Antibiotics in pouchitis bought six weeks, and traded resistance for them

Design

prospective longitudinal cohort with metagenomic and whole-genome sequencing over one year

Population

21 patients with active pouchitis after ileal pouch-anal anastomosis, 130 faecal samples

Primary outcome

microbial and clinical change after two weeks of metronidazole with ciprofloxacin or doxycycline

Effect

faecal calprotectin 728 to 265 µg/g (p<0.05), rebounding by 6 weeks; resistance gene abundance rose without added diversity

Antibiotics are first-line for pouchitis and they work, which has largely ended the enquiry into why relapse is so common. This prospective study recruited 21 patients after ileal pouch-anal anastomosis during a clinical flare, treated them with two weeks of metronidazole plus either ciprofloxacin or doxycycline, and followed them for a year, collecting 130 faecal samples across flares, recovery and relapse. Metagenomic sequencing was combined with whole-genome sequencing of Escherichia coli isolates to give strain-level virulence and resistance profiles.

Both regimens produced rapid clinical improvement, with faecal calprotectin falling from 728 to 265 micrograms per gram and a marked reduction in bacterial exotoxin genes. Both measures rebounded by six weeks after treatment stopped. Meanwhile antibiotic resistance gene abundance rose significantly during treatment without any expansion in resistance gene diversity — meaning the resistant strains already present had expanded, rather than new resistance arriving.

The mechanism this describes is a temporary suppression of exotoxin-producing bacteria that lets resistant, low-virulence strains dominate for a while. It explains the clinical pattern every pouch clinic sees: reliable short-term response, reliable relapse, and a gradual narrowing of which antibiotic still works. It does not give you an alternative, and 21 patients is a small series — but it does argue for treating the antibiotic course as a suppression measure with a known expiry, rather than as a treatment that failed when symptoms return.

• Set the expectation at the start that antibiotic remission in pouchitis is typically temporary
• Measure faecal calprotectin before and after — the fall and the rebound are both informative
• Record which agent was used each course; resistance here comes from expansion of strains already present
• Avoid reflexively extending or repeating courses to chase a durable remission
• Note this is 21 patients — the mechanism is better supported than any treatment implication

Why it matters

It explains why relapse after antibiotic remission is the rule, and why the same drugs stop working over time.

Don't overread it

Twenty-one patients with mechanistic sequencing — this characterises what happens, it does not test any alternative strategy.

The statistics, in plain English

The calprotectin fall from 728 to 265 micrograms per gram is a large drop, but note it does not reach the range usually taken as remission, and it is a group mean across a small cohort. The key observation is not any single p value but the shape: two independent measures, calprotectin and exotoxin gene abundance, falling together and rebounding together by six weeks. That resistance gene abundance rose without diversity rising is a clean finding — it distinguishes selection of existing strains from acquisition of new resistance.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.