Design: Phase 3, randomised, double-blind, placebo (own-stool)-controlled
Population: 450 adults aged 18–65 with moderate to severe IBS
Primary outcome: ≥75-point fall in IBS-SSS at 90 days
Effect: 40% vs 38%; difference 1.9 points (95% CI −8.1 to 12.0)

REFIT2, published in The Lancet, randomised 450 adults aged 18–65 with moderate to severe IBS at five Norwegian hospitals to a single rectal enema of donor stool or their own stool. Patients, clinicians and assessors were masked.

At 90 days, 40% of the donor group and 38% of the placebo group had a meaningful fall in symptom score, a difference of 1.9 percentage points with an interval from −8.1 to 12.0. Adverse events were similar.

Earlier smaller trials had produced mixed results and some enthusiasm. This adequately sized, well-blinded trial suggests that, at least with a single enema, changing the microbiota does not improve IBS symptoms. The high placebo response is itself a reminder of how much IBS symptoms fluctuate.

FMT is not a treatment for IBS outside research. Stay with established approaches: explanation, diet, gut–brain therapies and symptom-directed drugs.

Do not refer IBS patients for faecal microbiota transplant outside a trial
Expect a placebo response of about a third to two-fifths in IBS, and use it in counselling
Use established options: explanation, diet such as a supervised low-FODMAP trial, and gut-directed psychological therapy
Keep FMT for its established indication, recurrent Clostridioides difficile infection

Why it matters

It closes off a treatment many patients with IBS have been asking about.

The statistics, in plain English

The interval from −8.1 to 12.0 points includes no effect and rules out a large benefit. The upper end of the interval, 12 points, is the largest benefit still compatible with the data.