The last decade has been kind to physiology-guided PCI enthusiasts. FFR and iFR are no longer fringe tools, but part of our everyday vocabulary. These tools were established in primarily stable CAD patients, so my purist attendings would always ask me: Can we really trust microvascular physiology in the setting of plaque rupture, no-reflow, free radicals, and a storm of inflammation?
Two randomized trials FLOWER-MI and FRAME-AMI, asked essentially the same question:
In STEMI patients undergoing complete revascularization, should non-culprit lesions be assessed with FFR, or should we just rely on angiography?
FLOWER-MI (NEJM, 2021)
| Feature | Details |
|---|---|
| Design | 1,171 STEMI patients randomized after successful culprit PCI |
| Arms | FFR-guided PCI of non-culprit lesions (>50% angio stenosis) vs angiography-guided PCI (>50% angio stenosis, operator discretion) |
| Timing | Non-culprit PCI within 5 days of index PCI (majority staged 2–3 days later) |
| Primary Endpoint | Death, nonfatal MI, or urgent revascularization at 12 months |
| Results | No difference in composite outcome. Mean stents per patient: 1.0 (FFR) vs 1.5 (angio). |
| Imaging | Unknown |
| Lesion distribution | About half of non-culprit lesions were in the LAD |
Key nuance: FLOWER-MI’s neutral finding was largely driven by a higher rate of events in lesions deferred by FFR. In other words, some “FFR-negative” lesions went on to cause problems. This highlights an unresolved problem: in the inflammatory milieu post-MI, FFR may not fully capture lesion vulnerability.
FRAME-AMI (Korea, 2022)
| Feature | Details |
|---|---|
| Design | 562 STEMI patients randomized, median follow-up 3.5 years |
| Arms | FFR-guided PCI vs angiography-guided PCI of non-culprit lesions |
| Timing | 60% of non-culprit PCI performed during the same procedure as culprit PCI |
| Primary Endpoint | Death, nonfatal MI, unplanned revascularization |
| Results | Mortality: 2.1% (FFR) vs 8.5% (angio), HR 0.30. MI: 2.5% (FFR) vs 8.9% (angio), HR 0.32. No difference in unplanned revascularization. |
| Imaging | IVUS use ~26% in both groups |
| Lesion distribution | ~50% of non-culprit lesions were in the LAD |
Key nuance: FRAME-AMI diverged sharply from FLOWER-MI, showing mortality and MI benefit with FFR. With longer follow-up and more index-procedure complete revascularization, FFR guidance appeared protective. Looking through the detailed data in FRAME-AMI, in patients who had >70% stenosis, angiographically driven PCI had higher composite event rates. Is this because of poor placement of the stents? Perhaps pre-post FFR would have shown that the stents did not completely relieve the obstruction? I don’t know, it’s hard to explain that
Why Such Different Results?
Several factors likely explain the divergence:
- Timing of PCI
- FLOWER-MI staged non-culprit PCI days later.
- FRAME-AMI: >50% of patients had non-culprit PCI during the same procedure.
- Early physiology-based decisions may matter, particularly if vulnerable plaques are stabilized promptly.
- Follow-up duration
- FLOWER-MI: 1 year.
- FRAME-AMI: 3.5 years.
- The benefit of avoiding unnecessary stenting may only become apparent over longer horizons.
- Event drivers
- In FLOWER-MI, patients with FFR-deferred lesions were the ones who went on to have events.
- Suggests that “FFR-negative” in the post-MI inflammatory milieu isn’t the same as “truly benign.”
- System-level and practice differences
- FRAME-AMI’s lower overall event rates and higher rates of complete revascularization during the index procedure likely shaped outcomes.
- Plaque biology
- Both trials remind us that FFR is a hemodynamic tool, not a plaque biology tool. In ACS, vulnerable lesions may be physiologically silent but still dangerous.
How Does This Fit With Other Trials?
- COMPLETE trial: showed benefit of complete revascularization in STEMI, but “complete” was defined angiographically, not physiologically.
- FAME & FFR trials (stable CAD): taught us that deferring PCI in FFR-negative lesions is safe in chronic disease — but ACS is different.
- Imaging trials (COMBINE OCT-FFR, FORZA): suggest plaque morphology (e.g., thin-cap fibroatheromas) may be just as important as physiology.
Practical Pearls
- Don’t assume every >50% lesion post-STEMI should be stented.
- FFR and iFR can guide decision-making, but recognize their limits in ACS.
- Remember: COMPLETE supports revascularization of “bad” lesions post-ACS, but doesn’t tell us which lesions matter.
- Deferred lesions in ACS are not the same as deferred lesions in stable CAD — imaging may be the missing piece.
- The future may lie in hybrid strategies: physiology + IVUS/OCT to capture both flow limitation and plaque vulnerability.
My Take
| Feature | FLOWER-MI (2021) | FRAME-AMI (2022) |
| N | 1,171 | 562 |
| Follow-up | 1 year | 3.5 years |
| Timing of non-culprit PCI | Within 5 days (mostly staged 2–3 days later) | 60% during index procedure |
| Primary Endpoint | Death, MI, urgent revasc (no difference) | Death, MI, revasc (↓ death, ↓ MI) |
| Mortality (%) | 1.5 (FFR) vs 1.7 (angio) (no sig diff) | 2.1 (FFR) vs 8.5 (angio) (no sig diff |
| MI (%) | 3.1 (FFR) vs 1.7 (angio) (no sig diff) | 2.5 (FFR) vs 8.9 (angio) |
| Unplanned Revasc | 2.6 (FFR) vs 1.9 (angio) (no sig diff) | 4.1 (FFR) vs 9.2 (angio) (no sig diff) |
| IVUS use | Unknown | ~25-50% of patients |
| Lesion distribution | ~50% LAD | ~50% LAD |
| Notes | Events higher in deferred FFR lesions | Post-hoc analysis of a completed trialgoo |
What do I suspect drove the difference in outcomes? Likely a combination of:
- FRAME-AMI had longer followup to help identify more events / allow separation of curves. This makes sense because confidence intervals were wide in FLOWER-MI
- Lower event rates overall in FRAME-AMI reflecting systems of care / population differences
Both trials converge on a central lesson: stenting physiologically insignificant lesions does harm, but deferring PCI in ACS is not the same as deferring PCI in stable disease. The inflammatory milieu of STEMI changes the rules, and physiology alone may not be enough. This is where intravascular imaging comes in; whether IVUS, OCT, or future tools for thin-cap fibroatheroma detection. A hybrid approach that combines hemodynamics with plaque characterization may help us distinguish the truly dangerous lesions from the innocent bystanders.
TLDR;
- FLOWER-MI (2021): Neutral at 1 year, fewer stents in FFR arm, but more events in deferred lesions.
- FRAME-AMI (2022): Mortality and MI benefit with FFR at 3.5 years, majority of non-culprit PCI done during index procedure.
- Common message: Stenting physiologically insignificant lesions is harmful, but physiology may miss vulnerable plaques in ACS.