Subscribe free

General Surgery ·

Preventing Anastomotic Leak

The Leak You Can Prevent: 7 Moves That Work (and 5 That Don’t)

Most leaks are decided before the patient reaches the ward. The evidence finally tells us which decisions matter.

The Bottom Line:

Three measures have randomized evidence behind them: bowel preparation with oral antibiotics, ICG perfusion assessment, and a diverting stoma for the high-risk low anastomosis. Optimizing the patient and watching CRP add to them. Several popular habits, from the low IMA tie “for perfusion” to the transanal tube, do not reduce leaks at all.

The call every surgeon remembers

Day 4. Tachycardia. A CRP that should be falling but isn’t.

Then the CT confirms what you already knew.

An anastomotic leak turns a routine resection into weeks of drains, reoperations and, too often, a stoma that never goes away.

For years, prevention was a mix of habit and personal scars. “I always do it this way.”

That has changed. In the last two years, large trials and trial sequential analyses have separated what works from what only feels safe.

Here is what the evidence says, in the order you meet the patient.

7 moves that actually work

1. Prep the bowel, and add oral antibiotics. Mechanical preparation alone does not prevent leaks. Add oral antibiotics and it does: RR 0.60 in the Cochrane review, OR 0.45 across 9 RCTs in colorectal cancer (Willis et al., Cochrane 2023; Gosavi et al., Colorectal Dis 2025). In rectal surgery, the double-blind MOBILE2 trial cut leaks from 13.5% to 5.8% (Koskenvuo et al., JAMA Surg 2024).

2. Fix the patient before you build the anastomosis. The LekCheck study named six modifiable risk factors: anemia, hyperglycemia, hypothermia, mistimed antibiotics, vasopressors and epidurals. A bundle targeting them, mostly by fixing antibiotic timing and dropping epidurals, lowered leaks from 8.6% to 6.2% in a before–after design; the effect held for colonic, not rectal, anastomoses (de Wit et al., Ann Surg 2026). The catch: a checklist alone did not work. Without supervision and feedback, leaks stayed at 8.9% and complications rose (SmartCheck, Br J Surg 2025). Prehabilitation, IV iron and quitting smoking at least 4 weeks before surgery lower overall morbidity, but none has been proven to lower leaks.

3. Check perfusion with ICG. This one is now settled. Nine RCTs, 4,754 patients: RR 0.66, NNT 24, with high certainty and a conclusive trial sequential analysis. The benefit grows with BMI (Ryan et al., Lancet Gastroenterol Hepatol 2026). For low colorectal and coloanal anastomoses: RR 0.63 (Brucchi et al., Surgery 2026).

4. Test the anastomosis, and act on a positive test. The air-leak test has only two small RCTs behind it (OR 0.33). Its real value is what you do next: a simple suture repair of a bubbling anastomosis is linked to more leaks. Redo it or divert (Clifford et al., Colorectal Dis 2020).

5. Divert the high-risk low anastomosis. In the Matthiessen RCT, a defunctioning stoma cut symptomatic leaks from 28.0% to 10.3% and urgent reoperations from 25.4% to 8.6% (Ann Surg 2007). Across 9 RCTs: OR 0.36 (Emile et al., J Gastrointest Surg 2022). The price is real: a second operation and high output. But skipping the stoma doesn’t spare the patient one: at 42 months, 13.8% of diverted and 16.9% of non-diverted patients still had a stoma. Choose by risk: male, irradiated, low, obese, ASA ≥3.

6. Think twice about NSAIDs after a low anastomosis. The signal is observational but consistent: early non-selective NSAIDs carry an OR of about 1.7 for leak (Hu et al., Ann Surg Oncol 2026). No RCT exists. Until one does, avoid diclofenac and repeated non-selective doses after an undiverted low anastomosis (Kowal et al., ANZ J Surg 2026).

7. Use CRP to rule out, not to rule in. A CRP below 100 mg/L on day 2 has a negative predictive value of about 99% after colon cancer resection (Sinimäki et al., Ann Surg 2026). Above 100? Check again on day 3: if CRP is falling, the risk stays around 1%. With this two-step rule, nearly 9 in 10 patients could go home by day 2 or 3. A high value means CT, not relaparotomy: only about 1 in 12 patients above the cutoff actually leaked. After low anterior resection it is not validated: follow the trend and the patient.

5 habits that don’t hold up

  • Mechanical prep alone: no effect on leak. The benefit comes only with oral antibiotics (Willis et al., Cochrane 2023).
  • The low IMA tie “to save perfusion”: the newest RCT found symptomatic leaks in 4.9% vs 6.0%, RR 0.82 (0.31–2.13). Tie low for reach, not for leak (Kim et al., JAMA Surg 2026).
  • The transanal tube: in RCTs it did not significantly reduce leaks (OR 0.67, 0.42–1.05) and doubled bleeding (OR 2.36). It may soften a leak; it does not replace a stoma (Guo et al., Colorectal Dis 2022).
  • A bigger circular stapler: 31–33 mm vs 28–29 mm, OR 0.96. Size doesn’t matter here (Oppelt et al., Dis Colon Rectum 2026).
  • Goal-directed fluids to prevent leak: fewer complications overall (RR 0.84), no proven effect on leak. Avoid overload and unnecessary vasopressors instead (Xu et al., Int J Surg 2018).

Where do we go from here?

Leak prevention is no longer one trick. It’s a chain:

Prep. Optimize. Perfuse. Test. Protect. Watch.

The strongest links now have randomized proof. The weakest are the habits we never questioned.

The next challenge isn’t knowing what works. It’s making it stick. SmartCheck showed that a checklist on the wall changes nothing without a team that checks back.

No surgeon can promise an anastomosis won’t leak. The real question: did you do everything the evidence allows to stop it?

Further Reading:

  • Koskenvuo et al. JAMA Surg, 2024 (MOBILE2). Link
  • Willis et al. Cochrane Database Syst Rev, 2023. Link
  • Gosavi et al. Colorectal Dis, 2025. Link
  • de Wit et al. Ann Surg, 2026 (DoubleCheck). Link
  • SmartCheck study. Br J Surg, 2025. Link
  • Ryan et al. Lancet Gastroenterol Hepatol, 2026. Link
  • Brucchi et al. Surgery, 2026. Link
  • Clifford et al. Colorectal Dis, 2020. Link
  • Matthiessen et al. Ann Surg, 2007. Link
  • Emile et al. J Gastrointest Surg, 2022 Link
  • Hu et al. Ann Surg Oncol, 2026 Link
  • Kowal et al. ANZ J Surg, 2026 Link
  • Sinimäki et al. Ann Surg, 2026 Link
  • Kim et al. JAMA Surg, 2026. Link
  • Guo et al. Colorectal Dis, 2022 Link
  • Oppelt et al. Dis Colon Rectum, 2026 Link
  • Xu et al. Int J Surg, 2018 Link

Summaries are written to help you decide what to read. They do not replace the full paper.

Free · 5 minutes a day

Get the five that matter, every day.

The Daily Selection: the five most relevant papers of the day, each cut to its key points.

Specialty newsletters: every new paper in your field, as a list of titles linked to the source. Daily Selection: the five that matter, cut to their key points.

FreeNo spamOne-click unsubscribe

More in General Surgery

The latest stitches

All papers in General Surgery →