Complete Revascularization Guided by Functional Coronary Angiography in STEMI

Authors, Journal, Affiliations, Type, DOI

• Authors: Simone Biscaglia, Andrea Erriquez (co-first authors); I. Colaiori; A. Hakeem; F. Mantovani; M. Menozzi; M. Barbierato; M. Arioti; G. Bugani; R. Scarsini; D. D'Amario; L. Donazzan; A. Polimeni; M. Giordan; G. Benatti; G. Venturi; M. Ruozzi; F. Moretti; A. Monello; E. Moscarella; F. Versaci; J.A. Shah; V. Guiducci; C. Cavazza; F. Gallo; A. Rubboli; M. Borgi; A.A. Lakho; A. Capecchi; V. Lanzilotti; F. Ottani; G. Casella; A.M. Leone; M. Tebaldi; R. Pavasini; R. Piccolo; J. Farina; D. Martella; and G. Campo (senior author) — for the AIR-STEMI Trial Investigators
• Affiliations: Cardiology Unit, Azienda Ospedaliera Universitaria S. Anna, Cona (Ferrara), Italy; University of Ferrara; and further Italian and international centres
• Journal: New England Journal of Medicine 2026;395:972-82 (published online 29 August 2026)
• Type: Investigator-initiated, international, prospective, randomized, open-label superiority trial
• DOI: 10.1056/NEJMoa2605373
• Registration: NCT05818475
• Funding: Italian Health Ministry (primary sponsor; Ricerca Finalizzata 2021, GR-2021-12372516), administered through the University Hospital of Ferrara. Consorzio Futuro in Ricerca coordinated international centres and committees and received unrestricted funding from Sahajanand Medical Technologies and Siemens Healthineers. Pulse Medical Imaging Technology and Pie Medical Imaging provided software licences to the core laboratory for the angiography-derived FFR analyses

Overview

AIR-STEMI is the first randomized trial to show that guiding complete revascularization in STEMI by functional coronary angiography — angiography-derived FFR, without a pressure wire or hyperaemic agent — beats conventional angiography-guided decision making. In 1,823 patients with STEMI and multivessel disease whose culprit lesion had been successfully treated, physiology guidance reduced the composite of death, myocardial infarction, stroke/TIA or ischaemia-driven revascularization from 13.7% to 8.9% (HR 0.62; P<0.001) at a median 17.9 months, and reduced the safety composite of contrast-associated acute kidney injury or major bleeding (4.6% vs 7.1%; HR 0.63; P=0.02). The benefit was driven by myocardial infarction and repeat revascularization rather than mortality, and followed mechanically from treating fewer nonculprit vessels with fewer procedures and less contrast. The result runs opposite to the inference previously drawn from comparing angiography-guided COMPLETE with wire-FFR-guided FULL REVASC; the authors reconcile the discrepancy by timing and implementation rather than by the principle of physiological guidance.

Keywords

STEMI, multivessel coronary artery disease, complete revascularization, nonculprit lesion, functional coronary angiography, angiography-derived fractional flow reserve, quantitative coronary angiography, pullback analysis, pressure-wire-free physiology, contrast-associated acute kidney injury, major bleeding, ischaemia-driven revascularization

Key Takeaways

Design and Population

• Investigator-initiated, international, prospective, randomized, open-label superiority trial; randomization performed centrally with a secure web-based system
• Eligibility: STEMI, successful PCI of the culprit lesion, multivessel coronary artery disease with ≥1 nonculprit lesion of reference vessel diameter ≥2.5 mm and angiographically estimated diameter stenosis 50–99%
• Randomization within 48 hours of successful culprit treatment, 1:1, stratified by centre, sex, age (<75 vs ≥75 years) and nonculprit vessel location (LAD vs non-LAD)
• 1,823 patients randomized: 913 physiology-guided, 910 angiography-guided. Median age 66 years (IQR 58–76); 24% women
• Goal in both groups was complete revascularization during the initial hospitalization, either in the index procedure or staged at operator discretion
• Second-generation drug-eluting stents, with a recommendation for sirolimus-eluting, biodegradable-polymer, ultrathin-strut stents (Supraflex Cruz, Sahajanand Medical Technologies)
• Power calculation: assumed 9% event rate in the angiography group, 35% relative risk reduction, 1,718 patients for 80% power at two-sided alpha 0.05
• Primary analysis performed when the last enrolled patient reached 12 months (prespecified data cutoff); intention-to-treat; no interim analyses; no imputation of missing values
• No adjustment for multiplicity — the trial states explicitly that confidence-interval widths should not be used to infer treatment effects

The Intervention — What "Functional Coronary Angiography" Means Here

• Physiology-guided group: functional coronary angiography of all nonculprit lesions with angiographically estimated diameter stenosis ≥50%; this consisted of an estimation of the distal FFR value in the nonculprit vessel, with FFR ≤0.80 defining a functionally important lesion that was treated
• Performed without pressure wires or hyperaemic agents — physiology derived from the angiographic images themselves
• Analysed in a central core laboratory, with core-laboratory support for the strategy
• The strategy was used for two distinct purposes: (1) lesion selection — which nonculprit lesions to treat; and (2) PCI planning — pullback-based analysis to discriminate focal from diffuse disease, identify the segment responsible for the greatest pressure loss, and refine stent length and landing zones
• Comparator: conventional angiography-guided complete revascularization, i.e. visual assessment of lesion severity by the operator

Primary Outcome — Efficacy

• Primary outcome: composite of death from any cause, myocardial infarction, cerebrovascular accident (stroke or TIA), or ischaemia-driven revascularization; time-to-event analysis
• 8.9% (81/913) physiology-guided vs 13.7% (125/910) angiography-guided; HR 0.62 (95% CI 0.47–0.83); P<0.001
• Median follow-up 17.9 months
• Kaplan–Meier event counts show early and sustained separation: by day 180, 40 vs 71 events

Primary Safety Outcome

• Primary safety outcome: composite of contrast-associated acute kidney injury or major bleeding
• 4.6% (42) vs 7.1% (65); HR 0.63 (95% CI 0.43–0.93); P=0.02 — the physiology-guided strategy was safer, not merely non-inferior
• Components: contrast-associated acute kidney injury 2.2% vs 4.0%; BARC type 3, 4 or 5 bleeding 2.5% vs 3.6%
• The authors judge this biologically plausible given the lower number of vessels treated, fewer PCI procedures and reduced contrast exposure — while cautioning that the absolute number of safety events is small and longer follow-up is needed to define clinical relevance

Secondary Outcomes

• Cardiovascular death or myocardial infarction: 5.7% (52) vs 10.5% (96); HR 0.52 (95% CI 0.37–0.73) — the strongest secondary signal
• Death, whether from any cause or cardiovascular causes, did not appear to differ between groups — the benefit is not a mortality benefit
• The reduction was in myocardial infarction (both procedure-related and spontaneous) and in ischaemia-driven revascularization

Prespecified Subgroups (CIs unadjusted for multiplicity)

Benefit was consistent in direction across subgroups. The few whose intervals crossed 1 were the higher-risk groups:

The Authors' Proposed Mechanism — Three Parts

1. Physiological assessment during the index procedure allowed identification of nonculprit lesions without additional invasive diagnostic procedures, reducing the need for staged interventions
2. Restricting PCI to functionally important lesions substantially reduced the number of nonculprit vessels treated, with a consequent decrease in procedure- and stent-related events
3. Physiology was used not only for lesion selection but also to guide PCI. Pullback-based analysis discriminates focal from diffuse disease, identifies the segment responsible for the greatest pressure loss, and refines stent length and landing zones — enabling more targeted treatment of focal pressure drops and reducing the risk of leaving functionally important segments untreated. Angiography-guided PCI relies on visual assessment and may fail to identify that segment

Reconciling the Divergent Physiology Trials — The Central Contribution

The authors set AIR-STEMI against the two prior randomized comparisons of angiography-guided and physiology-guided strategies in myocardial infarction, and attribute the divergence to timing and implementation:

• The authors' conclusion: "these differences in timing and implementation may explain, at least in part, the divergent findings of the two trials"
• Broader position taken from the trial series: the clinical benefit of physiological assessment depends on it being integrated into a predefined physiological treatment strategy, not used as an isolated binary gate — and when so integrated it was associated with fewer procedure-related and spontaneous MIs and less ischaemia-driven revascularization
• On setting: evidence for functional coronary angiography had largely come from chronic coronary syndromes, but the authors argue STEMI is a particularly suitable context because it enables incorporation of physiological assessment while separating diagnostic assessment from procedural decision making

Limitations of the document

• Open-label design, and revascularization is itself a component of the primary outcome — the authors acknowledge this may have introduced ascertainment bias. Their defence is the consistency of the prespecified cardiovascular death or MI composite, which does not depend on operator decisions
• The strategy bundled two distinct interventions — lesion selection and PCI planning — so the relative contribution of each to the benefit cannot be determined. This is the trial's most important internal limitation: it demonstrates that the package works, not which element drives it
• Generalisability is limited to patients with coronary anatomy and angiographic image quality suitable for functional coronary angiography. Patients whose images are unsuitable cannot use the strategy
• Functional coronary angiography was analysed in a central core laboratory. Operator-performed analyses have been shown to be feasible and reproducible, but the trial does not demonstrate that routine, non-core-lab practice reproduces the result
• No adjustment for multiplicity, and the trial states confidence intervals for secondary and subgroup analyses should not be used to infer treatment effects — so the CV death/MI HR of 0.52 and all subgroup findings are hypothesis-generating
• The low overall use of intracoronary imaging may be relevant to interpretation of stent-related outcomes (its use was numerically greater in one group, per the truncated text)
• Mortality did not differ between groups, so the trial does not establish a survival benefit; the composite is driven by MI and repeat revascularization
• Industry involvement: unrestricted funding to the coordinating non-profit from Sahajanand Medical Technologies and Siemens Healthineers, and software licences for the core laboratory's angiography-derived FFR analyses from Pulse Medical Imaging Technology and Pie Medical Imaging — that is, from manufacturers of the devices and software under evaluation
• Absolute safety-event numbers are small, which the authors concede limits the precision of the safety conclusion

Key Concepts Mentioned

• concepts/Angiography-Based-Coronary-Physiology — the technique tested; AIR-STEMI is its first outcome-based randomized trial in STEMI, complementing the diagnostic-accuracy evidence from FAST III
• concepts/Multivessel-PCI-STEMI-Timing — the clinical question (how to select and when to treat nonculprit lesions); AIR-STEMI supplies the first direct head-to-head of physiology- vs angiography-guided completion
• concepts/Fractional-Flow-Reserve — the physiological index being estimated non-invasively; the ACS guidance question is directly affected
• concepts/Intracoronary-Imaging-Guided-PCI — the parallel but distinct question of imaging-guided rather than physiology-guided PCI
• concepts/Coronary-Microvascular-Dysfunction — relevant to the reliability of angiography-derived physiology in the acute setting
• entities/Heart-Failure — contrast-associated acute kidney injury as a safety outcome in an acute coronary syndrome population

Key Entities Mentioned

• Sahajanand Medical Technologies — manufacturer of the recommended Supraflex Cruz stent; provided unrestricted funding to the coordinating organisation
• Siemens Healthineers — unrestricted funding to the coordinating organisation
• Pulse Medical Imaging Technology and Pie Medical Imaging — provided the software licences used for the core laboratory's angiography-derived FFR analyses. None of these has a wiki page and none is warranted on a funding or licence disclosure alone

Wiki Pages Updated

• wiki/sources/AIR-STEMI-NEJM-2026.md — created (this page)
• wiki/concepts/Multivessel-PCI-STEMI-Timing.md — the standing position that physiology guidance is "questioned in the ACS setting" reframed; source_count 3→4
• wiki/concepts/Angiography-Based-Coronary-Physiology.md — first outcome RCT added; source_count 1→2
• wiki/concepts/Fractional-Flow-Reserve.md — the FFR-versus-angiography guidance section updated for the TIMING trial; source_count 3→4
• wiki/sourceindex.md — index row
• log.md — ingest entry