Design: Methodological survey of guideline-cited randomised trials
Population: 161 superiority RCTs cited in perioperative guidelines, 2012-2022
Primary outcome: Fragility index of the supporting binary outcome
Effect: Median fragility index 4 (IQR 2-8); paediatric median 1

A methodological survey in Anesthesiology (25 September 2026) calculated the fragility index for 161 randomised superiority trials with binary outcomes that were cited in support of North American and European perioperative guidelines from 2012 to 2022.

The median trial had 120 participants and a fragility index of 4: changing the outcome in four patients would remove statistical significance. Paediatric trials were most fragile (median 1), cardiovascular trials least (median 6). Single-centre trials had lower fragility indices than multicentre ones.

This does not mean the recommendations are wrong, but it does mean many rest on results that a few events could overturn. It is a reason to read the trial behind a recommendation before treating it as settled, especially in paediatric practice.

Check the size and fragility of the trial behind any recommendation you rely on
Give more weight to multicentre trials than to single-centre ones with similar results
Be especially cautious with paediatric recommendations built on small trials
Use fragility alongside, not instead of, GRADE certainty

Why it matters

Guideline strength and the stability of the evidence behind it are not the same thing.

The statistics, in plain English

The fragility index is the number of patients whose outcome would need to change to make a significant result non-significant. A median of 4, and of 1 in paediatrics, means many results are numerically fragile. It is crude, though, and does not capture bias or effect size.