Why This Debate Still Matters
For decades, CABG and PCI have been cast as rivals—durability versus convenience. CABG promised long-term completeness of revascularization; PCI offered speed, recovery, and adaptability.
With modern technology, that dichotomy is fading. The 5-year results of FAME 3, along with 10-year pooled data from SYNTAX and PRECOMBAT, show that physiology-guided, imaging-optimized PCI has closed much of the historical gap. CABG remains superior only in a narrow band of high-risk anatomy.
The Landmark: FAME 3 (5-Year Outcomes)
| Feature | Details |
|---|---|
| Population | 1,424 patients with three-vessel CAD, no left main involvement |
| Design | Randomized to FFR-guided PCI (zotarolimus DES) vs CABG |
| Primary Endpoint | Death, MI, or stroke |
| Follow-up | 5 years |
| Results | Composite endpoint: no significant difference between CABG and PCI |
| MI | Higher in PCI group |
| Repeat Revascularization | Higher in PCI group |
| Death / Stroke | Similar between groups |
| Interpretation | CABG advantage largely neutralized for death/stroke; PCI carries more reintervention and MI risk but offers less invasiveness and faster recovery |
At 5 years, death, MI, or stroke occurred in 18.6% of CABG and 20.6% of PCI patients (p = NS). The surgical edge in completeness of revascularization was offset by the PCI arm’s physiologic lesion selection, contemporary DES, and lower procedural morbidity (1, 2).
Ten-Year Perspective: Mortality Curves Converge
The individual-patient meta-analysis by Sabatine et al. (Lancet 2021) pooled SYNTAX and PRECOMBAT and demonstrated that all-cause mortality through 5 years was statistically identical between CABG and PCI (HR 1.10, 95% CI 0.91–1.32; p = 0.33). After the first postoperative year, the survival curves rise in parallel with no late separation.
[Insert Figure 1 – Five-Year All-Cause Mortality, CABG vs PCI (Sabatine MS et al., Lancet 2021; 398:2247–57)]
Interpretation: CABG no longer confers a universal survival advantage. The persistent trade-off is familiar—CABG delivers fewer MIs and reinterventions; PCI offers lower procedural morbidity and faster recovery.
FAME 3 and Diabetes: The Persistent Frontier
A prespecified analysis from FAME 3 (Takahashi K et al., JAMA Cardiol 2025, in press) examined outcomes by diabetic status.
| Group | Adjusted HR (PCI vs CABG) | 95% CI | p Value |
|---|---|---|---|
| Diabetes (DM) | 1.51 | 0.90 – 2.51 | 0.115 |
| Non-Diabetes (Non-DM) | 1.19 | 0.80 – 1.79 | 0.390 |
| Interaction p | – | – | 0.427 |
Interpretation:
- Numerically, diabetic patients fared worse after PCI than CABG, but this did not reach statistical significance.
- Non-diabetic outcomes were nearly superimposable between PCI and CABG.
- The absence of a significant interaction (p = 0.427) reinforces that the relative balance between PCI and CABG holds across glycemic states in this FFR-guided, contemporary-DES cohort.
Visual summary: Figure 2 – Composite of death, MI, or stroke by diabetic status (Takahashi K et al., JAMA Cardiol 2025).
[Insert Figure 2 – FAME 3 and Diabetes Subgroup Analysis]
Clinical take:
Even though diabetics remain the group most likely to derive incremental benefit from CABG, this figure illustrates how the magnitude of difference has narrowed in the FFR/IVUS era. What used to be a categorical rule (“CABG for all diabetics”) is now a nuanced decision:
- If anatomy is favorable and physiology-guided PCI achieves complete revascularization, outcomes approach parity.
- For diffuse small-vessel disease or long-segment involvement, CABG still wins on durability.
The SYNTAX Score and the Anatomy Factor
One recurring truth across every generation of data from SYNTAX and FREEDOM to FAME 3 is that anatomic complexity (and its ability to prevent complete revascularization) plays a significant role.
What FAME 3 Taught Us About SYNTAX
In FAME 3, the advantage of CABG was driven almost entirely by patients with high SYNTAX scores (> 32).
- In low-to-intermediate SYNTAX (≤ 32), event rates for death, MI, or stroke were numerically lower with PCI and statistically similar to CABG.
- In high SYNTAX, (>32) CABG maintained superiority for ischemic outcomes and reintervention rates.
This gradient confirms that physiology guidance and imaging optimization level the playing field in less complex anatomy
How This Fits with Sabatine 2021 (SYNTAX + PRECOMBAT)
Sabatine’s Lancet 2021 meta-analysis echoed the same pattern:
- Low/intermediate complexity: survival curves nearly overlap through 5 years.
- High complexity: CABG continues to show a modest but consistent survival edge.
When you account for the contemporary PCI arm: FFR selection, IVUS/OCT confirmation, and second-generation DES, the anatomical ceiling for PCI keeps rising, but not indefinitely.
Interplay with Diabetes
The FAME 3 diabetes analysis (Takahashi 2025) underscores that complexity and biology amplify each other.
- A diabetic with low SYNTAX may do just as well with physiology-guided PCI.
- A diabetic with high SYNTAX remains the archetype for CABG benefit.
This interaction explains why p-interaction = 0.427 in the FAME 3 diabetic subgroup: anatomy, not diabetes alone, determines where the lines diverge. This may also indicate that diabetics with high SYNTAX scores reflect poorer substrate (i.e. worse diabetic control, genetic predisposition, other uncontrolled risk factors, etc).
Clinical Translation
- SYNTAX < 22: PCI and CABG achieve comparable long-term outcomes – patient preference and operator expertise dominate.
- SYNTAX 23–32: true equipoise zone
- SYNTAX > 32: CABG remains the durable standard, especially in diabetics or those with LV dysfunction.
When CABG Still Holds the Edge
CABG remains clearly advantageous when:
- High anatomic complexity (e.g., SYNTAX > 32)
- Diffuse diabetic disease, especially with small distal targets
- Impaired LV function with viability, where multi-territory grafts enhance supply
- Concomitant surgical needs (valve or aortic intervention)
In these settings, the durable graft supply and anastomotic completeness of CABG still translate into fewer ischemic events over time (3–5).
Where PCI Has Caught Up
Modern PCI is not the PCI of the SYNTAX era. Several advances have closed the gap:
- FFR/iFR-based lesion selection reduces unnecessary stenting.
- Intravascular imaging (IVUS/OCT) ensures full expansion and apposition, limiting restenosis and stent thrombosis.
- Next-generation DES (ultrathin struts, biocompatible polymers) improve long-term patency.
- Shorter DAPT durations mitigate bleeding and compliance issues.
In low-to-intermediate SYNTAX anatomy, and for patients prioritizing less invasiveness and faster recovery, modern PCI now achieves similar survival and stroke rates to CABG—accepting more reinterventions as the trade-off (1, 2, 6).
How to Frame It Clinically
“Survival and stroke risk are essentially equivalent either way. CABG reduces MI and repeat procedures but requires greater recovery. PCI gets you back faster but may need a touch-up later. Your anatomy and goals decide the fit.”
Clinical Pearls
- FAME 3 (5 yrs): death/stroke/MI ≈ no difference; MI and reintervention ↑ with PCI.
- Sabatine 2021: mortality curves nearly overlap; HR 1.10 (NS).
- CABG advantage persists only in high-SYNTAX + diabetic, or LV-dysfunction subsets.
- Modern PCI bridged the gap through FFR guidance, IVUS/OCT, and DES evolution.
- Shared decision-making now replaces dogma.
My Take
The “CABG vs PCI” debate has matured into a conversation about how to achieve complete, durable revascularization for each individual. In skilled hands, imaging-optimized PCI now provides survival equivalent to surgery for most three-vessel or left-main cases of moderate complexity. CABG retains clear value in complex or diabetic anatomy, but for many patients, the question now is one that requires a true heart team approach, taking into consideration the patient’s values, as well as their concerns.
TLDR
- FAME 3 (5 yrs): death/stroke/MI ≈ no difference.
- Sabatine et al. (2021): 10-yr mortality curves converge.
- CABG > PCI in complex, diabetic, or LV-dysfunction cases.
- Modern PCI ≈ CABG for survival; trade-offs are MI vs invasiveness.
References
- Fearon WF et al. FAME 3 Trial – FFR-Guided PCI vs CABG in Three-Vessel Disease. Lancet. 2025. ACC Summary
- ACC.org. ESC 2025 Hot Line: FAME 3 Five-Year Results. ACC Coverage
- Sabatine MS et al. CABG vs PCI in Left Main and Multivessel Disease: Meta-analysis. Lancet. 2021;398:2247–2257. Full Text
- Park S-J et al. PRECOMBAT 10-Year Outcomes. Circulation. 2020;141:1437–1446. Article
- SYNTAXES Investigators. Ten-Year Mortality After PCI vs CABG. Lancet. 2019;394:1325–1334. Lancet Article
- Farkouh ME et al. FREEDOM Trial – CABG vs PCI in Diabetics. NEJM. 2012;367:2375–2384. NEJM Full Text
- Velazquez EJ et al. STICHES Trial – CABG in Ischemic Cardiomyopathy. NEJM. 2016;374:1511–1520. NEJM Article
- Bangalore S et al. BEST Trial – Everolimus Stents or Bypass Surgery. NEJM. 2015;372:1204–1212. NEJM Article
