Design: randomised, controlled, blinded, two-centre superiority trial
Population: 320 adults undergoing elective cardiac surgery with cardiopulmonary bypass (159 opioid-free, 161 sufentanil)
Primary outcome: composite of 30-day neurological, respiratory, cardiovascular or renal complications, or death
Effect: 75.4% versus 84.5%, risk ratio 0.90 (95% CI 0.80 to 0.99), fragility index 1; digestive complications 2.5% versus 11.2% (risk ratio 0.22, 95% CI 0.08 to 0.65)

Three hundred and twenty adults having elective cardiac surgery with cardiopulmonary bypass at two French university hospitals were randomised, blinded, to opioid-free anaesthesia — intravenous ketamine, dexamethasone, lidocaine and magnesium sulfate — or to sufentanil. The primary outcome was a 30-day composite of neurological, respiratory, cardiovascular or renal complications, or death.

The composite occurred in 120 of 159 (75.4%) with opioid-free anaesthesia against 136 of 161 (84.5%) with sufentanil, risk ratio 0.90 (95% CI 0.80 to 0.99; P=0.049). The authors report the fragility index as 1 — one patient changing group would have removed significance. Individual components: cardiovascular complications 64.2% against 75.2% (risk ratio 0.86, 95% CI 0.74 to 0.99), driven by postoperative myocardial damage (61.6% against 72.1%); digestive complications 2.5% against 11.2% (risk ratio 0.22, 95% CI 0.08 to 0.65). No deaths in the opioid-free group against six in the control group (P=0.014).

The digestive result is the most believable. A risk ratio of 0.22 with an interval well clear of 1 for a plausible mechanism — less opioid, less ileus — is the kind of finding that survives replication. The composite is not: a fragility index of 1 on a P value of 0.049 means the primary result should be read as a hypothesis, and the authors say so themselves.

The mortality difference is where discipline is needed. Zero against six deaths is arresting and it was not the primary outcome, in a trial of 320 with no stated plan to detect it. It is the sort of finding that has repeatedly failed to replicate in perioperative medicine. What the trial supports is running the confirmatory trial, and in the meantime taking the ileus benefit seriously.

Do not switch a cardiac list to opioid-free anaesthesia on this trial alone
Consider the opioid-sparing components where postoperative ileus is a recurrent problem in your practice
Note the regimen tested was a specific four-drug combination, not opioid avoidance in general
Watch for the confirmatory trial before quoting the mortality difference to anyone
Account for ketamine's emergence effects and lidocaine's toxicity ceiling if adopting any part of it

Don't overread it

A fragility index of 1 means the primary result turns on one patient — the authors themselves call it hypothesis-generating, and the mortality difference is not a finding.

The statistics, in plain English

A fragility index of 1 is the key number in this paper: reclassify a single patient and P=0.049 becomes non-significant. A composite of four organ systems occurring in three-quarters of both groups is also easy to move, because it is driven by the commonest and least severe components — here, biochemical myocardial damage. The digestive complication result, risk ratio 0.22 with an interval of 0.08 to 0.65, is the one whose confidence interval is nowhere near 1. Six deaths against none is a secondary comparison on tiny numbers, and P=0.014 in that context is not evidence of a mortality benefit.