Global Risk Score for Choosing the Best Revascularization Strategy in Patients With Unprotected Left Main Stenosis

Abstract: Background. Coronary artery bypass graft (CABG) is recommended for patients with unprotected left main stenosis (ULMS). Percutaneous coronary intervention (PCI) is only recommended in specific anatomic conditions as in patients with low/mid SYNTAX score (SS). The aim of this study was to assess if the clinical and anatomic global risk classification (GRC) can enhance the indication of both revascularization therapies. Methods. A total of 407 patients with ULMS treated with CABG (n = 285) or PCI (n = 122) were prospectively collected. The decision to treat with CABG or PCI was dependent on patient and physician’s choice. Patients with ST-elevation myocardial infarction, shock, or valve disease were excluded. Clinical follow-up was obtained at 3 years. Results. Patients with low GRC (n = 151) treated with CABG vs those with PCI had similar cardiac mortality (5.9% vs 0%, respectively; P=.17) and major adverse cardiac events (MACE; 18.5% vs 12.5%, respectively; P=.40). Patients classified as mid GRC (n = 175) had similar cardiac death (11.1% vs 10.3%; P=.85) and MACE rates (20.7% vs 22.4%; P=.92) with CABG or PCI, respectively. Patients with high GRC (n = 81) treated with CABG had numerically fewer cardiac deaths (16.3% vs 28.1%; P=.16) and lower MACE rates (24.5% vs 40.6%; P=.048) than with PCI. Statistical models using the GRC as a predictor of cardiac death showed better goodness-of-fit than the SS. Conclusion. Patients with low/mid GRC have similar mid-term outcomes with either CABG or PCI; patients with high GRC seem to benefit from CABG. Although further investigations are required, GRC is a better predictor of outcomes than SS. 

J INVASIVE CARDIOL 2013;25(12):650-658

Key words: unprotected left main stenosis, global risk classification, percutaneous coronary intervention, coronary artery bypass graft

_________________________

Unprotected left main stenosis (ULMS) has been classically associated with unfavorable clinical outcomes when not revascularized.¹ Coronary artery bypass graft (CABG) has been shown to significantly reduce rates of major adverse cardiac event (MACE) rates compared to medical treatment;^2,3 today, CABG is still the standard of care for patients with ULM stenosis.^4,5 

Percutaneous coronary intervention (PCI) with drug-eluting stent (DES) has become an alternative revascularization therapy in selected patients with ULMS suitable for both revascularization therapies.^4-6 Current guidelines only use anatomical criteria in order to determine which patients benefit the most from each revascularization modality.^4,5 These criteria include: isolated ULMS, ULMS plus single- or double-vessel disease, or low-mid SYNTAX score (<33).^4,5 However, none of these criteria take into account the patient’s clinical condition.^4,5 

The global risk classification (GRC) is a novel score system combining the clinical surgical additive EuroSCORE with the angiographic SYNTAX score (Table 1 shows the GRC algorithm). Recent publications suggest that the GRC is better able to discriminate and predict outcomes in patients with ULMS undergoing either CABG or PCI.^7,8 Nevertheless, the ability of this score system to predict outcomes has only been tested in two different populations and there are few data suggesting that the GRC can predict what type of revascularization therapy is best in those patients. 

The aims of this study are: (1) to assess the clinical outcomes of patients undergoing ULMS revascularization either with CABG or PCI according to the GRC; and (2) to assess the accuracy of the GRC to predict outcomes in patients undergoing CABG and PCI for ULMS.  

Methods 

Population. This study included all consecutive patients diagnosed with de novo ULMS at our institution from 2004 to 2009 who underwent coronary revascularization with CABG or PCI. Patients referred to coronariography for heart valve disease or ST-elevation myocardial infarction (STEMI) were excluded. Patients undergoing coronariography with cardiogenic shock (systolic blood pressure <90 mm Hg and signs of tissue hypoperfusion) were also excluded. Patients with previous CABG were not included.

All data were prospectively collected in a dedicated database. Additive EuroSCORE was estimated at the time of the coronary revascularization. SYNTAX score was estimated retrospectively by reviewing the angiograms of the diagnostic procedure by two independent, experienced analysts according to the SYNTAX score algorithm.⁹ All patients were regularly followed in the outpatient clinic or by telephone contact. All patients had a minimum follow-up of 3 years. 

Endpoints and definitions. The primary endpoint was cardiac death. The secondary endpoints included all-cause death, myocardial infarction (MI), target vessel revascularization (TVR), and MACE. Stent thrombosis was also investigated for the PCI population according to the Academic Research Consortium (ARC) definition.¹⁰ 

Cardiac death was defined as all deaths caused by sudden death, MI, or heart failure. All deaths with unexplained cause were also defined as cardiac. Major adverse cardiac events were defined as the composite of cardiac death, MI, and TVR. Myocardial infarction was defined as an increase of troponin or CK-MB values above the upper normal limits associated with electrocardiographic changes of myocardial ischemia or chest pain.¹⁰Target vessel revascularization (in cases of PCI) was defined as any repeat revascularization with PCI or CABG due to in-stent or in-segment restenosis with ≥50% diameter stenosis (including the first 5 mm of the ostial left anterior descending or circumflex arteries in cases of bifurcation treatment). Target vessel revascularization (in cases of CABG) was defined as any repeat revascularization with PCI or CABG to the left coronary tree due to graft stenosis (≥50%). Urgent indication was considered in cases of patients with acute coronary syndrome with refractory angina despite full medical treatment or intraaortic balloon pump. 

Statistical analysis. Continuous variables were expressed as mean ± 1 standard deviation (SD). Categorical variables were expressed as count (percentages). Comparisons of baseline clinical and angiographic characteristics between groups were performed with Chi-square test or Student’s t-test, as appropriate. Comparisons of clinical outcomes between groups were performed with log-rank test and data are presented with Kaplan-Meier survival curves. The accuracy of the GRC score to predict cardiac death was investigated with logistic regression for each cohort (CABG and PCI). The additive EuroSCORE and the SYNTAX score values where included in the model as covariates. The GRC score performance to predict cardiac death and MACE was evaluated with the Brier test and with the area under the receiver-operator characteristic curves (AUCs). All P-values were 2-tailed, with statistical significance set at a level of <.05. Statistical analysis was performed with the SPSS 20.0 for Mac (SPSS, Inc). Brier test was estimated manually. 

Results 

Population. A total of 638 patients were diagnosed of ULMS at our institution during the study period. After the exclusion of all patients with exclusion criteria, a total of 413 patients were included (290 patients treated with CABG and 123 with PCI). Figure 1 shows the study flow chart. Six patients (1.5%) were lost to follow-up and were also excluded from the study (5 treated with CABG and 1 treated with PCI). Finally, 285 patients treated with CABG and 122 patients treated with PCI were included.

The GRC score at the time of diagnosis allowed us to divide our cohort of patients into 3 groups: 151 patients (37.1%) were classified into the low-GRC group (119 CABG and 32 PCI), 175 patients (43.0%) were classified into the mid-GRC group (117 CABG and 58 PCI) and 81 patients (19.9%) into the high-GRC group (49 CABG and 32 PCI). 

Baseline clinical and angiographic characteristics. Table 2 shows the clinical and angiographic characteristics of patients included in the study according to the GRC score. In summary, patients treated with CABG vs those treated with PCI were younger (65.7 years vs 70.5 years; P<.001), were less often female (16.1% vs 24.6%; P=.045), had lower number of patients with renal dysfunction (20.4% vs 39.3%; P<.001), had lower number of patients with left ventricular dysfunction (18.6% vs 29.5%; P=.03), and had lower additive EuroSCORE value (5.1 vs 7.0; P<.001). 

Table 3 shows the procedural characteristics. In summary, in the CABG group, the number of grafts was 2.52 ± 0.73 per patient. The internal mammary artery was used to bypass the LAD artery in 89.5% of patients. In the PCI group, a total of 89.3% of patients were treated with DES and 16.4% of patients were treated with a two-stent technique for left-main bifurcation. 

Outcomes. Figure 2 shows the survival-free curves of all patients treated with CABG and PCI. Table 4 summarizes the clinical outcomes according to the revascularization strategy and to the GRC. At 3 years, low-GRC patients treated with CABG vs those treated with PCI had similar mortality (5.9% vs 0%; P=.17) and MACE rates (18.5% vs 12.5%; P=.40). Mid-GRC patients treated with CABG vs those treated with PCI had similar cardiac death (11.1% vs 10.3%; P=.85) and MACE (20.7% vs 22.4%; P=.92). High-GRC patients treated with CABG vs those treated with PCI presented with numerically fewer cardiac deaths (16.3% vs 28.1%; P=.16) and lower MACE rates (24.5% vs 40.6%; P=.048).

At 3 years, TVR was observed in 4.9% of patients treated with CABG and 8.2% of patients treated with PCI, which was not statistically significant (P=.17). There were no differences between CABG and PCI patients when taking into account the GRC groups. In the PCI population, there was no definite stent thrombosis. Three cases (2.5%) had probable stent thrombosis (2 patients in the mid-GRC group and 1 patient in the high-GRC group).

Predictive scores. Figure 5 shows the scatter plot of the AUCs and the Brier test results derived from the GRC, additive Euro-SCORE, and SYNTAX score in both groups. The GRC had the highest goodness-of-fit to predict cardiac death and MACE compared to additive EuroSCORE and SYNTAX score. 

Discussion

The main findings of the study are: (1) patients with low or mid GRC have similar outcomes at 3 years when treated with CABG or PCI; (2) patients with high GRC have better outcomes at 3 years with CABG than with PCI; and (3) the clinical and anatomic GRC offers a better prediction of cardiac death than the anatomic SYNTAX score either in patients treated with CABG or PCI. 

The randomized arm of the SYNTAX trial with ULMS showed no differences in mortality (8.4% vs 7.3%; P=.64) and MACE (22.3% vs 26.8%; P=.20) between CABG and PCI patients at 3 years, respectively.¹¹ However, patients with a high SYNTAX score (>33) treated with CABG presented with a lower number of MACE than when treated with PCI (37.3% vs 21.2%; P=.01).¹¹ Patients with low or mid SYNTAX score (0-33) had similar outcomes with both revascularization therapies.¹¹ For this reason, the current revascularization guidelines recommend either CABG or PCI in patients with ULMS and low-mid SYNTAX score, and CABG is the only revascularization therapy recommended in patients with high SYNTAX score.^4,5 

Nevertheless, it is remarkable that the SYNTAX score has only been shown to have a moderate predictive accuracy of cardiac events in patients treated with PCI (C-statistic between 0.650-0.750) and had no predictive effect in patients treated with CABG (C-statistic between 0.500-0.600).^7,8 In both studies, the GRC showed better discrimination than the SYNTAX score either in patients treated with PCI (C-statistic between 0.700 and 0.800) or CABG (C-statistic between 0.600 and 0.700). The present manuscript supports these results, with similar AUC values for the GRC, as previously reported in patients treated with PCI (0.865) and CABG (0.700). 

The addition of clinical data has been shown to increase the accuracy to predict outcomes with other predictive scores as well.¹² The SYNTAX score II found 6 clinical predictors of all-cause death in the SYNTAX trial population at 4 years of clinical follow-up: age, creatinine clearance, left ventricular ejection fraction, peripheral vascular disease, female sex, and chronic obstructive pulmonary disease. It is remarkable that the EuroSCORE already contains most of these predictive factors. Diabetes mellitus is not included in either the EuroSCORE or SYNTAX II score. This is important, since the presence of diabetes mellitus seems not to be associated with all-cause death.^12,13 In our study, a total of 36% of patients had diabetes mellitus and were similarly distributed according to the revascularization therapy and the GRC group. 

In our study, a total of 139 patients (34.2%) had a SYNTAX score >33 and therefore, should have undergone CABG according to the current guidelines. With the use of the GRC as guidance for the revascularization strategy, only 81 patients (19.9%) with high GRC would have undergone surgical revascularization. Therefore, the GRC implies the recruitment of 14.3% of patients suitable for both revascularization strategies. Figure 6 shows the reclassification of patients according to the GRC. In the randomized arm of the SYNTAX trial with ULMS, 40.4% of patients had a SYNTAX score >33.¹⁴ However, with the implementation of the GRC, only 13.3% of patients were classified as high GRC.⁸

In the randomized arm of the left-main SYNTAX trial, including a selected group of patients suitable for both CABG and PCI, patients with low GRC had a higher rate of all-cause mortality (7.5% vs 1.2%; P=.01) and MACE (15.8% vs 23.1%; P=.09) when treated with CABG compared to PCI.⁸ Similarly, different randomized studies and registries including low-risk patients with ULMS have shown higher mortality when treated with CABG than with PCI.^15,16 However, in the all-comers low-GRC population of the SYNTAX trial, including all patients screened for the study, these differences were not observed for all-cause mortality (5.3% vs 2.7%; P=.19) and MACE (18.0% vs 18.5%; P=.94), respectively.⁸ This discrepancy in clinical outcomes between the low-GRC randomized and all-comers populations may be caused by the difference in baseline clinical and angiographic characteristics of patients treated with CABG and PCI. In our study, with important differences in baseline clinical and angiographic characteristics between groups, patients with low or mid GRC had also similar outcomes when treated either with CABG or PCI.

Our study further showed that patients with high GRC treated with CABG had a numerically lower rate of cardiac death (16.3 vs 28.1%; P=.16) and MACE (24.5% vs 40.6%; P=.048) compared to PCI, respectively. These results are similar to those observed in the randomized arm of the SYNTAX trial with no differences in all-cause mortality (18.9% vs 18.6%; P=.92), and lower MACE (28.9% vs 38.5%; P=.29) in patients treated with CABG compared to PCI.⁸ These differences were mainly driven by a higher risk of TVR in the PCI group. Similar studies including a high-risk population also found similar results.^17,18

Study limitations. The present study has several limitations. First, this study is observational. Most of the patients were discussed among the Heart Team, and the revascularization strategy was decided taking into account clinical and anatomic factors, as well as patient preference. For this reason, the clinical and anatomic characteristics of both groups were different. Moreover, due to the low number of patients in each GRC group, no adjustment has been made for these clinical and anatomic differences. Second, CABG and PCI procedures were performed according to the operator’s decision. Although internal mammary artery bypass to the left anterior descending and use of DES for ULMS stenting was the goal in all CABG and PCI procedures, many patients were treated with different techniques and materials. 

Conclusions

Patients with low and mid GRC have similar outcomes when treated with CABG or PCI. However, patients with high GRC have a lower number of cardiac events when treated with CABG instead of PCI. The GRC is a better predictor of clinical outcomes than the merely anatomic SYNTAX score. Although further investigations are required, it seems advisable to use the GRC in the revascularization strategy algorithm for patients with ULMS.  

References

1. Takaro T, Pifarre R, Fish R. Veterans administration cooperative study of medical versus surgical treatment for stable angina — progress report. Section 3. Left main coronary artery disease. Prog Cardiovasc Dis. 1985;28(3):229-234.
2. Caracciolo EA, Davis KB, Sopko G, et al. Comparison of surgical and medical group survival in patients with left main coronary artery disease. Long-term cass experience. Circulation. 1995;91(9):2325-2334.
3. Jones RH, Kesler K, Phillips HR 3rd, et al. Long-term survival benefits of coronary artery bypass grafting and percutaneous transluminal angioplasty in patients with coronary artery disease. J Thorac Cardiovasc Surg. 1996;111(5):1013-1025.
4. Wijns W, Kolh P, Danchin N, et al. Guidelines on myocardial revascularization. Eur Heart J. 2010;31(20):2501-2555.
5. Patel MR, Dehmer GJ, Hirshfeld JW, Smith PK, Spertus JA. ACCF/SCAI/STS/AATS/AHA/ASNC/HFSA/SCCT 2012 appropriate use criteria for coronary revascularization focused update: a report of the american college of cardiology foundation appropriate use criteria task force, society for cardiovascular angiography and interventions, society of thoracic surgeons, american association for thoracic surgery, american heart association, american society of nuclear cardiology, and the society of cardiovascular computed tomography. J Am Coll Cardiol. 2012;59(9):857-881.
6. Lee MS, Yang T, Dhoot J, Liao H. Meta-analysis of clinical studies comparing coronary artery bypass grafting with percutaneous coronary intervention and drug-eluting stents in patients with unprotected left main coronary artery narrowings. Am J Cardiol. 2010;105(8):1070-1075.
7. Capodanno D, Caggegi A, Miano M, et al. Global risk classification and clinical syntax (synergy between percutaneous coronary intervention with taxus and cardiac surgery) score in patients undergoing percutaneous or surgical left main revascularization. JACC Cardiovasc Interv. 2011;4(3):287-297.
8. Serruys PW, Farooq V, Vranckx P, et al. A global risk approach to identify patients with left main or 3-vessel disease who could safely and efficaciously be treated with percutaneous coronary intervention: the SYNTAX trial at 3 years. JACC Cardiovasc Interv. 2012;5(6):606-617.
9. Sianos G, Morel MA, Kappetein AP, et al. The SYNTAX score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention. 2005;1(2):219-227.
10. Cutlip DE, Windecker S, Mehran R, et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation. 2007;115(17):2344-2351.
11. Kappetein AP, Feldman TE, Mack MJ, et al. Comparison of coronary bypass surgery with drug-eluting stenting for the treatment of left main and/or three-vessel disease: 3-year follow-up of the syntax trial. Eur Heart J. 2011;32(17):2125-2134.
12. Farooq V, van Klaveren D, Steyerberg EW, et al. Anatomical and clinical characteristics to guide decision making between coronary artery bypass surgery and percutaneous coronary intervention for individual patients: development and validation of SYNTAX score II. Lancet. 2013;381(9867):639-650.
13. Gomez-Hospital JA, Gomez-Lara J, Rondan J, et al. Long-term follow-up after percutaneous treatment of the unprotected left main stenosis in high risk patients not suitable for bypass surgery. Rev Esp Cardiol (Engl Ed). 2012;65(6):530-537.
14. Morice MC, Serruys PW, Kappetein AP, et al. Outcomes in patients with de novo left main disease treated with either percutaneous coronary intervention using paclitaxel-eluting stents or coronary artery bypass graft treatment in the synergy between percutaneous coronary intervention with taxus and cardiac surgery (SYNTAX) trial. Circulation. 2010;121(24):2645-2653.
15. Boudriot E, Thiele H, Walther T, et al. Randomized comparison of percutaneous coronary intervention with sirolimus-eluting stents versus coronary artery bypass grafting in unprotected left main stem stenosis. J Am Coll Cardiol. 2011;57(5):538-545.
16. Wu X, Chen Y, Liu H, et al. Comparison of long-term (4-year) outcomes of patients with unprotected left main coronary artery narrowing treated with drug-eluting stents versus coronary-artery bypass grafting. Am J Cardiol. 2010;105(12):1728-1734.
17. Palmerini T, Barlocco F, Santarelli A, et al. A comparison between coronary artery bypass grafting surgery and drug eluting stent for the treatment of unprotected left main coronary artery disease in elderly patients (aged > or =75 years). Eur Heart J. 2007;28(22):2714-2719.
18. Kang SH, Park KH, Choi DJ, et al. Coronary artery bypass grafting versus drug-eluting stent implantation for left main coronary artery disease (from a two-center registry). Am J Cardiol. 2010;105(3):343-351.