Key words: residual SYNTAX score; SYNTAX score; unprotected left main coronary artery disease

Transcription

1 CORONARY ARTERY DISEASE Catheterization and Cardiovascular Interventions 82: (2013) Editor s Choice Objectifying the Impact of Incomplete Revascularization by Repeat Angiographic Risk Assessment With the Residual SYNTAX Score After Left Main Coronary Artery Percutaneous Coronary Intervention Davide Capodanno, 1,2 * MD, PHD, Alberto Chisari, 1 MD, Daniele Giacoppo, 1 MD, Salvatore Bonura, 1 MD, Vincenzo Lavanco, 1 MD, Piera Capranzano, 1 MD, Anna Caggegi, 1 MD, Margherita Ministeri, 1 MD, and Corrado Tamburino, 1,2 MD, PHD Objectives: We investigated the prognostic accuracy of a standardized quantification of incomplete revascularization after percutaneous coronary intervention (PCI) of the unprotected left main coronary artery (ULMCA) named residual SYNTAX score (rss). Background: Prognostic implications of coronary lesions left untreated after ULMCA PCI are confounded by the lack of a uniform definition of incomplete revascularization. Methods: Baseline SYNTAX score (bss), rss, and the difference between bss and rss (D SS ) were assessed in predicting the risk of 2-year cardiac mortality of 400 patients undergoing ULMCA PCI. Results: The rss and bss showed comparable discrimination (rss area under the curve [AUC] 0.72, 95% confidence interval [95% CI] ; bss AUC 0.73, 95% CI ). Hosmer Lemeshow statistics were 0.60 for rss (P ) and 2.45 (P ) for bss, reflecting better calibration ability of the rss. The D SS provided the worst discrimination and calibration characteristics (AUC 0.55; 95% CI ; Hosmer Lemeshow statistic 3.13, P ). The rss was independently associated with the 2-year adjusted-risk of cardiac mortality (hazard ratio 1.07, 95% CI , P ). The risk information from both the rss and bss slightly improved the discrimination ability compared with risk information from each single risk assessment (AUC 0.74, 95% CI ) with a net reclassification improvement of 114.2% and 113.6% over rss and bss alone, respectively. Conclusions: The rss carries a prognostic value as independent predictor of 2-year cardiac mortality. Compared with a single assessment of the SYNTAX score, information coming from repeat assessment of the angiographic risk may improve the ability to discriminate and reclassify patients undergoing ULMCA PCI. VC 2013 Wiley Periodicals, Inc. Key words: residual SYNTAX score; SYNTAX score; unprotected left main coronary artery disease INTRODUCTION In patients with unprotected left main coronary artery (ULMCA) disease, percutaneous coronary intervention (PCI) is associated with nonsignificantly different rates of death and myocardial infarction (MI), a lower risk of stroke, and a higher risk of target vessel revascularization compared with coronary artery bypass grafting (CABG) [1]. One of the intuitive differences between PCI and CABG for the treatment of ULMCA concerns the ability to achieve complete anatomical revascularization, particularly in real world scenarios [2,3]. Notably, incomplete anatomical revascularization 1 Cardiovascular Department, Ferrarotto Hospital, Catania, Italy 2 ETNA Foundation, Catania, Italy Conflict of interest: Nothing to report. *Correspondence to: Davide Capodanno, MD, PHD, Cardiology Department, Ferrarotto Hospital, University of Catania, via Citelli 6, Catania, Italy. dcapodanno@gmail.com Received 24 August 2012; Revision accepted 28 August 2012 DOI /ccd Published online 8 April 2013 in Wiley Online Library (wileyonlinelibrary.com) VC 2013 Wiley Periodicals, Inc.

2 334 Capodanno et al. has been linked to mortality [4 6]. In patients with ULMCA plus additional vessels involvement, CABG offers the unique ability to address multiple coronary lesions with relatively few grafts, while protecting the distal coronary circulation if lesions proximal to the site of anastomosis continue to develop. In contrast, only target lesions are addressed during PCI and ULMCA revascularization sometimes diverts attention from other lesions, which may be perceived as technically demanding or clinically less important. Whether incomplete revascularization after ULMCA PCI carries a prognostic implication has been poorly explored. Recently, systematic characterization and quantification of residual atherosclerosis after multivessel PCI by means of the so-called residual SYNTAX score (rss) have been proposed [7]. Different from the baseline SYNTAX score (bss), a semi-quantitative angiographic model aimed at characterizing the individual coronary severity and complexity [8], the rss could be calculated after PCI to provide an objective and potentially useful quantification of incomplete revascularization. Whether this approach has a prognostic value in the setting of ULMCA PCI is currently undefined. was obtained and the final decision was made by consensus. All data were assessed for quality and entered into a dedicated computerized database. In the case of staged PCI procedures, the final planned procedure was used as the entry point for this study. Follow Up, End Points, and Definitions Information about in-hospital outcome was obtained from an electronic centralized clinical database. After discharge, all clinical follow-up data were prospectively collected by scheduled clinic evaluations or direct telephone interviews. Referring cardiologists, general practitioners, and patients were contacted whenever necessary for further information. All repeat coronary intervention and re-hospitalization data were prospectively collected during follow-up and entered into the centralized computer system of our institution or by directly contacting the hospitals where the patients were admitted or referred. Clinical events were adjudicated by an independent clinical events committee. The primary objective was the incidence of 2-year cardiac death considered as any death with a demonstrable cardiovascular cause or any death that was not clearly attributable to a non-cardiovascular cause. MATERIALS AND METHODS Patient Population We used data from the CUSTOMIZE (Appraise a CUSTOMIZEd strategy for left main revascularization) registry, a large cohort which was originally set up to investigate the effect of treatment by PCI or CABG in consecutive patients with ULMCA (defined as the presence of lesions with stenosis of at least 50% of vessel diameter) [3]. The analysis was limited to individuals undergoing PCI with complete data for calculating both the bss and the rss. The local ethics committee at each center approved the use of clinical data for this study, and all patients provided written informed consent. The authors are responsible for the completeness and accuracy of data gathering and analysis. SXscore Calculation The bss and rss were derived from the summation of the individual scorings for each lesion (defined as 50% stenosis in vessel 1.5 mm) on angiograms obtained before and after the procedure, respectively, as previously described [7 9]. All angiographic variables pertinent to bss and rss calculation were computed by two of the three experienced cardiologists, trained for SYNTAX score assessment, who were blinded to procedural data and clinical outcome. In case of disagreement, the opinion of the third observer Statistical Analysis A classification tree procedure was used to identify the optimal predictive cut off of rss in terms of cardiac mortality, as previously described [10,11]. Briefly, this method involved the segregation of different values of classification variables through a decision tree composed of progressive binary splits. Every value of each predictor variable was considered as a potential split, and the optimal split was selected based on impurity criterion (the reduction in the residual sum of squares due to a binary split of the data at that tree node). This procedure (data not shown) yielded a cutoff value of 8, consistent with Généreux et al [7]. Patients with incomplete revascularization (rss > 0) were therefore grouped into two strata of rss for analysis (rss LOW and rss HIGH ) and compared with those with complete revascularization (rss ¼ 0). For all analyses, a two-sided P < 0.05 was considered statistically significant. All data were processed using the statistical Package for Social Sciences, version 15 (SPSS, Chicago, IL). Continuous variables are presented as mean standard deviations or as median and inter-quartile range, and were compared using Student s unpaired t test or Mann-Whitney Rank Sum test, as appropriate. The normality assumption for continuous variables was evaluated by the Kolmogorov Smirnov test. Categorical variables are presented as counts and percentages and were compared with the chi-

3 Residual SYNTAX Score and LM Revascularization 335 Fig. 1. Nomograms of Models 1 and 2. In Model 1, patients at high risk (H) are those with two criteria of high risk identified by rss and bss; patients at intermediate risk (I) are those with only one criterion of high risk; patients at low risk (L) are those without any criterion of high risk. In Model 2, patients at high risk (H) are those with two criteria of high risk identified by rss and D SS ; patients at intermediate risk (I) are those with only one criterion of high risk; patients at low risk (L) are those without any criterion of high risk. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] square test when appropriate (expected frequency >5). Otherwise, the Fisher s exact test was used. The Spearman s test was used to assess the correlation between the rss and the bss. Two-year cumulative rates of cardiac mortality were estimated by the Kaplan Meier method, and the log-rank test was used to evaluate differences between groups. Cardiac mortality rates for groups characterized by different magnitude of change in the SYNTAX score over the procedure (D SS ¼ bss - rss) were also calculated. Although follow-up extended beyond 2 years in a proportion of patients at the time of data analysis, we restricted the follow-up to 2 years in all patients to account for bias introduced by incomplete follow-up. Patients lost to follow-up were considered at risk until the date of last contact, at which point they were censored. In order to adjust for a range of potential confounders, multivariate analysis of independent predictors of cardiac death was performed with a Cox proportional hazard regression model. The assumption of the proportional hazard was verified by a visual examination of the log (minus log) curves and the linearity assumption was assessed by plotting the Martingale residuals against continuous covariates. The variables considered as possible predictors included age, diabetes, left ventricular ejection fraction (LVEF), EuroSCORE, and three-vessel disease as independent control variables and rss (or any different model) as the independent study variable of interest. The selection in the final model was based on a plausible association with the primary end point or a significant P value on univariate analysis. Crude and adjusted hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were reported. The global accuracy of rss, bss, and D SS was evaluated in terms of calibration and discrimination, as previously described [8,12]. Briefly, calibration evaluates the degree of correspondence between the estimated probabilities produced by a model and the actual observation. For each score, it was measured by the Hosmer Lemeshow test. Discrimination is the probability that the score will assign higher values of risk to patients who will go on to have events compared with those who will not. It was measured by using the areas under the Receiver Operating Characteristic curve (AUC), which range from 0.50 (no discrimination) to 1.0 (perfect discrimination) and the index of separation (IoS), defined as P worst - P best, assuming P worst as the predicted P of event for a patient in the group with the highest score and P best as the predicted P of the same event for a patient in the group with the lowest score [8,12]. Net reclassification improvement (NRI) was calculated as previously described [13]. To investigate the added value of using information from both risk assessment, two further models were developed, which combined information from (i) the bss and rss (Model 1) and (ii) the rss and D SS (Model 2) (Fig. 1). This was done by fitting (i) the strata of the bss and rss or (ii) the strata of rss and D SS in a new matrix for prediction of cardiac mortality. We then computed measures of global accuracy, as described above, comparing combined information from the two new models with risk information from the bss and the rss alone. Finally, a measure of incremental prognostic value was obtained by adding into the Cox regression model for rss the bss and D SS strata, as reported in Models 1 and 2 (Fig. 1). A statistically significant increase in the global chi-square of

4 336 Capodanno et al. Fig. 2. Completeness of revascularization stratified by rss and bss. Distribution of completeness of revascularization by rss based on strata of bss. Complete revascularization was less likely to be achieved in the upper risk groups of the bss (P < 0.001). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] the model after the addition of the bss and the D SS defined the incremental prognostic value. RESULTS Patients and Baseline Characteristics Paired rss and bss metrics were available in 400 patients undergoing ULMCA PCI. The rss and the bss showed a strong positive correlation (R ¼ 0.73, P < 0.001). Patients with complete revascularization (rss ¼ 0), rss LOW (rss ¼ 1 8), and rss HIGH (rss > 8) were 48.8%, 22.8%, and 28.5%, respectively. Patients with bss LOW,bSS INTER- MEDIATE, and bss HIGH based on the historically accepted cutoffs of the SYNTAX score were 45%, 26.3%, and 28.7%, respectively. A total of 131 patients (32.8%) were differently categorized on the basis of the two scores. PCI yielded a mean decrease in SYNTAX score from bss to rss of 20 8 with a significant gradient across bss strata (15 5, 23 5, and 27 8 for patients in the bss LOW, bss INTERMEDIATE, and bss HIGH groups, respectively, P < 0.001). Figure 2 shows the level of completeness of revascularization stratified by rss and bss. Baseline demographic, clinical, and angiographic characteristics of patients stratified by completeness of revascularization are summarized in Table I. Patients with rss HIGH were older and more likely to have diabetes, lower ejection fraction, and higher EuroSCORE compared with patients with rss LOW and those with complete revascularization. Patients with rss HIGH also presented with higher angiographic complexity, reflected by higher prevalence of three-vessel disease involvement, and higher SYNTAX score at baseline. Prediction of 2-Year Cardiac Mortality The cumulative incidences of cardiac mortality at 2 years were 3.3%, 4.5%, and 19.8% in the complete revascularization, rss LOW, and rss HIGH strata, respectively (P < 0.001, Fig. 3), and 3.9%, 0.8%, and 15.9% in the bss LOW, bss INTERMEDIATE, and bss HIGH strata, respectively (P < 0.001). There were no significant differences in 2-year cardiac mortality in increasing tertiles of change in SYNTAX score from pre- to post- PCI (3.9%, 11.1%, and 6.9% for D SS-LOW, D SS-INTER- MEDIATE, D SS-HIGH, respectively, P ¼ 0.18). After adjustment for potential confounding factors, rss (HR 1.07, 95% CI , P ¼ 0.001) and LVEF (HR 0.95, 95% CI , P ¼ 0.02) were shown to be independently associated with the 2-year risk of cardiac mortality (Table II). Forcing bss in the model, rss was no longer an independent predictor of cardiac mortality, whereas LVEF was an independent predictor. The rss and the bss showed comparable discrimination (rss: AUC 0.72, 95% CI ; bss: AUC 0.73, 95% CI ), as also reflected by the similar IoS (Table III). Hosmer Lemeshow statistics were 0.60 for rss (P ¼ 0.44) and 2.45 (P ¼ 0.12) for bss, reflecting better calibration ability of the rss. The D SS provided the worst discrimination and calibration characteristics (AUC 0.55, 95% CI , Hosmer Lemeshow statistic 3.13, P ¼ 0.08) with a trivial IoS. The NRI was 2.4% for rss over bss and -58.4% for D SS over bss. Added Value of Repeat Risk Assessment Adjusted HRs and 95% CIs for different angiographic risk models are shown in Table II. When

5 Residual SYNTAX Score and LM Revascularization 337 TABLE I. Baseline Characteristics Stratified by Residual SYNTAX Score Residual SYNTAX score 0(N ¼ 195) 1 8 (N ¼ 91) >8 (N ¼ 114) P value Age, years SD Male, % 153 (79) 69 (76) 86 (75) 0.79 Risk factors, % Systemic hypertension 130 (67) 67 (74) 80 (70) 0.48 Hypercolesterolemia 103 (53) 55 (60) 65 (57) 0.46 Smoking habitus 93 (48) 36 (40) 41 (36) 0.11 Diabetes mellitus 50 (26) 30 (33) 45 (40) 0.04 Creatinine >2 mg/dl 19 (10) 4 (4) 13 (11) 0.19 Medical history, % Previous myocardial infarction 60 (31) 30 (33) 50 (44) 0.06 Peripheral artery disease 29 (15) 21 (23) 22 (19) 0.22 Previous PCI 60 (31) 24 (26) 22 (19) 0.09 Clinical presentation, % Stable angina 81 (42) 34 (37) 37 (33) 0.28 UA/NSTEMI 102 (52) 49 (54) 71 (62) 0.22 STEMI 12 (6) 8 (9) 6 (5) 0.57 LVEF, % <0.001 EuroSCORE, % <3 47 (24) 21 (23) 17 (15) (48) 38 (42) 41 (36) 0.11 >6 54 (28) 32 (35) 56 (49) Lesion location, % Ostium 76 (39) 27 (30) 33 (29) 0.12 Shaft 22 (11) 14 (15) 18 (16) 0.45 Distal 97 (50) 50 (55) 63 (55) 0.56 Extent of coronary artery disease, % Isolated ULMCA disease 28 (14) 0 (0) 0 (0) <0.001 ULMCA plus 1-vessel disease 76 (39) 20 (22) 8 (7) <0.001 ULMCA plus 2-vessel disease 42 (22) 27 (30) 26 (23) 0.31 ULMCA plus 3-vessel disease 11 (6) 17 (19) 46 (40) <0.001 SYNTAX score <0.001 LVEF ¼ left ventricular ejection fraction; NSTEMI ¼ non-st segment elevation myocardial infarction; PCI ¼ percutaneous coronary intervention; SD ¼ standard deviation; STEMI ¼ ST segment elevation myocardial infarction; UA ¼ unstable angina; ULMCA ¼ unprotected left main coronary artery. TABLE II. Adjusted HRs and 95% CIs for Different Angiographic Risk Models HR 95% LCL 95% UCL P As continuous variables a rss bss D SS As categorical variables rss bss D SS Model Model HR ¼ hazard ratio; LCL ¼ lower confidence limit; UCL ¼ higher confidence limit. Other abbreviations as in Table II. a HRs per unit increase. Fig. 3. Risk prediction of residual SYNTAX score. Kaplan Meier estimates of 2-year cardiac mortality by residual SYN- TAX score (rss) risk categories. forced in the multivariable model for statistical adjustment, both Models 1 and 2 were shown to be independent predictors of cardiac mortality at 2 years. However, although using risk information from both the rss and bss (Model 1) marginally improved the

6 338 Capodanno et al. discrimination ability compared with risk information from each single risk assessment (AUC 0.74, 95% CI ) at the price of a slightly poorer calibration (Hosmer Lemeshow statistic 1.47, P ¼ 0.23), incorporation of the change in SYNTAX score across PCI in the rss model (Model 2) yielded a worse discrimination (AUC 0.66, 95% CI ) with some improvement of the global model fit (Hosmer Lemeshow statistic 0.07, P¼0.79, Table III). In aggregate, the graphical comparison of the five models characteristics (rss, bss, D SS, Model 1, and Model 2) showed that the best balance in terms of discrimination and calibration for cardiac mortality was offered by the rss and the Model 1, with the first one displaying better calibration and the second one displaying better discrimination (Fig. 4). However, the NRI for the use of Model 1 over rss and bss were þ14.2% and þ13.6%, respectively, outlining a sensible reclassification advantage in using the combined risk TABLE III. Calibration and Discrimination Parameters HL statistic (P value) AUC (95% CI) IoS a rss 0.60 (0.44) 0.72 ( ) 0.17 bss 2.45 (0.12) 0.73 ( ) 0.15 D SS 3.13 (0.08) 0.55 ( ) 0.03 Model (0.23) 0.74 ( ) 0.21 Model (0.79) 0.66 ( ) 0.14 D SS ¼ changes in SYNTAX score from pre- to post-percutaneous coronary intervention; AUC ¼ Area under the curve; bss ¼ baseline SYN- TAX score; HL ¼ Hosmer Lemeshow; rss ¼ residual SYNTAX score. a The index of separation (IoS) is defined as P worst P best, assuming P worst as the predicted P of event for a patient in the group with the highest score and P best as the predicted P of the same event for a patient in the group with the lowest score. information over each individual score. On the other hand, the use of Model 2 was associated with an NRI of -35% over the rss and -35.6% over the bss. Consistent with the above findings, significant incremental information was obtained from the addition of bss to rss in the Cox regression model (P ¼ 0.04), while no significant incremental information was obtained from the addition of D SS (P ¼ 0.59). DISCUSSION This study adds to the growing evidence on prognostic scores in ULMCA revascularization with the following observations. First, in ULMCA patients treated with PCI, the rss calculated after the procedure is an objective measure of incomplete revascularization with prognostic significance as independent predictor of 2- year cardiac mortality. Second, compared with the bss, the rss has similar discrimination and better calibration. Third, compared with risk information from a single risk assessment, information from both the bss and the rss yields better risk classification. The prognostic implications of complete revascularization in patients with multivessel disease have been extensively documented in surgical and PCI series with mixed results [4 6,14 20]. Notably, the above studies did not address the specific subset of patients with ULMCA disease, a population with distinct prognostic implications, and expected benefits from PCI [1]. In addition, these studies lacked a uniform, standardized definition of complete revascularization and focused on different patient populations, leading to conflicting results [21]. Specifically, the incompleteness of coronary Fig. 4. Prediction accuracy of rss, bss, D SS, Model 1, and Model 2. The position of the score in the graph depends on the balance between the score calibration (assessed by the Hosmer Lemeshow statistic) and the discrimination (assessed by the area under the receiver-characteristic curve). The prognostic accuracy of each score increases from top to bottom and from left to right.

7 revascularization was established as a binary outcome (yes/no) and patients with extensive unrevascularized territories or small amounts of residual ischemic myocardium were pooled together. The advantage of better population characterization, baseline disease stratification, and residual disease quantification in patients with multivessel disease and moderate-to-high risk acute coronary syndromes enrolled in the ACUITY (Acute Catheterization and Urgent Intervention Triage StrategY) trial has been recently emphasized, suggesting that the rss is an independent predictor of 1-year mortality, cardiac mortality, MI, and unplanned revascularization, with good discriminatory power and the potential for identifying patients with acute coronary syndromes who could benefit from further revascularization (i.e., those with rss > 8) [7]. In our study, the rss linearly correlated with the bss: the higher the bss, the higher the rss. This finding reinforces the idea that the SYNTAX score is a sensitive marker of the ability of PCI to achieve the goal of complete revascularization [3]. Interestingly, patients with high residual coronary artery disease after PCI were not only those with more complex angiographic features, such as three-vessel involvement, but also clinical features that likely discouraged or prevented the operators from attempting more complete revascularization (i.e., older age, diabetes, lower LVEF, higher EuroSCORE). This may contribute to explain why scoring systems that encompass both angiographic and clinical information are globally more accurate than scores that rely on angiographic or clinical information only, since they provide a more accurate identification of patients who are susceptible of incomplete revascularization following PCI [8,13]. Because a link between rss and cardiac mortality was demonstrated, the notion that a patient is unlikely to obtain complete revascularization from PCI (i.e., a patient with high bss) may shift the operator s treatment decision toward alternative revascularization strategies, such as CABG. Importantly, since the bss and rss carry a similar prognostic significance (bss and rss are no longer independent predictors of cardiac mortality when forced simultaneously in the multivariable model, probably due to collinearity), one could argue that calculating both scores is useless and time consuming. However, a new model integrating information from bss and rss showed to correctly reclassify about 14% of patients, thereby suggesting that repeating risk assessment in the peri-pci period improves prediction, compared with single risk assessment, and accurately reclassifies a not negligible proportion of patients into more appropriate risk categories. Repeat assessment of angiographic risk may have benefits in terms of patients and cardiologists behavior change. In fact, a clinical implication of this model is Residual SYNTAX Score and LM Revascularization 339 the identification of a sizeable cohort of patients who do not derive a significant prognostic benefit despite PCI and are therefore to be followed-up strictly. While this study demonstrates that a standardized and quantitative assessment of residual angiographic disease following ULMCA PCI may add to the prognostic information deriving from the preprocedural SYNTAX score, we were not able to observe any prognostic significance linked to the extent of revascularization by PCI. In fact, the D SS was consistently associated with weak prognostic accuracy and did not show to be predictive of cardiac mortality. In addition, a model integrating rss and D SS incorrectly reclassified about one third of patients. Although consistent with multivessel disease patients [7], the conclusions that can be drawn from these findings are limited by a number of factors. First, a minimal variation of the SYNTAX score (i.e., low D SS )includes both patients with low and high bss, with the first needing optimal medical therapy and/or few stents and the latter frequently receiving less-aggressive revascularization due to operator s choices or technical reasons. Therefore, relying on the D SS only might be puzzling. Second, the expected benefit of aggressive, extensive revascularization could be in theory counterbalanced by well-know disadvantages of PCI, especially when stents are placed on non-ischemia-producing lesions (i.e., acute kidney injury, periprocedural MI, stent thrombosis, restenosis [21,22]). Third, subsets typically represented in the highest tertiles of rss (i.e., chronic total occlusions, calcified lesions, complex bifurcations and trifurcations, small vessels) may be not ideal candidates to PCI. Finally, more patients for each bss group should be analyzed to elucidate a putative differential effect of complete revascularization across different risk strata. Overall, these findings suggest that the extent of lesions manageable by PCI alone should not be the deciding factor in selecting the best revascularization therapy for ULMCA patients. Prospectively designed studies are needed to better elucidate the role of complete revascularization and rss in patients submitted to ULMCA revascularization across different bss strata. Study Limitations This study is limited by its post-hoc nature. The CUS- TOMIZE registry is an observational study run over a long period, in which the treatment strategies for the PCI cohort might have changed over time. We cannot tell what impact these changes might have had on the results. However, time-dependent variables did not prove to be necessary to improve the accuracy of any model. Second, the results of this study may have varied if the bss and rss were assessed by core lab technicians with the aid of quantitative coronary analysis. Third, risk models such as

8 340 Capodanno et al. the SYNTAX score rely on the assumption that angiographic complete revascularization is always the goal of intervention, which, in reality, is often not achievable through either PCI or CABG. Therefore, both the bss and rss may theoretically be improved upon by knowledge of ischemia or viability. Finally, statistical adjustment was attempted by means of a parsimonious multivariable model with good discrimination and calibration abilities. However, the impact of unidentified confounders cannot be entirely ruled out. CONCLUSIONS A quantitative assessment of the incompleteness of revascularization following ULMCA PCI by means of the rss carries a prognostic significance as independent predictor of 2-year cardiac mortality. Individuals with high estimated risk both pre- and post-ulmca PCI have higher risk of cardiac mortality than those who remain in the lowest risk category at both assessment. Compared with a single assessment of the SYNTAX score, information coming from repeat assessment of the angiographic risk may improve the ability to correctly reclassify patients undergoing ULMCA PCI. REFERENCES 1. Capodanno D, Stone GW, Morice MC, et al. Percutaneous coronary intervention versus coronary artery bypass graft surgery in left main coronary artery disease: a meta-analysis of randomized clinical data. J Am Coll Cardiol 2011;58: Chieffo A, Meliga E, Latib A, et al. Drug eluting stent for left main coronary artery disease: the DELTA Registry a multicenter registry evaluating PCI vs. CABG for left main treatment. JACC Cardiovasc Interv 2012;5: Capodanno D, Capranzano P, Di Salvo ME, et al. Usefulness of SYNTAX score to select patients with left main coronary artery disease to be treated with coronary artery bypass graft. JACC Cardiovasc Interv 2009;2: Kim YH, Park DW, Lee JY, et al. Impact of angiographic complete revascularization after drug-eluting stent implantation or coronary artery bypass graft surgery for multivessel coronary artery disease. Circulation 2011;123: Tamburino C, Angiolillo DJ, Capranzano P, et al. Complete versus incomplete revascularization in patients with multivessel disease undergoing percutaneous coronary intervention with drugeluting stents. Catheter Cardiovasc Interv 2008;72: Hannan EL, Wu C, Walford G, et al. Incomplete revascularization in the era of drug-eluting stents: impact on adverse outcomes. JACC Cardiovasc Interv 2009;2: Généreux P, Palmerini T, Caixeta A, et al. Quantification and impact of untreated coronary artery disease after percutaneous coronary intervention: The Residual SYNTAX (Synergy Between PCI With Taxus and Cardiac Surgery) score. J Am Coll Cardiol 2012;59: Capodanno D, Tamburino C. Integrating the Synergy between percutaneous coronary intervention with Taxus and Cardiac Surgery (SYNTAX) score into practice: use, pitfalls, and new directions. Am Heart J 2011;161: Sianos G, Morel MA, Kappetein AP, et al. The SYNTAX Score: an angiographic tool grading the complexity of coronary artery disease. EuroIntervention 2005;1: Valgimigli M, Serruys PW, Tsuchida K, et al. Cyphering the complexity of coronary artery disease using the syntax score to predict clinical outcome in patients with three-vessel lumen obstruction undergoing percutaneous coronary intervention. Am J Cardiol 2007;99: Capodanno D, Di Salvo ME, Cincotta G, et al. Usefulness of the SYNTAX score for predicting clinical outcome after percutaneous coronary intervention of unprotected left main coronary artery disease. Circ Cardiovasc Interv 2009;2: 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: Capodanno D, Miano M, Cincotta G, et al. EuroSCORE refines the predictive ability of SYNTAX score in patients undergoing left main percutaneous coronary intervention. Am Heart J 2010;159: Kleisli T, Cheng W, Jacobs MJ, et al. In the current era, complete revascularization improves survival after coronary artery bypass surgery. J Thorac Cardiovasc Surg 2005;129: Vander Salm TJ, Kip KE, Jones RH, et al. What constitutes optimal surgical revascularization? Answers from the bypass angioplasty revascularization investigation (BARI). J Am Coll Cardiol 2002;39: Van den Brand MJ, Rensing BJ, Morel MA, et al. The effect of completeness of revascularization on event-free survival at one year in the ARTS trial. J Am Coll Cardiol 2002;19: Head SJ, Mack MJ, Holmes DR Jr., et al. Incidence, predictors and outcomes of incomplete revascularization after percutaneous coronary intervention and coronary artery bypass grafting: a subgroup analysis of 3-year SYNTAX data. Eur J Cardiothorac Surg 2012;41: Aggarwal V, Rajpathak S, Singh M, et al. Clinical outcomes based on completeness of revascularisation in patients undergoing percutaneous coronary intervention: a meta-analysis of multivessel coronary artery disease studies. EuroIntervention 2012;7: Ijsselmuiden AJ, Ezechiels J, Westendorp IC, et al. Complete versus culprit vessel percutaneous coronary intervention in multivessel disease: a randomized comparison. Am Heart J 2004;148: De Bruyne B. Multivessel disease: from reasonably incomplete to functionally complete revascularization. Circulation 2012;125: Rihal CS, Textor SC, Grill DE, et al. Incidence and prognostic importance of acute renal failure after percutaneous coronary intervention. Circulation 2002;105: Holmes DR Jr, Kereiakes DJ, Garg S, et al. Stent thrombosis. J Am Coll Cardiol 2010;56: