Ann Thorac Surg 2009;87:1981-1992. doi:10.1016/j.athoracsur.2008.12.063
© 2009 The Society of Thoracic Surgeons
Reviews
Biochemical Injury Markers and Mortality After Coronary Artery Bypass Grafting: A Systematic Review
Liisa Petäjä, MDa,*,
Markku Salmenperä, MD, PhDa,
Kari Pulkki, MD, PhDb,
Ville Pettilä, MD, PhDa
a Department of Anesthesiology and Intensive Care Medicine of Helsinki University Central Hospital, Helsinki, Finland
b Eastern Finland Laboratory Center and University of Kuopio, Kuopio, Finland
* Address correspondence to Dr Petäjä, Department of Anesthesiology and Intensive Care Medicine, Helsinki University Central Hospital, PO Box 340, Helsinki, FIN-00029 HUS, Finland (Email: liisa.petaja{at}hus.fi).
| Dr Pulkki discloses that she has a financial relationship with Roche and Siemens.
|
 |
Abstract
|
|---|
The strength of the association between cardiac biomarker release and prognosis is uncertain. We performed a systematic literature search to find articles regarding these markers and death after coronary surgical interventions, and evaluated the results with meta-analytic methods. We found 23 articles concerning 29,483 patients that reported the MB fraction of creatine kinase (CK-MB) and troponin T and I. Heterogeneity of existing studies prevented the pooling of the results of troponin studies. The pooled data of the CK-MB studies suggest that after coronary artery bypass grafting, CK-MB release of more than five to eight times the upper limit of the reference range is associated with an increased risk of death during the next 40 months.
The significance of cardiac injury marker release after coronary interventions has long been debated. Especially after coronary artery bypass grafting (CABG), cardiac biomarker release is multifactorial [1]. First, after the operation, the elevation of serum concentrations of cardiac enzymes is caused partly by enzyme release from tissues other than the heart, and cannulation of the heart may lead to biomarker release not caused by ischemia [2]. Second, at least two types of ischemic phenomena—global ischemia during aortic cross-clamping and regional ischemia due to graft or native vessel occlusion—may cause cardiac marker release after the operation, and it is conceivable that they are associated with prognosis to differing extents.
Despite these distracting factors, the amount of cardiac biomarker released seems to be associated with short-term and long-term survival after CABG. Recent articles report varying figures on the actual magnitude and extension of this association. This review aims to evaluate systematically the available data on the association of postoperative cardiac injury marker release with patient mortality after CABG.
|
Material and Methods
|
|---|
Selection of Studies
A systematic search was performed with the assistance of information specialists in the Ovid Medline and Embase databases covering January 1, 1990, to May 30, 2008. Search strategies were planned to find articles concerning myocardial injury markers and cardiac operations, excluding pediatric operations, and they differed slightly due to database differences (Appendix 1
*). A study was included if it was a peer-reviewed article published in English that reported deaths in relation to one of the injury biomarkers—MB fraction of creatine kinase (CK-MB), troponin T (TnT), troponin I (TnI), fatty acid-binding protein, or myoglobin—detected after CABG. The reference lists of the included articles were manually searched for additional eligible articles.
Quality Assessment
Predefined criteria to assess internal validity of the studies comprised 12 questions and were adapted from Altman [3]. These criteria were modified according to the subject. They also covered external validity; that is, whether the results were applicable to the topic of this review—perioperative injury in CABG patients (Appendix 2
* [4]). Two reviewers (L. P. and V. P.) assessed the quality of the eligible studies independently of each other. Each item was scored as 1 if the study met the criteria and as 0 if the criteria were not met.
Data Extraction
The following data were extracted from individual studies:
- 1 publishing data (first author, year published, years during which patients were collected);
- 2 study design, including prospective vs retrospective, whether cutoff values were determined a priori, sample size, follow-up interval;
- 3 data related to the patient and operation, including gender, mean age, risk category (graded as low, average, or high according to inclusion and exclusion criteria), type of operation (CABG or other intervention, off-pump or on-pump procedure), and inclusion or exclusion of patients with preoperative myocardial infarction and renal insufficiency;
- 4 laboratory data, including name and manufacturer of the assay (whether mass or activity was used for CK-MB) and timing of samples (peak value or later value used in analyses), and cutoff values; and
- 5 effect size data, comprising crude dichotomous data on patient survival in low and high cardiac marker groups, and results of multivariate analysis.
In addition, when necessary, authors of the original articles were asked to supply additional data.
Data Processing
Studies were categorized according to the biomarker and short-term vs long-term follow-up interval. When possible, categorization according to the use of cardiopulmonary bypass (CPB) was also performed. CK-MB values were treated as multiples of the upper limit of the reference range (ULRR), whereas troponin values were treated as absolute figures. Troponin values were also transformed into multiples of ULRR according to:
- 1 the 99th percentile concentration of the reference population reported by the manufacturer (TnI),
- 2 the clinical cutoff value for acute myocardial infarction without cardiac operation at each institution reported in the article (TnT), and
- 3 the concentration at which the coefficient of variation was 10% or less for the assay used in each study (TnI) [5]. The assays for TnI were also classified according to their imprecision at the lower end of the detection threshold according to Panteghini and colleagues [5] (Table 1).
Publication bias was assessed by one-tailed Egger's test (p < 0.05 was considered statistically significant), indicating bias only among studies reporting CK-MB and long-term mortality. Owing to considerable heterogeneity revealed by I
2 statistics, random effects model and methods of moments for meta-regression were chosen to pool and analyze the results. Risk ratio (RR) was used as an effect size measure. Potential explanatory factors for the diversity of the RRs between studies were first searched by subgrouping the studies according to dichotomous and categoric variables and then by meta-regression according to continuous and ordered categoric variables. CK-MB studies with only off-pump patients (one study and one subgroup) were excluded from these analyses (ie, subgrouping and meta-regression). All calculations were performed with Comprehensive Meta Analysis 2.2.027 software (Biostat, Englewood, NJ).
|
Results
|
|---|
Search Results
Of the 3129 articles found, 23 studies reporting data on 29,483 patients fulfilled the inclusion criteria (Fig 1). The biomarkers studied were only CK-MB in nine studies, only TnT in three studies, both of CK-MB and TnT in two studies, only TnI in seven studies, and both CK-MB and TnI in two studies (Table 1). No eligible studies of fatty acid-binding protein or myoglobin were found.
Of the 13 CK-MB studies (24,089 patients), nine were multicenter trials that were not primarily designed to examine the association between CK-MB and survival [6–9] or single-center trials that received mortality data from registries [10–14]. Three of these single-center studies had, however, gathered their patient data prospectively [12–14]. Only four CK-MB studies, which assessed both TnT or TnI and CK-MB, concentrated primarily on biomarkers and prognosis [15–18].
The 14 studies of troponin (7201 patients) in contrast, were single-center or double-center studies mostly evaluating the relationship between troponin levels and survival [15–25]; only three studies concentrated on other issues—perioperative myocardial infarction [26], cardiac dysfunction [27], and risk factors for troponin release [28]. One TnT study [16] reported a subgroup of another study [15] in which the patients were merely monitored for a longer period.
Methodological Quality
Scoring disagreements were resolved after a precise definition of the questions. The studies had median score of 7 (range, 3 to 8) of the maximum 12. The most common weakness of the studies was the absence of blinding. All but five of the studies failed to analyze predictive variables appropriately. Less than half of the studies reported and adjusted the analysis for the most important prognostic factors. Small sample size or inclusion of heterogeneous patient material was typical for studies on troponin, as was the failure to report the details of analytic methods for studies on CK-MB. In addition, more than half of the studies failed to include over 90% of eligible patients.
Mortality and CK-MB
Ten of the 13 CK-MB studies provided data to calculate crude RRs of mortality—7 studies (15,185 patients) for short-term and 10 studies (21,657 patients) for long-term mortality (Table 1, Fig 2). These studies consisted only of CABG patients, and their RRs were pooled. Cutoff values varied widely, but data according to a cutoff of five to eight times ULRR was available in all of the studies. One study [18] and a subgroup of another study [14] included only off-pump procedures, one study [10] had less than 20% of off-pump procedures, whereas in one multicenter study [9], information about the use of CPB was unavailable.
View larger version (22K):
[in this window]
[in a new window]
|
Fig 2. (A) The plots of risk ratios of the studies on creatine kinase (CK)-MB and short-term mortality. Off-pump and on-pump patients are analyzed separately. A CK-MB level five times the upper limit of the reference range (ULRR) was used as a cutoff value in most studies. The cutoff was six times the ULRR in Kuduvalli and colleagues and seven times in Paparella and colleagues. (B) The plots of risk ratios of the studies on CK-MB and long-term mortality. Off-pump and on-pump patients are analyzed separately. A CK-MB level five times ULRR was used as a cutoff value in most studies. The cutoff was six times ULRR in Kuduvalli and colleagues, seven times in Paparella and colleagues, and eight times in Engoren and colleagues. (CI = confidence interval.)
|
|
Of the potential explanatory factors, the use of a cutoff value obtained by a data-dependent process (vs defined a priori) was associated with a higher RR (p = 0.056) for short-term mortality. RRs for long-term mortality diminished with lengthening follow-up (p = 0.00011).
In CK-MB studies examining short-term mortality, there were 112 deaths among the 3116 patients (3.6%) with CK-MB values above the cutoff and 142 deaths among the 12,060 patients (1.2%) with CK-MB values below cutoff. The pooled RR was 3.69 (95% confidence interval [CI], 2.17 to 6.26). Among on-pump patients, these figures were 78 of 2613 (3.0%), 102 of 9936 (1.0%), and RR of 3.69 (95% CI, 1.72 to 7.95), respectively; whereas among off-pump patients they were 7 of 88 (8.0%), 17 of 1133 (1.5%), and RR of 4.56 (95% CI, 1.61 to 17.93), respectively (Fig 2A).
In CK-MB studies reporting long-term mortality data (6 to 60 months), there were 522 deaths among the 4369 patients (11.9%) with CK-MB values above the cutoff, and 820 deaths among the 17,288 patients (4.7%) with CK-MB values below the cutoff. The pooled RR was 2.66 (95% CI, 1.95 to 3.63). Among on-pump patients, these figures were 396 of 3225 (12.3%), 781 of 12,026 (6.5%), and RR of 2.33 (95% CI, 1.60 to 3.39), respectively; and among off-pump patients, 13 of 78 (16.7%), 39 of 1110 (3.5%), and RR of 4.87 (95% CI, 2.12 to 11.21), respectively; and in the two studies with both on-pump and off-pump patients, 113 of 1066 (10.6%), 269 of 4152 (6.5%), and RR of 1.84 (95% CI, 1.00 to 3.38), respectively (Fig 2B). The RRs of on-pump patients decreased from 2.78 (95% CI, 2.00 to 3.86) for 6 months of follow-up to 1.78 (95% CI, 1.02 to 4.01) for 40 months of follow-up.
Mortality and TnT
Four of the five TnT studies (2547 patients) provided data to calculate crude RRs [16, 19, 20, 26] (Table 1, Fig 3). Two of these studies reported only short-term mortality [20, 26], one reported only long-term mortality [16], and one reported both short-term and long-term mortality [19]. Absolute cutoff values varied more than sevenfold (0.46 to 3.4 µg/L), but after transforming the cutoff values into multiples of ULRR, they varied only less than threefold (7 to 16 times ULRR). There was variance in patient populations and partly in the analytic properties of the assays.
View larger version (13K):
[in this window]
[in a new window]
|
Fig 3. (A) The plots of risk ratios of the studies on troponin (Tn) T and short-term mortality. No pooled estimate is presented due to considerable heterogeneity. (B) The plots of risk ratios of the studies on TnT and long-term mortality. *Figures were estimated from the Kaplan-Meier curve. No pooled estimate is presented due to considerable heterogeneity. (CI = confidence interval.)
|
|
In TnT studies examining short-term mortality, there were 24 deaths among the 752 patients (3.2%) with TnT values above the cutoff and 8 deaths among the 1456 patients (0.5%) with TnT values below the cutoff. In TnT studies reporting long-term mortality (12 to 28 months), there were 19 deaths among the 118 patients (16.1%) with TnT values above the cutoff and 5 deaths among the 221 patients (2.3%) with TnT values below the cutoff. Because of the small number of studies and the differences in patient populations and cutoff values, no pooling or any further analysis was performed. Only the RRs obtained from separate studies are presented (Fig 3).
Mortality and TnI
Six of the nine TnI studies (3271 patients) provided data to calculate crude RRs for short-term mortality [18, 21, 22, 24, 25, 28] and seven studies (2438 patients) for long-term mortality [18, 22–25, 27, 28] (Table 1, Fig 4). Cutoff values were variable, and they were not even on the same scale: absolute figures of cutoff values varied fivefold (2.5 to 13 µg/L), but after transforming the cutoff values into multiples of ULRR, depending on the reference value chosen (see Methods), they varied up to 12-fold (21 to 260 times ULRR). The studies included varying numbers of CABG and other cardiac procedures, as well as on-pump and off-pump procedures. Only a few studies offered the data separately for different patient groups; hence, data of most studies were analyzed as a whole.
View larger version (23K):
[in this window]
[in a new window]
|
Fig 4. (A) The plots of risk ratios of the studies on troponin (Tn) I and short-term mortality. No pooled estimate is presented due to considerable heterogeneity. (B) The plots of risk ratios of the studies on TnI and long-term mortality. No pooled estimate is presented due to considerable heterogeneity. (CI = confidence interval.)
|
|
In TnI studies reporting short-term mortality, there were 50 deaths among the 611 patients (8.1%) with TnI values above the cutoff and 40 deaths among the 2660 patients (1.5%) with TnI values below the cutoff. In TnI studies examining long-term mortality (6 to 36 months), there were 55 deaths among the 517 patients (10.6%) with TnI values above the cutoff and 59 deaths among the 1921 patients (3.1%) with TnI values below the cutoff. RRs varied considerably. They were higher in prospective vs retrospective studies (p = 0.058) and in studies with later TnI sampling rather than in studies using peak value (p = 0.035). Variation in the cutoff values failed to explain the variation in RRs, even after the cutoff values were transformed into multiples of ULRR with different methods (see Methods). Because of fundamental differences in assays and patient populations, no pooled estimate was considered reasonable.
|
Comment
|
|---|
Methodological Quality
According to our criteria, the methodological level of the studies appeared intermediate. Some reasons for this are obvious. First, many studies may have been appropriately designed for the purpose they were meant for, but eight of the studies were designed for investigating other issues, such as perioperative myocardial infarction, cardiac dysfunction instead of death, or cardiac operations instead of only coronary operations. Four studies were post hoc analyses of previous randomized controlled trials (Table 1).
Second, the known association between cardiac marker release and prognosis, and the standard practice of using cardiac markers in clinical decision making has probably hindered blinding.
Finally, the importance of various risk factors of mortality has emerged over time; therefore, when the oldest studies were planned, not all risk factors were yet known. Regardless of the reason, at least the absence of blinding, and failure to include over 90% of eligible patients may attenuate the association between biomarkers and mortality.
Biomarkers and Mortality
According to this meta-analysis, postoperative CK-MB release is associated with survival up to 40 months postoperatively. The RRs vary widely in separate studies, and this is probably the result of a complex interplay between different study designs, different populations, and different institutions. Troponin levels could be presumed to be better predictors of mortality than CK-MB because of their specificity to the myocardium. Owing to the smaller number of patients and the heterogeneity of troponin studies, this hypothesis could not be confirmed by these meta-analytic data. Nevertheless, two studies investigating CK-MB combined with TnI [17, 18] and one study examining CK-MB combined with TnT [15] actually suggest troponins are stronger predictors.
Timing of the Sample
In TnI studies, the 20- to 24-hour sample was a stronger predictor of mortality than the peak value. The kinetics of biomarkers after CABG is complex and affected by the extent of preoperative myocardial injury and reperfusion injury, graft failure, retransfusion of mediastinal blood, skeletal muscle injury (CK-MB), continuous extended release from degrading myofibrils (troponins), renal failure (troponins), and ischemic modification of molecules (TnI). Despite this complexity, the literature seems to support our finding: even though some studies have suggested meaningful early biomarker release to predict length of stay in the hospital [29] and adverse outcome [30], most studies show either benign early releases or more meaningful late releases of the biomarkers [19, 26, 31–39].
Standardization of the Assays
Calibration of the CK-MB assays by common reference material suggested in 1999 reduced between-manufacturer bias from 40% to 13% [40]. One study [14] included in this meta-analysis used activity assay whereas others used mass assay. CK-MB values were expressed as multiples of the local ULRR, and for the purposes of this study, they were accepted as comparable; nevertheless, some bias caused by interassay variability is probably present.
Because of patent protection, TnT assays are available from only one manufacturer. New versions of the assay have been developed. The initial 1% to 2% cross-reactivity with skeletal muscle [41] was solved by choice of new antibodies for the second-generation assay [42]. The third-generation assay used recombinant human cardiac TnT instead of bovine TnT as a standard material [43], and the fourth-generation assays allowed the use of plasma as sample material in addition to serum [44]. The recently released fifth-generation TnT assay is highly sensitive, with an extremely low detection limit [45]. Cutoff values between assays of different generations vary. One TnT study in this meta-analysis used a first-generation assay, and other TnT studies used third-generation assays.
Unlike with TnT, several commercial TnI assays are available. These assays use varying antibodies and detect varying epitopes. Because of this and the extensive modification of TnI molecules in ischemic myocardium, including complex formation with other troponins [46], proteolytic degradation [47], phosphorylation, and oxidation [48], current assays detect TnI unevenly [49]. Furthermore, interfering factors, such as rheumatoid factor, human antianimal antibodies, autoantibodies, heparin, or ethylenediaminetetraacetic acid in the test tube, all have influenced the assays to varying degrees [49]. Thus, variation of the results of the assays has been up to 100-fold among first-generation assays and is still fivefold among current assays. Standardization efforts are under way. The nine TnI studies in this meta-analysis used seven different types of TnI assay. Patient data were collected between 1998 and 2004; hence, the generations of the assays also differ. Nevertheless, TnI values obtained with different assays are not comparable. Thus, if postoperative TnI is used as quality control, an institutional cutoff value is recommended to be determined [50].
Different Patient Populations
Aortic cross-clamping [32] and even mere CPB [2] cause biomarker release. These meta-analytic data fail to give the answer whether this is negligible noise or evidence of cell destruction. RRs were higher for off-pump patients, but the difference was not significant. Also, most patients underwent on-pump operations (Fig 2). According to the study by Kuduvalli and colleagues [14], however, the same extent of CK-MB release has higher prognostic value among off-pump patients than it does among on-pump patients. This is supported by studies using magnetic resonance imaging [37, 51], according to which perioperative myocardial infarctions of the same size were associated with greater releases of biomarkers when CPB was used.
Also, cardiac operations other than CABG might cause greater biomarker release unrelated to ischemia. This meta-analysis is unable to either confirm or contradict this issue because of varying numbers of CABG patients and patients undergoing other cardiac operations in the studies. In the study by Croal and colleagues [25], however, the association between TnI and death was comparable in CABG patients and patients undergoing other cardiac operations. On the other hand, a study [52] not included in this review (adverse events as an end point) found a lower prognostic threshold of TnI for CABG than for other cardiac operations.
If our finding that the underlying risk does not affect the association between CK-MB and postoperative mortality holds true, the usefulness of a biomarker varies depending on a patient's underlying risk. A RR of 3.69 (for on-pump patients and short-term death) would mean an additional death for every 37 patients whose postoperative CK-MB value is above cutoff, if the underlying risk is 1%, but an additional death for every 3 patients in case of a 10% underlying risk. Thus, biomarkers should be analyzed only of the patients with considerable risk.
Independence of Prediction
All but one [27] of the multivariate analyses presented in the studies suggest that biomarkers provide independent prognostic data after coronary artery operations, even though only half of them were adjusted for the required factors. This independence in prediction is also supported by the finding involving isolated troponin release by Onorati and colleagues [28]. Patients with both TnI elevation and electrocardiographic or echocardiographic changes suggestive of ischemia had the poorest survival rate, but the survival of patients with isolated TnI elevation was also less likely than that of patients with low TnI values.
Reasons for the Association
The most natural explanation to the question why cardiac biomarkers associate with mortality is, that regardless of the reason, the amount of myocardial cell loss affects prognosis. Also, TnI has been able to distinguish between patients with and without graft occlusion within a subpopulation of patients having considerable TnI elevation [53]. However, TnI probably failed to reveal most graft occlusions because the incidence of graft occlusion was low, 1.7%, compared with 10% in a previous multicenter study [54]. Furthermore, the overlap between CK-MB and TnT release, graft occlusion rate, and clinical course has been substantial in another study [34].
Interestingly, TnI release into the circulation might also have harmful effects. Autoantibodies against TnI have been shown to cause myocardial inflammation, fibrosis, heart failure, and increased mortality in mice [55]. In patients these autoantibodies have been associated with worse recovery of left ventricular ejection fraction 6 to 9 months after acute myocardial infarction [56].
Limitations of the Review
The limitations of this study include the relatively small number of studies on a single biomarker and heterogeneity of the studies. Also, meta-analysis of observational studies is prone to confounding [57]. For example, some of the interinstitutional differences causing heterogeneity are impossible to explore, and there are competing explanations to the variability of RRs, such as inclusion of only low-risk patients [6] or off-pump patients, selection bias due to a large proportion of missing biomarker results [9, 10], or a very long, 5-year follow-up with inclusion of only patients surviving 30 days [11]. Also, we assessed the underlying risk by simply categorizing studies into three groups. This crude method may have influenced the absence of an association of RR and underlying risk. In addition, the search may not have found all of the eligible studies because we accepted only articles published in English. Asymmetry of the funnel plots suggests that publication bias may have affected the studies concerning CK-MB and long-term death, but, on the other hand, the asymmetry could have been caused by differences in study design as mentioned.
Conclusions
This review evaluates and interprets the published reports that examine the association between postoperative cardiac injury marker release and mortality. CK-MB release of more than five to eight times ULRR after CABG is associated with an increased risk of mortality extending beyond 3 years postoperatively. Troponins might serve as better predictors than CK-MB. However, fewer troponin studies have been published, and the patient populations in troponin studies are more heterogeneous. Lack of standardization of TnI assays prevents generalization and pooling of the results of the studies. Off-pump operations probably require different cutoff values from the ones used for on-pump procedures.
Standardized statistical methods and TnI analyses, as well as higher numbers of better-categorized patients, are warranted in future studies. Furthermore, whether the prognostic association is caused by perioperative, possibly preventable ischemic injury, or whether the marker merely serves as a surrogate for a patient population less tolerant to coronary surgery remains unclear.
|
Appendix 1A
|
|---|
Search Strategy in Ovid Medline Database
| exp |
explosion, ie, expanding the search by the terms under the one searched |
| su |
viewpoint of surgery |
| mp |
title, original title, abstract, name of substance word, subject heading word |
| $ |
including words with similar beginnings |
| 1 |
exp Cardiac Surgical Procedures/ |
| 2 |
Cardiac Surgical Procedures/or Cardiomyoplasty/or exp Heart Arrest, Induced/or exp Heart Bypass, Right/or Heart Bypass, Left/or Heart Valve Prosthesis Implantation/or Myocardial Revascularization/or exp Coronary Artery Bypass/ |
| 3 |
Cardiac Surgical Procedures/or Cardiomyoplasty/or exp Heart Arrest, Induced/or exp Heart Bypass, Right/or Heart Bypass, Left/or Heart Valve Prosthesis Implantation/ |
| 4 |
Myocardial Revascularization/or exp Coronary Artery Bypass/ |
| 5 |
exp Heart Diseases/su |
| 6 |
Heart Diseases/su or exp Arrhythmias, Cardiac/su or Carcinoid Heart Disease/su or exp Cardiomyopathies/su or exp Endocarditis/su or Heart Aneurysm/su or exp Heart Arrest/su or exp Heart Failure/su or Heart Neoplasms/su or exp Heart Rupture/su or exp Heart Valve Diseases/su or exp Myocardial Ischemia/su or Pulmonary Heart Disease/su or Rheumatic Heart Disease/su or exp Ventricular Dysfunction/su or exp Ventricular Outflow Obstruction/su |
| 7 |
Heart Diseases/su or exp Arrhythmias, Cardiac/su or Carcinoid Heart Disease/su or exp Cardiomyopathies/su or exp Endocarditis/su or Heart Aneurysm/su or exp Heart Arrest/su or exp Heart Failure/su or Heart Neoplasms/su or exp Heart Rupture/su or exp Heart Valve Diseases/su or Pulmonary Heart Disease/su or Rheumatic Heart Disease/su or exp Ventricular Dysfunction/su or exp Ventricular Outflow Obstruction/su |
| 8 |
exp Myocardial Ischemia/su |
| 9 |
1 or 5 |
| 10 |
2 or 6 |
| 11 |
3 or 7 |
| 12 |
4 or 8 |
| 13 |
Creatine kinase, MB Form/ |
| 14 |
(creatin$ adj1 (kinase or phosphokinase) adj5 (MB$ or myocardi$)).ti,ab. |
| 15 |
(CK-MB$ or CKMB$ or CPK-MB$ or CPKMB$).ti,ab. |
| 16 |
13 or 14 or 15 |
| 17 |
exp Troponin/ |
| 18 |
troponin$.mp. |
| 19 |
17 or 18 |
| 20 |
Fatty acid-binding proteins/ |
| 21 |
Fatty acid binding protein$.mp. |
| 22 |
fabp$.mp. |
| 23 |
20 or 21 or 22 |
| 24 |
Myoglobin/ |
| 25 |
myoglobin$.mp. |
| 26 |
24 or 25 |
| 27 |
9 and 16 |
| 28 |
9 and 19 |
| 29 |
9 and 23 |
| 30 |
9 and 26 |
| 31 |
27 or 28 or 29 or 30 |
| 32 |
10 and (16 or 19 or 23 or 26) |
| 33 |
11 and (16 or 19 or 23 or 26) |
| 34 |
12 and (16 or 19 or 23 or 26) |
| 35 |
limit 31 to yr="1990 - 2008" |
| 36 |
limit 32 to yr="1990 - 2008" |
| 37 |
limit 33 to yr="1990 - 2008" |
| 38 |
limit 34 to yr="1990 - 2008" |
|
|
Appendix 1B
|
|---|
Search Strategy in Embase Database
| DE |
Descriptor |
| MJ |
Major Descriptor |
| W |
Descriptor consisting of only one word |
| # |
Expanding the search by the descriptors under the major descriptor in the hierarchy of descriptors in Embase |
| $ |
including words with similar beginnings ADJ-operator "Words next to each other and in this order" |
| 1 |
emzz heart-surgery.mj. or cardiomyoplasty.w..mj. or cardioplegia.w..mj. or coronary-artery-surgery.mj. or coronary-artery-bypass-graft.mj. or coronary-artery-bypass-surgery.mj. or heart-left-ventricle-bypass.mj. or heart-muscle-revascularization.mj. or off-pump-coronary-surgery.mj. or heart-aneurysmectomy.mj. |
| 2 |
emzz heart-valve-surgery.mj. or annuloplasty.w..mj. or heart-valve-prosthesis#.mj. or heart-valve-replacement#.mj. or mitral-valve-commissurotomy.mj. or valvuloplasty.w..mj. or valvulotomy.w..mj. or heart-ventricle-remodeling.mj. or minimally-invasive-cardiac-surgery.mj. or open-heart-surgery.mj. |
| 3 |
emzz 1 or 2 |
| 4 |
emzz heart-disease-su.mj. or ecg-abnormality-su.mj. or endocardial-disease-su.mj. or endocarditis-su#.w..mj. or heart-abscess-su.mj. or heart-aneurysm-su#.mj. or heart-arrhythmia-su#.mj. or heart-death-su.mj. or heart-failure-su#.mj. or heart-injury-su#.mj. or heart-tumor-su#.mj. or ischemic-heart-disease-su#.mj. |
| 5 |
emzz cardiomyopathy-su#.w..mj. or heart-dilatation-su.mj. or heart-hypertrophy-su#.mj. or heart-muscle-injury-su.mj. or heart-muscle-ischemia-su.mj. or heart-rupture-su.mj. or heart-ventricle-septum-rupture-su.mj. or rheumatic-heart-disease-su.mj. or silent-myocardial-ischemia-su.mj. or stunned-heart-muscle-su.mj. or valvular-heart-disease-su#.mj. |
| 6 |
emzz 4 or 5 |
| 7 |
emzz 3 or 6 |
| 8 |
emzz creatine-kinase-mb.mj. or creatine-kinase-isoenzyme.mj. |
| 9 |
emzz (creatin$4 adj (kinase or phosphokinase) near (mb$4 or myocardi$3)).ti,ab. |
| 10 |
emzz (ck-mb$4 or ckmb$4 or cpk-mb$4 or cpkmb$4).ti,ab. |
| 11 |
emzz 8 or 9 or 10 |
| 12 |
emzz troponin#.w..mj. or troponin$3.ti,ab. |
| 13 |
emzz fatty-acid-binding-protein.mj. |
| 14 |
emzz (fatty adj acid adj binding adj protein$2).ti,ab. or fabp$4.ti,ab. |
| 15 |
emzz 13 or 14 |
| 16 |
emzz myoglobin.w..mj. or myoglobin$2.ti,ab. |
| 17 |
emzz 7 and 11 |
| 18 |
emzz 7 and 12 |
| 19 |
emzz 7 and 15 |
| 20 |
emzz 7 and 16 |
| 21 |
emzz 17 or 18 or 19 or 20 |
| 22 |
emzz limit set 21 year > 1989 |
| 23 |
emzz ischemic-heart-disease-su#.mj. |
| 24 |
emzz coronary-artery-surgery#.mj. |
| 25 |
emzz atherectomy#.w..mj. |
| 26 |
emzz 23 or 24 or 25 |
| 27 |
emzz creatine-kinase-isoenzyme.mj. |
| 28 |
emzz creatine-kinase-mb.mj. |
| 29 |
emzz ck-mb or ckmb |
| 30 |
emzz mb adj isoenzyme adj creatine adj kinase |
| 31 |
emzz 27 or 28 or 29 or 30 |
| 32 |
emzz troponin#.w..mj. |
| 33 |
emzz troponin$2 |
| 34 |
emzz 32 or 33 |
| 35 |
emzz fabp |
| 36 |
emzz fatty adj acid adj binding adj protein$2 |
| 37 |
emzz 35 or 36 |
| 38 |
emzz myoglobin.w..mj. |
| 39 |
emzz myoglobin$2 |
| 40 |
emzz 38 or 39 |
| 41 |
emzz 26 and 31 |
| 42 |
emzz 26 and 34 |
| 43 |
emzz 26 and 37 |
| 44 |
emzz 26 and 40 |
| 45 |
emzz 41 or 42 or 43 or 44 |
| 46 |
emzz limit set 45 year > 1989 |
|
|
Appendix 2
|
|---|
Framework for Assessing Validity of Articles Dealing With Prognosis (Adapted From Altman [3])
- Sample of patients
- 1 Were the inclusion criteria well defined and the sample selection clearly explained?
- 2 Was the sample of patients planned to be representative and assembled at a similar point in the course of the disease?
- • inclusion of only coronary artery bypass grafting (CABG) patients
- • sample size large enough to enable detection of 30 end points (mortality); if no power calculation was provided, mortality rates were drawn from meta-analysis of 13 randomized trials [37] and 2% was added to those studies with high-risk patientsa
- 3 Were all the eligible patients included?b
- • more than 90% of all the on-pump CABG patients, except those with preoperative myocardial infarction, included; number of excluded patients clearly reported
- 4 Were the clinical and demographic characteristics fully described?
- • at least seven variables included in more than 50% (10 of 19) of the available risk scores [38] (age, gender, renal function, peripheral vascular disease, prior cardiac operation, left ventricle ejection fraction, emergency operation).
- II Follow-Up
- 5 Was the follow-up sufficiently long and complete?
- • no shorter than in-hospital mortalityc
- • more than 90% of the patients
- III Prognostic variable
- 6 Was the prognostic variable fully defined, including the details of the measurement methods?
- • CKMB: manufacturer's name, and mass or activity
- • TnT: manufacturer's name and generation of the analysis or data-collection years
- • TnI: manufacturer's name and specific analysis
- 7 Was the prognostic variable precisely measured?
- • coefficient of variation reported and magnitude approximating a maximum 10%
- IV Outcome
- 8 Was the outcome fully defined, objectively measured, and appropriate?
- • mortality assessed at a clearly defined time-point
- 9 Was the outcome unbiased, eg, assessment blinded to prognostic information?
- • reported whether physicians were blinded to biomarker results
- V Analysis
- 10 Was the continuous predictive variable analyzed appropriately?
- • area under the curve analysis performed and one of the ratios (hazard ratio, risk ratio, or likelihood ratio) clearly reported with its 95% confidence intervals
- 11 Was predictability statistically adjusted for all other important factors?
- • multivariable analysis performed (logistic regression or Cox regression)
- • at least six of the seven risk factors included in the three risk scores with best predictive ability among CABG patients [38] (age, serum creatinine, chronic obstructive pulmonary disease, cerebrovascular disease, peripheral vascular disease, prior cardiac surgery, left ventricle ejection fraction)
- VI Treatment
- 12 Was the treatment subsequent to inclusion in the cohort fully described and standardized or randomized?
- • surgical technique, including mode of cardioplegia, reported
|
|
| a The required sample size was calculated according to 30 end-points to be in harmony with the requirement for the adjustment for all the other important factors (Item V.11.). In regression models, the number of variables should be restricted due to multiple significance tests [4]. An intermediate restriction was selected—5 per covariate, ie, 30 deaths per 6 factors in the regression model—from the recommendations that range from 3 to 10 dependent variables per covariate. |
| b Eligible patients were defined as the patient group providing the most coherent but generalizable data on biomarkers reflecting perioperative injury after coronary artery bypass grafting; hence, the exclusion of patients with preoperative acute myocardial infarction was allowed, as was surgery without cardiopulmonary bypass. |
| c The standard minimum follow-up interval in intensive care research is in-hospital, or 30-day mortality, which were our requirements as well. |
|
CK-MB = Creatine kinase MB subunit; TnI = troponin I; TnT = troponin T.
|
Acknowledgments
|
|---|
The expenses of the literature search in the Embase database were covered with a Helsinki University Hospital Grant. We wish to express our gratitude to information specialists Terhi Sandgren, Eeva-Liisa Aatola, Ulla Neuvonen, and Katri Larmo for performing the literature searches, and to Peter Raivio, MD, PhD, for reviewing the manuscript.
|
Footnotes
|
|---|
The Appendices are available online only. To access them, please visit http://ats.ctsnetjournals.org and search for the article by Petäjä et al, Vol 87, pages 1981–1992.
* See note at end of article regarding e-only Appendices. 
|
References
|
|---|
- Wu AH, Jaffe AS, Apple FS, et al. National Academy of Clinical Biochemistry laboratory medicine practice guidelines: use of cardiac troponin and B-type natriuretic peptide or N-terminal proB-type natriuretic peptide for etiologies other than acute coronary syndromes and heart failure Clin Chem 2007;53:2086-2096.[Abstract/Free Full Text]
- Rastan AJ, Bittner HB, Gummert JF, et al. On-pump beating heart versus off-pump coronary artery bypass surgery—evidence of pump-induced myocardial injury Eur J Cardiothorac Surg 2005;27:1057-1064.[Abstract/Free Full Text]
- Altman DG. Systematic reviews of evaluations of prognostic variables BMJ 2001;323:224-228.[Free Full Text]
- Peduzzi P, Concato J, Kemper E, Holford T, Feinstein A. A simulation study of the number of events per variable in logistic regression analysis J Clin Epidemiol 1996;49:1373-1379.[Medline]
- Panteghini M, Pagani F, Yeo KT, et al. Evaluation of imprecision for cardiac troponin assays at low-range concentrations Clin Chem 2004;50:327-332.[Abstract/Free Full Text]
- Costa MA, Carere RG, Lichtenstein SV, et al. Incidence, predictors, and significance of abnormal cardiac enzyme rise in patients treated with bypass surgery in the Arterial Revascularization Therapies Study (ARTS) Circulation 2001;104:2689-2693.[Abstract/Free Full Text]
- 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;38:1070-1077.[Abstract/Free Full Text]
- 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;129:300-306.[Abstract/Free Full Text]
- Mahaffey KW, Roe MT, Kilaru R, et al. Creatine kinase-MB elevation after coronary artery bypass grafting surgery in patients with non-ST-segment elevation acute coronary syndromes predict worse outcomes: results from four large clinical trials Eur Heart J 2007;28:425-432.[Abstract/Free Full Text]
- Brener SJ, Lytle BW, Schneider JP, Ellis SG, Topol EJ. Association between CK-MB elevation after percutaneous or surgical revascularization and three-year mortality J Am Coll Cardiol 2002;40:1961-1967.[Abstract/Free Full Text]
- Steuer J, Horte LG, Lindahl B, Stahle E. Impact of perioperative myocardial injury on early and long-term outcome after coronary artery bypass grafting Eur Heart J 2002;23:1219-1227.[Abstract/Free Full Text]
- Marso SP, Bliven BD, House JA, Muehlebach GF, Borkon AM. Myonecrosis following isolated coronary artery bypass grafting is common and associated with an increased risk of long-term mortality Eur Heart J 2003;24:1323-1328.[Abstract/Free Full Text]
- Engoren MC, Habib RH, Zacharias A, et al. The association of elevated creatine kinase-myocardial band on mortality after coronary artery bypass grafting surgery is time and magnitude limited Eur J Cardiothorac Surg 2005;28:114-119.[Medline]
- Kuduvalli M, Newall N, Stott A, Grayson AD, Fabri BM. Impact of avoiding cardiopulmonary bypass for coronary surgery on perioperative cardiac enzyme release and survival Eur J Cardiothorac Surg 2006;29:729-735.[Abstract/Free Full Text]
- 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;39:1518-1523.[Abstract/Free Full Text]
- 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;94:879-881.[Medline]
- Adabag AS, Rector T, Mithani S, et al. Prognostic significance of elevated cardiac troponin I after heart surgery Ann Thorac Surg 2007;83:1744-1750.[Abstract/Free Full Text]
- Paparella D, Cappabianca G, Malvindi P, et al. Myocardial injury after off-pump coronary artery bypass grafting operation Eur J Cardiothorac Surg 2007;32:481-487.[Abstract/Free Full Text]
- 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;50:1560-1567.[Abstract/Free Full Text]
- Nesher N, Alghamdi AA, Singh SK, et al. Troponin after cardiac surgery: a predictor or a phenomenon? Ann Thorac Surg 2008;85:1348-1354.[Abstract/Free Full Text]
- Lasocki S, Provenchere S, Benessiano J, et al. Cardiac troponin I is an independent predictor of in-hospital death after adult cardiac surgery Anesthesiology 2002;97:405-411.[Medline]
- Fellahi J, Gue X, Richomme X, Monier E, Guillou L, Riou B. Short- and long-term prognostic value of postoperative cardiac troponin I concentration in patients undergoing coronary artery bypass grafting Anesthesiology 2003;99:270-274.[Medline]
- Leal JC, de Paula Neto A, Avanci LE, Braile MC, de Godoy MF, Braile DM. Risk stratification with troponin I in patients undergoing myocardial revascularization surgery Arq Bras Cardiol 2003;80:279-288.[Medline]
- Paparella D, Cappabianca G, Visicchio G, et al. Cardiac troponin I release after coronary artery bypass grafting operation: effects on operative and midterm survival Ann Thorac Surg 2005;80:1758-1764.[Abstract/Free Full Text]
- Croal BL, Hillis GS, Gibson PH, et al. Relationship between postoperative cardiac troponin I levels and outcome of cardiac surgery Circulation 2006;114:1468-1475.[Abstract/Free Full Text]
- Carrier M, Pellerin M, Perrault LP, Solymoss BC, Pelletier LC. Troponin levels in patients with myocardial infarction after coronary artery bypass grafting Ann Thorac Surg 2000;69:435-440.[Abstract/Free Full Text]
- Provenchere S, Berroeta C, Reynaud C, et al. Plasma brain natriuretic peptide and cardiac troponin I concentrations after adult cardiac surgery: association with postoperative cardiac dysfunction and 1-year mortality Crit Care Med 2006;34:995-1000.[Medline]
- Onorati F, De Feo M, Mastroroberto P, et al. Determinants and prognosis of myocardial damage after coronary artery bypass grafting Ann Thorac Surg 2005;79:837-845.[Abstract/Free Full Text]
- Baggish AL, MacGillivray TE, Hoffman W, et al. Postoperative troponin-T predicts prolonged intensive care unit length of stay following cardiac surgery Crit Care Med 2004;32:1866-1871.[Medline]
- Eigel P, van Ingen G, Wagenpfeil S. Predictive value of perioperative cardiac Troponin I for adverse outcome in coronary artery bypass surgery Eur J Cardiothorac Surg 2001;20:544-549.[Abstract/Free Full Text]
- Benoit MO, Paris M, Silleran J, Fiemeyer A, Moatti N. Cardiac troponin I: its contribution to the diagnosis of perioperative myocardial infarction and various complications of cardiac surgery Crit Care Med 2001;29:1880-1886.[Medline]
- Fransen EJ, Maessen JG, Hermens WT, Glatz JF. Demonstration of ischemia-reperfusion injury separate from postoperative infarction in coronary artery bypass graft patients Ann Thorac Surg 1998;65:48-53.[Abstract/Free Full Text]
- Fransen EJ, Diris JHC, Maessen JG, Hermens WT, van Dieijen-Visser MP. Evaluation of "new" cardiac markers for ruling out myocardial infarction after coronary artery bypass grafting Chest 2002;122:1316-1321.[Abstract/Free Full Text]
- Holmvang L, Jurlander B, Rasmussen C, Thiis JJ, Grande P, Clemmensen P. Use of biochemical markers of infarction for diagnosing perioperative myocardial infarction and early graft occlusion after coronary artery bypass surgery Chest 2002;121:103-111.[Abstract/Free Full Text]
- Etievent J, Chocron S, Toubin G, et al. Use of cardiac troponin I as a marker of perioperative myocardial ischemia Ann Thorac Surg 1995;59:1192-1194.[Abstract/Free Full Text]
- Dahlin LG, Kågedal B, Nylander E, Olin C, Rutberg H, Svedjeholm R. Unspecific elevation of plasma troponin-T and CK-MB after coronary surgery Scand Cardiovasc J 2003;37:283.[Medline]
- Selvanayagam JB, Pigott D, Balacumaraswami L, Petersen SE, Neubauer S, Taggart DP. Relationship of irreversible myocardial injury to troponin I and creatine kinase-MB elevation after coronary artery bypass surgery: insights from cardiovascular magnetic resonance imaging J Am Coll Cardiol 2005;45:629-631.[Free Full Text]
- Salamonsen RF, Schneider HG, Bailey M, Taylor AJ. Cardiac troponin I concentrations, but not electrocardiographic results, predict an extended hospital stay after coronary artery bypass graft surgery Clin Chem 2005;51:40-46.[Abstract/Free Full Text]
- Fellahi JL, Hanouz JL, Gue X, Monier E, Guillou L, Riou B. Kinetic analysis of cardiac troponin I release is no more accurate than a single 24-h measurement in predicting in-hospital outcome after cardiac surgery Eur J Anaesthesiol 2008;25:490-497.[Medline]
- Christenson RH, Vaidya H, Landt Y, et al. Standardization of creatine kinase-MB (CK-MB) mass assays: the use of recombinant CK-MB as a reference material Clin Chem 1999;45:1414-1423.[Abstract/Free Full Text]
- Katus HA, Remppis A, Neumann FJ, et al. Diagnostic efficiency of troponin T measurements in acute myocardial infarction Circulation 1991;83:902-912.[Abstract/Free Full Text]
- Muller-Bardorff M, Hallermayer K, Schroder A, et al. Improved troponin T ELISA specific for cardiac troponin T isoform: assay development and analytical and clinical validation Clin Chem 1997;43:458-466.[Abstract/Free Full Text]
- Hallermayer K, Klenner D, Vogel R. Use of recombinant human cardiac troponin T for standardization of third generation troponin T methods Scand J Clin Lab Invest Suppl 1999;230:128-131.[Medline]
- Hermsen D, Apple F, Garcia-Beltran L, et al. Results from a multicenter evaluation of the 4th generation Elecsys troponin T assay Clin Lab 2007;53:1-9.[Medline]
- Kurz K, Giannitsis E, Zehelein J, Katus HA. Highly sensitive cardiac troponin T values remain constant after brief exercise- or pharmacologic-induced reversible myocardial ischemia Clin Chem 2008;54:1234-1238.[Abstract/Free Full Text]
- Katrukha AG, Bereznikova AV, Esakova TV, et al. Troponin I is released in bloodstream of patients with acute myocardial infarction not in free form but as complex Clin Chem 1997;43:1379-1385.[Abstract/Free Full Text]
- Katrukha AG, Bereznikova AV, Filatov VL, et al. Degradation of cardiac troponin I: implication for reliable immunodetection Clin Chem 1998;44:2433-2440.[Abstract/Free Full Text]
- Gaze DC, Collinson PO. Multiple molecular forms of circulating cardiac troponin: analytical and clinical significance Ann Clin Biochem 2008;45:349-355.[Abstract/Free Full Text]
- Tate JR. Review: troponin revisited 2008: assay performance Clin Chem Lab Med 2008;46:1489-1500.[Medline]
- Apple FS, Wu AH, Jaffe AS. European Society of Cardiology and American College of Cardiology guidelines for redefinition of myocardial infarction: how to use existing assays clinically and for clinical trials Am Heart J 2002;144:981-986.[Medline]
- Selvanayagam J, Petersen S, Francis J, et al. Effects of off-pump versus on-pump coronary surgery on reversible and irreversible myocardial injury: a randomized trial using cardiovascular magnetic resonance imaging and biochemical markers Circulation 2004;109:345-350.[Abstract/Free Full Text]
- Fellahi J, Hedoire F, Le Manach Y, Monier E, Guillou L, Riou B. Determination of the threshold of cardiac troponin I associated with an adverse postoperative outcome after cardiac surgery: a comparative study between coronary artery bypass graft, valve surgery, and combined cardiac surgery Crit Care 2007;11:R106.[Medline]
- Thielmann M, Massoudy P, Marggraf G, et al. Role of troponin I, myoglobin, and creatine kinase for the detection of early graft failure following coronary artery bypass grafting Eur J Cardiothorac Surg 2004;26:102-109.[Abstract/Free Full Text]
- Alderman EL, Levy JH, Rich JB, et al. Analyses of coronary graft patency after aprotinin use: results from the International Multicenter Aprotinin Graft Patency Eexperience (IMAGE) Trial J Thorac Cardiovasc Surg 1998;116:716-730.[Abstract/Free Full Text]
- Goser S, Andrassy M, Buss SJ, et al. Cardiac troponin I but not cardiac troponin T induces severe autoimmune inflammation in the myocardium Circulation 2006;114:1693-1702.[Abstract/Free Full Text]
- Leuschner F, Li J, Goser S, et al. Absence of auto-antibodies against cardiac troponin I predicts improvement of left ventricular function after acute myocardial infarction Eur Heart J 2008;29:1949-1955.[Abstract/Free Full Text]
- Egger M, Schneider M, Davey Smith G. Spurious precision?. Meta-analysis of observational studies. BMJ 1998;316:140-144.[Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
A. M. Ranasinghe, D. W. Quinn, M. Richardson, N. Freemantle, T. R. Graham, J. Mascaro, S. J. Rooney, I. C. Wilson, D. Pagano, and R. S. Bonser
Which Troponometric Best Predicts Midterm Outcome After Coronary Artery Bypass Graft Surgery?
Ann. Thorac. Surg.,
June 1, 2011;
91(6):
1860 - 1867.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
