Design: Multicentre, double-blind, placebo-controlled randomised trial
Population: 1,210 adults having elective curative surgery for digestive cancer, 23 French centres
Primary outcome: Major postoperative complications within 30 days
Effect: 23% (139/594) methylprednisolone vs 20% (119/594) placebo; P = .16

This double-blind, placebo-controlled trial across 23 French centres randomised 1,210 patients having curative elective surgery for digestive cancer with an anastomosis to methylprednisolone 20 mg/kg intravenously at induction or placebo. The idea was that damping the inflammatory surge would cut complications.

It did not. Major complications within 30 days occurred in 23% with methylprednisolone and 20% with placebo (139 vs 119 of 594 each; P = .16). Postoperative infections (31% vs 30%), intra-abdominal infections (24% vs 21%), unsatisfactory wound healing (15% vs 12%) and length of stay (about 13 days in both) did not differ either. The authors conclude it should not be used routinely.

This is a high-dose pulse, not the single low dose of dexamethasone many anaesthetists give for nausea, which the trial did not test. The numerically higher complication and wound figures in the steroid arm were not significant, but they point the same direction and give no reason to try the strategy outside a trial.

Do not add a pulse dose of methylprednisolone to digestive cancer surgery hoping to cut complications.
Keep dexamethasone for nausea prophylaxis a separate decision; this trial did not test it.
Put effort into measures with evidence behind them: prehabilitation, ERAS elements and infection prevention bundles.
Expect about one in five patients to have a major complication after these operations, and plan monitoring accordingly.

Why it matters

It closes a long-running question about steroid pulses in major gastrointestinal surgery with a large, blinded trial.

The statistics, in plain English

Complications were 23% versus 20%; the P value of .16 means a difference this size could easily arise by chance, so there is no evidence of benefit, and if anything the trend favoured placebo. With nearly 600 patients per arm, a meaningful benefit would probably have shown.