Clinical SYNTAX score predicts outcomes of patients

Clinical Investigation

Clinical SYNTAX score predicts outcomes of patients undergoing coronary artery bypass grafting Giovanni Melina, MD, PhD, a Emiliano Angeloni, MD, PhD, a Simone Refice, MD, PhD, a Francesco Monti, MD, b Roberto Serdoz, MD, b Stefano Rosato, MS, c Fulvia Seccareccia, MS, c Furio Colivicchi, MD, d Roberta Serdoz, MD, e Francesco Paneni, MD, f,g and Riccardo Sinatra, MD a Roma, Rieti, Italia; University of Zurich, and University Hospital Zurich, Switzerland

Background The SYNTAX score (SS) is a determinant of outcome in patients undergoing percutaneous coronary intervention. In addition, it has been recently shown that the clinical SYNTAX score (cSS), obtained by adding clinical variables to the SS, improves the predictive power of the resulting risk model. We assessed the hypothesis that the use of the cSS may predict outcomes of patients undergoing coronary artery bypass grafting (CABG). Methods We measured the SYNTAX score in 874 patients undergoing isolated first time on-pump CABG. The clinical SYNTAX score was calculated at the time of the study using age, creatinine clearance and ejection fraction, the modified ACEF score, and analyses performed for major adverse cardiac and cerebrovascular events (MACCE) and all-cause mortality at 3-year follow-up. Results

The mean age of the study population was 70.9 ± 8.1 years, and the median cSS 14.2 (range 2.1–286.5). The ROC curve analysis showed that a cSS N14.5 (81.4% sensitivity and 67.8% specificity) was a reliable tool in discrimination of patients for the occurrence of MACCE (AUC 0.78) and all-cause mortality (AUC 0.74). Kaplan-Meier survival analysis confirmed that patients belonging to higher cSS quartiles have poorer 3-year survival (P = .0001) and MACCE-free survival (P = .0001), with respect to those with lower cSS.

Conclusions This observational study has shown that the clinical SYNTAX score, incorporating the lesion-based SS and clinical-based ACEF score, predicted mid-term adverse outcomes of patients undergoing CABG and may play an important role in the risk stratification of this population. Further studies are needed to confirm these findings. (Am Heart J 2017;188:118-26.)

Currently, cardiovascular disease is still the leading cause of death in the United States and Europe. 1-3 Coronary artery disease (CAD) is the most common type of cardiovascular disease and despite the good results achieved by preventive measures, projections show that in less than 10 years its prevalence may increase by 16.6%. 3 It is then expected that revascularization procedures (coronary artery bypass grafting

From the aDepartment of Cardiac Surgery, Ospedale Sant'Andrea, Roma, Italia, b Department of Cardiology, Ospedale San Pietro Fatebenefratelli, Roma, Italia, cNational Centre for Epidemiology, Surveillance and Health Promotion, Istituto Superiore di Sanità, d Roma, Italia, Department of Cardiology, Ospedale San Filippo Neri, Roma, Italia, e Catheterization Laboratory, Ospedale San Camillo De Lellis, Rieti, Italia, fCenter for Molecular Cardiology, University of Zurich, Switzerland, and gUniversity Heart Center, Cardiology, University Hospital Zurich, Switzerland. Submitted October 29, 2016; accepted March 22, 2017. Reprint requests: Giovanni Melina, MD, PhD, Department of Cardiac Surgery, Ospedale Sant'Andrea, Via di Grottarossa, 1035, Roma 00189, Italia. E-mail: [email protected] 0002-8703 © 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2017.03.016

(CABG) or percutaneous coronary intervention (PCI)) to treat such an increasing population may rise as well. Therefore, risk stratification is an important and essential component in appropriating informing patients and may help clinicians to guide decision-making in this setting. 4,5 Within cardiac surgical practice, the use of risk models to stratify patients is well-established, and they are predominantly related to clinical variables alone, with scores such as the EuroSCORE 6,7 or the Society of Thoracic Surgeons (STS) Score 8 being in widespread contemporary use. The SYNTAX score (SS) is an anatomical tool that was originally designed to establish the optimal revascularization approach in patients with complex CAD. 9 Within the surgical population, it has not been shown to provide any additional predictive benefit compared to the clinical scores. 10 Conversely, within interventional cardiology practice, the SYNTAX score has emerged as a useful risk model in identifying a subset of patients who could be safely and efficaciously treated with PCI. 10 However, since clinical factors are not being taken into account in the risk stratification of these patients, potentially

American Heart Journal Volume 188

important prognostic and morbidity information might be missing. 10 We and others have previously shown that a high complexity CAD (assessed by means of SS) predicts worse outcomes in patients undergoing surgical revascularization with left ventricular dysfunction, 11 or off-pump CABG, 12 but these investigations, conducted among specific patients subsets, did not consider anatomical and clinical factors combined into a single score. Recently, Girasis et al. have developed a more accurate tool to stratify risk of patients undergoing PCI integrating the SYNTAX score with a clinical scoring model based upon age, creatinine clearance and left ventricular ejection fraction (LVEF) known as the ACEF score, 13 obtaining the clinical SYNTAX Score (cSS). 14 This was found to accurately predict the risk of adverse events in the long-term outcome of patients undergoing PCI. 14 In addition, most data addressing the concept of combined anatomical-based risk score with clinical variables in predicting adverse events come from series of patients undergoing interventional cardiology procedures. Since there are few data on the CABG population, aim of the present study was to investigate the predictive power of cSS with regards to adverse outcomes following surgical myocardial revascularization.

Methods We studied a consecutive series of patients undergoing first-time elective isolated on-pump CABG at one Institution from May 2004 till May 2013, for whom complete 3-year follow-up was available. No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents. All data were prospectively collected in an electronic database. Eligible patients had one or more of the following American College of Cardiology/American Heart Association class I indications for isolated CABG surgery: (1) left main stenosis N50%; (2) left main equivalent disease (proximal left anterior descending coronary artery [LAD] and left circumflex stenosis N70%); (3) 3-vessel disease; 4) proximal LAD N50% and left ventricular ejection fraction b50%; (5) 2- or 3-vessel disease and LVEF b50%; or (6) 2-vessel disease including proximal LAD with either angina or demonstrable ischemia on stress testing. The only exclusion criterion was the need for combined procedures. Severe left ventricular dysfunction was not an exclusion criterion. Prior to surgery, LVEF was calculated in all patients by echocardiography using the Simpson's biplane method, by means of a Sonos 7500 systems (Philips Medical Ultrasound). All patients after a full median sternotomy underwent on-pump CABG using standard techniques. The left internal mammary artery was used to graft the left anterior descending artery and revascularization completed using saphenous vein grafts to the right coronary

Melina et al 119

and left circumflex artery segments. The right internal mammary artery was used in patients aged less than 60 years at surgeon discretion to complete revascularization to the right or left circumflex [territories. Aspirin was prescribed indefinitely for all patients after the operation. Clinical follow-up was completed with routine outpatient clinics. Patients who did not present at the visit were contacted by telephone and all symptoms, mortality and any complications that occurred during follow-up were recorded. Major adverse cardiovascular events (MACCE) were defined as a composite of death, cerebrovascular accident, any revascularization (percutaneous or surgical) and myocardial infarction at 3-year follow-up. At the time of the study, the SYNTAX score algorithm, which is described in full elsewhere 15 and is available on the SYNTAX score website (www.syntaxscore.com), was employed to score all coronary lesions deemed to have a percentage diameter stenosis ≥50%, in vessels ≥1.5 mm. All angiographic variables pertinent to SYNTAX score calculation were computed by two experienced interventional cardiologists on diagnostic angiograms obtained before the procedure. In case of disagreement, the opinion of a third analyst was obtained and the final decision was made by consensus. Analysts were blinded to procedural data and clinical outcome. The final score was calculated on a patient basis from the individual lesion scores, which were saved in a dedicated database, and was not made available to the analysts until after the completion of the study. The inter-observer variability for calculation of the SYNTAX score - based on reanalyzing 50 cases at 6-month interval by a fourth interventional cardiologist - showed excellent reproducibility (kappa statistics = 0.84). At the time of the study, the modified ACEF score was calculated according to a previously described method, 13 based on the patients' left ventricular ejection fraction, age, and creatinine clearance derived using the Cockcroft-Gault equation. 16 The clinical SYNTAX score was then calculated multiplying the value of the SYNTAX score by the modified ACEF score. Both SS and cSS were calculated on the basis of pre-surgery data by those unaware of post-surgery outcomes. The statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS Inc, Chicago, IL), version 11.0. Continuous variables are presented as mean ± 1 SD or median values (25th-75th percentile) as appropriate; categorical variables are displayed as counts and percentages. Complete follow-up data were available for all patients at 3 years postoperatively, which was chosen as the fixed-time of follow-up. Calibration and discrimination of cSS predicting 3-year mortality and MACCE was explored with calculation of Hosmer-Lemesow (H-L) goodness of fit test and of the areas under the receiver-operating characteristics curves (ROC); an area (AUC) of 1.0 would indicate perfect discrimination, whereas an area of 0.5 indicates

American Heart Journal June 2017

120 Melina et al

the total absence of discriminatory power. As the highest accuracy points of those ROC curves fall around the 50th percentile, quartiles distribution was then used to identify subgroups of cSS. Areas-under-the-curves (AUCs) for SS, cSS, EuroSCORE II and ACEF score were compared by the DeLong test 17 using MedCalc Statistical Software version 15.11.0 (MedCalc Software bvba, Ostend, Belgium). Cumulative 3-year event rates were estimated by the Kaplan–Meier method and the comparison between cSS quartiles was made using the log-rank test. The survival time of a patient started at the time of surgery and ended at death/MACCE (event) or after 3 years of follow-up (censoring). All statistical tests were two-sided and a P b .05 was considered statistically significant. Finally, in order to formally assess whether cSS improved the risk stratification over the ACEF score, a net reclassification improvement (NRI) analysis was performed. 18

Results The baseline characteristics of the entire sample are reported in Table I. The study population consisted of 874 patients (mean age: 70.9 ± 8.1 years), with 86% (754/ 874) being male. The clinical SYNTAX score ranged from 2.1 to 286.5 with a median of 14.2 (25th percentile 5.6, 75th percentile 42.1). Overall, 51 (5.8%) of 874 patients did not receive complete revascularization (defined as if each significantly diseased territory received at least one graft). Reasons for incomplete revascularization (N = 51) were small vessel diameter in 60.8% (31/51), severely calcified target vessel in 25.5% (13/51), vessel not found in 5.9% (3/51), other in 7.8% (4/51). A total of 237/874 patients (27.1%) received a bilateral internal mammary artery, and 30-day mortality was 3.0% (26/874 patients). Mortality and morbidity rates at 3-year follow-up stratified for cSS quartiles are reported in Table II. At 3-year follow-up, both SS and cSS showed a good predictive power with respect to all-cause mortality and MACCE (Table III). Moreover, standing on ROC curve analysis, the highest accuracy point for cSS in predicting 3-year mortality was a score N19.5 points (73.4% sensitivity and 65.5% specificity); and the highest accuracy point for cSS in predicting 3-year MACCE was a score N14.5 points (81.4% sensitivity and 67.8% specificity). Given that both the latter cut-off values fall above the 50th percentile of study group distribution, we divided our sample in quartiles to further investigate follow-up end-points. The 25th, 50th, and 75th percentile used for quartiles stratification of cSS were 5.6, 14.5, and 42.1, respectively. Kaplan–Meier analysis showed 3-year survival rates of 89.1 ± 3.1%, 89.3 ± 4.2%, 67.4 ± 5.2%, and 54.5 ± 4.8%, respectively for cSS quartiles 1, 2, 3, and 4 (log-rank P b .0001; Figure 1). Rates of 3-year MACCE-free survival were 83.3 ± 4.6%, 83.7 ± 5.5%, 56.9 ± 4.2%, and 41.8 ±

Table I. Baseline characteristics of the study population (N = 874) VARIABLE

N

%

Male gender Diabetes mellitus Dyslipidemia Hypertension Chronic obstructive pulmonary disease Smoking history Unstable angina Previous cerebrovascular accident Previous myocardial infarction Previous percutaneous coronary intervention Preoperative drugs Beta-blockers ACE-inhibitors Nitrates Aspirin NYHA class I II III IV Complete revascularization 3-Vessel disease Left main disease 30-day mortality

754 318 637 772 113 506 650 100 505 141

86.3 36.4 72.9 88.3 12.9 57.9 74.4 11.4 57.8 16.1

567 441 274 494

64.9 50.5 31.4 56.5

233 409 213 19 823 689 264 26

26.7 46.8 24.4 2.2 94.2 78.8 30.2 3.0

Variable

Mean

SD

Age Left ventricular ejection fraction (%) Creatinine clearance (mL/min) Cardiopulmonary bypass time (min) Cross-clamp time (min) Mean number of arterial anastomoses Mean number of total anastomoses SYNTAX score ACEF score Clinical SYNTAX score EuroSCORE II

70.9 47.7 74.7 110.7 80.6 1.4 2.6 12.5 2.1 32.5 3.6

8.1 14.0 27.3 59.2 36.0 0.6 0.9 11.3 1.4 23.7 4.1

Categorical variables are presented as absolute number and percentage, continuous variables as mean and standard deviation (SD). NYHA: New York Heart Association.

6.8%, respectively for cSS quartiles 1, 2, 3, and 4 (Log rank P b .0001; Figure 2). In addition, comparison of ROC curves (Figures 3, and 4) demonstrated a statistically significant superiority of cSS in predicting both mortality and MACCE with respect to either SS or EuroSCORE II (Table III). Highest accuracy point of cSS and ACEF score predicting 3-year mortality and MACCE were used to perform reclassification analysis between the latter scores, which showed larger AUCs. With respect to 3-year mortality, net gain in reclassification was 15.6% among 218/874 deaths, whereas among 620/874 survivors net gain was −0.3%. Therefore, the NRI for cSS over ACEF in predicting 3-year mortality was estimated to be 15.3%. With respect to 3-year MACCE, net gain in reclassification was 18.7% among 294/874 patients with MACCE, whereas among 580 of 874 patients without MACCE net


Melina et al 121

Table II. Outcomes incidence at 3-year follow-up stratified for clinical SYNTAX score (cSS) quartiles

cSS range Sample size Mortality MACCE

Q1

Q2

Q3

Q4

TOTAL

P

cSS ≤ 5.6 217 16 (7.4%) 17 (5.8%)

5.6 b cSS ≤ 14.5 221 24 (10.9%) 46 (15.7%)

14.5 b cSS ≤ 42.1 220 81 (36.8%) 102 (34.7%)

cSSN 42.1 216 97 (44.9%) 123 (41.8%)

2.1–286.5 874 218 (24.9%) 294 (33.6%)

n/a n/a b.0001 b.0001

MACCE: Major adverse cardiac and cerebrovascular events.

Table III. Comparison of receiver operating characteristic (ROC) curves for different scoring systems predicting 3-year major adverse cardiac and cerebrovascular events (MACCE) and all-cause mortality 3-year MACCE ROC Model

AUC

Clinical SYNTAX score ACEF score SYNTAX score EuroSCORE II

0.769 0.742 0.733 0.605

H-L P

95% Wald CL 0.76 0.71 0.70 0.57

0.80 0.77 0.76 0.64

Comparison

Difference

95% Wald CL

Clinical SYNTAX score vs. ACEF score Clinical SYNTAX score vs. SYNTAX score Clinical SYNTAX score vs. EuroSCORE II

+0.027 +0.036 +0.164

−0.0001 0.01 0.11

.23 .27 .31 .30

0.05 0.07 0.22

χ2

P

9.4 19.7 48.4

.19 b.0001 b.0001

3-year mortality ROC model

AUC

Clinical SYNTAX score ACEF score SYNTAX score EuroSCORE II

0.731 0.710 0.688 0.609

H-L P

95% Wald CL 0.70 0.68 0.64 0.58

Comparison

Difference

95% Wald CL

Clinical SYNTAX score vs. ACEF score Clinical SYNTAX score vs. SYNTAX score Clinical SYNTAX score vs. EuroSCORE II

+0.021 +0.043 +0.122

−0.01 0.01 0.08

0.76 0.74 0.70 0.72

0.05 0.08 0.31

.25 .33 .29 .37 χ2

P

6.22 23.05 16.8

0.26 b.0001 .0004

AUC: Area under the curve; CL: Confidence limit; H-L: Hosmer-Lemeshow.

gain was −0.7%. Therefore, the NRI for cSS over ACEF in predicting 3-year MACCE was estimated to be 18.0%.

Discussion The present study has shown that the clinical SYNTAX score, by adding clinical variables to the SYNTAX score, predicts mid-term outcomes of patients with coronary artery disease undergoing surgical myocardial revascularization. Due to the aging population and the increased co-morbidities, 1-3 risk stratification has become a major issue in the preoperative assessment of CABG candidates. Consequently, assessing the appropriate risk for patients

who more often present with multi-vessel CAD, advanced age, and several comorbidities 1-3 is a key issue in guiding the decision-making process to delineate the best treatment strategy. In cardiac surgery, the EuroSCORE 6,7 and STS score 8 have shown good accuracy in predicting postoperative mortality and morbidity, but factors as the degree of native CAD complexity quantification are not included in such risk scores. The SYNTAX score is an angiographic method to objectively measure the complexity of coronary artery disease. 9 In the context of the SYNTAX trial, 19 it has been found to significantly predict outcomes following PCI, but to date, has failed to demonstrate a similar relationship

122 Melina et al


Figure 1

Kaplan-Meier analysis of survival stratified for quartiles of clinical SYNTAX Score. Q1, Q2, Q3, and Q4 are the subgroups of patients in the first, second, third, and fourth quartiles of the cSS distribution.

Figure 2

Kaplan-Meier analysis of MACCE-free survival stratified for quartiles of clinical SYNTAX Score. Q1, Q2, Q3, and Q4 are the subgroups of patients in the first, second, third, and fourth quartiles of the cSS distribution.


Melina et al 123

Figure 3

Comparison of ROC curves analyses of the predictive power of cSS, SS, ACEF Score, and EuroSCORE II with respect to all-cause 3-year mortality.

among CABG surgery patients, 20 even though this has been demonstrated for specific subgroups. 11,12 Vice-versa, the so-called “parsimonious” ACEF score, a simple method to stratify risk using only age, creatinine clearance and left ventricular ejection fraction in patients undergoing cardiac operations, 13 does not take into account the anatomical component of native coronary artery disease. In a recent study, Girasis et al. 14 have demonstrated that the clinical SYNTAX score, obtained by adding clinical and anatomical based scores, 21 translated into more refined risk stratification after PCI. 14 Due to the lack of data on surgical series, we have formally designed the present study to investigate the predictive power of the clinical SYNTAX score with regards to adverse outcomes following CABG. The SYNTAX score in the present investigation was related to the outcomes investigated. Recent studies using the SYNTAX score report different findings about the capacity of the SYNTAX score to predict adverse events among patients undergoing CABG. 15,22-24 Potential explanations include different therapies, limited follow-up period, and the inter-observer and intra-observer variability of the SYNTAX score may explain different conclusions. 24

To further assess the variability of the SYNTAX score calculation, we have reanalyzed a sample of angiograms and found an excellent inter-observer reproducibility. In addition, the mean SS here was substantially lower than in the original prospective SYNTAX trial. We believe that this can be partially explained by the different inclusion criteria for the original study (3-vessel and/or left main coronary artery disease) as compared to our retrospective investigation, which included all patients with coronary artery disease undergoing isolated CABG. The main finding of the present investigation was that combining the evaluation of coronary disease (assessed by means of the SYNTAX score) and the clinical characteristics of the patient (as with the ACEF score), one could obtain a reliable and accurate tool to predict and stratify the risk of postoperative mortality and morbidity. Therefore, the resulting clinical SYNTAX score, which accounts for the quality of the coronary vessels, definitively completes the preoperative evaluation also among surgical series of CAD patients. Indeed, we have demonstrated the association of high clinical SYNTAX scores with progressive increase of cardiovascular morbidity and all-cause

124 Melina et al


Figure 4

Comparison of ROC curves analyses of the predictive power of cSS, SS, ACEF Score, and EuroSCORE II with respect to 3-year MACCE.

mortality, and statistical analysis showed a significantly increased incidence of adverse outcomes among patients with a cSS N14.5 at 3-year follow-up. Long-term adverse events are expected to be dependent on well-known predictors of outcomes following CABG, such as age and diabetes. Age is included in the cSS, while diabetes mellitus is known to impact on renal function and cardiovascular outcomes but it is not considered in the cSS. In fact, as previously shown, 13,25-29 the end organ manifestation of diabetes affected long-term mortality in CABG and PCI populations, and not the actual presence of the risk factor. In a recent paper, Shahian et al made a study from real-world clinical data to construct a robust long-term CABG survival prediction calculator. 30 This large investigation, the ASCERT study, included 348,341 patients aged ≥65 years that underwent isolated CABG between 2002 and 2007. Even if we could not apply such a model due to a proportion of patients aged b65 in our study population, they found similar rates of mortality both early (3.2% at 30 days) and late (23.3% at 3 years). 30 In this study, they have used a vast number of clinical variables

from the STS adult cardiac surgery database and vital status from Medicare claims. 30 Intuitively, the use of multiple clinical variables should improve the accuracy of a risk model. However, this accuracy may ultimately be contaminated by the desire to create “the perfect model”. 21,31 In practice, Ranucci et al 13,26,27 developed a simple risk model consisting of only 3 clinical variables (the ACEF score), for assessing risk in elective cardiac operations. This was based on the “law of parsimony” or “the Ockham razor” concept, whereby a simple model can explain a phenomenon with the same level of accuracy as complex models. In addition, our investigation has shown that the clinical SYNTAX score predicted outcomes better than EuroSCORE II or the SYNTAX and ACEF scores separately. We cannot exclude that the poor performance of EuroSCORE II may be due to the long-term follow-up of the study population regarding mortality—whilst the EuroSCORE is designed for predicting 30-day mortality—and that it is not commonly used to predict the risk of major adverse cardiovascular events. Finally, a major strength of our analysis relies on the fact that the ACEF score was developed and validated in


patients undergoing CABG. In the present investigation, the ACEF score was indeed appropriately used in a surgical population to measure the clinical SYNTAX score, which has been previously applied to analyze PCI series of patients. 14,21 However, the present study was designed to test whether—among CABG patients— clinical SS may stratify risk of cardiovascular events overtime. Our findings should be interpreted only in the context of a series of patients who underwent CABG surgery. We do not presume to extend our SS cut-off to all CAD patients. Although our study provides a new tool of measuring the risk for patients undergoing CABG, there are some limitations. First, this is a retrospective study of a relatively small population of patients, and therefore unidentified confounders may have influenced the analysis. Second, the SYNTAX score, which is used to build the clinical SS, predominately provides an assessment of lesion characteristics, consistent with its purpose to assess risk in PCI, which of course is a treatment to the lesion. The SYNTAX score does not include extensive consideration of the quality of the vessel distal to the important lesion. This part of the artery is the target for the distal anastomosis in CABG and perhaps not well assessed by the SYNTAX score and therefore data could be not completely accurate. Third, missing data regarding the percentage of patients receiving appropriate secondary prevention medications and/or the degree of compliance may have affected results. Finally, comparisons to other risk calculators, such as the STS score, would definitely add more information, but due to the retrospective nature of the present study, data for its calculation were not available for all patients. Further ad-hoc studies are warranted to better characterize the individual performance of different risk scores.

Conflicts of interest None.

References 1. Sidney S, Quesenberry Jr CP, Jaffe MG, et al. Recent Trends in Cardiovascular Mortality in the United States and Public Health Goals. JAMA Cardiol 2016;1:594-9. 2. Nichols M, Townsend N, Scarborough P, et al. Cardiovascular disease in Europe 2014: epidemiological update. Eur Heart J 2014;35:2950-9. 3. Heidenreich PA, Trogdon JG, Khavjou OA, et al. Forecasting the future of cardiovascular disease in the United States. A policy statement from the American Heart Association. Circulation 2011;123:933-44. 4. 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;381:639-50.

Melina et al 125

5. Weisel RD, Nussmeier N, Newman MF, et al. Predictors of contemporary coronary artery bypass grafting outcomes. J Thorac Cardiovasc Surg 2014;148:2720-6. 6. Nashef SA, Roques F, Michel P, et al. European system for cardiac operative risk evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9-13. 7. Nashef SA, Roques F, Sharples LD, et al. EuroSCORE II. Eur J Cardiothorac Surg 2012;41:734-45. 8. Shahian DM, O'Brien SM, Filardo G, et al. Society of Thoracic Surgeons Quality Measurement Task Force.: The Society of Thoracic Surgeons 2008 cardiac surgery risk models: Part 1- Coronary artery bypass grafting surgery. Ann Thorac Surg 2009;88:S2-22. 9. Sianos G, Morel MA, Kappetein AP, et al. The SYNTAX Score: An angiographic tool grading the complexity of coronary artery disease. EuroIntervention 2005;1:219-27. 10. Farooq V, Brugaletta S, Serruys PW. The SYNTAX score and SYNTAX-based clinical risk scores. Semin Thorac Cardiovasc Surg 2011;23:99-105. 11. Melina G, Angeloni E, Benedetto U, et al. Complexity of coronary artery disease affects outcome of patients undergoing coronary artery bypass grafting with impaired left ventricular function. J Thorac Cardiovasc Surg 2013;146:656-61. 12. Carnero-Alcázar M, Maroto Castellanos LC, Silva Guisasola JA, et al. SYNTAX score is associated with worse outcomes after off-pump coronary artery bypass grafting surgery for three-vessel or left main complex coronary disease. J Thorac Cardiovasc Surg 2011;142: e123-32. 13. Ranucci M, Castelvecchio S, Menicanti L, et al. Risk of assessing mortality in elective cardiac operations: age, creatinine, ejection fraction, and the law of parsimony. Circulation 2009;119:3053-61. 14. Girasis C, Garg S, Räber L, et al. SYNTAX score and Clinical SYNTAX score as predictors of very long-term clinical outcomes in patients undergoing percutaneous coronary interventions: a substudy of SIRolimus-eluting stent compared with pacliTAXel-eluting stent for coronary revascularization (SIRTAX) trial. Eur Heart J 2011;32: 3115-27. 15. Mohr FW, Rastan AJ, Serruys PW, et al. Complex coronary anatomy in coronary artery bypass graft surgery: Impact of complex coronary anatomy in modern bypass surgery? Lessons learned from the SYNTAX trial after two years. J Thorac Cardiovasc Surg 2011;141:130-40. 16. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16:31-41. 17. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing areas under two or more correlated receiver operating characteristics curves: a nonparamentric approach. Biometrics 1988;44:837-45. 18. Pencina MJ, D'Agostino Sr RB, D'Agostino Jr RB, et al. Evaluating the added predictive ability of a new marker: from area under the curve to reclassification and beyond. Stat Med 2008;27:157-72. 19. Ong AT, Serruys PW, Mohr FW, et al. The SYnergy between percutaneous coronary intervention with TAXus and cardiac surgery (SYNTAX) study: design, rationale, and sign-in phase. Am Heart J 2006;151:1194-204. 20. Serruys PW, Morice MC, Kappetein AP, et al. Mohr FW; SYNTAX Investigators. SYNTAX Investigators. Percutaneous coronary intervention versus coronary artery bypass grafting for severe coronary artery disease. N Engl J Med 2009;360:961-72. 21. Garg S, Sarno G, Garcia-Garcia HM, et al. A new tool for the risk stratification of patients with complex coronary artery disease: the Clinical SYNTAX score. Circ Cardiovasc Interv 2010;3:317-26. 22. Birim O, van Gameren M, Bogers AJ, et al. Complexity of coronary vasculature predicts outcome of surgery for left main disease. Ann Thorac Surg 2009;87:1097-105.

126 Melina et al

23. Lemesle G, Bonello L, de Labriolle A, et al. Prognostic value of the Syntax score in patients undergoing coronary artery bypass grafting for three-vessel coronary artery disease. Catheter Cardiovasc Interv 2009;73:612-7. 24. Ikeno F, Brooks MM, Nakagawa K, et al. BARI-2D Study Group. SYNTAX Score and Long-Term Outcomes: The BARI-2D Trial. J Am Coll Cardiol 2017;69:395-403. 25. Farooq V, Head SJ, Kappetein AP, et al. Widening clinical applications of the SYNTAX Score. Heart 2014;100:276-87. 26. Ranucci M, Castelvecchio S, Conte M, et al. The easier, the better: age, creatinine, ejection fraction score for operative mortality risk stratification in a series of 29,659 patients undergoing elective cardiac surgery. J Thorac Cardiovasc Surg 2011;142:581-6. 27. Ranucci M, Castelvecchio S. The ACEF score one year after: a skeleton waiting for muscles, skin, and internal organs. EuroIntervention 2010;6:549-53.


28. Farooq V, Serruys PW, Bourantas C, et al. Incidence and multivariable correlates of long-term mortality in patients treated with surgical or percutaneous revascularization in the synergy between percutaneous coronary intervention with taxus and cardiac surgery (SYNTAX) trial. Eur Heart J 2012;33:3105-13. 29. Farooq V, Vergouwe Y, Räber L, et al. Combined anatomical and clinical factors for the long-term risk stratification of patients undergoing percutaneous coronary intervention: the Logistic Clinical SYNTAX score. Eur Heart J 2012;33:3098-104. 30. Shahian DM, O'Brien SM, Sheng S, et al. Predictors of long-term survival after coronary artery bypass grafting surgery. Results from the Society of Thoracic Surgeons Adult Cardiac Database (The ASCERT Study). Circulation 2012;125:1491-500. 31. Concato J, Feinstein AR, Holford TR. The risk of determining risk with multivariable models. Ann Intern Med 1993;118:201-10.