A three-year-old girl in western Japan spent months on antibiotics that worked and then stopped working. What finally broke the cycle was 150 cubic centimetres of liquid made from her father's stool.

Kagawa University Hospital announced the case on July 2, 2026, describing it as the first fecal transplant performed in Japan on a child with recurrent Clostridioides difficile colitis. A year on, the hospital reported, the girl had suffered no relapse and no complications.

A father's stool, filtered through gauze

C. diff is a gut bacterium that flourishes when antibiotics clear out its competition. It causes severe diarrhoea, and the standard treatment is more antibiotics, which is where the trap closes. Each course knocks back the infection and also knocks back the microbes that would otherwise hold it down.

Roughly 20% of patients relapse after a first course of treatment, the hospital says, and once someone has relapsed even once, the chance of relapsing again runs to about 60%. The girl had been round that loop several times. Every time doctors stopped the antibiotics, the diarrhoea came back.

In February 2025, after the hospital's ethics committee signed off and the family gave consent, doctors collected stool from her father, who was in good health. They dissolved it in saline, filtered it through sterile gauze, and adjusted the volume to 150 cc. A colonoscope then delivered it directly into her large intestine.

The hospital reports that her gut bacteria diversified quickly and shifted toward a profile resembling her father's. Local broadcaster KSB reported the improvement was visible within two or three days. Later, when she needed antibiotics for an unrelated infection, the transplanted community recovered on its own.

Why a three-year-old's gut was this fragile

The girl had been born extremely underweight and spent her first stretch of life in intensive care. She had heart and lung complications, and because of them she was fed through a nasal tube for a long period.

Assistant Professor Takeo Kondo of Kagawa University's pediatrics department pointed to that combination as a likely cause. Antibiotics after birth, plus months of nutritionally uniform formula going straight into her stomach, may have left her gut with too little microbial variety to defend itself.

Feeding tubes and antibiotics are what kept her alive. They may also have set up the infection that would not leave.

What fecal transplants are actually proven to do

Fecal microbiota transplantation, or FMT, sounds crude, and mechanically it is. For recurrent C. diff, though, it has an unusually strong record. The logic is simple: instead of killing an overgrown bacterium, you reintroduce a working ecosystem that crowds it out.

Kagawa University cites cure rates above 90% in adults. In children the picture is similar. A retrospective study across 18 pediatric centres in the United States covered 372 patients ranging from 11 months to 23 years old. Of the 335 followed for two months, 81% had no recurrence after a single transplant, rising to roughly 87% once repeat transplants were counted.

That record does not carry over. The American Gastroenterological Association's 2024 guideline backs fecal microbiota therapies for recurrent C. diff. For ulcerative colitis, Crohn's disease, pouchitis and irritable bowel syndrome it advises against conventional FMT outside clinical trials, and says plainly that FMT cannot yet be recommended for other gut conditions. Even within C. diff, success rates fall in children with serious underlying illness, which is the category the Kagawa patient fell into.

Japan waited a long time

Japan is not new to fecal transplants. Juntendo University has run FMT clinical research since 2014 and reported more than 240 cases as of August 2024. Keio University has used the procedure for ulcerative colitis and intestinal Behçet's disease.

Nor is Kagawa the first Japanese team to give a transplant to a child. A registered clinical study run through a group of private clinics gave FMT to children aged 5 to 12 with autism spectrum disorder and published its results in Frontiers in Pediatrics in March 2026. The doctors chosen to run it qualified precisely because they had already been giving children the procedure as private, out-of-pocket care. What Kagawa can claim is narrower: the first Japanese child treated for the condition where the evidence is strongest.

That gap is a funding gap. FMT is not reimbursed under national health insurance, so it happens either inside research protocols or as private care. Hospital case numbers stay low, and low case numbers keep the procedure from becoming routine, while the out-of-pocket market grows around conditions with far thinner evidence. Kondo noted his hospital sees only two or three C. diff colitis cases a year, and that ethical review means not every institution can offer a transplant even when one is warranted.

He also said that after presenting the case at two pediatric conferences, in September 2025 and April 2026, other doctors came to him with the same message: they had patients like this, no options left, and wanted to know exactly how it had been done.

America closed the door it opened

The United States built the infrastructure that made pediatric FMT routine. OpenBiome, a non-profit stool bank founded in 2012, eventually supplied more than 1,300 hospitals and clinics and shipped 72,507 treatments. It operated under a policy of enforcement discretion: the Food and Drug Administration classified FMT as a drug, but chose not to enforce the paperwork for C. diff patients who had run out of options. In June 2019 the FDA warned that two immunocompromised adults had developed invasive infections traced to donor stool that was not tested for drug-resistant E. coli before use, and that one of them had died.

Then the FDA approved two commercial products, Rebyota in November 2022 and Vowst in April 2023, and ended that discretion. OpenBiome shipped its final clinical treatment in December 2024 and reorganised as a foundation.

Both approved products are licensed only for people aged 18 and over, and Rebyota's labelling states it has not been studied in patients under 18. American children with recurrent C. diff therefore sit in a gap: the conventional transplant that worked for them has become far harder to obtain, and the products that replaced it are not approved for them. By its own account, OpenBiome had flagged exactly this risk before it wound down distribution.

Europe is writing the rules for 2027

The European Union has taken a third path. Regulation 2024/1938, covering substances of human origin, brings intestinal microbiota into its scope from August 2027, setting shared standards for donor screening, traceability and follow-up.

Until then the picture is patchy. The European Medicines Agency notes there is no agreed EU-wide classification for FMT, which leaves member states to treat it variously as a medicinal product, as tissue and cells, or as a therapeutic procedure. Rather than forcing the practice to be either unsupervised or a commercial drug, the EU wants to regulate the material itself. Neither Japan nor the United States has managed that.

Kondo says he wants to accumulate cases and establish FMT as a standard option for children. One success in Kagawa will not get him there.

Every country has landed somewhere different on the same awkward question: when the treatment is another person rather than a product, who gets to give it, and to whom? Where has your country landed?

参照