Design: international randomised trial, physiology-guided vs angiography-guided complete revascularisation
Population: 1,823 patients with STEMI and multivessel disease after successful culprit treatment; median age 66, 24% women
Primary outcome: composite of death, myocardial infarction, stroke or TIA, and ischaemia-driven revascularisation at median 17.9 months
Effect: 8.9% vs 13.7%, hazard ratio 0.62 (95% CI 0.47-0.83), P < 0.001; safety composite 4.6% vs 7.1%, HR 0.63 (0.43-0.93)

AIR-STEMI, reported on 29 August 2026, randomised 1,823 patients with STEMI and multivessel disease whose culprit lesion had been treated successfully to complete revascularisation guided either by functional coronary angiography or by conventional angiography. Median age was 66, and 24% were women. At a median 17.9 months, the composite of death, myocardial infarction, stroke or TIA and ischaemia-driven revascularisation occurred in 81 of 913 (8.9%) in the physiology-guided arm against 125 of 910 (13.7%) with angiography alone — hazard ratio 0.62 (95% CI 0.47-0.83, P < 0.001).

The safety result is what makes this unusual. Contrast-associated acute kidney injury or major bleeding occurred in 4.6% with physiology guidance against 7.1% without (HR 0.63, 95% CI 0.43-0.93, P = 0.02). A strategy that adds an assessment step reduced harm as well as events, which tells you the mechanism: measuring the non-culprit lesions means treating fewer of them, and the stents not placed are the benefit.

Completeness of revascularisation after STEMI is already recommended; what has been uncertain is how to choose which non-culprit lesions qualify. The visual estimate overtreats. Functional angiography is derived from the images already acquired rather than requiring a pressure wire, which matters where wire cost and procedure time are the constraint — and that makes this more portable into Indian cath labs than a wire-based strategy would be. If your lab treats angiographically severe-looking non-culprit lesions on sight, this is the trial that argues for a measured threshold instead.

Assess non-culprit lesions functionally before deciding to stent them, rather than by visual severity.
Expect to place fewer stents; the benefit runs partly through contrast and bleeding avoided.
Check whether your lab's angiography platform supports an image-derived physiology tool — no pressure wire is required.
Keep the staged-versus-immediate question separate; this trial addressed how to select lesions, not when to treat them.
Audit how many non-culprit lesions your lab currently treats on visual assessment alone — that is the size of the change.

Why it matters

The non-culprit lesion that looks severe on the screen is being stented on an impression that this trial shows is wrong often enough to cost outcomes.

The statistics, in plain English

A hazard ratio of 0.62 with a 95% CI of 0.47 to 0.83 means the benefit is clear — the interval sits well below 1.0 — and translates to about five fewer primary events per hundred patients over roughly 18 months. The primary outcome is a composite, and ischaemia-driven revascularisation is its softest component and the one most easily influenced in an open procedural trial; the safety composite moving in the same direction is what makes the result harder to explain away.