Remember me
ORIGINAL ARTICLE
Year : 2023 | Volume
: 24
| Issue : 3 | Page : 141-147
Percutaneous coronary intervention versus coronary artery bypass grafting in complex coronary artery disease: Long-term clinical outcomes from a high-volume center
Kerrick Hesse, Mohaned Egred, Azfar Zaman, Mohammad Alkhalil, Mohamed Farag
Department of Cardiothoracic, Freeman Hospital, Newcastle Upon Tyne, Tyne and Wear, UK
Correspondence Address:
Mohamed Farag
Department of Cardiothoracic, Freeman Hospital, Newcastle Upon Tyne, Tyne and Wear, NE7 7DN
UK
Mohammad Alkhalil
Department of Cardiothoracic, Freeman Hospital, Newcastle Upon Tyne, Tyne and Wear, NE7 7DN
UK
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/heartviews.heartviews_116_22
Abstract
Background: Clinical equipoise between a percutaneous coronary intervention (PCI) and coronary artery bypass grafting surgery (CABG) in the treatment of complex coronary artery disease (CAD), including unprotected left main coronary artery (LMCA) and/or three-vessel disease (3VD), remains debatable.
Methods: A retrospective analysis of an unselected cohort undergoing contemporary PCI versus CABG at a large center in 2015. Patients who received nonemergent treatment of unprotected LMCA and/or 3VD were included. The primary study endpoint was all-cause mortality at 5 years. Secondary endpoints included a composite of all-cause mortality, spontaneous myocardial infarction (MI), or ischemia-driven repeat revascularization at 30 days and 1 year.
Results: Four hundred and thirty patients met the inclusion criteria, 225 had PCI, and 205 had CABG. PCI patients were older with frequent LMCA involvement and higher EuroSCORE yet they had a fourfold shorter in-hospital stay compared to CABG patients. At 5 years, there was no significant difference in the primary endpoint between CABG and PCI (adjusted Hazard ratios 0.68, 95% confidence interval: 0.38–1.22, P = 0.19). Likewise, there was no significant difference in the incidence of the secondary composite endpoint or its components at 30 days or 1 year. A propensity score-matched analysis in 220 patients revealed similar outcomes.
Conclusions: In real-world long-term contemporary data, survival after PCI was comparable to CABG at 5 years in patients with unprotected LMCA and/or 3VD. At 1 year, the incidence of spontaneous MI and ischemia-driven repeat revascularization did not differ between the two cohorts. The mode of revascularization in these complex patients should be guided by the heart team.
Keywords: Coronary artery bypass grafting surgery, coronary artery disease, myocardial infarction, outcome, percutaneous coronary intervention, revascularization
Introduction 
Coronary artery disease (CAD) mortality has fallen drastically over the past 30 years since the advent of percutaneous coronary intervention (PCI) for the treatment of acute coronary syndromes (ACS), especially ST-elevation myocardial infarction (STEMI).[1]
In the United Kingdom, the growth in the number of PCI procedures has been near exponential.[1] In contrast, coronary artery bypass grafting surgery (CABG), previously recognized as the gold standard for complete revascularization in complex CAD, has steadily declined since its peak in the late 1990s.[1],[2]
However, the wide variability in PCI to CABG ratios across centers in Europe and the United Kingdom suggests selection and treatment bias.[2] Despite multiple large randomized trials comparing PCI to CABG in unprotected left main coronary artery (LMCA) and/or multi-vessel CAD, an equipoise in mortality risk is observed.[2] Nonetheless, a difference in other important clinical outcomes, such as recurrent infarction or revascularization remains a subject of much debate.
The percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease (SYNTAX) trial was the pioneer randomized trial comparing the two revascularization modalities in patients with complex CAD. The 1-year, 3-year, and 5-year results demonstrated that CABG was superior to PCI for the combined clinical endpoint of major adverse cardiac and cerebrovascular events, including death from any cause, myocardial infarction (MI), stroke or repeat revascularization.[3],[4],[5] This was mainly driven by increased MI and repeat revascularization events.[3],[4],[5]
The Randomized Comparison of Coronary Artery Bypass Surgery and Everolimus-Eluting Stent Implantation in the Treatment of Patients with Multivessel CAD (BEST) trial, the Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) trial, the Nordic-Baltic-British left main revascularization study, and the Evaluation of XIENCE versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial largely corroborated the principal findings from the SYNTAX trial.[6],[7],[8],[9],[10],[11] Important differences included excess mortality at 5 years in the PCI cohorts of the FREEDOM and EXCEL trials.[6],[9],[10] Of note, extended (>10-year) follow-up of BEST and SYNTAX trials revealed no difference in all-cause mortality.[12],[13]
The challenge in translating trial evidence into routine clinical practice is at least threefold. First, continued advances in technology and procedural techniques in PCI and CABG continue to weaken the validity of observed results in contemporary practice.[14] Second, the value that trialists, clinicians, and patients put on different endpoints likely differ. The most robust endpoint is death, but the need for repeat interventions due to MI or ongoing angina in a resource-limited health-care service may garner greater importance to patients and physicians, especially in the 1st year following the intervention. Third, trial populations are often less comorbid with better background medical therapy than the general population. They are also deemed to benefit equally from both revascularization strategies, which is increasingly not the case in an older and frailer real-life population.[15] In addition, surgical ineligible patients who end up having PCI are more frequent in daily practice and often contribute to worse PCI outcomes independent from predicted clinical scores through factors such as limited mobility, cognitive dysfunction, poor distal coronary targets, or morbid obesity. Ultimately, the mode of revascularization in these complex patients should be guided by the heart team.[16],[17],[18]
This study aimed to compare the long-term clinical outcomes of contemporary PCI versus CABG in a real-world unselected population with complex CAD.
Methods Study population and design
We undertook a single-center, retrospective observational study. The study was conducted following the Declaration of Helsinki and Good Clinical Practice guidelines. We identified all patients with LMCA and/or three-vessel disease (3VD) treated with either PCI or CABG at the Freeman Hospital, a tertiary clinical center of excellence in the Northeast of England. The scope of the study was limited to between January 01, 2015, and December 31, 2015, to capture relatively contemporary PCI and CABG practices with robust long-term outcome data. Eligible patients presented with either intractable angina or ACS. They were all discussed by the heart team and offered either PCI or CABG. Exclusion criteria were a history of previous CABG and unstable patients requiring emergency intervention, including rescue/primary PCI (PPCI) and salvage CABG.
Complex CAD was defined as ≥50% unprotected LMCA disease, 3VD, i.e., ≥70% stenotic disease in all three major epicardial arteries (left anterior descending [LAD], left circumflex [LCx], and right coronary [RCA] arteries) or their major tributaries, or a combination thereof. Patients were classified into LMCA disease and/or 3VD based on the vessels intervened on during PCI or grafted during surgery. During PCI, this meant the delivery of a stent or balloon to the LMCA and/or to the LAD, LCx, and RCA or their major tributaries. In CABG, ≥3 grafts defined 3VD with grafts to the proximal LAD used as surrogate markers for LMCA disease.
Data were sourced from Intellect, a national live online database hosted by the National Institute of Cardiovascular Outcomes Research and the British Cardiovascular Interventional Society. Further data were collected from electronic medical records. Data were prospectively collected, validated, and entered into the database, which was analyzed retrospectively. Ethical approval was waived by our institution as only anonymized patient data were obtained retrospectively.
Patient-level data included age, sex, smoking status, cardiovascular comorbidities, previous cardiac events, and interventions, left ventricular ejection fraction with an EF ≤35% representing severe left ventricular systolic dysfunction (LVSD), hemoglobin (Hb) level (g/L), creatinine level (μmol/L), additive/logistic EuroSCOREs, and SYNTAX score.[19],[20],[21],[22] The additive and logistic EuroSCOREs and SYNTAX scores were calculated by two investigators using the formulas from www.euroscore.org and www.syntaxscore.org.
Procedure-specific data included the number and length of stents deployed, the use of pressure-wire technology, intravascular ultrasound, optical coherence tomography, and atherectomy devices in PCI, and the duration of cardiopulmonary bypass, the type and size of grafts in CABG.
Procedure details
All patients underwent either PCI or CABG with the aim of complete revascularization of all vessels at least 2.5 mm in diameter with ≥50% stenosis. PCI was performed by the interventional cardiologist on the local protocol and personal experience, including the type and number of stent platforms, and the type and duration of antiplatelet therapy. Intravascular imaging and adjunctive intracoronary calcium modification techniques were used at the discretion of the operator. Similarly, the CABG surgical technique followed standard practice.
Study endpoints
The primary study endpoint was all-cause mortality at 5 years. Secondary endpoints included a composite of all-cause mortality, spontaneous MI, or ischemia-driven repeat revascularization, or its components, at 30 days and 1 year. Death could be of any cause with the date of the last hospital encounter used as a surrogate for the date of death if it was unavailable. Spontaneous MI was defined by the fourth and most recent universal definition of MI.[21] In general, it required a rise in cardiac troponin above the 99th percentile with evidence of ischemia based on symptoms, electrocardiogram changes, imaging, and/or angiography.[21] Ischemia-driven repeat revascularization included any repeat PCI or CABG for intractable angina or ACS.
Statistical analysis
Data presented as mean ± standard deviation (SD) or median (Q1–Q3) for continuous variables and as numbers with percentages for binary variables. Continuous variables were compared with the Student's t-test and Mann–Whitney U-test for parametric and nonparametric distributions, respectively. Categorical variables were compared with the Chi-square or Fisher's exact tests, as appropriate.
We used the Kaplan–Meier methodology with a log-rank test to compare time to first event rates between PCI and CABG. Hazard ratios (HR) and their 95% confidence interval (CI) were estimated with the Cox Proportional Hazards model after adjusting for significant differences in baseline patient characteristics. The bootstrap technique using one thousand samples was used to account for final multivariable model uncertainty. All study variables were first analyzed with univariate analysis and those that showed a significant interaction (P < 0.1) were entered into the final multivariable analysis (i.e., age, gender, hypertension, diabetes, previous cerebrovascular accidents, chronic obstructive pulmonary disease, previous MI or PCI, clinical presentation, coronary anatomy, Hb level, and EuroSCORE).
To further account for confounding baseline variables, we applied propensity score matching methodology. Propensity scores were calculated using multivariate logistic regression based on the significantly different baseline patient characteristics mentioned earlier. Matching was performed without replacement in a 1:1 manner and with a caliper of 0.1 of the propensity score. A two-sided P < 0.05 was considered statistically significant. Statistical analysis was performed with SPSS version 29.0 (Armonk, NY: IBM Corp).
Results Patients and procedures
A systematic search of the Intellect database identified 3093 revascularization procedures performed at the Freeman Hospital between January 01, 2015, and December 31, 2015. We excluded 1868 (60.4%) procedures that did not fulfill the criteria for unprotected LMCA and/or 3VD revascularization. A further 775 (25.1%) were emergent procedures, including 766 (98.8%) PPCIs. In the end, we identified 430 unique patients with complex CAD, of whom 205 had CABG and 225 had PCI [Figure 1].
Figure 1: Study flowchart. BCIS: British cardiovascular intervention society, PCI: percutaneous coronary intervention, CABG: coronary artery bypass grafting surgery, PPCI: primary percutaneous coronary intervention, STEMI: ST-elevation myocardial infarction, LMCA: left main coronary artery, CAD: coronary artery diseasePatients who received PCI were older (70.1 ± 11.7 vs. 67.5 ± 9.3 years; P = 0.011), frequently presented with LMCA involvement (41.3% vs. 16.1%; P < 0.001), and were more likely to have a history of obstructive CAD including previous MI (36.2% vs. 22.9%; P = 0.003), or PCI (28.9% vs. 20.5%; P = 0.045), compared to patients who received CABG [Table 1]. However, PCI patients had a lower burden of diabetes (30.7% vs. 40.8%; P = 0.033), previous cerebrovascular accidents (11.4% vs. 19.0%; P = 0.031), and chronic obstructive pulmonary disease (15.9% vs. 24.4%; P = 0.039) compared to CABG patients. There was no difference in the calculated SYNTAX score between PCI and CABG (27.9 ± 8.7 vs. 26.0 ± 7.0 points; P = 0.44). Likewise, there was no difference in the proportion of patients with severe LVSD (13.7% vs. 11.6%; P = 0.62). As expected, the PCI cohort had a higher perioperative risk, as calculated by the additive (5.2 ± 3.2 vs. 4.7 ± 2.6 points; P = 0.047) and logistic (6.1 ± 6.3 vs. 4.9 ± 4.8 points; P = 0.029) EuroSCOREs.
The PCI and CABG procedural details are presented in [Table 2] and [Table 3], respectively. In comparison to CABG patients, PCI patients had a fourfold shorter length of hospital stay (3.1 ± 7.0 vs. 12.9 ± 13.6 days; P < 0.001). On average, three lesions were treated during PCI and CABG, represented by the number of vessels attempted and drug-eluting stents and balloons delivered in PCI and the number of grafts used in CABG. The mean ± SD longest segment of contiguous stenting was 46.8 ± 26.5 mm in PCI. The left internal mammary artery was grafted in 98.7% of CABG cases.
Table 3: Coronary artery bypass grafting surgery procedural data (n=205)Primary endpoint
At 5 years, the primary endpoint of all-cause mortality occurred in 26/205 patients (12.7%) who underwent CABG and in 53/225 patients (23.6%) who had PCI (adjusted HR 0.68, 95% CI: 0.38–1.22, P = 0.19) [Table 4].
Table 4: Primary and secondary clinical endpoints after percutaneous coronary intervention and coronary artery bypass grafting in complex coronary artery diseaseSecondary endpoints
At 30 days and 1 year, there were no significant differences noted between CABG and PCI for the treatment of complex CAD either in the composite endpoint of all-cause mortality, spontaneous MI, ischemia-driven repeat revascularization, or its components [Table 4].
Propensity score-matched analysis
Similar outcomes were observed at all-time points in a propensity score-matched cohort of 220 patients [Table 4]. The primary endpoint of all-cause mortality at 5 years occurred in 13.6% after CABG and 20.9% after PCI (HR 0.61, 95% CI: 0.32–1.17, P = 0.14) [Table 4] and [Figure 2].
Figure 2: Five-year cumulative survival after PCI and CABG in propensity score-matched cohort. Survival after PCI was comparable to CABG in 220 patients with unprotected left main coronary artery and/or three-vessel disease. CABG: coronary artery bypass grafting surgery, PCI: percutaneous coronary intervention Discussion PCI and CABG are two distinctive revascularization strategies for obstructive CAD. Both modalities have specific indications including PPCI for STEMI and CABG where concomitant cardiac surgery is required.[2] However, their role in complex CAD remains heavily debated, especially as PCI technology and operator experience continue to evolve. In our study involving patients with unprotected LMCA and/or 3VD undergoing contemporary PCI and CABG, mortality after PCI was comparable to CABG at 5 years. Moreover, the incidence of spontaneous MI and ischemia-driven repeat revascularization did not differ between the two cohorts at 30 days and up to 1 year. Of note, patients who underwent PCI had a fourfold shorter length of hospital stay compared to patients who had CABG.
Our principle finding that survival is similar after PCI and CABG in patients with complex CAD involving LMCA is in keeping with data from landmark trials, including the SYNTAX and BEST trials, and more recently the Fractional Flow Reserve versus Angiography for Multivessel Evaluation 3 trial.[3],[4],[5],[11],[22],[23] It is noteworthy that our PCI cohort had a higher risk compared to the CABG cohort. They were on average at least 5 years older, frequently presented with LMCA involvement, more likely to have a history of obstructive CAD, and considered at greater perioperative risk as assessed by the EuroSCORE. Furthermore, importantly, some PCI patients (approximately 25%) were turned down for surgery (i.e., CABG surgical ineligibility) due to unfavorable factors known to relate to poor long-term clinical outcomes and not accounted for in surgical risk scores, such as frailty, poor distal coronary target, and morbid obesity. Improved stent platforms, routine coronary imaging, effective dual antiplatelet regime, and improved PCI techniques may facilitate more durable revascularization but are less likely to affect the greater risk of an overall more comorbid and frail population.[2],[14]
Our outcomes after real-life contemporary PCI at a large-volume center are consistent with and perhaps better than other observational data. Approximately 30%–50% of screened patients in the aforementioned trials were excluded as one revascularization strategy was deemed more suitable; CABG was often the preferred modality for very complicated diseases and PCI was reserved for high-risk patients.[3],[4],[5],[6],[7],[8],[9],[10],[11] Ineligible patients that entered a nested registry parallel to the SYNTAX trial were phenotypically similar to our PCI cohort; the mean age was 71.2 years, 40.4% had prior MI, 5.7% had severe LVSD, and the mean EuroSCORE was 5.8.[3],[4],[5] Their 5-year all-cause mortality was 30.0% which is higher than the death rates observed in our PCI cohort.[3],[4],[5],[24] In a propensity-matched cohort of 636 diabetic patients with multivessel CAD from an Italian registry, the 5-year mortality rate remained significantly higher after PCI than CABG (24.8% vs. 18.2%, P < 0.055).[25] In another Chinese single-center cohort of 478 LMCA disease patients, 5-year mortality rates after PCI and CABG were comparable (27.4% vs. 31.5%, P = 0.36).[26] Our 5-year death rate following PCI and CABG was not significantly different.
Study limitations
To the best of our knowledge, our study is one of the largest studies with long-term follow-up data comparing relatively contemporary PCI and CABG.[27],[28] However, it has several limitations. First, this was a retrospective observational study, and therefore, has inherent limitations of the retrospective design. Second, the Freeman Hospital is a tertiary specialist center of a large catchment area with only 14% of our surgically treated patients local to Newcastle compared to 24% of PCI-treated patients. Nonetheless, all patients usually receive follow-ups at the Freeman for up to 1 year. Further clinical events beyond 1 year can be easily missed if patients presented and received treatment at their local hospitals. Therefore, we were unable to include MI and revascularization events after 1 year. Third, although this is a single-center study, it is still subjected to differences in individual PCI and surgical techniques. Fourth, approximately a quarter of the PCI cohort was ineligible for CABG surgery due to factors not necessarily captured by clinical risk scores, such as frailty, poor distal coronary target, and/or morbid obesity.[29] These factors were not possible to adjust for in the analysis and likely contributed to the observed numerical increase in clinical events after PCI compared to CABG. Fifth, we did not have data on the stroke endpoint. In addition, we did not have data on patients with complex CAD treated conservatively during the same period. It would have been interesting to add that group as a comparator to patients treated invasively with either PCI or CABG. Finally, our long-term analysis reports all-cause mortality, and thus noncardiac deaths might have been included. Nonetheless, all-cause mortality is considered an appropriate endpoint to follow in the long term because it accounts for both cardiac and systemic diseases and is unaffected by any reporting or misclassification bias.
Conclusions Our study with long-term real-world contemporary data from a large tertiary center suggests that survival after PCI is comparable to CABG at 5 years in patients with nonemergent unprotected LMCA and/or 3VD. At 1 year, the incidence of spontaneous MI and ischemia-driven repeat revascularization did not differ between the two cohorts. These complex patients must be discussed by the heart team to guide the best revascularization strategy.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
Comments (0)