Design: Multicentre, double-blind, placebo-controlled randomised trial
Population: 1,210 adults having elective curative digestive cancer surgery with anastomosis, 23 French centres
Primary outcome: Major postoperative complications within 30 days
Effect: 23% methylprednisolone vs 20% placebo (P = .16)

This double-blind, placebo-controlled trial at 23 French reference centres randomised 1,210 adults having elective curative surgery for digestive cancer with an anastomosis to 20 mg/kg intravenous methylprednisolone at induction or an identical placebo. Mean age was 66.

Major complications within 30 days occurred in 23% with methylprednisolone and 20% with placebo (P = .16). None of the secondary outcomes moved in steroids' favour: postoperative infections 31% vs 30%, intra-abdominal infections 24% vs 21%, poor wound healing 15% vs 12%, and length of stay about 13 days in both arms.

The theory — damping the inflammatory surge to reduce complications — has driven years of mixed small studies. This adequately sized, blinded trial found no benefit, and the point estimates for infection and wound healing leaned slightly the wrong way. The authors advise against routine use, and nothing here supports adding it to an enhanced recovery pathway. This is distinct from low-dose dexamethasone for nausea, which the trial did not test.

Do not add high-dose methylprednisolone at induction to reduce complications after digestive cancer surgery.
Remove it from local enhanced recovery protocols where it was added on earlier, smaller studies.
Keep low-dose dexamethasone for nausea prophylaxis separate; this trial did not test it.
Note that infection and wound-healing figures leaned slightly against steroids, though not significantly.

Why it matters

It closes a long-running question with a blinded trial big enough to answer it.

The statistics, in plain English

A P value of .16 means the 3-point difference in major complications could easily be chance. Because the trial was large and blinded, a meaningful benefit is unlikely to have been missed. The differences that did appear pointed slightly towards harm rather than benefit.