We are all familiar with the expression “reality surpasses fiction.” Often, we find a significant alignment between outcomes in clinical practice and registry data. A consistency that is not always reflected in the findings of clinical trials. Highly restrictive selection criteria, non-inferiority analyses, varied definitions of clinical outcomes, and their interpretation as composite events have been among the methods used to influence results towards predetermined tendencies, often in line with the sponsor’s policies. The generation of “fictional evidence” extends even further with successive post-hoc analyses or derived meta-evidence, torturing the data to recount the same alternative reality repeatedly until it becomes dogma.
Revascularization of left main coronary artery disease (LMCA) is not exempt from this controversy. The distortion begins with the consideration of LMCA disease as separate from multi-vessel disease, although it is subsequently classified based on anatomical complexity via the SYNTAX score, which reflects the extent of disease in the remaining coronary tree. This should be considered only when LMCA disease is the sole significant coronary lesion, which represents less than 15% of cases. Consequently, the proportion of LMCA patients with a SYNTAX score indicating surgery as superior to percutaneous coronary intervention (PCI) per current guidelines (>22 points) is over half of cases, around 35% with scores of 22–32 points and approximately 25% with scores >32 points. While considering LMCA disease is relevant for prognosis and the subsequent revascularization recommendation, it should not determine the therapeutic indication unless, as noted, it is the only lesion present.
The SWEDENHEART registry includes patients treated at 28 centers from 2005 to 2015, providing substantial long-term follow-up data. A total of 11,337 patients with LMCA disease were included, with 84% undergoing surgical revascularization and 16% receiving PCI. These differences reflect the registry’s alignment with guideline-recommended indications, capturing the characteristics of coronary artery disease presentation described above. Patients with prior coronary surgery, presentation as STEMI and/or cardiogenic shock were excluded from the study. Patient characteristics included a mean age of 72.8 years for PCI versus 69.6 years for surgery; a leaner profile compared to our region with a BMI of 26.2 kg/m²; smoking rates >50%; hypertension >60%; and lower rates of diabetes mellitus around 20%, compared to close to 50% in our region. The mode of presentation was in the context of unstable angina or acute coronary syndrome (ACS) in 60-70% of cases, necessitating urgent revascularization during the index admission. Most data, therefore, reflect this urgent context, with only 30% of stable coronary disease cases undergoing elective revascularization. In PCI cases, 1 or 2 stents were used in >60% of cases, nearly 80% if the stent score was 3 or less. For surgery, more than 75% of cases had more than 3 distal anastomoses, though only 2.1% utilized a second internal mammary artery.
After balancing groups for confounding factors and performing Cox regression analysis, PCI in LMCA disease treatment showed higher mortality (HR = 1.5), increased risk of recurrent MI (HR = 6.1), need for repeat revascularization (HR = 14), and major cardiovascular and cerebrovascular events (HR = 2.8). There were no significant differences in the incidence of stroke analyzed independently. Although current clinical guidelines do not influence the preference for surgical over percutaneous treatment in the presence of diabetes, a favorable interaction in terms of survival was observed, translating into an average 3.6-year increase if patients were treated surgically (p = 0.014). Such favorable interactions for surgery were also observed in subgroup analysis of younger patients <70 years, and as expected, in those with complex LMCA disease involving two or three vessels (compared to >70 years and isolated LMCA disease with/without a single vessel).
The authors concluded that surgical revascularization in patients with LMCA disease is associated with better survival and lower rates of major cardiovascular events compared to PCI.
COMMENTARY:
The SWEDENHEART registry results contribute evidence on the revascularization of LMCA disease, a significant controversy in the treatment of ischemic heart disease.
To date, evidence is based on the four clinical trials NOBLE, EXCEL, PRECOMBAT, and SYNTAX, which randomized mostly stable coronary artery disease patients 1:1 to PCI or surgery. The stents used were drug-eluting, specifically biolimus, everolimus, sirolimus, and paclitaxel, respectively. The real-world context of revascularized patients, mostly undergoing urgent procedures, is poorly represented in these trials. PRECOMBAT excluded patients with an MI in the previous week, NOBLE applied the same criterion in the last 24 hours, and in EXCEL, patients with elevated CK-MB levels post-MI but in decline were excluded. SYNTAX systematically excluded all patients with prior STEMI or NSTEMI prior to revascularization.
A meta-analysis by Gaba et al. contemporaneous with this study explored PCI versus surgery outcomes in LMCA disease revascularization, emphasizing the urgent context following ACS and unstable angina. After pooling patient-level data from the four trials, only 1466 patients were included since the remaining 2927, approximately two-thirds of the sample, had stable coronary artery disease, a population notably different from that of the SWEDENHEART registry. Indeed, stable coronary artery disease was present in 82% and 53% of the NOBLE and EXCEL samples, respectively. The average follow-up was 5 years. In PCI cases, the mean stent score was 2. In surgery cases, although dual mammary artery usage was above 20%, the mean conduit count was only 2. This was due to a significant number of sequential anastomoses for complete revascularization as reflected in the SWEDENHEART registry and lower anatomical complexity of the disease. In NOBLE, complex lesions such as coronary occlusions, bifurcation lesions excluding LMCA, or severely tortuous or calcified vessels were systematically excluded, showing proportions that do not align with the real-world LMCA disease casuistry and single-vessel disease cases. The EXCEL trial, on its part, excluded patients with a SYNTAX score >33 points.
Five-year mortality rates in Gaba et al.’s study were PCI vs. surgery 10.9% vs. 11.5% in ACS patients and 11.3% vs. 9.6% in stable disease patients, all without significant differences. Recurrent MI rates were HR = 1.74 with prior ACS and 3.03 in stable disease; the need for repeat revascularization HR 1.57 with prior ACS and 1.9 in stable disease; all without statistically significant differences. The authors concluded their work with the fallacy of equivalence between the therapeutic options at 5-year outcomes, clearly influenced by the low statistical power and lack of representativeness in the real-world casuistry due to the nature of the analyzed data.
Finally, despite the influence of clinical trial evidence, the SWEDENHEART experience is not unique. Another Canadian registry by Tam et al., using Ontario’s administrative database, demonstrated surgical benefits, with a propensity-score-adjusted analysis showing lower mortality (HR = 1.6) and major cardiovascular events (HR = 1.7).
In an era when the EACTS withdrew its endorsement of the current revascularization guidelines in 2018, studies like this are essential. The foundation of the previous consensus document on trials like EXCEL, which introduced a deliberate bias destined to be a milestone in the specialty’s history, led to EACTS’s stance. Till of today, there is no reply or amendment by the corresponding cardiology society. Although EXCEL showed better survival with surgery versus PCI, this was obscured by adopting a change in the definition of peri-procedural MI to favor PCI in the composite outcome and justify class I and IIa indications for PCI in SYNTAX scores <22 and 22–32 points, respectively. This imposed fiction should not overshadow rigor and reality, and we hope that with the publication of the new consensus document in the coming months, we will not be “asked to swallow a bitter pill.”
REFERENCE:
Persson J, Yan J, Angerås O, Venetsanos D, Jeppsson A, Sjögren I, et al. PCI or CABG for left main coronary artery disease: the SWEDEHEART registry. Eur Heart J. 2023 Jun 8. doi: 10.1093/eurheartj/ehad369.
Gaba P, Christiansen EH, Nielsen PH, Murphy SA, O’Gara PT, Smith PK, et al. Percutaneous Coronary Intervention vs Coronary Artery Bypass Graft Surgery for Left Main Disease in Patients With and Without Acute Coronary Syndromes: A Pooled Analysis of 4 Randomized Clinical Trials. JAMA Cardiol. 2023 May 31. doi: 10.1001/jamacardio.2023.1177.
Tam DY, Fang J, Rocha RV, Rao SV, Dzavik V, Lawton J, et al. Real-World Examination of Revascularization Strategies for Left Main Coronary Disease in Ontario, Canada. JACC Cardiovasc Interv. 2023 Feb 13;16(3):277-288. doi: 10.1016/j.jcin.2022.10.016.
