed up in all three. In the US data, about one adult in twelve reported leaking mucus or stool at least once a month. Each one-point rise in the inflammatory score went with about 7.5% higher odds of leakage, and the quarter of people with the highest scores had about 1.5 times the odds of the lowest quarter. The result was much the same in each of the three survey periods. In England, using a slimmed-down 15-nutrient version of the score so it could be compared, the odds rose by about 14% per point. In China, where only a simple food-frequency score was available and the question covered bladder and bowel control together, they rose by 44% for each step of one standard deviation.

Then came the test that matters for anyone hoping diet could prevent the problem. In England and China, the researchers took the people who had no leakage at the start and checked who developed it over the following years. In England 430 of 2,837 people did; in China 920 of 8,021. Their starting diet score made no clear difference: the odds ratios were 0.97 and 1.04, both with ranges that comfortably include 'no effect'. The authors write that baseline scores 'showed no clear associations with incident outcomes'.

The most likely reading, which the authors raise themselves, is that the arrow can point the other way. People who already have leakage often change what they eat to cope with it. In interviews with older American women with incontinence, they described skipping certain foods, changing how food was cooked and eating smaller or fewer meals. Cutting back on fruit, vegetables or dairy would push an inflammatory score up, so a diet recorded after the symptoms began may partly be a result of them rather than a cause. The link was also fragile in places: in England it weakened to the edge of chance once the sample was weighted to represent the population, and in China it shrank and lost statistical significance once the analysis allowed for memory and thinking problems, which affect both what people eat and whether they can get to a toilet in time.

Some details narrow what the study can say. Of the three kinds of leakage asked about in the US survey, only liquid stool held up as linked to the score after correcting for multiple comparisons; mucus and solid stool did not. The American diet data are 15 to 20 years old, because the bowel questions were only asked between 2005 and 2010. The Chinese score was built on assumptions that the authors flag as debatable, such as counting eggs, milk and tea as anti-inflammatory. And the study measured no inflammation at all: the index is a way of scoring a diet, not a test of what is happening in the gut. The authors disclose that ChatGPT and Codex helped with writing and with reviewing the analysis code during revision, and that this AI-assisted code review changed the analytic sample and several of the reported estimates from the version first submitted.

What does help with leakage has been tested more directly. In a 2014 randomised trial from the University of Minnesota, 189 people with loose or liquid stool leakage took 16 g a day of one of three fibre supplements or a dummy powder for 32 days. Psyllium, the husk sold in India as isabgol, cut episodes to about 2.5 a week, against 5.5 on the dummy powder. Gum arabic came in at about 4.3 a week, and carboxymethylcellulose, a cellulose-based fibre, made leakage more frequent at 6.2. Only psyllium formed a gel in the stool. The new paper cites that trial as a reminder that the type of fibre matters more than a general score.

The study's own conclusion is careful: its findings 'do not establish that lowering DII improves continence'. Eating plenty of vegetables, fruit, pulses and whole grains has many good reasons behind it, but on this evidence, preventing bowel leakage is not yet one of them. Anyone who has leakage, especially if it is new, should see a doctor: it has many causes, some of them treatable, and it is worth checking before changing diet or starting supplements.