JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 68, NO. 21, 2016 ª 2016 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 68, NO. 21, 2016 ª 2016 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION PUBLISHED BY ELSEVIER ISSN 0735-1097/$36.00 http://dx.doi.org/10.1016/j.jacc.2016.08.059 High-Sensitivity Troponin T and Mortality After Elective Percutaneous Coronary Intervention Gjin Ndrepepa, MD, a Roisin Colleran, MB, BCH, a Siegmund Braun, MD, b Salvatore Cassese, MD, a Julia Hieber, MD, a Massimiliano Fusaro, MD, a Sebastian Kufner, MD, a Ilka Ott, MD, a Robert A. Byrne, MB, BCH, PHD, a Oliver Husser, MD, a Christian Hengstenberg, MD, a Karl-Ludwig Laugwitz, MD, c,d Heribert Schunkert, MD, a,d Adnan Kastrati, MD a,d ABSTRACT BACKGROUND The prognostic value of high-sensitivity troponin T (hs-tnt) elevation after elective percutaneous coronary intervention (PCI) in patients with or without raised baseline hs-tnt levels is unclear. OBJECTIVES The goal of this study was to assess whether the prognostic value of post-procedural hs-tnt level after elective PCI depends on the baseline hs-tnt level. METHODS This study included 5,626 patients undergoing elective PCI who had baseline and peak post-procedural hs-tnt measurements available. The primary outcome was 3-year mortality (with risk estimates calculated per SD increase of the log hs-tnt scale). RESULTS Patients were divided into 4 groups: nonelevated baseline and post-procedural hs-tnt levels (hs-tnt #0.014 mg/l; n ¼ 742); nonelevated baseline but elevated post-procedural hs-tnt levels (peak post-procedural hs-tnt >0.014 mg/l; n ¼ 2,721); elevated baseline hs-tnt levels (hs-tnt >0.014 mg/l) with no further rise post-procedure (n ¼ 516); and elevated baseline hs-tnt levels with a further rise post-procedure (n ¼ 1,647). A total of 265 deaths occurred: 6 (1.6%) in patients with nonelevated baseline and post-procedural hs-tnt levels; 54 (3.8%) in patients with nonelevated baseline but elevated post-procedural hs-tnt levels; 50 (16.0%) in patients with elevated baseline hs-tnt levels with no further rise post-procedure; and 155 (18.2%) in patients with elevated baseline hs-tnt levels with a further rise post-procedure (p < 0.001). After adjustment, baseline hs-tnt levels (hazard ratio [HR]: 1.22; 95% confidence interval [CI]: 1.09 to 1.38; p < 0.001) but not peak post-procedural hs-tnt levels (HR: 1.04; 95% CI: 0.85 to 1.28; p ¼ 0.679) were associated with an increased risk of mortality. Peak post-procedural hs-tnt findings were not associated with mortality in patients with nonelevated (HR: 0.93; 95% CI: 0.69 to 1.25; p ¼ 0.653) or elevated (HR: 1.24; 95% CI: 0.91 to 1.69; p ¼ 0.165) baseline hs-tnt levels. CONCLUSIONS In patients with coronary artery disease undergoing elective PCI, an increase in post-procedural hs-tnt level did not offer prognostic information beyond that provided by the baseline level of the biomarker. (J Am Coll Cardiol 2016;68:2259 68) 2016 by the American College of Cardiology Foundation. Cardiac troponins are the most commonly used biomarkers for the diagnosis of myocardial damage. High-sensitivity troponin T (hs-tnt) assays enable the detection of even minor myocardial damage occurring spontaneously or after percutaneous coronary intervention (PCI) (1,2). Although elevation of cardiac troponin levels after elective PCI is common (3 5), theclinicalsignificance Listen to this manuscript s audio summary by JACC Editor-in-Chief Dr. Valentin Fuster. From the a Department of Adult Cardiology, Deutsches Herzzentrum München, Technische Universität, Munich, Germany; b Department of Laboratory Medicine, Deutsches Herzzentrum München, Technische Universität, Munich, Germany; c 1. Medizinische Klinik, Klinikum rechts der Isar, Technische Universität, Munich, Germany; and d DZHK (German Centre for Cardiovascular Research), partner site Munich Heart Alliance, Munich, Germany. Dr. Byrne has received lecture fees from B. Braun Melsungen, Biotronik, and Boston Scientific; and research grants to the institution from Boston Scientific and HeartFlow. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Manuscript received June 21, 2016; revised manuscript received August 5, 2016, accepted August 31, 2016.

2260 Ndrepepa et al. JACC VOL. 68, NO. 21, 2016 Troponin and Mortality After Elective PCI NOVEMBER 29, 2016:2259 68 ABBREVIATIONS AND ACRONYMS of this finding remains unclear. A rise in hs- TnT level above the 99th percentile upper reference limit (URL) after elective PCI was CAD = coronary artery disease recently reported in up to 80% of patients CI = confidence interval who had baseline hs-tnt levels within HR = hazard ratio hs-tnt = high-sensitivity normal limits (6). Raised levels of baseline circulating troponin in patients with stable troponin T coronary artery disease (CAD) are also PCI = percutaneous coronary intervention commonly found (7). TIMI = Thrombolysis In SEE PAGE 2269 Myocardial Infarction URL = upper reference limit Several previous studies have shown that baseline, but not post-procedural, troponin elevation was associated with a poor outcome after PCI (7,8). Thus, elevated baseline troponin is a great confounder that may modulate the association between post-procedural troponin rise and outcome after PCI. Notably, whether the prognostic value of post-procedural troponin is dependent on the baseline troponin level remains unknown. Moreover, the magnitude of change and factors predisposing to elevated post-procedural levels after elective PCI in patients with or without elevated baseline troponin are unclear. Theaimofthepresentstudywas2-fold:first, to investigate whether the prognostic value of postprocedural hs-tnt level after elective PCI depends on the baseline hs-tnt level; and second, to determine the factors that predispose to post-procedural hs-tnt elevation in patients with and without elevated baseline hs-tnt levels. PATIENTS AND METHODS STUDY PATIENTS. The present study was a retrospective analysis of 5,626 patients with stable CAD who underwent elective PCI in our hospitals between October 2009 and January 2015, with the last day of follow-up at the end of May 2016. The indication for the index intervention was stable CAD (if significant coronary stenosis was found on diagnostic angiography in patients presenting with symptoms) or restenosis diagnosed in the setting of scheduled angiography after previous stent implantation. All patients had baseline (pre-procedural) and peak post-procedural hs-tnt measurements available for analysis. Each patient was included in the analysis only once (i.e., at his or her first PCI procedure). Patients with acute coronary syndromes, acute infections, pregnancy, advanced impairment of renal function (serum creatinine $2.0 mg/dl), or a known malignancy were excluded. The study was performed in accordance with the Declaration of Helsinki. DIAGNOSTIC CRITERIA AND PCI. Patients with stable angina, defined as chest pain that had not changed in intensity, character, frequency, or threshold over the preceding 2 months, underwent elective PCI if significant stenoses ($70% lumen obstruction) were documented on diagnostic coronary angiography. Patients undergoing PCI in the setting of a scheduled examination after stent implantation underwent PCI if a significant restenosis in the stented coronary segment or a significant stenosis in a nonstented segment was documented on coronary angiography. Personnel blinded to patient clinical or follow-up data performed angiographic analysis in the core laboratory using an automated edge detection system (CMS, Medis Medical Imaging Systems, Neuen, the Netherlands). Cardiovascular risk factors, including diabetes, arterial hypertension, hypercholesterolemia, and current smoking, were defined according to accepted criteria. Epicardial blood flow pre- and post-pci was graded by using the Thrombolysis In Myocardial Infarction (TIMI) group angiographic criteria. Left ventricular ejection fraction was calculated by using the area length method on left ventricular angiography. Body mass index was calculated by using patient weight and height measured during the index hospitalization, and glomerular filtration rate was calculated according to the Cockcroft-Gault formula. Coronary angiography and PCI were performed according to standard practices. Before PCI, patients received aspirin (325 to 500 mg) and clopidogrel (loading dose of 600 mg) and anticoagulation therapy in the form of unfractionated heparin or bivalirudin. After PCI, patients received clopidogrel 150 mg/day until hospital discharge, followed by 75 mg/day for at least 1 month after bare-metal stent implantation or at least 6 months after drug-eluting stent implantation, in addition to aspirin 200 mg/day indefinitely. Other drugs (comprising predominantly statins, angiotensin-converting enzyme inhibitors, or betablockers) were prescribed at the discretion of the treating physician. BIOCHEMICAL MEASUREMENTS. Blood samples were collected in tubes containing a lithium-heparin anticoagulant. Blood samples for hs-tnt measurements were obtained before the procedure (on admission), 6 h after PCI, and on a daily basis thereafter during the hospital stay (usually 48 h). Two or more postprocedural hs-tnt measurements were performed in 98.7% of patients. The peak level was defined as the highest post-procedural hs-tnt concentration. Troponin T was measured by a high-sensitivity assay in a cobas e 411 immunoanalyzer using electrochemiluminescence technology (Roche Diagnostics,

JACC VOL. 68, NO. 21, 2016 NOVEMBER 29, 2016:2259 68 Ndrepepa et al. Troponin and Mortality After Elective PCI 2261 TABLE 1 Baseline Data hs-tnt Elevated at Baseline (n ¼ 2,163) hs-tnt Not Elevated at Baseline (n ¼ 3,463) Further Elevation Post-PCI (n ¼ 1,647) No Further Elevation Post-PCI (n ¼ 516) pvalue Elevation Post-PCI (n ¼ 2,721) No Elevation Post-PCI (n ¼ 742) Age, yrs 75.0 [68.5 81.1] 72.7 [66.7 78.8] <0.001 69.3 [61.6 74.4] 66.1 [58.3 72.2] <0.001 Female 332 (20.2) 105 (20.3) 0.925 619 (22.7) 197 (26.6) 0.030 Diabetes 624 (37.9) 206 (39.9) 0.407 712 (26.2) 238 (32.1) 0.001 On insulin therapy 245 (14.9) 87 (16.9) 0.275 180 (6.6) 64 (8.6) 0.058 Body mass index, kg/m 2 27.0 [24.3 30.0] 27.5 [24.7 30.4] 0.130 27.3 [24.8 30.1] 27.6 [25.2 30.5] 0.018 History of arterial hypertension 1,226 (74.4) 354 (68.6) 0.009 2,089 (76.8) 527 (71.0) 0.001 Current smoker 214 (13.0) 78 (15.1) 0.218 388 (14.3) 129 (17.4) 0.034 Hypercholesterolemia 1,256 (76.3) 387 (75.0) 0.559 2,218 (81.5) 581 (78.3) 0.049 History of myocardial infarction 583 (35.4) 172 (33.3) 0.391 819 (30.1) 248 (33.4) 0.082 History of coronary artery bypass grafting 266 (16.2) 70 (13.6) 0.157 317 (11.7) 82 (11.1) 0.651 Baseline hs-tnt, mg/l 0.022 [0.020 0.040] 0.030 [0.020 0.067] <0.001 0.010 [0.010 0.010] 0.010 [0.009 0.010] <0.001 Peak post-procedural hs-tnt, mg/l 0.079 [0.040 0.171] 0.030 [0.020 0.060] <0.001 0.040 [0.020 0.090] 0.010 [0.010 0.010] <0.001 Glomerular filtration rate, ml/min 59.5 [43.5 79.1] 63.6 [47.3 84.6] 0.010 82.0 [64.7 102.8] 87.1 [70.8 107.4] <0.001 No. of stenosed coronary arteries 0.068 0.002 1 140 (8.5) 61 (11.8) 434 (16.0) 144 (19.4) 2 370 (22.5) 117 (22.7) 698 (25.7) 218 (29.4) 3 1,137 (69.0) 338 (65.5) 1,589 (58.3) 380 (51.2) Multivessel disease 1,507 (91.5) 455 (88.2) 0.023 2,287 (84.0) 598 (80.6) 0.025 Multilesion intervention 785 (47.7) 167 (32.4) <0.001 1,291 (47.4) 203 (27.4) <0.001 Left ventricular ejection fraction, %* 52.0 [40.0 60.0] 53.0 [43.8 60.0] 0.200 60.0 [52.0 62.0] 60.0 [52.0 62.0] 0.400 Therapy on discharge Statin 1,487 (90.6) 463 (89.9) 0.631 2,552 (93.9) 687 (92.7) 0.229 Beta-blocker 1,407 (85.7) 447 (86.8) 0.547 2,386 (87.9) 652 (88.0) 0.956 ACE inhibitor 1,090 (66.6) 352 (68.8) 0.363 1,845 (67.9) 498 (67.2) 0.728 pvalue Values are median [25th to 75th percentile] or n (%). *Available in 3,799 patients. ACE ¼ angiotensin-converting enzyme; hs-tnt ¼ high-sensitivity troponin T; PCI ¼ percutaneous coronary intervention. Risch-Rotkreuz, Switzerland). The limit of blank for this assay (the concentration below which analyte-free samples are found with 95% probability) is #0.003 mg/l. The functional sensitivity (the lowest analyte concentration that can be reproducibly measured with a coefficient of variation #10%) is #0.013 mg/l.the99thpercentileurlis0.014mg/l. Baseline and peak post-procedural hs-tnt values were used for analysis. Creatinine was measured with a kineticcolorimetricassayusingthecompensatedjaffe method. Laboratory personnel unaware of patient clinical or follow-up data measured other biochemical parameters using standard laboratory methods. OUTCOME AND FOLLOW-UP. The primary outcome measure was all-cause mortality up to 3 years after PCI. Follow-up was performed by telephone interview at 1, 6, and 12 months after PCI in the first year and yearly thereafter. Data on mortality were obtained from hospital charts, death certificates, telephone contact with relatives of the patient or family physicians, insurance companies, or the registration of address office. Medical personnel unaware of patient clinical or laboratory data performed followup and adjudication of events. STATISTICAL ANALYSIS. Data are presented as medians with 25th to 75th percentiles, proportions (%), or Kaplan-Meier estimates (%). The distribution of continuous data was tested by using the 1-sample Kolmogorov-Smirnov test. Because all continuous data exhibited a non-gaussian distribution pattern, the Kruskal-Wallis rank sum test was used for intergroup comparisons. Categorical data were compared by using the chi-square test. The correlates of increased hs-tnt after PCI were assessed by using the multiple linear regression model. All variables in Tables 1 and 2, except for left ventricular ejection fraction (due to incomplete data) and therapy at discharge, were entered into the model. Due to the skewed distribution of the hs-tnt levels, baseline and post-procedural hs-tnt measurements were entered into the model after logarithmic transformation. Survival analysis was performed by using the Kaplan-Meier method, and differences in survival rates were assessed according to the univariable Cox

2262 Ndrepepa et al. JACC VOL. 68, NO. 21, 2016 Troponin and Mortality After Elective PCI NOVEMBER 29, 2016:2259 68 TABLE 2 Procedural Data (Lesion-Based Analysis) hs-tnt Elevated at Baseline (n ¼ 3,500) hs-tnt Not Elevated at Baseline (n ¼ 5,654) Further Elevation Post-PCI (n ¼ 2,764) No Further Elevation Post-PCI (n ¼ 736) pvalue Elevated Post-PCI (n ¼ 4,665) Not Elevated Post-PCI (n ¼ 989) Pre-procedural TIMI flow grade* 0.749 0.470 0 127 (4.6) 35 (4.8) 166 (3.6) 36 (3.6) 1 58 (2.1) 12 (1.6) 94 (2.0) 24 (2.4) 2 149 (5.4) 45 (6.1) 274 (5.9) 47 (4.8) 3 2,424 (87.9) 644 (87.5) 4,126 (88.5) 882 (89.2) Vessel treated 0.384 <0.001 Left main coronary artery 114 (4.1) 31 (4.1) 192 (4.1) 36 (3.6) Left anterior descending coronary artery 1,116 (40.4) 311 (42.3) 1,986 (42.6) 426 (43.1) Left circumflex artery 669 (24.2) 152 (20.7) 1,121 (24.0) 193 (19.5) Right coronary artery 796 (28.8) 218 (29.6) 1,300 (27.9) 314 (31.7) Bypass graft 69 (2.5) 24 (3.3) 66 (1.4) 20 (2.1) Complex lesion (AHA B2/C class) 2,103 (76.1) 516 (70.1) <0.001 3,422 (73.4) 629 (63.6) <0.001 Bifurcation lesion 923 (33.8) 211 (28.7) 0.008 1,649 (35.5) 286 (29.0) <0.001 Restenotic lesion 262 (9.5) 77 (10.5) 0.423 445 (9.5) 124 (12.5) 0.004 Balloon diameter, mm 3.00 [3.00 3.50] 3.00 [2.75 3.50] 0.200 3.00 [2.99 3.50] 3.00 [2.75 3.50] 0.010 Maximal balloon pressure, atm 15.0 [12.8 17.0] 14.0 [12.0 16.0] 0.008 15.0 [12.0 17.0] 14.0 [12.0 16.0] <0.001 Total stented length, mm 24.0 [18.0 35.0] 23.0 [18.0 30.8] 0.002 23.0 [18.0 33.0] 23.0 [18.0 28.0] <0.001 Post-procedural TIMI flow grade 0.672 0.927 0 7 (0.2) 3 (0.4) 8 (0.2) 1 (0.1) 1 10 (0.4) 1 (0.1) 3 (0.1) 1 (0.1) 2 44 (1.6) 13 (1.8) 85 (1.8) 17 (1.7) 3 2,699 (97.8) 716 (97.7) 4,563 (97.9) 970 (98.1) pvalue Values are n (%) of lesions or median [25th to 75th percentiles]. *Available for 3,494 lesions in patients with elevated baseline troponin levels and 5,649 lesions in patients with baseline troponin levels within normal limits. Available for 3,493 lesions in patients with elevated baseline troponin levels and 5,648 lesions in patients with baseline troponin levels within normal limits. AHA ¼ American Heart Association; TIMI ¼ Thrombolysis In Myocardial Infarction; other abbreviations as in Table 1. proportional hazards model. Independent correlates of all-cause mortality were assessed by using the multivariable Cox proportional hazards model. All variables in Tables 1 and 2 were tested in the univariable analysis, and those exhibiting a significant association with mortality were entered into the model to assess factors associated with mortality. The baseline and post-procedural hs-tnt measurements were entered into the model after logarithmic transformation. The generalized estimating equation method was applied in all analyses (including multivariable analyses) that involved lesion characteristics to account for clustering of the data in the same patient. Statistical analysis was performed by using the R 2.15.1 Statistical Package (R Foundation for Statistical Computing, Vienna, Austria). A 2-sided pvalue<0.05 was considered to indicate statistical significance. RESULTS PATIENT CLASSIFICATION AND BASELINE DATA. Overall, the study included 5,626 patients. Using the 99th percentile URL of baseline hs-tnt (0.014 mg/l) as the cutoff, patients were divided into a group with baseline hs-tnt level within normal limits (hs-tnt #0.014 mg/l [n ¼ 3,463]) and a group with elevated baseline hs-tnt level (hs-tnt >0.014 mg/l [n ¼ 2,163]). Using the peak post-procedural hs-tnt level, patients with normal baseline hs-tnt were divided into a group with elevated post-procedural hs-tnt level (hs-tnt >0.014 mg/l [n ¼ 2,721]) and another group with normal post-procedural hs-tnt level (hs-tnt #0.014 mg/l [n ¼ 742]). Patients with elevated baseline hs-tnt levels were also divided into 2 groups: 1 with a further post-procedural rise in hs-tnt (peak post-procedural hs-tnt level higher than baseline hs-tnt level [n ¼ 1,647]) and another group with no further rise in post-procedural hs-tnt (peak post-procedural hs-tnt equal to or lower than thebaselinehs-tntlevel[n¼516]). Baseline data are shown in Table 1, and procedural characteristics are displayed in Table 2. CORRELATES OF POST-PROCEDURAL HS-TnT. Baseline and peak post-pci hs-tnt levels in the whole group of patients are shown in the Central Illustration. In patients with nonelevated baseline hs-tnt levels,

JACC VOL. 68, NO. 21, 2016 NOVEMBER 29, 2016:2259 68 Ndrepepa et al. Troponin and Mortality After Elective PCI 2263 CENTRAL ILLUSTRATION Troponin and Mortality After Elective PCI: Baseline and Post-Procedural hs-tnt Levels Ndrepepa, G. et al. J Am Coll Cardiol. 2016;68(21):2259 68. The graph in the left upper corner shows the distribution of paired baseline and peak post-procedural high-sensitivity troponin T (hs-tnt) values in each patient. The graph in the right upper corner shows the median and 25th to 75th percentiles of baseline (0.01 mg/l [0.01 to 0.02 mg/l]) and peak post-procedural (0.04 mg/l [0.02 to 0.10 mg/l]) hs-tnt levels. The median and 25th percentile values of baseline hs-tnt level are the same (0.01 mg/l). The lower graph shows the cumulative distribution curves of the baseline (blue line) and post-procedural (orange line) hs-tnt concentration. PCI ¼ percutaneous coronary intervention. the mean baseline and post-procedural hs-tnt values were 0.0095 mg/l and 0.098 mg/l. In patients with elevated baseline hs-tnt levels, mean baseline and post-procedural hs-tnt values were 0.113 mg/l and 0.244 mg/l. The mean change in hs-tnt (peak post-procedural value baseline value) was 0.088 mg/l in patients with nonelevated baseline hs-tnt levels (median [25th to 75th percentiles]: 0.020 [0.010 to 0.060] mg/l), and 0.131 mg/l (median: 0.023 [0.004 to 0.080] mg/l) in those with elevated baseline hs-tnt levels. The multiple linear regression model with the generalized estimating equation method was used to assess independent correlates of peak post-procedural hs-tnt in patients with and without elevated baseline hs-tnt levels (2 separate models). Independent correlates of post-procedural

2264 Ndrepepa et al. JACC VOL. 68, NO. 21, 2016 Troponin and Mortality After Elective PCI NOVEMBER 29, 2016:2259 68 TABLE 3 Independent Correlates of Post-Procedural hs-tnt Concentration Obtained From the Multiple Linear Regression Model Patients With hs-tnt Elevated at Baseline (n ¼ 2,163) Patients With hs-tnt Not Elevated at Baseline (n ¼ 3,463) Coefficient* p Value Coefficient p Value Diabetes 0.111785651 0.022 0.191225601 <0.001 Multivessel disease 0.281562811 <0.001 0.287904096 <0.001 Arterial hypertension 0.125431940 0.019 Baseline troponin 0.654213201 <0.001 0.832773399 <0.001 Previous myocardial infarction 0.128152726 0.011 Restenotic lesion 0.189879905 <0.001 0.238671558 <0.001 ACC/AHA B2/C class lesion 0.147118526 <0.001 0.178565892 <0.001 Multivessel disease 0.281562811 <0.001 Glomerular filtration rate 0.002781484 0.006 Body mass index 0.013267508 0.037 Bifurcation lesion 0.087846055 0.044 0.142993244 <0.001 Left anterior descending artery 0.159625100 <0.001 0.171036095 <0.001 intervention Maximal balloon pressure 0.016835049 <0.001 Post-procedural TIMI flow 0.291569192 0.003 0.334634960 0.005 Total stented length 0.008934003 <0.001 0.009691764 <0.001 *Coefficients denote the change in log post-procedural high-sensitivity troponin T (hs-tnt) per unit change in the independent variable. The negative sign before the coefficients shows the inverse correlation between the independent variable and the post-procedural troponin concentration. ACC ¼ American College of Cardiology; other abbreviations as in Tables 1 and 2. hs-tnt are shown in Table 3. The elevated baseline hs- TnT level and increased CAD severity and/or procedure complexity were independently associated with elevated hs-tnt levels after PCI. Notably, restenotic lesions were associated with reduced odds of an elevation in hs-tnt after PCI in patients from both groups. POST-PROCEDURAL HS-TnT AND MORTALITY. Overall, there were 265 deaths during the follow-up period: 6 deaths (1.6%) in patients with normal baseline and post-procedural hs-tnt levels; 54 deaths (3.8%) in patients with normal baseline and raised postprocedural hs-tnt levels; 50 deaths (16.0%) in patients with raised baseline levels but no further elevation in post-procedural hs-tnt; and 155 deaths (18.2%) among patients with elevated baseline hs-tnt levels and further elevation in postprocedural hs-tnt (overall log-rank test p < 0.001) (Figure 1). In patients with normal baseline hs-tnt levels, elevated post-procedural hs-tnt measurements were associated with increased risk of mortality compared with patients with a post-procedural hs-tnt level within normal limits (univariable hazard ratio [HR]: 2.38; 95% confidence interval [CI]: 1.03 to 5.54; p ¼ 0.043). In patients with elevated baseline FIGURE 1 Kaplan-Meier Curves Showing Probability of 3-Year Mortality for Patient Subgroups According to Baseline and Post-Procedural High-Sensitivity Troponin T Levels 30.0 Probability of Mortality (%) 25.0 20.0 15.0 10.0 5.0 HR=1.09 [0.79-1.51]; P=0.575 18.2% 16.0% HR=2.38 [1.03-5.54]; P=0.043 3.8% Troponin groups/patients at risk 0.0 Raised at baseline/further raised after PCI Raised at baseline/not raised after PCI Not raised at baseline/raised after PCI Not raised at baseline/not raised after PCI 0 1647 516 2721 742 1 2 3 912 325 1832 484 Years 529 198 1142 309 1.6% 248 106 588 144 The curves are hierarchically ordered according to the order of the numbers of patients at risk. Percentages show the Kaplan-Meier estimates of mortality. Numbers in brackets denote 95% confidence intervals. HR ¼ hazard ratio; PCI ¼ percutaneous coronary intervention.

JACC VOL. 68, NO. 21, 2016 NOVEMBER 29, 2016:2259 68 Ndrepepa et al. Troponin and Mortality After Elective PCI 2265 hs-tnt levels, further elevation in post-procedural hs-tnt was not associated with increased risk of mortality compared with patients with no further elevation in post-procedural hs-tnt (univariable HR: 1.09;95%CI:0.79to1.51;p¼0.575). In the univariable analysis, age (p < 0.001), female sex (p ¼ 0.002), diabetes (p ¼ 0.009),arterialhypertension (p < 0.001), body mass index (p ¼ 0.019), hypercholesterolemia (p ¼ 0.005), multivessel disease (p < 0.001), previous coronary artery bypass surgery (p < 0.001), baseline hs-tnt level (p < 0.001), peak post-procedural hs-tnt level (p < 0.001), glomerular filtration rate (p < 0.001), left ventricular ejection fraction (p < 0.001), vessel treated (p ¼ 0.004), restenotic lesions (p < 0.001), baseline TIMI flow grade (p ¼ 0.014), balloon diameter (p ¼ 0.003), maximal balloon pressure (p ¼ 0.007), and post-procedural TIMI flow grade (p ¼ 0.042) were independently associated with mortality risk. All of these variables, except for post-procedural TIMI flow (due to strong correlation with post-procedural hs- TnT), were entered into the Cox proportional hazards model with the generalized estimating equation method. When the post-procedural hs-tnt-mortality association was tested in the model in the entire group of patients and adjusted for the factors associatedwithmortalityrisklistedpreviously(excluding baseline hs-tnt), there was a trend for an association between post-procedural hs-tnt level and the risk of mortality (adjusted HR: 1.22; 95% CI: 0.98 to 1.41; p ¼ 0.082 for each SD increase in the logarithmic scale of post-procedural hs-tnt). When baseline hs-tnt level was entered into the model, the baseline hs- TnT level (HR: 1.22; 95% CI: 1.09 to 1.38; p < 0.001), but not peak post-procedural hs-tnt level (HR: 1.04; 95% CI: 0.85 to 1.28; p ¼ 0.679), was associated with increased risk of mortality (both risk estimates calculated per SD increase in the logarithmic scale of hs-tnt) (Table 4). The association between peak postprocedural hs-tnt and mortality was not significant in the group with normal baseline hs-tnt level (HR: 0.93; 95% CI: 0.69 to 1.25; p ¼ 0.653) or in the group with elevated baseline hs-tnt level (HR: 1.24; 95% CI: 0.91 to 1.69; p ¼ 0.165), with both risk estimates calculated per SD increase in the logarithmic scale of peak post-procedural hs-tnt. MORTALITY ACCORDING TO VARIOUS CUTOFFS OF POST-PROCEDURAL HS-TnT. The association between post-procedural hs-tnt level and mortality was assessed over several multiples of 99th percentile URL cutoffs in patients with and without elevations in baseline hs-tnt. The results of this analysis are shown in Table 5. Notably, the hs-tnt elevation TABLE 4 Results of Multivariable Cox Proportional Hazards Model Applied to Assess Predictors of All-Cause Mortality With Baseline and Post-Procedural hs-tnt Entered Into the Model HR (95% CI) p Value Peak post-procedural hs-tnt (for 1-SD increase in 1.04 (0.85 1.28) 0.679 logarithmic scale of hs-tnt) Baseline hs-tnt (for 1-SD increase in 1.22 (1.09 1.38) <0.001 logarithmic scale of hs-tnt) Age (for 10-yr increase) 1.92 (1.31 2.81) <0.001 Female 0.57 (0.34 0.91) 0.028 Diabetes 1.09 (0.72 1.65) 0.693 Arterial hypertension 0.54 (0.37 0.79) 0.002 Body mass index (for 5 kg/m 2 increase) 1.22 (0.96 1.55) 0.112 Hypercholesterolemia 0.73 (0.48 1.12) 0.149 Multivessel disease 1.26 (0.63 2.53) 0.505 Previous coronary artery bypass surgery 1.03 (0.64 1.65) 0.897 GFR (for 30 ml/min decrease) 1.69 (1.14 2.51) 0.008 LVEF (for 10% decrease) 1.41 (1.23 1.62) <0.001 Vessel treated 1.54 (0.92 2.59) 0.102 Restenotic lesions 1.32 (0.87 1.99) 0.187 Baseline TIMI flow grade (for 1 grade decrease) 1.26 (0.98 1.64) 0.076 Balloon diameter (for 0.5-mm increase) 1.05 (0.94 1.17) 0.387 Maximal balloon pressure (5 atm increase) 1.16 (0.96 1.39) 0.120 CI ¼ confidence interval; GFR ¼ glomerular filtration rate; HR ¼ hazard ratio; LVEF ¼ left ventricular ejection fraction; other abbreviations as in Tables 1 and 2. >70 the99thpercentileurlcutoff,comparedwith post-procedural hs-tnt #70 the 99th percentile URL, was associated with increased risk of mortality in patients with elevated baseline hs-tnt level (45.1% vs. 17.1%; univariable HR: 4.20; 95% CI: 1.98 to 8.44; p < 0.001) but not in those with nonelevated levels of the biomarker (5.3% vs. 3.3%; HR: 1.42; 95% CI: 0.03 to 8.72; p ¼ 0.772). In patients with elevated baseline hs-tnt, the association between the >70 the 99th percentile URL cutoff and mortality was tested in the multivariable Cox model, with hs-tnt entered as a categorical variable (dichotomized at the 70 the 99th percentile cutoff). The analysis showed a trend toward an independent association between hs-tnt TABLE 5 Mortality According to Different Cutoffs of Post-Procedural hs-tnt Elevated Baseline hs-tnt (n ¼ 2,163) Baseline hs-tnt Within Normal Limits (n ¼ 3,463) Patients (%) Mortality (%)* Patients (%) Mortality (%)* Not increased (vs. baseline) 516 (23.9) 50 (16.0) 742 (21.4) 6 (1.6) 1to3 99th percentile URL 595 (27.5) 45 (17.0) 1,475 (42.6) 26 (3.4) >3to5 99th percentile URL 362 (16.8) 35 (17.1) 453 (13.1) 12 (5.4) >5 to10 99th percentile URL 295 (13.6) 28 (20.8) 380 (11.0) 7 (2.7) >10 to 35 99th percentile URL 286 (13.2) 28 (17.2) 310 (8.9) 7 (4.6) >35 to 70 99th percentile URL 65 (3.0) 6 (15.5) 62 (1.8) 1 (1.8) >70 99th percentile URL 44 (2.0) 13 (45.1) 41 (1.2) 1 (5.3) *Numbers in parentheses in the mortality columns represent Kaplan-Meier estimates of mortality. hs-tnt ¼ high-sensitivity troponin T; URL ¼ upper reference limit.

2266 Ndrepepa et al. JACC VOL. 68, NO. 21, 2016 Troponin and Mortality After Elective PCI NOVEMBER 29, 2016:2259 68 >70 the 99th percentile URL and mortality (p ¼ 0.093), and a trend toward an interaction between post-procedural hs-tnt >70 the 99th percentile URL and baseline level of the biomarker with regard to the prediction of mortality (p for interaction ¼ 0.073). DISCUSSION The main findings of this study can be summarized as follows: 1) in patients with CAD undergoing elective PCI, a procedure-related hs-tnt rise occurred in the majority of patients (77.6%); 2) post-procedural hs-tnt elevation was not associated with an increased risk of mortality for up to 3 years, regardless of the pre-procedural hs-tnt level; 3) the preprocedural hs-tnt elevation was independently associated with an increased risk of mortality; and 4) pre-procedural hs-tnt level and angiographic characteristics underlying procedure complexity were independent correlates of elevated post-procedural hs-tnt level in patients both with and without elevated baseline levels of the biomarker. Circulating troponin after PCI consists of 2 fractions: a pre-procedural or baseline fraction and a fraction that reflects PCI-related troponin rise. Continuous microscopic loss of cardiomyocytes during normal life (9) and cardiomyocyte renewal (10) are 2 processes that contribute to physiological baseline concentrations of circulating troponin. Multiple additional factors can increase levels of circulating cardiac troponin in clinical scenarios other than acute coronary syndromes (11). Prasadetal.(7) showed that 37% of patients undergoing elective PCI have elevated pre-procedural troponin levels. The elevated circulating troponin level in these patients may be explained by a less favorable cardiovascular and metabolic risk profile (11), more extensive CAD (12), or clinically silent complicated atherosclerotic plaques (13). It is proposed that elevated baseline circulating troponin level in patients with stable CAD may be caused by cardiometabolic risk related stress on the myocardium (14,15) or by cycles of silent atherosclerotic plaque rupture and sealing, leading to repeated myocardial ischemia (13). Both of these conditions are associated with increased cardiovascular risk, and they may explain the association between elevated baseline hs-tnt and mortality. The principal finding of the present study was that post-procedural hs-tnt elevation was not associated with an increased risk of mortality, regardless of baseline hs-tnt level. Although elevated baseline hs- TnT in itself was strongly associated with an increased risk of mortality, it had no impact on the association between the PCI-related rise in hs-tnt and the risk of subsequent mortality up to 3 years after PCI. A PCI-related hs-tnt elevation 6 times the URL in patients with normal baseline hs-tnt and >9 timesthe URL in patients with elevated baseline levels of the biomarker was not associated with increased mortality risk. The striking separation of the Kaplan-Meier curves for mortality (showing mortality differences predominantly secondary to baseline hs-tnt levels), in addition to the results of the multivariable analysis, lend further support to this finding. Miller et al. (8) also showed that long-term prognosis was most often related to the baseline troponin level, rather than to the biomarker response to PCI. However, the study by Miller et al. differs from the present study in that it included not only patients with stable CAD but also those with acute coronary syndromes, a conventional troponin assay was used, and fewer deaths were reported. Notwithstanding these differences, the current study corroborates the findings of Miller et al. in a large series of patients with stable CAD by using a contemporary high-sensitivity cardiac troponin assay. Furthermore, by findingatrendtowardaninteraction between the post-procedural hs-tnt >70 the 99th percentile URL and baseline hs-tnt level regarding prediction of mortality, the present study may offer support to the definition of clinically relevant myocardial infarction after coronary revascularization in the Consensus Document of the Society for Cardiovascular Angiography and Interventions (16). Although factors responsible for baseline hs-tnt elevation help to explain the increased mortality risk in patients with elevated baseline levels of the biomarker, the relationship between factors associated with troponin rise post-pci and mortality is less clear. Procedure-related factors or complications increasing the risk of myocardial damage via distal embolization, side-branch occlusion, or suboptimal myocardial flow are proven risk factors for postprocedural troponin rise (3,17,18). The present study also found that increased procedural complexity was associated with elevated hs-tnt levels after PCI, regardless of the baseline hs-tnt level. However, if the troponin response to PCI was assessed by using less sensitive troponin assays, as was the case in multiple previous studies, relatively extensive myocardial damage would be needed to result in a procedure-related troponin rise. The currently used high-sensitivity troponin assays, including that used in the present study, allow detection of troponin release from minuscule damage of myocardial tissue. Thus, it is plausible that post-procedural elevation of the magnitude observed in the present study may reflect subtle PCI-related myocardial damage that is too small (or transient) to have clinical sequelae.

JACC VOL. 68, NO. 21, 2016 NOVEMBER 29, 2016:2259 68 Ndrepepa et al. Troponin and Mortality After Elective PCI 2267 AstudybyLimetal.(19) found that only a small minority (5 of 26 patients) fulfilling troponin criteria for PCI-related (type 4a) myocardial infarction had evidence of peri-procedural necrosis on cardiac magnetic resonance imaging. The investigators suggest that current troponins are oversensitive for the diagnosis of PCI-related myocardial injury. In the present study, despite post-procedural hs-tnt elevations in the majority of patients and identification of several indexes of procedural complexity as predisposing factors for this rise, optimal postprocedural TIMI flow (grade 3) was restored in >97% of patients. Thus, a combination of factors (including the use of a high-sensitivity troponin assay) capable of detecting troponin elevation caused by subtle PCIrelated myocardial damage, adjustment for baseline hs-tnt level, and the capacity of current-day PCI to achieve optimal revascularization, even in the setting of high procedural complexity, may have attenuated the association of post-procedural troponin rise with mortality. The inverse association between restenotic lesions and post-procedural hs-tnt may offer evidence for the role of distal embolization in the elevation of post-procedural hs-tnt level. Intervention in restenotic lesions, which have a higher fibrotic/atherosclerotic content ratio compared with native atherosclerotic plaques, may be associated with less distal embolization, myocardial injury, and troponin elevation due to this factor. STUDY LIMITATIONS. First, serial testing of hs-tnt levels before PCI was not performed, resulting in a lack of data on biomarker stability at baseline. If hs- TnT levels are unstable before PCI, the ability to discriminate between a spontaneous and procedurerelated hs-tnt elevation is limited. This scenario is particularly relevant in the case of patients presenting with acute coronary syndromes. Nonetheless, all patients included in the present study had clinically stable CAD at the time of the index PCI. Moreover, given that the proportion of patients with hs-tnt elevation post-pci seems to differ little between the groups with or without elevated baseline hs-tnt levels, we believe that PCI was responsible for the biomarker increase in the vast majority of patients in both groups. The collection of blood samples after the PCI follows the common practice in our center as well as others regarding post-procedural troponin measurements. We are aware of the possibility that the precise peak value of post-procedural troponin may have been missed in a number of patients. This would have required multiple measurements in a short time interval. Finally, these data are from the hs-tnt assay and may not be extrapolated to assess the performance of other conventional or high-sensitivity troponin assays. In the present study, a 0.014-mg/l cutoff was used to define the hs-tnt elevation, which differs from the 0.03-mg/lcutoffweusedpreviously to detect troponin elevations with conventional troponin assays. CONCLUSIONS In patients with CAD undergoing elective PCI, postprocedural hs-tnt elevation was not associated with an increased risk of mortality for up to 3 years in patients with or without elevated baseline levels of the biomarker. Although there was a strong and independent association between baseline hs-tnt and mortality, the PCI-related elevation in hs-tnt did not offer prognostic information beyond that provided by baseline hs-tnt levels. REPRINT REQUESTS AND CORRESPONDENCE: Dr. Gjin Ndrepepa, Deutsches Herzzentrum, Lazarettstrasse 36, 80636 München, Germany. E-mail: ndrepepa@dhm.mhn.de. PERSPECTIVES COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In patients with CAD undergoing elective PCI, the baseline level of hs-tnt was strongly and independently associated with 3-year mortality, whereas post-procedural hs-tnt elevations did not provide additional prognostic information beyond the baseline hs-tnt level. TRANSLATIONAL OUTLOOK: Further studies in larger numbers of patients are needed to determine whether certain thresholds of post-procedural hs-tnt elevation carry independent prognostic value in patients with particular patterns of CAD undergoing PCI. REFERENCES 1. Apple FS, Collinson PO. IFCC Task Force on Clinical Applications of Cardiac Biomarkers. Analytical characteristics of high-sensitivity cardiac troponin assays. Clin Chem 2012;58:54 61. 2. Jaffe AS, Ordonez-Llanos J. High-sensitivity cardiac troponin: from theory to clinical practice. Rev Esp Cardiol (Engl Ed) 2013;66: 687 91. 3. Loeb HS, Liu JC. Frequency, risk factors, and effect on long-term survival of increased troponin I following uncomplicated elective percutaneous coronary intervention. Clin Cardiol 2010;33:E40 4.

2268 Ndrepepa et al. JACC VOL. 68, NO. 21, 2016 Troponin and Mortality After Elective PCI NOVEMBER 29, 2016:2259 68 4. De Labriolle A, Lemesle G, Bonello L, et al. Prognostic significance of small troponin I rise after a successful elective percutaneous coronary intervention of a native artery. Am J Cardiol 2009; 103:639 45. 5. Cavallini C, Savonitto S, Ardissino D. Impact of the elevation of biochemical markers of myocardial damage on long-term mortality after percutaneous coronary intervention: results of the CK-MB and PCI study: reply. Eur Heart J 2005; 26:2206 7. 6. Ndrepepa G, Braun S, Cassese S, et al. Prognostic value of high-sensitivity troponin T after percutaneous coronary intervention in patients with stable coronary artery disease. Rev Esp Cardiol (Engl Ed) 2016;69:746 53. 7. Prasad A, Rihal CS, Lennon RJ, et al. Significance of periprocedural myonecrosis on outcomes after percutaneous coronary intervention: an analysis of preintervention and postintervention troponin T levels in 5487 patients. Circ Cardiovasc Interv 2008;1:10 9. 8. Miller WL, Garratt KN, Burritt MF, et al. Baseline troponin level: key to understanding the importance of post-pci troponin elevations. Eur Heart J 2006;27:1061 9. 9. Olivetti G, Giordano G, Corradi D, et al. Gender differences and aging effects on the human heart. J Am Coll Cardiol 1995;26:1068 79. 10. Bergmann O, Bhardwaj RD, Bernard S, et al. Evidence for cardiomyocyte renewal in humans. Science 2009;324:98 102. 11. Ndrepepa G, Braun S, Mehilli J, et al. Prognostic value of sensitive troponin T in patients with stable and unstable angina and undetectable conventional troponin. Am Heart J 2011;161:68 75. 12. Ndrepepa G, Braun S, Schulz S, et al. Highsensitivity troponin T level and angiographic severity of coronary artery disease. Am J Cardiol 2011;108:639 43. 13. Korosoglou G, Lehrke S, Mueller D, et al. Determinants of troponin release in patients with stable coronary artery disease: insights from CT angiography characteristics of atherosclerotic plaque. Heart 2011;97:823 31. 14. Unger RH. Lipotoxic diseases. Annu Rev Med 2002;53:319 36. 15. Hessel MH, Michielsen EC, Atsma DE, et al. Release kinetics of intact and degraded troponin I and T after irreversible cell damage. Exp Mol Pathol 2008;85:90 5. 16. Moussa ID, Klein LW, Shah B, et al. Consideration of a new definition of clinically relevant myocardial infarction after coronary revascularization: an expert consensus document from the Society for Cardiovascular Angiography and Interventions (SCAI). Catheter Cardiovasc Interv 2014;83:27 36. 17. Ricciardi MJ, Davidson CJ, Gubernikoff G, et al. Troponin I elevation and cardiac events after percutaneous coronary intervention. Am Heart J 2003;145:522 8. 18. Lansky AJ, Stone GW. Periprocedural myocardial infarction prevalence, prognosis, and prevention. Circ Cardiovasc Interv 2010;3: 602 10. 19. Lim CC, van Gaal WJ, Testa L, et al. With the universal definition, measurement of creatine kinase-myocardial band rather than troponin allows more accurate diagnosis of periprocedural necrosis and infarction after coronary intervention. J Am Coll Cardiol 2011;57:653 61. KEY WORDS angina, biomarkers, coronary artery disease, coronary stenosis, prognosis, risk factors, stable