- Email: [email protected]
Troponin T is a better predictor than creatine kinase-MB of long-term mortality after coronary artery bypass graft surgery
Surgery
Troponin T is a better predictor than creatine kinase-MB of long-term mortality after coronary artery bypass graft surgery Camilla Lund Søraas, MD, a Charlotte Friis, MD, b Kristin Victoria Tunheim Engebretsen, MD, b Leiv Sandvik, PhD, c,d Sverre Erik Kjeldsen, MD, PhD, a,d and Theis Tønnessen, MD, PhD b,d Oslo, Norway
Background Elevations of creatine kinase-MB (CK-MB) and cardiac troponin T (cTnT) have an uncertain long-term prognostic value after coronary artery bypass graft (CABG) surgery. We aimed to test the hypothesis that CK-MB and cTnT are predictors of long-term survival after CABG and to assess which of these 2 biomarkers is the better predictor. Methods A total of 1,350 consecutive patients undergoing isolated on-pump CABG had CK-MB and cTnT measured at 7, 20, and 44 hours, postoperatively. The end point was all-cause mortality, and during the median follow-up time of 6.1 years, 207 patients (15.3%) died. Results Both peak CK-MB and peak cTnT independently predicted long-term mortality (hazard ratio [HR] 1.003, 95% confidence interval [CI] 1.001-1.005, P = .007, and HR 1.31, 95% CI 1.17-1.46, P b.001, respectively) when analyzed in separate multivariate Cox models, adjusting for baseline demographic characteristics and perioperative risk factors. However, when analyzed simultaneously in the same Cox model, cTnT was a significant predictor (HR 1.31, 95% CI 1.13-1.51, P b.001), whereas CK-MB was not (P = .99). Similar results were found when the biomarkers were analyzed together in a Cox model adjusting for European System for Cardiac Operative Risk Evaluation. The differences in mortality between the biomarker groups were consistent also when analyzing strict quartiles of peak values of CK-MB and cTnT (P = .81 and P = .001, respectively). Conclusions
Both CK-MB and cTnT are predictors of mortality after CABG surgery; however, our data suggest that cTnT is a better predictor of long-term mortality after CABG surgery than CK-MB. (Am Heart J 2012;164:779-85.)
Elevations of the biochemical markers of myocardial damage, creatine kinase–myocardial band (CK-MB), and troponins are common after coronary artery bypass graft (CABG) surgery and have been associated with increased inhospital and short-term mortality. 1-4 However, few studies have assessed long-term mortality, especially regarding cardiac troponin T (cTnT), and larger studies are needed. In a recent meta-analysis, peak CK-MB taken from 7 studies predicted mortality at 1 and 12 months postoperatively but not significantly after 5 years. 5 Another systematic review found that postoperative CKMB release was associated with mortality up to 40 months
From the aDepartment of Cardiology, Oslo University Hospital, Ullevaal, Oslo, Norway, b Department of Cardiothoracic Surgery, Oslo University Hospital, Ullevaal, Oslo, Norway, c Section of Biostatistics and Epidemiology, Oslo University Hospital, Ullevaal, Oslo, Norway, and dFaculty of Medicine, University of Oslo, Oslo, Norway. Submitted January 21, 2012; accepted May 7, 2012. Reprint requests: Theis Tønnessen, MD, PhD, Department of Cardiothoracic Surgery, Oslo University Hospital, Ullevaal, Pb. 4956 Nydalen, 0424 Oslo, Norway. E-mail: [email protected] 0002-8703/$ - see front matter © 2012, Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2012.05.027
postoperatively but could not conclude regarding troponins due to few published studies. 6 Hence, the prognostic value of these biomarkers is still unclear because few large, long-term studies of cardiac biomarkers after CABG have been conducted. Previous studies have largely investigated either CK-MB or troponins as predictors postoperatively. Few studies have assessed both CK-MB and troponins simultaneously as predictors of mortality in the same patient population, and these suffer from limited study size. 7-9 Levels of troponins could presumably be better predictors of mortality than CK-MB because of their increased sensitivity and specificity to the myocardium and measurements thereof represent the “gold standard” in patients with acute coronary syndrome. 10 Thus, we aimed to test the hypothesis that CK-MB and cTnT are long-term predictors of mortality after CABG and to determine which of these 2 markers of myonecrosis is the better predictor.
Materials and methods Patients One thousand three hundred fifty patients operated on consecutively for isolated CABG, in the time period from
American Heart Journal November 2012
780 Søraas et al
January 1, 2003, to December 31, 2006, at Oslo University Hospital, Ullevaal, were included in the study. Patients undergoing additional surgery to CABG were excluded. Data, registered prospectively, were obtained from the hospital database and from patient files. The study was approved by an institutional review board. Few data were missing, and these were coded as absent. All-cause mortality status on all patients through June 30, 2011, was established through the Norwegian National Death Registry. A few patients had emigrated, and these were censored from the emigration date or date of last contact. Median (interquartile range) follow-up time was 6.1 (5.1-7.2) years. Five patients were excluded because they could not be identified in the National Registry, and 3 patients were excluded because of missing postoperative blood samples. Thus, a total of 1,342 patients underwent further analysis. One thousand two hundred ninety-four of these patients had undergone elective surgery and 48 had undergone emergent surgery (3.6%), according to the definition by European System for Cardiac Operative Risk Evaluation (EuroSCORE). 11 Hypertension was defined as clinically recognized and under medical treatment. Smoking was defined as smoking until the month before operation. Renal failure was defined as preoperative creatinine N200 μmol/L (2.26 mg/dL). Unstable angina was defined as rest angina pectoris requiring intravenous nitrate until arrival in the anesthetic room. The EuroSCORE was assessed preoperatively according to a scoring system based on validated preoperative risk factors that indicate the 30-day mortality risk for each patient undergoing open heart surgery. 11
Surgical procedure All patients were operated on with median sternotomy and using the standard technique of cardiopulmonary bypass with ascending aorta cannulation, single venous cannulation, and moderate systemic hypothermia (32-34°C), as previously described. 12
Measurements of biomarkers Venous blood samples were routinely collected at 7, 20, and 44 hours postoperatively. Both CK-MB and cTnT were immediately analyzed by the Department of Clinical Chemistry at Oslo University Hospital, Ullevaal. Serum cTnT concentration was determined by using the third- or fourth-generation TnT assay (Troponin T STAT) on the Roche Elecsys 2010 immunoassay analyzer (Roche Diagnostics, Mannheim, Germany) according to the manufacturer recommendations. Serum CKMB concentration was determined by using 2 monoclonal antibodies (CKMB STAT) on an Elecsys 2010 analyzer (Roche diagnostics) by electrochemiluminescence immunoassay. The upper normal reference limit is 0.10 μg/L for cTnT and 5 μg/L for CK-MB.
Statistical analysis Data are presented as mean ± SD for continuous variables with normal distribution, as median with interquartile range in variables with nonnormal distribution, and as proportions for categorical variables. Mean values were compared using Student t test for normally distributed data and Mann-Whitney U test for nonparametric data. Categorical variables were compared using
the χ 2 test or Fisher exact test when appropriate. Event rates were calculated and plotted according to the Kaplan-Meier product limit method. The relation of CK-MB and cTnT to mortality was assessed using Cox proportional hazards models. 13 Log minus log of survival plotted against survival times verified the proportional hazards assumption. The independence of the relationship of cTnT and CK-MB to mortality was evaluated in univariate and multivariate Cox regression analyses. All clinical relevant baseline and procedural variables were tested in Cox univariate analyses. Variables with a univariate P b .10 were included in the multivariate Cox model. Collinearity was checked by calculating variance inflation factor (VIF), and VIF b5 was considered acceptable. The variable EuroSCORE gave VIF N6 and thus was excluded from the multivariate analyses with other clinical variables. The relationship of CK-MB to mortality was also evaluated in a multivariate Cox model with baseline variables, but not with cTnT as covariate. The results from the Cox analyses are presented as hazard ratios (HRs) with 95% CIs. For all tests, 2-tailed P b .05 was required for statistical significance. Data management and analysis were performed with IBM SPSS version 19.0 (SPSS Inc, Chicago, IL) software. No extramural funding was used to support this work, and the authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of this manuscript, and its final contents.
Results Patient characteristics A total of 1,342 patients were available for analysis, and of these, 1314 patients (97.9%) had complete cTnT and CK-MB values for all 3 postoperative time points, and 28 patients (2.1%) had values for 2 postoperative time points. The mean age ± SD for the cohort was 64.9 ± 10.0 years, and there were 1,089 men (81.1%). The all-cause mortality at 30 days was 0.67%. Totally, 207 patients (15.4%) died during a follow-up period of a median (interquartile range) of 6.1 (5.1-7.2) years. Patients who deceased were older with lower body mass index and had, at the time of surgery, higher prevalences of hypertension, peripheral vascular disease, renal failure and higher creatinine, lower left ventricular ejection fraction (LVEF), and higher EuroSCORE compared with patients who survived (Table I). Perioperatively, they had longer cardiopulmonary bypass time, and postoperatively, they had higher creatinine and peak cTnT levels (P b.001) but not different peak CK-MB (P = .21). Biomarker values The cardiac biomarkers were measured at 3 time points postoperatively, and the median (interquartile range) postoperative values for cTnT were 0.48 (0.32-0.75) μg/L at 7 hours, 0.31 (0.19-0.57) μg/L at 20 hours, and 0.20 (0.10-0.41) μg/L at 44 hours postoperatively. The corresponding CK-MB values were 26 (19-35) μg/L at 7 hours, 21 (14-31) μg/L at 20 hours, and 5 (4-8) μg/L at 48 hours postoperatively. Both biomarkers peaked rapidly within 7 hours post-CABG in most patients. The peak
American Heart Journal Volume 164, Number 5
Søraas et al 781
Table I. Patient characteristics according to mortality status Baseline characteristics Preoperative parameters Demographics Age (y) Male gender, n (%) Body mass index (kg/m2) Medical history Hypertension, n (%) Recent myocardial infarction (b90 d preop), n (%) Diabetes mellitus, n (%) COPD, n (%) Apoplexia cerebri, n (%) Peripheral vascular disease, n (%) Renal failure, n (%) Neurologic disease, n (%) Current smokers, n (%) NYHA class III or IV, n (%) LVEF preop (%) Left main stem stenosis, n (%) Creatinine preop (mg/dL) EuroSCORE Perioperative/postoperative parameters Total distals, n Unstable angina, n (%) Emergent procedure, n (%) CPB time (min) Cross-clamp time (min) Creatinine day 2 postop (mg/dL) Peak postop cTnT (μg/L) Peak postop CK-MB (μg/L)
All patients (N = 1342)
Survivor (n = 1135)
Nonsurvivor (n = 207)
P⁎
64.9 ± 10.0 1089 (81.1) 26.7 ± 3.9
63.9 ± 9.8 926 (81.6) 26.8 ± 3.8
70.6 ± 9.4 163 (78.7) 26.2 ± 4.3
b.001 .39 .020
556 (41.5) 474 (35.3) 213 (15.9) 88 (6.6) 26 (1.9) 125 (9.3) 18 (1.3) 18 (1.3) 366 (27.3) 1040 (77.5) 60.6 ± 14.5 265 (19.7) 0.93 (0.81-1.07) 3 (2-5)
452 (39.9) 389 (34.3) 175 (15.4) 68 (6.0) 23 (2.0) 85 (7.5) 7 (0.6) 15 (1.3) 307 (27.1) 870 (81.0) 61.2 ± 14.4 218 (19.2) 0.91 (0.80-1.04) 3 (1-5)
104 (50.2) 85 (41.1) 38 (18.4) 20 (9.7) 3 (1.4) 40 (19.3) 11 (5.3) 3 (1.4) 59 (28.5) 170 (85.0) 57.3 ± 15.0 47 (22.7) 1.02 (0.84-1.36) 5 (4-7)
.007 .072 .34 .070 .79 b.001 b.001 .75 .75 .22 b.001 .29 b.001 b.001
3.1 ± 0.9 23 (1.7) 48 (3.6) 75.4 ± 22.2 38.8 ± 13.6 0.85 (0.74-1.02) 0.53 (0.34-0.83) 27 (20-38)
3.1 ± 0.9 18 (1.6) 41 (3.6) 74.9 ± 22.0 38.6 ± 13.6 0.84 (0.73-0.98) 0.51 (0.34-0.79) 27 (20-38)
3.2 ± 0.9 5 (2.4) 7 (3.4) 78.3 ± 22.9 39.8 ± 14.1 0.97 (0.80-1.47) 0.63 (0.39-1.02) 29 (20-41)
.076 .38 1.00 .040 .27 b.001 b.001 .21
⁎ P values refer to nonsurvivor vs survivor. Data are provided as means ± SD, number (%), or median (interquartile range). Preop, Preoperatively; COPD, chronic obstructive pulmonary disease; NYHA, New York Heart Association; CPB, cardiopulmonary bypass; Postop, postoperatively.
postoperative cTnT value was measured at 7 hours postoperatively in 1,044 (77.8%) of the patients, at 20 hours postoperatively in 205 (15.3%), and at 44 hours postoperatively in 113 (8.4%) of the patients. Peak cTnT was a median of 0.53 (interquartile range 0.34-0.83) μg/ L, and peak CK-MB was 27 (20-38) μg/L. Sixteen patients (1.2%) had peak cTnT values within the normal reference area, and 2 patients (0.1%) had normal peak CK-MB values.
Biomarker quartile analysis Figure 1 shows the Kaplan-Meier survival curves for patients grouped by (a) quartile of peak cTnT and (b) quartile of peak CK-MB. There were significant differences in mortality between the quartiles of cTnT (χ 2 = 17.15, P = .001), but not between the CK-MB quartiles (χ 2 = 0.98, P = .81). The highest mortality was observed among patients with a cTnT level in the top quartile (quartile 1 vs quartile 2, χ 2 = 1.54, P = .21; quartile 1 vs quartile 3, χ 2 = 1.90, P = .17; and quartile 1 vs quartile 4, χ 2 = 6.92, P = .009). For CK-MB levels, no quartile differed significantly from the others (quartile 1 vs quartile 2, χ 2 = 0.13, P = .72; quartile 1 vs quartile 3, χ 2 = 0.33, P = .57; and quartile 1 vs quartile 4, χ 2 = 0.15, P = .70).
Mortality in relationship to quartiles of cTnT and CK-MB is shown in Table II. For each quartile of cTnT (horizontal row), there is a trend that the percentage mortality is rather constant, independent of the CK-MB quartiles. For CK-MB, however, the percentage mortality seems more associated with the cTnT quartile, that is, the percentage mortality increases vertically within each CK-MB quartile dependent of the matching cTnT quartile. This indicates that CK-MB gives little additional prognostic information apart from cTnT.
Univariate and multivariate predictors of mortality In univariate analyses, age, body mass index, hypertension, chronic obstructive pulmonary disease, peripheral vascular disease, LVEF, preoperative creatinine, creatinine at day 2 postoperatively, EuroSCORE, and peak postoperative cTnT and CK-MB were significantly associated with mortality. Age, peripheral vascular disease, LVEF, serum creatinine at day 2 postoperatively, and peak cTnT remained significant predictors in multivariate analysis, and age was the strongest predictor (Table III). Creatine kinase-MB was not an independent predictor of mortality in the multivariate model including cTnT. However, in multivariate analysis with the same clinical
American Heart Journal November 2012
782 Søraas et al
Figure 1
Table II. Mortality stratified by quartiles of CK-MB and cTnT CK-MB CK-MB CK-MB CK-MB quartile 1 quartile 2 quartile 3 quartile 4 cTnT quartile 1 cTnT quartile 2 cTnT quartile 3 cTnT quartile 4
13% (26/205) 13% (13/104) 21% (10/48) 36% (5/14)
13% (10/80) 7% (8/111) 20% (17/87) 30% (8/27)
13% (6/45) 12% (12/98) 19% (20/108) 22% (18/83)
20% (2/10) 8% (3/37) 12% (9/77) 19% (40/208)
Mortality data are expressed as percentages with the absolute numbers of patients who have died in each group divided by the total number of patients in each group in parentheses.
prognostic significance in univariate analyses (data not shown) and when adjusted for EuroSCORE (Figure 2).
Discussion
A, Kaplan-Meier survival curves for mortality associated with quartiles of peak cTnT, P = .001. B, Kaplan-Meier survival curves for mortality associated with quartiles of peak CK-MB, P = .81.
variables but excluding cTnT, peak CK-MB was a significant independent predictor (P = .007). The biomarkers were also analyzed adjusting for EuroSCORE (Table IV). Peak cTnT, but not peak CK-MB, was a significant independent predictor of mortality when separately analyzed. Likewise, when these 2 biomarkers were included together in the same multivariate model, cTnT was highly significant, whereas CK-MB was nonsignificant. There was no significant interaction between CK-MB and cTnT. The predictive values of CK-MB and cTnT at different time points varied in our population. For both CK-MB and cTnT, the values at 44 hours postoperatively had the greatest
The present study evaluated the long-term prognostic significance of myocardial damage defined by CK-MB and cTnT release after CABG. We found that both CK-MB and cTnT were predictors of long-term all-cause mortality after CABG in our patient population. However, when analyzing the 2 cardiac markers in the same Cox model, cTnT was a highly significant predictor of mortality, whereas CK-MB was nonsignificant. Troponin T remained a highly significant independent predictor of mortality after adjustment for a wide range of baseline and perioperative variables. Nearly all patients had elevated cTnT and CK-MB postoperatively. The biomarker value at 44 hours postoperatively had a greater predictive value than the values at 7 or 20 hours or the peak value. Both CK-MB and cTnT have been shown to be predictors of inhospital and short-term mortality after CABG, 1-4 but few studies have examined the relationship to long-term mortality. We found that both CK-MB and cTnT were independent predictors of long-term mortality when examined in separate multivariate Cox analyses adjusting for a large set of widely accepted baseline and perioperative prognostic factors. Interestingly, however, when both biomarkers were analyzed together in the same model, CK-MB gave no additional prognostic information beyond cTnT. A similar result was found when the biomarkers were analyzed together in a model adjusting for the commonly used preoperative risk score EuroSCORE. This may be explained by the higher sensitivity and specificity of cTnT for myocardial damage. Skeletal muscle may contain up to 6% CK-MB. 14 Consequently, skeletal muscle trauma from dissection and internal mammary artery harvest during CABG may increase the levels of CK-MB postoperatively. Few studies have compared CK-MB and cTnT regarding mortality after CABG. Vikenes et al. 8compared cardiac
American Heart Journal Volume 164, Number 5
Søraas et al 783
Table III. Predictors of mortality in univariate and multivariate analyses Univariate analysis Variables
HR (95% CI)
Age (per year) Male gender Body mass index (per 1 kg/m 2) History of hypertension History of myocardial infarction (b90 d preoperatively) History of diabetes mellitus History of COPD History of peripheral vascular disease Current smoker NYHA class III or IV (yes/no) LVEF (per 1 %) Left main stem stenosis Creatinine preop (per 1 mg/dL) EuroSCORE (per unit) Total distals Emergent procedure CPB time (per 1 min) Cross-clamp time (per 1 min) Unstable angina Creatinine day 2 postop (per 1 mg/dL) Peak postoperative cTnT (per 1 μg/L) Peak postoperative CK-MB (per 1 μg/L)
1.08 0.85 0.96 1.53 1.31 1.22 1.63 2.64 1.07 1.37 0.98 1.26 1.29 1.29 1.14 0.89 1.005 1.003 1.28 1.56 1.38 1.003
(1.06-1.10) (0.61-1.19) (0.92-0.99) (1.17-2.01) (1.00-1.73) (0.86-1.74) (1.03-2.59) (1.87-3.73) (0.79-1.45) (0.93-2.02) (0.97-0.99) (0.91-1.74) (1.19-1.38) (1.23-1.35) (0.98-1.33) (0.42-1.88) (1.00-1.01) (0.99-1.01) (0.53-3.11) (1.42-1.71) (1.25-1.52) (1.001-1.006)
Multivariate analysis
χ2
P
79.5 0.9 5.7 9.4 3.7 1.3 4.3 30.4 0.2 2.5 14.7 1.9 39.6 119.9 2.9 0.1 2.7 0.5 0.3 90.8 43.7 7.7
b.001 .34 .02 .002 .054 .26 .037 b.001 .65 .115 b.001 .169 b.001 b.001 .087 .75 .102 .49 .59 b.001 b.001 .005
χ2
P
1.07 (1.06-1.09)
62.6
b.001
0.99 (0.95-1.03) 1.19 (0.90-1.56) 0.95 (0.70-1.27)
0.2 1.5 0.1
.62 .23 .71
1.40 (0.88-2.24) 1.66 (1.15-2.40)
2.0 7.4
.161 .006
0.98 (0.97-0.99)
12.6
b.001
1.04 (0.89-1.21)
0.3
.60
45.1 13.6 0.0
b.001 b.001 .99
HR (95% CI)
1.45 (1.30-1.62) 1.31 (1.13-1.51) 1.00 (0.997-1.003)
COPD, Chronic obstructive pulmonary disease; NYHA, New York Heart Association; Preop, preoperatively; CPB, cardiopulmonary bypass; Postop, postoperatively.
Table IV. The value of adding peak biomarkers to a clinical model or EuroSCORE in Cox regression analyses for predicting mortality Model cTnT only CK-MB only Clinical model + cTnT Clinical model + CK-MB Clinical model + CKMB + cTnT EuroSCORE + cTnT EuroSCORE + CKMB EuroSCORE + CKMB + cTnT
HR (95% CI)
P⁎
1.34 1.14 1.27 1.13
(1.23-1.45) (1.04-1.25) (1.15-1.41) (1.03-1.24)
b.001 .005 b.001 .007
1.00 1.28 1.22 1.09
(0.88-1.14) (1.12-1.45) (1.11-1.34) (0.99-1.20)
.99 b.001 b.001 .07
0.93 (0.80-1.09) 1.28 (1.12-1.46)
.38 b.001
⁎ P denotes significance of the added biomarker variable in the model. HR and 95% CIs are shown per 1 SD for each biomarker peak value (0.90 μg/L for cTnT and 38.4 μg/L for CK-MB). The clinical model contains clinical predictors with P b .10 in univariate analysis (age, body mass index, hypertension, prior myocardial infarction b90 days preoperatively, chronic obstructive pulmonary disease, peripheral vascular disease, preoperative LVEF, total distals, creatinine at day 2 postoperatively).
troponin I and cTnT with CK-MB and concluded that CK-MB was superior to the cardiac troponins in predicting event-free survival in low-risk patients with a median follow-up time of 7.7 years However, this study included 156 patients with isolated CABG only, and the discrepancy between the findings of this study and ours may be due to the limited number of patients
included. Our results are, on the other hand, supported by 2 recent studies that found that cardiac troponin I was better than CK-MB for quantifying myocyte necrosis as assessed by magnetic resonance imaging after CABG. 15,16 Clearly, taken together with other previous studies of the prognostic importance of troponins, 2,7,17 our finding with cTnT suggests that these more specific markers of myocardial necrosis have stronger prognostic power than the less cardiac specific enzyme CK-MB and should be considered the preferred assay. The mechanism of how small leakages of troponins are associated with increased mortality in the post-CABG setting is unknown. However, the same relationships are found after percutaneous coronary intervention, 18,19 noncardiac surgery, 20 and in the general population, 21 indicating common pathophysiologic pathways, regardless of the type of initiating stimulus. Another possibility is that these markers are just surrogate markers of other, yet undiscovered, markers of worse prognosis in these patients. Currently, guidelines do not address the prognostic influence of postoperative biomarker levels. An interesting debate is whether postoperative biomarker levels could be incorporated into the widely used EuroSCORE or whether they are sufficiently powerful to remain independent. However, because the EuroSCORE is a preoperative scoring system, incorporating postoperative biomarker levels seems to be somewhat conflicting. Our data, in line with other studies, show that biomarkers have independent prognostic importance.
American Heart Journal November 2012
784 Søraas et al
Figure 2
Forest plot demonstrating the predictive value of biomarkers at various time points postoperatively for all-cause mortality. Cox multivariable proportional hazards model, adjusted for EuroSCORE, is shown for each individual variable. Hazard ratio and 95% CIs are shown per 1 SD for each variable. All biomarkers were entered as continuous variables. 7, 7 hours postoperatively; 20, 20 hours postoperatively; 44, 44 hours postoperatively; peak, peak value postoperatively. Significances for mortality: cTnT 7 (HR 1.13, 95% CI 1.03-1.24, P = .009), cTnT 20 (HR 1.16, 95% CI 1.05-1.29, P = .004), cTnT 44 (HR 1.24, 95% CI 1.151.34, P b.001), peak cTnT (HR 1.22, 95% CI 1.11-1.34, P b.001), CKMB 20 (HR 1.10, 95% CI 1.00-1.21, P = .05), and CK-MB 44 (HR 1.13, 95% CI 1.05-1.22, P = .002).
Therefore, in our opinion, patients with elevated biomarkers should have a more intensive follow-up schedule after hospital discharge. Compliance with evidence-based guidelines documented to improve survival from coronary heart disease has been shown to be less after CABG than in nonsurgical populations. 22 Thus, in patients with increased postoperative biomarker levels, there should be put especially emphasis on implementing evidencebased medical therapies. Perhaps these patients also should have a lower threshold for angiography in the case of new angina symptoms. However, these proposals for future guidelines may need to be validated in prospective studies. Our study was strengthened by the long follow-up, equal to N7,900 patient-years. Hypothetically, the long follow-up and use of all-cause mortality (rather than cardiovascular death) might have diluted the prognostic importance of cTnT and CK-MB, yet we found that the biochemical markers still carried significant prognostic information. Most of the published studies to date have examined quite well-defined homogenous CABG cohorts with strict inclusion criteria. Our study is strengthened by the heterogeneous population; all consecutive patients undergoing isolated CABG in our center during 3 years, hence, increasing the applicability of our results to the clinical setting. Nonetheless, despite our heterogenic
study population, the all-cause mortality at 30 days was 0.67%, confirming the low risk of this study population. The 2011 American College of Cardiology Foundation/ American Heart Association CABG guidelines state that measurement of biomarkers of myonecrosis is reasonable in the first 24 hours after CABG, 23 and several studies have examined biomarkers within this time frame postoperatively. Interestingly, we found that biomarker release for both CK-MB and cTnT at 44 hours postoperatively had the greatest prognostic power. Lehrke et al 24 showed that cTnT measurement at 48 hours postoperatively was the best predictor for mortality, consistent with our data. Moreover, a recent study demonstrated that the troponin I value at 72 hours was the greatest predictor of mid-term mortality. 25 Taken together, these studies and our findings support extended measurements of troponins during the first postoperative days. Possibly, the later release of biochemical markers may be of greater discriminatory value due to gradual release of structurally bound troponin representing greater irreversible injury and necrosis. Although our study showed that biomarkers at 44 hours postoperatively had slightly greater prognostic power compared with the peak values, the peak values are easier to follow in the clinical setting and should be recommended used for clinicians.
Study limitations First, some patients with recent myocardial infarction may have had preoperatively elevated biochemical markers that could influence marker levels postoperatively. Second, we lack follow-up data regarding reinterventions and complications after the operation, which inevitably may affect long-term outcome. Third, we used the well-recognized EuroSCORE as an externally valid assessment of individual risk of long-term mortality although the EuroSCORE is not constructed for predicting mortality exceeding 30 days postoperatively and is shown to overestimate mortality. 11,26 Another limitation is that we examined all-cause mortality and did not differ between cardiac or non-cardiac cause of death. However, the accuracy of death certificates for coding coronary heart disease as the cause of death is questionable and thus, all-cause mortality is a more strict and objective parameter. 27 Finally, this is a single-center study with a limited number of patients, so delineation of quantitative relationships should be done with caution.
Conclusions In conclusion, both CK-MB and cTnT are predictors of mortality after CABG surgery; however, our data suggest that cTnT is a better predictor of long-term mortality after CABG surgery than CK-MB. Further studies are warranted to identify the best follow-up regime for patients with postoperatively elevated cTnT.
American Heart Journal Volume 164, Number 5
Acknowledgements The authors are grateful for the help provided by Dr Stein Koldsland, Heidi Thorstensen, and the staff at the Department of Cardiothoracic Surgery, Oslo University Hospital, Ullevaal.
Disclosures None.
References 1. Nesher N, Alghamdi AA, Singh SK, et al. Troponin after cardiac surgery: a predictor or a phenomenon? Ann Thorac Surg 2008;85:1348-54. 2. Croal BL, Hillis GS, Gibson PH, et al. Relationship between postoperative cardiac troponin I levels and outcome of cardiac surgery. Circulation 2006;114:1468-75. 3. Ramsay J, Shernan S, Fitch J, et al. Increased creatine kinase MB level predicts postoperative mortality after cardiac surgery independent of new Q waves. J Thorac Cardiovasc Surg 2005; 129:300-6. 4. Klatte K, Chaitman BR, Theroux P, et al. Increased mortality after coronary artery bypass graft surgery is associated with increased levels of postoperative creatine kinase–myocardial band isoenzyme release: results from the GUARDIAN trial. J Am Coll Cardiol 2001;38:1070-7. 5. Domanski MJ, Mahaffey K, Hasselblad V, et al. Association of myocardial enzyme elevation and survival following coronary artery bypass graft surgery. JAMA 2011;305:585-91. 6. Petäjä L, Salmenperä M, Pulkki K, et al. Biochemical injury markers and mortality after coronary artery bypass grafting: a systematic review. Ann Thorac Surg 2009;87:1981-92. 7. Januzzi JL, Lewandrowski K, MacGillivray TE, et al. A comparison of cardiac troponin T and creatine kinase-MB for patient evaluation after cardiac surgery. J Am Coll Cardiol 2002;39:1518-23. 8. Vikenes K, Andersen KS, Melberg T, et al. Long-term prognostic value of cardiac troponin I and T versus creatine kinase-MB mass after cardiac surgery in low-risk patients with stable symptoms. Am J Cardiol 2010;106:780-6. 9. Kathiresan S, Servoss SJ, Newell JB, et al. Cardiac troponin T elevation after coronary artery bypass grafting is associated with increased one-year mortality. Am J Cardiol 2004;94:879-81. 10. Thygesen K, Alpert JS, White HD, et al. Universal definition of myocardial infarction. Circulation 2007;116:2634-53. 11. Nashef SA, Roques F, Michel P, et al. European System for Cardiac Operative Risk Evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9-13. 12. Larstorp AC, Lund Søraas C, Tønnessen T, et al. Scintigraphic demonstration of myocardial perfusion and ischaemia associated with coronary artery bypass grafting. Scand Cardiovasc J 2006;40: 354-62.
Søraas et al 785
13. Cox DR. Regression models and life-tables. J R Stat Soc, B 1972;34: 187-220. 14. Tsung JS, Tsung SS. Creatine kinase isoenzymes in extracts of various human skeletal muscles. Clin Chem 1986;32:1568-70. 15. Lim CC, Cuculi F, van Gaal WJ, et al. Early diagnosis of perioperative myocardial infarction after coronary bypass grafting: a study using biomarkers and cardiac magnetic resonance imaging. Ann Thorac Surg 2011;92:2046-53. 16. Pegg TJ, Maunsell Z, Karamitsos TD, et al. Utility of cardiac biomarkers for the diagnosis of type V myocardial infarction after coronary artery bypass grafting: insights from serial cardiac MRI. Heart 2011;97:810-6. 17. Adabag AS, Rector T, Mithani S, et al. Prognostic significance of elevated cardiac troponin I after heart surgery. Ann Thorac Surg 2007;83:1744-50. 18. Ohlmann P, Jaquemin L, Morel O, et al. Prognostic value of C-reactive protein and cardiac troponin I in primary percutaneous interventions for ST-elevation myocardial infarction. Am Heart J 2006;152: 1161-7. 19. Feldman DN, Kim L, Rene AG, et al. Prognostic value of cardiac troponin-I or troponin-T elevation following nonemergent percutaneous coronary intervention: a meta-analysis. Catheter Cardiovasc Interv 2011;77:1020-30. 20. Levy M, Heels-Ansdell D, Hiralal R, et al. Prognostic value of troponin and creatine kinase muscle and brain isoenzyme measurement after noncardiac surgery: a systematic review and meta-analysis. Anesthesiology 2011;114:796-806. 21. de Lemos JA, Drazner MH, Omland T, et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304:2503-12. 22. Hiratzka LF, Eagle KA, Liang L, et al. Atherosclerosis secondary prevention performance measures after coronary bypass graft surgery compared with percutaneous catheter intervention and nonintervention patients in the get with the guidelines database. Circulation 2007;116:I207-12. 23. Hillis LD, Smith PK, Anderson JL, et al. 2011 ACCF/AHA guideline for coronary artery bypass graft surgery: executive summary: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2011; 124:2610-42. 24. Lehrke S, Steen H, Sievers HH, et al. Cardiac troponin T for prediction of short- and long-term morbidity and mortality after elective open heart surgery. Clin Chem 2004;50:1560-7. 25. Ranasinghe AM, Quinn DW, Richardson M, et al. Which troponometric best predicts midterm outcome after coronary artery bypass graft surgery? Ann Thorac Surg 2011;91:1860-7. 26. Engebretsen KV, Friis C, Sandvik L, et al. Survival after CABG—better than predicted by EuroSCORE and equal to the general population. Scand Cardiovasc J 2009;43:123-8. 27. Lauer MS, Blackstone EH, Young JB, et al. Cause of death in clinical research: time for a reassessment? J Am Coll Cardiol 1999;34: 618-20.
Troponin T is a better predictor than creatine kinase-MB of long-term mortality after coronary artery bypass graft surgery Camilla Lund Søraas, MD, a Charlotte Friis, MD, b Kristin Victoria Tunheim Engebretsen, MD, b Leiv Sandvik, PhD, c,d Sverre Erik Kjeldsen, MD, PhD, a,d and Theis Tønnessen, MD, PhD b,d Oslo, Norway
Background Elevations of creatine kinase-MB (CK-MB) and cardiac troponin T (cTnT) have an uncertain long-term prognostic value after coronary artery bypass graft (CABG) surgery. We aimed to test the hypothesis that CK-MB and cTnT are predictors of long-term survival after CABG and to assess which of these 2 biomarkers is the better predictor. Methods A total of 1,350 consecutive patients undergoing isolated on-pump CABG had CK-MB and cTnT measured at 7, 20, and 44 hours, postoperatively. The end point was all-cause mortality, and during the median follow-up time of 6.1 years, 207 patients (15.3%) died. Results Both peak CK-MB and peak cTnT independently predicted long-term mortality (hazard ratio [HR] 1.003, 95% confidence interval [CI] 1.001-1.005, P = .007, and HR 1.31, 95% CI 1.17-1.46, P b.001, respectively) when analyzed in separate multivariate Cox models, adjusting for baseline demographic characteristics and perioperative risk factors. However, when analyzed simultaneously in the same Cox model, cTnT was a significant predictor (HR 1.31, 95% CI 1.13-1.51, P b.001), whereas CK-MB was not (P = .99). Similar results were found when the biomarkers were analyzed together in a Cox model adjusting for European System for Cardiac Operative Risk Evaluation. The differences in mortality between the biomarker groups were consistent also when analyzing strict quartiles of peak values of CK-MB and cTnT (P = .81 and P = .001, respectively). Conclusions
Both CK-MB and cTnT are predictors of mortality after CABG surgery; however, our data suggest that cTnT is a better predictor of long-term mortality after CABG surgery than CK-MB. (Am Heart J 2012;164:779-85.)
Elevations of the biochemical markers of myocardial damage, creatine kinase–myocardial band (CK-MB), and troponins are common after coronary artery bypass graft (CABG) surgery and have been associated with increased inhospital and short-term mortality. 1-4 However, few studies have assessed long-term mortality, especially regarding cardiac troponin T (cTnT), and larger studies are needed. In a recent meta-analysis, peak CK-MB taken from 7 studies predicted mortality at 1 and 12 months postoperatively but not significantly after 5 years. 5 Another systematic review found that postoperative CKMB release was associated with mortality up to 40 months
From the aDepartment of Cardiology, Oslo University Hospital, Ullevaal, Oslo, Norway, b Department of Cardiothoracic Surgery, Oslo University Hospital, Ullevaal, Oslo, Norway, c Section of Biostatistics and Epidemiology, Oslo University Hospital, Ullevaal, Oslo, Norway, and dFaculty of Medicine, University of Oslo, Oslo, Norway. Submitted January 21, 2012; accepted May 7, 2012. Reprint requests: Theis Tønnessen, MD, PhD, Department of Cardiothoracic Surgery, Oslo University Hospital, Ullevaal, Pb. 4956 Nydalen, 0424 Oslo, Norway. E-mail: [email protected] 0002-8703/$ - see front matter © 2012, Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2012.05.027
postoperatively but could not conclude regarding troponins due to few published studies. 6 Hence, the prognostic value of these biomarkers is still unclear because few large, long-term studies of cardiac biomarkers after CABG have been conducted. Previous studies have largely investigated either CK-MB or troponins as predictors postoperatively. Few studies have assessed both CK-MB and troponins simultaneously as predictors of mortality in the same patient population, and these suffer from limited study size. 7-9 Levels of troponins could presumably be better predictors of mortality than CK-MB because of their increased sensitivity and specificity to the myocardium and measurements thereof represent the “gold standard” in patients with acute coronary syndrome. 10 Thus, we aimed to test the hypothesis that CK-MB and cTnT are long-term predictors of mortality after CABG and to determine which of these 2 markers of myonecrosis is the better predictor.
Materials and methods Patients One thousand three hundred fifty patients operated on consecutively for isolated CABG, in the time period from
American Heart Journal November 2012
780 Søraas et al
January 1, 2003, to December 31, 2006, at Oslo University Hospital, Ullevaal, were included in the study. Patients undergoing additional surgery to CABG were excluded. Data, registered prospectively, were obtained from the hospital database and from patient files. The study was approved by an institutional review board. Few data were missing, and these were coded as absent. All-cause mortality status on all patients through June 30, 2011, was established through the Norwegian National Death Registry. A few patients had emigrated, and these were censored from the emigration date or date of last contact. Median (interquartile range) follow-up time was 6.1 (5.1-7.2) years. Five patients were excluded because they could not be identified in the National Registry, and 3 patients were excluded because of missing postoperative blood samples. Thus, a total of 1,342 patients underwent further analysis. One thousand two hundred ninety-four of these patients had undergone elective surgery and 48 had undergone emergent surgery (3.6%), according to the definition by European System for Cardiac Operative Risk Evaluation (EuroSCORE). 11 Hypertension was defined as clinically recognized and under medical treatment. Smoking was defined as smoking until the month before operation. Renal failure was defined as preoperative creatinine N200 μmol/L (2.26 mg/dL). Unstable angina was defined as rest angina pectoris requiring intravenous nitrate until arrival in the anesthetic room. The EuroSCORE was assessed preoperatively according to a scoring system based on validated preoperative risk factors that indicate the 30-day mortality risk for each patient undergoing open heart surgery. 11
Surgical procedure All patients were operated on with median sternotomy and using the standard technique of cardiopulmonary bypass with ascending aorta cannulation, single venous cannulation, and moderate systemic hypothermia (32-34°C), as previously described. 12
Measurements of biomarkers Venous blood samples were routinely collected at 7, 20, and 44 hours postoperatively. Both CK-MB and cTnT were immediately analyzed by the Department of Clinical Chemistry at Oslo University Hospital, Ullevaal. Serum cTnT concentration was determined by using the third- or fourth-generation TnT assay (Troponin T STAT) on the Roche Elecsys 2010 immunoassay analyzer (Roche Diagnostics, Mannheim, Germany) according to the manufacturer recommendations. Serum CKMB concentration was determined by using 2 monoclonal antibodies (CKMB STAT) on an Elecsys 2010 analyzer (Roche diagnostics) by electrochemiluminescence immunoassay. The upper normal reference limit is 0.10 μg/L for cTnT and 5 μg/L for CK-MB.
Statistical analysis Data are presented as mean ± SD for continuous variables with normal distribution, as median with interquartile range in variables with nonnormal distribution, and as proportions for categorical variables. Mean values were compared using Student t test for normally distributed data and Mann-Whitney U test for nonparametric data. Categorical variables were compared using
the χ 2 test or Fisher exact test when appropriate. Event rates were calculated and plotted according to the Kaplan-Meier product limit method. The relation of CK-MB and cTnT to mortality was assessed using Cox proportional hazards models. 13 Log minus log of survival plotted against survival times verified the proportional hazards assumption. The independence of the relationship of cTnT and CK-MB to mortality was evaluated in univariate and multivariate Cox regression analyses. All clinical relevant baseline and procedural variables were tested in Cox univariate analyses. Variables with a univariate P b .10 were included in the multivariate Cox model. Collinearity was checked by calculating variance inflation factor (VIF), and VIF b5 was considered acceptable. The variable EuroSCORE gave VIF N6 and thus was excluded from the multivariate analyses with other clinical variables. The relationship of CK-MB to mortality was also evaluated in a multivariate Cox model with baseline variables, but not with cTnT as covariate. The results from the Cox analyses are presented as hazard ratios (HRs) with 95% CIs. For all tests, 2-tailed P b .05 was required for statistical significance. Data management and analysis were performed with IBM SPSS version 19.0 (SPSS Inc, Chicago, IL) software. No extramural funding was used to support this work, and the authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of this manuscript, and its final contents.
Results Patient characteristics A total of 1,342 patients were available for analysis, and of these, 1314 patients (97.9%) had complete cTnT and CK-MB values for all 3 postoperative time points, and 28 patients (2.1%) had values for 2 postoperative time points. The mean age ± SD for the cohort was 64.9 ± 10.0 years, and there were 1,089 men (81.1%). The all-cause mortality at 30 days was 0.67%. Totally, 207 patients (15.4%) died during a follow-up period of a median (interquartile range) of 6.1 (5.1-7.2) years. Patients who deceased were older with lower body mass index and had, at the time of surgery, higher prevalences of hypertension, peripheral vascular disease, renal failure and higher creatinine, lower left ventricular ejection fraction (LVEF), and higher EuroSCORE compared with patients who survived (Table I). Perioperatively, they had longer cardiopulmonary bypass time, and postoperatively, they had higher creatinine and peak cTnT levels (P b.001) but not different peak CK-MB (P = .21). Biomarker values The cardiac biomarkers were measured at 3 time points postoperatively, and the median (interquartile range) postoperative values for cTnT were 0.48 (0.32-0.75) μg/L at 7 hours, 0.31 (0.19-0.57) μg/L at 20 hours, and 0.20 (0.10-0.41) μg/L at 44 hours postoperatively. The corresponding CK-MB values were 26 (19-35) μg/L at 7 hours, 21 (14-31) μg/L at 20 hours, and 5 (4-8) μg/L at 48 hours postoperatively. Both biomarkers peaked rapidly within 7 hours post-CABG in most patients. The peak
American Heart Journal Volume 164, Number 5
Søraas et al 781
Table I. Patient characteristics according to mortality status Baseline characteristics Preoperative parameters Demographics Age (y) Male gender, n (%) Body mass index (kg/m2) Medical history Hypertension, n (%) Recent myocardial infarction (b90 d preop), n (%) Diabetes mellitus, n (%) COPD, n (%) Apoplexia cerebri, n (%) Peripheral vascular disease, n (%) Renal failure, n (%) Neurologic disease, n (%) Current smokers, n (%) NYHA class III or IV, n (%) LVEF preop (%) Left main stem stenosis, n (%) Creatinine preop (mg/dL) EuroSCORE Perioperative/postoperative parameters Total distals, n Unstable angina, n (%) Emergent procedure, n (%) CPB time (min) Cross-clamp time (min) Creatinine day 2 postop (mg/dL) Peak postop cTnT (μg/L) Peak postop CK-MB (μg/L)
All patients (N = 1342)
Survivor (n = 1135)
Nonsurvivor (n = 207)
P⁎
64.9 ± 10.0 1089 (81.1) 26.7 ± 3.9
63.9 ± 9.8 926 (81.6) 26.8 ± 3.8
70.6 ± 9.4 163 (78.7) 26.2 ± 4.3
b.001 .39 .020
556 (41.5) 474 (35.3) 213 (15.9) 88 (6.6) 26 (1.9) 125 (9.3) 18 (1.3) 18 (1.3) 366 (27.3) 1040 (77.5) 60.6 ± 14.5 265 (19.7) 0.93 (0.81-1.07) 3 (2-5)
452 (39.9) 389 (34.3) 175 (15.4) 68 (6.0) 23 (2.0) 85 (7.5) 7 (0.6) 15 (1.3) 307 (27.1) 870 (81.0) 61.2 ± 14.4 218 (19.2) 0.91 (0.80-1.04) 3 (1-5)
104 (50.2) 85 (41.1) 38 (18.4) 20 (9.7) 3 (1.4) 40 (19.3) 11 (5.3) 3 (1.4) 59 (28.5) 170 (85.0) 57.3 ± 15.0 47 (22.7) 1.02 (0.84-1.36) 5 (4-7)
.007 .072 .34 .070 .79 b.001 b.001 .75 .75 .22 b.001 .29 b.001 b.001
3.1 ± 0.9 23 (1.7) 48 (3.6) 75.4 ± 22.2 38.8 ± 13.6 0.85 (0.74-1.02) 0.53 (0.34-0.83) 27 (20-38)
3.1 ± 0.9 18 (1.6) 41 (3.6) 74.9 ± 22.0 38.6 ± 13.6 0.84 (0.73-0.98) 0.51 (0.34-0.79) 27 (20-38)
3.2 ± 0.9 5 (2.4) 7 (3.4) 78.3 ± 22.9 39.8 ± 14.1 0.97 (0.80-1.47) 0.63 (0.39-1.02) 29 (20-41)
.076 .38 1.00 .040 .27 b.001 b.001 .21
⁎ P values refer to nonsurvivor vs survivor. Data are provided as means ± SD, number (%), or median (interquartile range). Preop, Preoperatively; COPD, chronic obstructive pulmonary disease; NYHA, New York Heart Association; CPB, cardiopulmonary bypass; Postop, postoperatively.
postoperative cTnT value was measured at 7 hours postoperatively in 1,044 (77.8%) of the patients, at 20 hours postoperatively in 205 (15.3%), and at 44 hours postoperatively in 113 (8.4%) of the patients. Peak cTnT was a median of 0.53 (interquartile range 0.34-0.83) μg/ L, and peak CK-MB was 27 (20-38) μg/L. Sixteen patients (1.2%) had peak cTnT values within the normal reference area, and 2 patients (0.1%) had normal peak CK-MB values.
Biomarker quartile analysis Figure 1 shows the Kaplan-Meier survival curves for patients grouped by (a) quartile of peak cTnT and (b) quartile of peak CK-MB. There were significant differences in mortality between the quartiles of cTnT (χ 2 = 17.15, P = .001), but not between the CK-MB quartiles (χ 2 = 0.98, P = .81). The highest mortality was observed among patients with a cTnT level in the top quartile (quartile 1 vs quartile 2, χ 2 = 1.54, P = .21; quartile 1 vs quartile 3, χ 2 = 1.90, P = .17; and quartile 1 vs quartile 4, χ 2 = 6.92, P = .009). For CK-MB levels, no quartile differed significantly from the others (quartile 1 vs quartile 2, χ 2 = 0.13, P = .72; quartile 1 vs quartile 3, χ 2 = 0.33, P = .57; and quartile 1 vs quartile 4, χ 2 = 0.15, P = .70).
Mortality in relationship to quartiles of cTnT and CK-MB is shown in Table II. For each quartile of cTnT (horizontal row), there is a trend that the percentage mortality is rather constant, independent of the CK-MB quartiles. For CK-MB, however, the percentage mortality seems more associated with the cTnT quartile, that is, the percentage mortality increases vertically within each CK-MB quartile dependent of the matching cTnT quartile. This indicates that CK-MB gives little additional prognostic information apart from cTnT.
Univariate and multivariate predictors of mortality In univariate analyses, age, body mass index, hypertension, chronic obstructive pulmonary disease, peripheral vascular disease, LVEF, preoperative creatinine, creatinine at day 2 postoperatively, EuroSCORE, and peak postoperative cTnT and CK-MB were significantly associated with mortality. Age, peripheral vascular disease, LVEF, serum creatinine at day 2 postoperatively, and peak cTnT remained significant predictors in multivariate analysis, and age was the strongest predictor (Table III). Creatine kinase-MB was not an independent predictor of mortality in the multivariate model including cTnT. However, in multivariate analysis with the same clinical
American Heart Journal November 2012
782 Søraas et al
Figure 1
Table II. Mortality stratified by quartiles of CK-MB and cTnT CK-MB CK-MB CK-MB CK-MB quartile 1 quartile 2 quartile 3 quartile 4 cTnT quartile 1 cTnT quartile 2 cTnT quartile 3 cTnT quartile 4
13% (26/205) 13% (13/104) 21% (10/48) 36% (5/14)
13% (10/80) 7% (8/111) 20% (17/87) 30% (8/27)
13% (6/45) 12% (12/98) 19% (20/108) 22% (18/83)
20% (2/10) 8% (3/37) 12% (9/77) 19% (40/208)
Mortality data are expressed as percentages with the absolute numbers of patients who have died in each group divided by the total number of patients in each group in parentheses.
prognostic significance in univariate analyses (data not shown) and when adjusted for EuroSCORE (Figure 2).
Discussion
A, Kaplan-Meier survival curves for mortality associated with quartiles of peak cTnT, P = .001. B, Kaplan-Meier survival curves for mortality associated with quartiles of peak CK-MB, P = .81.
variables but excluding cTnT, peak CK-MB was a significant independent predictor (P = .007). The biomarkers were also analyzed adjusting for EuroSCORE (Table IV). Peak cTnT, but not peak CK-MB, was a significant independent predictor of mortality when separately analyzed. Likewise, when these 2 biomarkers were included together in the same multivariate model, cTnT was highly significant, whereas CK-MB was nonsignificant. There was no significant interaction between CK-MB and cTnT. The predictive values of CK-MB and cTnT at different time points varied in our population. For both CK-MB and cTnT, the values at 44 hours postoperatively had the greatest
The present study evaluated the long-term prognostic significance of myocardial damage defined by CK-MB and cTnT release after CABG. We found that both CK-MB and cTnT were predictors of long-term all-cause mortality after CABG in our patient population. However, when analyzing the 2 cardiac markers in the same Cox model, cTnT was a highly significant predictor of mortality, whereas CK-MB was nonsignificant. Troponin T remained a highly significant independent predictor of mortality after adjustment for a wide range of baseline and perioperative variables. Nearly all patients had elevated cTnT and CK-MB postoperatively. The biomarker value at 44 hours postoperatively had a greater predictive value than the values at 7 or 20 hours or the peak value. Both CK-MB and cTnT have been shown to be predictors of inhospital and short-term mortality after CABG, 1-4 but few studies have examined the relationship to long-term mortality. We found that both CK-MB and cTnT were independent predictors of long-term mortality when examined in separate multivariate Cox analyses adjusting for a large set of widely accepted baseline and perioperative prognostic factors. Interestingly, however, when both biomarkers were analyzed together in the same model, CK-MB gave no additional prognostic information beyond cTnT. A similar result was found when the biomarkers were analyzed together in a model adjusting for the commonly used preoperative risk score EuroSCORE. This may be explained by the higher sensitivity and specificity of cTnT for myocardial damage. Skeletal muscle may contain up to 6% CK-MB. 14 Consequently, skeletal muscle trauma from dissection and internal mammary artery harvest during CABG may increase the levels of CK-MB postoperatively. Few studies have compared CK-MB and cTnT regarding mortality after CABG. Vikenes et al. 8compared cardiac
American Heart Journal Volume 164, Number 5
Søraas et al 783
Table III. Predictors of mortality in univariate and multivariate analyses Univariate analysis Variables
HR (95% CI)
Age (per year) Male gender Body mass index (per 1 kg/m 2) History of hypertension History of myocardial infarction (b90 d preoperatively) History of diabetes mellitus History of COPD History of peripheral vascular disease Current smoker NYHA class III or IV (yes/no) LVEF (per 1 %) Left main stem stenosis Creatinine preop (per 1 mg/dL) EuroSCORE (per unit) Total distals Emergent procedure CPB time (per 1 min) Cross-clamp time (per 1 min) Unstable angina Creatinine day 2 postop (per 1 mg/dL) Peak postoperative cTnT (per 1 μg/L) Peak postoperative CK-MB (per 1 μg/L)
1.08 0.85 0.96 1.53 1.31 1.22 1.63 2.64 1.07 1.37 0.98 1.26 1.29 1.29 1.14 0.89 1.005 1.003 1.28 1.56 1.38 1.003
(1.06-1.10) (0.61-1.19) (0.92-0.99) (1.17-2.01) (1.00-1.73) (0.86-1.74) (1.03-2.59) (1.87-3.73) (0.79-1.45) (0.93-2.02) (0.97-0.99) (0.91-1.74) (1.19-1.38) (1.23-1.35) (0.98-1.33) (0.42-1.88) (1.00-1.01) (0.99-1.01) (0.53-3.11) (1.42-1.71) (1.25-1.52) (1.001-1.006)
Multivariate analysis
χ2
P
79.5 0.9 5.7 9.4 3.7 1.3 4.3 30.4 0.2 2.5 14.7 1.9 39.6 119.9 2.9 0.1 2.7 0.5 0.3 90.8 43.7 7.7
b.001 .34 .02 .002 .054 .26 .037 b.001 .65 .115 b.001 .169 b.001 b.001 .087 .75 .102 .49 .59 b.001 b.001 .005
χ2
P
1.07 (1.06-1.09)
62.6
b.001
0.99 (0.95-1.03) 1.19 (0.90-1.56) 0.95 (0.70-1.27)
0.2 1.5 0.1
.62 .23 .71
1.40 (0.88-2.24) 1.66 (1.15-2.40)
2.0 7.4
.161 .006
0.98 (0.97-0.99)
12.6
b.001
1.04 (0.89-1.21)
0.3
.60
45.1 13.6 0.0
b.001 b.001 .99
HR (95% CI)
1.45 (1.30-1.62) 1.31 (1.13-1.51) 1.00 (0.997-1.003)
COPD, Chronic obstructive pulmonary disease; NYHA, New York Heart Association; Preop, preoperatively; CPB, cardiopulmonary bypass; Postop, postoperatively.
Table IV. The value of adding peak biomarkers to a clinical model or EuroSCORE in Cox regression analyses for predicting mortality Model cTnT only CK-MB only Clinical model + cTnT Clinical model + CK-MB Clinical model + CKMB + cTnT EuroSCORE + cTnT EuroSCORE + CKMB EuroSCORE + CKMB + cTnT
HR (95% CI)
P⁎
1.34 1.14 1.27 1.13
(1.23-1.45) (1.04-1.25) (1.15-1.41) (1.03-1.24)
b.001 .005 b.001 .007
1.00 1.28 1.22 1.09
(0.88-1.14) (1.12-1.45) (1.11-1.34) (0.99-1.20)
.99 b.001 b.001 .07
0.93 (0.80-1.09) 1.28 (1.12-1.46)
.38 b.001
⁎ P denotes significance of the added biomarker variable in the model. HR and 95% CIs are shown per 1 SD for each biomarker peak value (0.90 μg/L for cTnT and 38.4 μg/L for CK-MB). The clinical model contains clinical predictors with P b .10 in univariate analysis (age, body mass index, hypertension, prior myocardial infarction b90 days preoperatively, chronic obstructive pulmonary disease, peripheral vascular disease, preoperative LVEF, total distals, creatinine at day 2 postoperatively).
troponin I and cTnT with CK-MB and concluded that CK-MB was superior to the cardiac troponins in predicting event-free survival in low-risk patients with a median follow-up time of 7.7 years However, this study included 156 patients with isolated CABG only, and the discrepancy between the findings of this study and ours may be due to the limited number of patients
included. Our results are, on the other hand, supported by 2 recent studies that found that cardiac troponin I was better than CK-MB for quantifying myocyte necrosis as assessed by magnetic resonance imaging after CABG. 15,16 Clearly, taken together with other previous studies of the prognostic importance of troponins, 2,7,17 our finding with cTnT suggests that these more specific markers of myocardial necrosis have stronger prognostic power than the less cardiac specific enzyme CK-MB and should be considered the preferred assay. The mechanism of how small leakages of troponins are associated with increased mortality in the post-CABG setting is unknown. However, the same relationships are found after percutaneous coronary intervention, 18,19 noncardiac surgery, 20 and in the general population, 21 indicating common pathophysiologic pathways, regardless of the type of initiating stimulus. Another possibility is that these markers are just surrogate markers of other, yet undiscovered, markers of worse prognosis in these patients. Currently, guidelines do not address the prognostic influence of postoperative biomarker levels. An interesting debate is whether postoperative biomarker levels could be incorporated into the widely used EuroSCORE or whether they are sufficiently powerful to remain independent. However, because the EuroSCORE is a preoperative scoring system, incorporating postoperative biomarker levels seems to be somewhat conflicting. Our data, in line with other studies, show that biomarkers have independent prognostic importance.
American Heart Journal November 2012
784 Søraas et al
Figure 2
Forest plot demonstrating the predictive value of biomarkers at various time points postoperatively for all-cause mortality. Cox multivariable proportional hazards model, adjusted for EuroSCORE, is shown for each individual variable. Hazard ratio and 95% CIs are shown per 1 SD for each variable. All biomarkers were entered as continuous variables. 7, 7 hours postoperatively; 20, 20 hours postoperatively; 44, 44 hours postoperatively; peak, peak value postoperatively. Significances for mortality: cTnT 7 (HR 1.13, 95% CI 1.03-1.24, P = .009), cTnT 20 (HR 1.16, 95% CI 1.05-1.29, P = .004), cTnT 44 (HR 1.24, 95% CI 1.151.34, P b.001), peak cTnT (HR 1.22, 95% CI 1.11-1.34, P b.001), CKMB 20 (HR 1.10, 95% CI 1.00-1.21, P = .05), and CK-MB 44 (HR 1.13, 95% CI 1.05-1.22, P = .002).
Therefore, in our opinion, patients with elevated biomarkers should have a more intensive follow-up schedule after hospital discharge. Compliance with evidence-based guidelines documented to improve survival from coronary heart disease has been shown to be less after CABG than in nonsurgical populations. 22 Thus, in patients with increased postoperative biomarker levels, there should be put especially emphasis on implementing evidencebased medical therapies. Perhaps these patients also should have a lower threshold for angiography in the case of new angina symptoms. However, these proposals for future guidelines may need to be validated in prospective studies. Our study was strengthened by the long follow-up, equal to N7,900 patient-years. Hypothetically, the long follow-up and use of all-cause mortality (rather than cardiovascular death) might have diluted the prognostic importance of cTnT and CK-MB, yet we found that the biochemical markers still carried significant prognostic information. Most of the published studies to date have examined quite well-defined homogenous CABG cohorts with strict inclusion criteria. Our study is strengthened by the heterogeneous population; all consecutive patients undergoing isolated CABG in our center during 3 years, hence, increasing the applicability of our results to the clinical setting. Nonetheless, despite our heterogenic
study population, the all-cause mortality at 30 days was 0.67%, confirming the low risk of this study population. The 2011 American College of Cardiology Foundation/ American Heart Association CABG guidelines state that measurement of biomarkers of myonecrosis is reasonable in the first 24 hours after CABG, 23 and several studies have examined biomarkers within this time frame postoperatively. Interestingly, we found that biomarker release for both CK-MB and cTnT at 44 hours postoperatively had the greatest prognostic power. Lehrke et al 24 showed that cTnT measurement at 48 hours postoperatively was the best predictor for mortality, consistent with our data. Moreover, a recent study demonstrated that the troponin I value at 72 hours was the greatest predictor of mid-term mortality. 25 Taken together, these studies and our findings support extended measurements of troponins during the first postoperative days. Possibly, the later release of biochemical markers may be of greater discriminatory value due to gradual release of structurally bound troponin representing greater irreversible injury and necrosis. Although our study showed that biomarkers at 44 hours postoperatively had slightly greater prognostic power compared with the peak values, the peak values are easier to follow in the clinical setting and should be recommended used for clinicians.
Study limitations First, some patients with recent myocardial infarction may have had preoperatively elevated biochemical markers that could influence marker levels postoperatively. Second, we lack follow-up data regarding reinterventions and complications after the operation, which inevitably may affect long-term outcome. Third, we used the well-recognized EuroSCORE as an externally valid assessment of individual risk of long-term mortality although the EuroSCORE is not constructed for predicting mortality exceeding 30 days postoperatively and is shown to overestimate mortality. 11,26 Another limitation is that we examined all-cause mortality and did not differ between cardiac or non-cardiac cause of death. However, the accuracy of death certificates for coding coronary heart disease as the cause of death is questionable and thus, all-cause mortality is a more strict and objective parameter. 27 Finally, this is a single-center study with a limited number of patients, so delineation of quantitative relationships should be done with caution.
Conclusions In conclusion, both CK-MB and cTnT are predictors of mortality after CABG surgery; however, our data suggest that cTnT is a better predictor of long-term mortality after CABG surgery than CK-MB. Further studies are warranted to identify the best follow-up regime for patients with postoperatively elevated cTnT.
American Heart Journal Volume 164, Number 5
Acknowledgements The authors are grateful for the help provided by Dr Stein Koldsland, Heidi Thorstensen, and the staff at the Department of Cardiothoracic Surgery, Oslo University Hospital, Ullevaal.
Disclosures None.
References 1. Nesher N, Alghamdi AA, Singh SK, et al. Troponin after cardiac surgery: a predictor or a phenomenon? Ann Thorac Surg 2008;85:1348-54. 2. Croal BL, Hillis GS, Gibson PH, et al. Relationship between postoperative cardiac troponin I levels and outcome of cardiac surgery. Circulation 2006;114:1468-75. 3. Ramsay J, Shernan S, Fitch J, et al. Increased creatine kinase MB level predicts postoperative mortality after cardiac surgery independent of new Q waves. J Thorac Cardiovasc Surg 2005; 129:300-6. 4. Klatte K, Chaitman BR, Theroux P, et al. Increased mortality after coronary artery bypass graft surgery is associated with increased levels of postoperative creatine kinase–myocardial band isoenzyme release: results from the GUARDIAN trial. J Am Coll Cardiol 2001;38:1070-7. 5. Domanski MJ, Mahaffey K, Hasselblad V, et al. Association of myocardial enzyme elevation and survival following coronary artery bypass graft surgery. JAMA 2011;305:585-91. 6. Petäjä L, Salmenperä M, Pulkki K, et al. Biochemical injury markers and mortality after coronary artery bypass grafting: a systematic review. Ann Thorac Surg 2009;87:1981-92. 7. Januzzi JL, Lewandrowski K, MacGillivray TE, et al. A comparison of cardiac troponin T and creatine kinase-MB for patient evaluation after cardiac surgery. J Am Coll Cardiol 2002;39:1518-23. 8. Vikenes K, Andersen KS, Melberg T, et al. Long-term prognostic value of cardiac troponin I and T versus creatine kinase-MB mass after cardiac surgery in low-risk patients with stable symptoms. Am J Cardiol 2010;106:780-6. 9. Kathiresan S, Servoss SJ, Newell JB, et al. Cardiac troponin T elevation after coronary artery bypass grafting is associated with increased one-year mortality. Am J Cardiol 2004;94:879-81. 10. Thygesen K, Alpert JS, White HD, et al. Universal definition of myocardial infarction. Circulation 2007;116:2634-53. 11. Nashef SA, Roques F, Michel P, et al. European System for Cardiac Operative Risk Evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16:9-13. 12. Larstorp AC, Lund Søraas C, Tønnessen T, et al. Scintigraphic demonstration of myocardial perfusion and ischaemia associated with coronary artery bypass grafting. Scand Cardiovasc J 2006;40: 354-62.
Søraas et al 785
13. Cox DR. Regression models and life-tables. J R Stat Soc, B 1972;34: 187-220. 14. Tsung JS, Tsung SS. Creatine kinase isoenzymes in extracts of various human skeletal muscles. Clin Chem 1986;32:1568-70. 15. Lim CC, Cuculi F, van Gaal WJ, et al. Early diagnosis of perioperative myocardial infarction after coronary bypass grafting: a study using biomarkers and cardiac magnetic resonance imaging. Ann Thorac Surg 2011;92:2046-53. 16. Pegg TJ, Maunsell Z, Karamitsos TD, et al. Utility of cardiac biomarkers for the diagnosis of type V myocardial infarction after coronary artery bypass grafting: insights from serial cardiac MRI. Heart 2011;97:810-6. 17. Adabag AS, Rector T, Mithani S, et al. Prognostic significance of elevated cardiac troponin I after heart surgery. Ann Thorac Surg 2007;83:1744-50. 18. Ohlmann P, Jaquemin L, Morel O, et al. Prognostic value of C-reactive protein and cardiac troponin I in primary percutaneous interventions for ST-elevation myocardial infarction. Am Heart J 2006;152: 1161-7. 19. Feldman DN, Kim L, Rene AG, et al. Prognostic value of cardiac troponin-I or troponin-T elevation following nonemergent percutaneous coronary intervention: a meta-analysis. Catheter Cardiovasc Interv 2011;77:1020-30. 20. Levy M, Heels-Ansdell D, Hiralal R, et al. Prognostic value of troponin and creatine kinase muscle and brain isoenzyme measurement after noncardiac surgery: a systematic review and meta-analysis. Anesthesiology 2011;114:796-806. 21. de Lemos JA, Drazner MH, Omland T, et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304:2503-12. 22. Hiratzka LF, Eagle KA, Liang L, et al. Atherosclerosis secondary prevention performance measures after coronary bypass graft surgery compared with percutaneous catheter intervention and nonintervention patients in the get with the guidelines database. Circulation 2007;116:I207-12. 23. Hillis LD, Smith PK, Anderson JL, et al. 2011 ACCF/AHA guideline for coronary artery bypass graft surgery: executive summary: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. Circulation 2011; 124:2610-42. 24. Lehrke S, Steen H, Sievers HH, et al. Cardiac troponin T for prediction of short- and long-term morbidity and mortality after elective open heart surgery. Clin Chem 2004;50:1560-7. 25. Ranasinghe AM, Quinn DW, Richardson M, et al. Which troponometric best predicts midterm outcome after coronary artery bypass graft surgery? Ann Thorac Surg 2011;91:1860-7. 26. Engebretsen KV, Friis C, Sandvik L, et al. Survival after CABG—better than predicted by EuroSCORE and equal to the general population. Scand Cardiovasc J 2009;43:123-8. 27. Lauer MS, Blackstone EH, Young JB, et al. Cause of death in clinical research: time for a reassessment? J Am Coll Cardiol 1999;34: 618-20.