- Email: [email protected]
Diagnostic value of ischemia-modified albumin in patients with suspected acute coronary syndrome
American Journal of Emergency Medicine (2010) 28, 170–176
www.elsevier.com/locate/ajem
Original Contribution
Diagnostic value of ischemia-modified albumin in patients with suspected acute coronary syndrome☆ Søren Hjortshøj MD a,⁎, Søren Risom Kristensen MD, DMSc b , Jan Ravkilde MD, DMSc a a
Department of Cardiology, Cardiovascular Research Centre, Aalborg Hospital, Aarhus University Hospital, DK-9000 Alborg, Denmark b Department of Clinical Biochemistry, Cardiovascular Research Centre, Aalborg Hospital, Aarhus University Hospital, DK-9000 Alborg, Denmark Received 17 September 2008; accepted 26 October 2008
Abstract Introduction: Ischemia-modified albumin (IMA) has been proposed as a useful rule-out marker for the diagnosis of acute coronary syndrome (ACS) in the emergency department. This study evaluated the ability of IMA to predict the acute myocardial infarction (AMI) diagnosis in a population of chest pain patients. Methods: The study population comprised 107 subjects (men, 62%; women, 38%) admitted with suspected ACS. None of the patients had ST-segment elevations that qualified for immediate revascularization. Ischemia-modified albumin was determined from serum with albumin cobalt binding test (Inverness Medical Innovations Inc, Stirling, UK). Furthermore, cardiac troponin T, creatinine kinase MB mass, myoglobin, and heart-type fatty acid binding protein (H-FABP) were determined on arrival, after 6 to 9 hours, and after 12 to 24 hours. All patients had at least 2 blood samples taken to exclude/verify the AMI. AMI was defined by a cardiac troponin T level greater than 0.03 μg/L. Results: Thirty-three percent of the patients (n = 35) had a final diagnosis of AMI. The sensitivity of admission IMA for a final diagnosis of ACS was 0.86 (95% confidence interval [95% CI], 0.69-0.95). Specificity was 0.49 (95% CI, 0.36-0.60). Negative predictive value was 0.88 (95% CI, 0.72-0.95). The optimal cutoff threshold derived from the receiver operating characteristics (ROC) curve (ROC analysis) was determined as 91 U/mL. The area under the ROC curve was 0.73. Ischemia-modified albumin did not, at any time, provide superior sensitivity or specificity compared with other biomarkers.We do not find the data supportive of IMA as a standard marker in the emergency department. © 2010 Elsevier Inc. All rights reserved.
1. Introduction Biochemical diagnosis of acute coronary syndrome (ACS) and myocardial infarction (MI) relies heavily on biomarkers of myocardial necrosis, that is, cardiac troponins ☆ Grants and support: This study was supported with assays, reagents, and technical support by Inverness Medical Inc, Stirling, UK. ⁎ Corresponding author. Tel.: +45 9932 2178 or +45 2293 9778. E-mail address: [email protected] (S. Hjortshøj).
0735-6757/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2008.10.038
and creatine kinase MB (CKMB) fraction [1]. The routine use of these markers has been shown to improve risk stratification, therapy, and outcomes [2,3]. Ischemia-modified albumin is a new ischemia marker and has been proposed as a useful marker for the diagnosis of severe ischemic heart disease and ACS. Ischemia-modified albumin represents a new class of markers that depend on the ischemia preceding myocardial necrosis. The goal is to have a biomarker challenging the electrocardiogram (ECG) as the fastest diagnostic tool in cardiac ischemia.
Ischemia modified albumin in acute coronary syndrome
171
Ischemia-modified albumin is considered to be a fraction of human serum albumin (HSA) where the Nterminal end has undergone chemical changes mediated by ischemia. This, in turn, decreases the binding capacity of HSA for exogenous cobalt. Initial studies suggested that damage by reactive ischemia-dependent oxidative species was responsible for damage to the N-terminal of HSA [4]. Although recent studies have investigated the underlying mechanisms in depth, the formation and clearance of IMA still remain largely unexplained [5]. The albumin cobalt binding test (ACB test) is the first commercially available assay of IMA and has been approved by the US Food and Drug Administration for the detection of cardiac ischemia. Ischemia-modified albumin has been investigated in different study populations. In chest pain patients with suspected ACS, IMA has been shown to be a useful adjunct to electrocardiography and troponins, increasing sensitivity for the diagnosis when used in combination [6-12]. A high negative predictive value (NPV) found in some studies has led to speculations that IMA may be a useful rule-out marker in the emergency department [8,13]. Others, however, have found IMA to be of limited practical value, with a number of false-negatives that could undermine the safety of the test [14]. Furthermore, there have been reports that increases in IMA may also reflect conditions other than cardiac ischemia [15-20]. The aim of this study was to evaluate the ability of IMA to predict the AMI diagnosis as well as its NPV in a population of subjects presenting with chest pain to the emergency department.
onset of symptoms as accurately as possible (ie, description of chest pain, pulmonary edema, severe dyspnea, and syncope), a general patient history, clinical examination, 12-lead ECG, and the laboratory tests described below.
2. Material and methods 2.1. Patient specimens The study population comprised 107 subjects who were admitted with chest pain and suspected of ACS during a 4-month period (February to May 2007). The study population included 66 men (62%) and 41 women (38%). Patients were referred by a general practitioner or the ambulance service. They were subsequently admitted and included by the physician on duty. Patients admitted more than once during the investigation period were included only on their first admission. 12-Lead ECG was performed in all patients. None of the patients had ST-segment elevations that qualified for immediate angiography and primary percutaneous coronary intervention (PCI) or thrombolytic therapy.
2.2. Routine diagnostic procedures At the emergency department and in the coronary care unit (CCU), routine procedures comprise establishing time of
2.3. Electrocardiogram An ECG was taken on admission to the emergency department, on arrival at the CCU, and at least once daily thereafter. Patients with ECGs readily diagnostic for STsegment elevation MI went to primary PCI and were excluded. Electrocardiogram changes were classified as progressive, nonprogressive, or uncodable. Electrocardiogram changes were considered transient if new specific STsegment (ST depression or ST elevation N0.1 mV) and Twave changes would subside in later recordings. Changes were considered progressive if ST-segment changes progressed in 2 or more leads from earlier recordings. Nonprogressive changes comprise ST-segment and T-wave changes that did not change in appearance in serial recordings. Electrocardiograms were considered uncodable in the presence of left bundle branch block, pacemaker artifacts, or complete atrioventricular block.
2.4. Medical therapy All patients were transferred to the CCU with a standard regimen for patients suspected of having ACS. On arrival, 300 mg of aspirin was administered to all patients (thereafter 75 mg OD). If the ECG showed ST depressions, T-wave inversions, or other signs that increased the likelihood of ACS, the physician on duty administered low-molecularweight heparin (enoxaparin 1 mg/kg body weight BD) and clopidogrel (300 mg bolus, 75 mg OD). Otherwise, results of troponin T analysis were awaited. If ACS was diagnosed by increased cardiac troponin T (cTnT) or if the patient was otherwise considered unstable, treatment with low-molecular-weight heparin, clopidogrel, aspirin, and, in many cases, statins and β-blockers was instituted.
2.5. Cardiac biomarkers Ischemia-modified albumin was determined from serum with the ACB test (Inverness Medical Innovations Inc, Stirling, UK). The principle of the assay is described elsewhere [5]. Analyses were performed on a Cobas MIRA Plus instrument (Roche Diagnostics, Mannheim, Germany). The within-series coefficient of variation (CV) was 8.3%, and the between-day imprecision was 11%. A reference level was previously determined from a control reference population of 258 healthy blood donors. Analyses from the reference population were performed on a Hitachi 911 instrument (Roche Diagnostics, Mannheim, Germany). Blood samples were handled and analyses were performed according to the manufacturer's protocols.
172 Cardiac troponin T, CKMB mass (CKMBmass), and myoglobin were determined from serum at 37°C on Elecsys 2010 (Roche Diagnostics, Mannheim, Switzerland). For cTnT (fourth-generation assay), the lowest concentration exhibiting an imprecision CV of less than 10% was 0.03 μg/L [21]. The upper reference value for CKMBmass was 4.0 μg/L for women and 7.0 μg/L for men [22], and that for myoglobin was 51 μg/L for women and 72 μg/L for men [23]. Heart-type fatty acid binding protein (H-FABP) was analyzed from dipotassium ethylenediamine tetraacetic acid (K-EDTA) plasma using ELISA technique (HyCult Biotechnology, Uden, The Netherlands); upper reference value is 6.0 μg/L with a CV of 10% or less [23].
2.6. Blood sampling Blood sampling was scheduled according to the department's routine procedures, where blood samples for cTnT and CKMBmass were obtained on arrival, after 6 to 9 hours,
S. Hjortshøj et al. and after 12 to 24 hours. All patients had at least 2 blood samples taken to exclude/verify the AMI diagnosis. On arrival, patients also had blood samples drawn for the evaluation of electrolytes, renal parameters, and lipid metabolism. Blood samples for determination of IMA and H-FABP were obtained together with the above-mentioned samples. Blood samples for determination of cTnT, CKMBmass, and IMA were drawn into dry tubes without anticoagulants. Samples for determination of H-FABP were drawn into K-EDTA tubes. After coagulation, samples were centrifuged at 4000g for 10 min. The serum was stored at −80°C until analysis. The total processing time from obtainment of samples to storage did not exceed 2.5 hours as recommended by the manufacturer. Frozen samples were mixed after thawing and recentrifuged before analysis. Samples did not undergo repeat freeze-thaw cycles. To evaluate sensitivity of the assays, relative concentrations were calculated as measured concentration divided by the 99th percentile given a CV of 10% or less. For IMA, relative
Fig. 1 Control reference population. A, IMA in control reference population followed a normal distribution. B, F test showing normal distribution. C, IMA levels were not influenced by age. D, An inverse relationship was seen with albumin levels. However, this did not influence other results.
Ischemia modified albumin in acute coronary syndrome Table 1
Baseline characteristics and levels of biomarkers
No. of patients admitted Sex, male:female, n Age, median (range), y Medical history before admission, n (%) Angina pectoris MI Coronary artery bypass grafting (CABG) Diabetes mellitus Biochemical markers, median (range) Peak IMA (U/mL) Peak cTnT (μg/L) Peak CKMBmass (μg/L) Peak H-FABP (μg/L) Electrocardiogram, no. (%) Q waves Progressive ST-segment changes Nonprogressive ST-segment changes Normal Uncodable
Chest pain with AMI (cTnT N0.03 μg/L)
Chest pain without AMI Peak IMA N88.2 U/mL
Peak IMA b88.2 U/mL
35 23:12 62 (25-92)
34 20:14 60 (30-88)
38 27:11 63 (34-84)
35 (100) 8 (23) 4 (11) 10 (29)
9 (26) 12 (35) 3 (9) 3 (9)
7 6 1 6
96.3 (58.5-117) 0.52 (0.03-2.82) 38 (1.7-387) 48.1 (0.1-151.6)
97.0 (88.5-110) 0.001 (0-0.02) 17 (1-128) 1.6 (0.1-9)
75.5 (63-88) 0.002 (0-0.02) 18 (1-91) 1.35 (0.4-6.8)
3 9 6 9 8
4 (11) 0 (0) 8 (24) 19 (56) 3 (9)
3 (8) 0 (0) 7 (18) 23 (61) 5 (13)
(9) (26) (16) (26) (23)
concentrations were calculated as measured concentrations divided by 95th percentile.
2.7. Definition of acute MI Patients were classified as having acute MI (AMI) according to the new universal definition of AMI [1]. These included detection of rise and/or fall in cTnT above the 99th percentile of the upper reference limit (N 0.03 μg/L) together with signs indicative of the following: • • • •
173
Relevant clinical symptoms of ischemia New ST-segment changes in the ECG Development of new Q waves in the ECG Echocardiographic signs of new regional motion defects.
(18) (16) (3) (16)
2.8. Exclusion criteria Patients were excluded, if they had a documented MI within the last week before admission. Patients who were on hemodialysis or had jaundice were also excluded from the study.
2.9. Statistics Results from the control reference population were evaluated with the F-test method for a normal distribution fit. Comparison between groups was performed by using Fischer's exact test. Medians and ranges were used for descriptive purposes. Receiver operating characteristics (ROC) analysis was performed by the method of Hanley and McNeil. Wilcoxon rank sum test was used for comparison of patients who were positive and negative for cTnT. The level of significance chosen was .05. The statistical analyses were performed using STATA statistical software package (StataCorp, College Station, Tex).
2.10. Ethics The study was approved by the ethical committee of North Jutland and Viborg Counties. Patients gave informed consent.
3. Results 3.1. Reference population Fig. 2 IMA in patients with normal and increased cTnT on admission. Cutoff for cTnT is 0.03 μg/L. There was significant difference between the 2 groups (P b .0001).
The ACB test values for the reference population were normally distributed (Fig. 1). Values for the control
174
S. Hjortshøj et al.
reference population were 49 to 136.5 U/mL (mean, 74.1 U/mL; median, 73.5 U/mL). The age distribution was 19 to 65 years (mean, 43 years; median, 43.5 years). The assay exhibited a CV of 10% or less at both the upper 95th and 99th percentile (88.2 and 111.8 U/mL, respectively). The 95th percentile has been used in other studies [8], and 88.2 U/mL was therefore defined as upper limit of normal. Because of a slight inverse relationship between IMA and levels of HSA, correction for albumin level was performed as proposed by Lippi et al [24] [(individual serum albumin concentration/median albumin concentration of the population) × IMA value = corrected IMA value]. This did not, however, lead to significant changes at all, and results are therefore given without correction.
3.2. Chest pain patients
Fig. 3 ROC curve for the ability of admission IMA to predict a cTnT level of 0.03 μg/L or greater (from admission to 24 hours).
The baseline characteristics of enrolled patients are shown in Table 1. Thirty-three percent of the patients (n = 35) had a final diagnosis of AMI with a cTnT level greater than 0.03 μg/L. Fig. 2 shows levels of admission IMA in patients with cTnT level greater than 0.03 μg/L and 0.03 μg/L or less. The sensitivity of admission IMA for a final diagnosis of AMI was 0.86 (95% confidence interval [95% CI], 0.69-0.95) (Table 2). Specificity was 0.49 (95% CI, 0.360.60). The NPV was 0.88 (95% CI, 0.72-0.95). Figs. 3 and 4 show the ROC curves (ROC analysis). The area under the ROC curve for admission IMA was 0.73. The optimal cutoff threshold derived from the ROC analysis was determined to be 91 U/mL. This corresponds to approximately the 97.25th percentile in the control reference material. Sensitivity, specificity, and NPV for a cutoff of 91 U/mL are also shown in Table 2. Sensitivity was decreased whereas specificity increased at this cutoff. Table 2
For comparison of different biomarkers, Table 2 shows the sensitivity, specificity, NPV, and area under the ROC curve for IMA, CKMBmass, myoglobin, H-FABP, and ECG.
4. Discussion This study was intended to evaluate IMA in a population of chest pain patients. The main results show that the admission sample of IMA identified only 86% (sensitivity) of patients who were otherwise diagnosed as having AMI according to current recommendations from ESC/ACC [1]. Likewise, the test had a low specificity of 49% and an NPV of 88%. Also, the ROC curve (1 − specificity vs sensitivity) showed an area under the curve of 0.73 (Fig. 3). Compared
Overview of IMA diagnostic performance of admission samples in different markers
Marker Admission IMA cutoff 88.2 U/mL IMA cutoff 91 U/mL CKMBmass Myoglobin H-FABP IMA at 6 to 9 h or peak value IMA 6-9 h, 88.2 U/mL Peak IMA, 88 U/mL Combinations with ECG ECG on admission IMA + ECG CKMB + ECG Myoglobin + ECG H-FABP + ECG
Sensitivity
Specificity
NPV
ROC analysis (area under curve)
0.86 (0.69-0.95) 0.71 (0.53-0.85) 0.85 (0.67-0.94) 0.64 (0.48-0.78) 0.83 (0.65-0.93)
0.49 (0.36-0.60) 0.65 (0.53-0.75) 0.67 (0.81-0.96) 0.88 (0.77-0.94) 0.92 (0.83-0.97)
0.88 (0.72-0.95) 0.82 (0.69-0.91) 0.93 (0.84-0.97) 0.79 (0.68-0.88) 0.92 (0.83-0.97)
0.73 (0.62-0.83) — 0.93 (0.87-0.99) 0.90 (0.82-0.98) 0.91 (0.83-0.99)
0.66 (0.48-0.80) 0.86 (0.69-0.95)
0.65 (0.53-0.76) 0.51 (0.39-0.63)
0.80 (0.67-0.89) 0.88 (0.74-0.96)
0.73 (63-0.83) 0.78 (0.69-0.87)
0.31 (0.21-0.43) 0.91 (0.76-0.98) 0.94 (0.79-0.99) 0.91 (0.71-0.98) 0.89 (0.72-0.96)
0.49 (0.32-0.67) 0.46 (0.32-0.55) 0.32 (0.22-0.44) 0.41 (0.27-0.58) 0.67 (0.54-0.77)
0.27 (0.16-0.38) 0.91 (0.79-0.98) 0.92 (0.72-0.98) 0.90 (0.79-0.98) 0.92 (0.81-0.98)
0.60 (0.50-0.70) 0.67 (0.58-0.76) 0.81 (0.74-0.89) 0.77 (0.72-0.86) 0.78 (0.70-0.85)
Data shown with 95% confidence intervals in parentheses. AMI is defined by the presence of a cTnT level greater than 0.03 μg/L according to the ESC/ACC consensus report.
Ischemia modified albumin in acute coronary syndrome
175
Fig. 4 ROC curves for the ability of IMA to predict the following: A, AMI; B, CKMBmass; C, myoglobin; and D, and H-FABP. AMI is defined by the presence of cTnT level greater than 0.03 mg/L according to the ESC/ACC consensus report.
with all other early biomarkers of myocardial injury, IMA exhibited lower values for diagnostic performance (Table 2). There have been confusing results regarding IMA and its value in the diagnosis of ACS. Some studies have found the test of possible use, and in particular, a high NPV has attracted attention [8,13]. However, we found an NPV of only 0.88 (95% CI, 0.72-0.95), which means that one cannot safely rely on a negative IMA as a single test for excluding IMA. To explore further into this particular patient population, we determined an optimal cutoff value of 91 U/mL from the ROC analysis. This value is between the 95th and the 97.5th percentile of our control reference material. The use of 91 U/mL led to a decreased sensitivity (0.71; 95% CI, 0.530.85) and increased specificity (0.65; 95% CI, 0.53-0.75). The NPV at this cutoff was even lower (0.82; 95% CI, 0.690.91). This patient cohort was enrolled at a tertiary hospital. One has to assume that some form of selection has occurred and that the prevalence of cTnT-positive patients would be larger than in many other settings in the health service. When applying classic Bayesian thinking, the prevalence of AMI in the population where the ACB test is introduced becomes very important for the test's ability to distinguish between
diseased and nondiseased individuals. The mediocre sensitivity, specificity, and NPV of the test make the ACB test particularly vulnerable to changes in the prevalence of AMI. An example illustrating this can be calculated. In a setting of 1000 patients with an AMI prevalence of 33%, for example, a CCU, a sensitivity of 0.91 and a specificity of 0.46 for the ACB test (+ECG) will lead to a positive predictive value of 0.45 and an NPVof 0.91. The test will thus lead to 30 patients being false-negative, whereas 362 patients would be falsepositive. In another setting, for example, an emergency department, with a prevalence of AMI of 5%, the test would generate a positive predictive value of 8% and an NPV of 99%. This translates into only 4 patients being false-negative but an overwhelming 513 patients who are false-positive and thus need further testing. It is therefore not easy balancing the scales and deciding where to apply the ACB test. No consensus on exact cutoff levels for IMA has yet been achieved from the studies that have been performed. Some studies have used 80 U/mL [25], whereas others have found even lower values of 75 U/mL [8,12]. Other groups have used the manufacturer's recommendations of 85 U/mL [9-11]. Although the cardiology and biochemistry communities have
176 decided that the optimal cutoff for other cardiac biomarkers (cTnT and CKMBmass) should be at the 99th percentile with a CV of 10% or less, there seems to be no consensus on this important point regarding IMA. Although our data from the control reference population indicate that an acceptable CV% could be achieved at the 99th percentile, the ROC analysis clearly shows that a cutoff at this level would decrease the specificity of the assay to an unacceptable degree. The limitations of this study are, first and foremost, the size of the patient group. On the other hand, our control reference population is the largest published normal material to this date and therefore provides a solid basis. In this study, we test the diagnostic performance of IMA in the setting of possible ACS. This may be an unfair comparison because no standardized test apart from the ECG at present measures ischemia. However, as a proposed ischemia marker, IMA must be able to detect the severest forms of ischemic heart disease, namely, AMI. Several issues have slowed the progress of IMA as a standard marker in coronary care. First, the general concept of marker of ischemia and its possible flaws has been regarded with suspicion within the cardiology community. Second, complicated issues regarding handling and storage of samples have limited widespread use in automated laboratories and clinical trials [14]. In conclusion, in our population of chest pain patients, we do not find the data supportive of IMA as a standard marker in the emergency department.
Acknowledgments The staff at the Departments of Cardiology and Clinical Chemistry, Aalborg Hospital, Aarhus University Hospital, is greatly thanked for their help and support during sampling and analysis of samples. The study was financially supported by The Danish Heart Foundation and the County of North Jutland Research Foundation. Also, Inverness Medical Innovations, Inc, is thanked for providing assays and technical support.
References [1] Thygesen K, Alpert JS, White HD, et al. Universal definition of myocardial infarction. Circulation 2007;116(22):2634-53. [2] Ravkilde J. Risk stratification in acute coronary syndrome using cardiac troponin I. Clin Chem 2000;46(4):443-4. [3] Wallentin L, Lagerqvist B, Husted S, Kontny F, Stahle E, Swahn E. Outcome at 1 year after an invasive compared with a non-invasive strategy in unstable coronary-artery disease: the FRISC II invasive randomised trial. FRISC II Investigators. Fast Revascularisation During Instability in Coronary Artery Disease. Lancet 2000;356 (9223):9-16. [4] Bar-Or D, Curtis G, Rao N, Bampos N, Lau E. Characterization of the Co(2+) and Ni(2+) binding amino-acid residues of the N-terminus of human albumin. An insight into the mechanism of a new assay for myocardial ischemia. Eur J Biochem 2001;268(1):42-7. [5] Bar-Or D, Rael LT, Bar-Or R, et al. The cobalt-albumin binding assay: insights into its mode of action. Clin Chim Acta 2008;387(1-2):120-7.
S. Hjortshøj et al. [6] Abadie JM, Blassingame CL, Bankson DD. Albumin cobalt binding assay to rule out acute coronary syndrome. Ann Clin Lab Sci 2005;35 (1):66-72. [7] Bar-Or D, Lau E, Winkler JV. A novel assay for cobalt-albumin binding and its potential as a marker for myocardial ischemia—a preliminary report. J Emerg Med 2000;19(4):311-5. [8] Christenson RH, Duh SH, Sanhai WR, et al. Characteristics of an albumin cobalt binding test for assessment of acute coronary syndrome patients: a multicenter study. Clin Chem 2001;47(3):464-70. [9] Collinson PO, Gaze DC, Bainbridge K, et al. Utility of admission cardiac troponin and “ischemia modified albumin” measurements for rapid evaluation and rule out of suspected acute myocardial infarction in the emergency department. Emerg Med J 2006;23(4):256-61. [10] Roy D, Quiles J, Aldama G, et al. Ischemia modified albumin for the assessment of patients presenting to the emergency department with acute chest pain but normal or non-diagnostic 12-lead electrocardiograms and negative cardiac troponin T. Int J Cardiol 2004;97(2):297-301. [11] Sinha MK, Roy D, Gaze DC, Collinson PO, Kaski JC. Role of “ischemia modified albumin”, a new biochemical marker of myocardial ischaemia, in the early diagnosis of acute coronary syndromes. Emerg Med J 2004;21(1):29-34. [12] Wu AH, Morris DL, Fletcher DR, Apple FS, Christenson RH, Painter PC. Analysis of the albumin cobalt binding (ACB) test as an adjunct to cardiac troponin I for the early detection of acute myocardial infarction. Cardiovasc Toxicol 2001;1(2):147-51. [13] Peacock F, Morris DL, Anwaruddin S, et al. Meta-analysis of ischemia-modified albumin to rule out acute coronary syndromes in the emergency department. Am Heart J 2006;152(2):253-62. [14] Keating L, Benger JR, Beetham R, et al. The PRIMA study: presentation ischaemia-modified albumin in the emergency department. Emerg Med J 2006;23(10):764-8. [15] Can M, Demirtas S, Polat O, Yildiz A. Evaluation of effects of ischaemia on the albumin cobalt binding (ACB) assay in patients exposed to trauma. Emerg Med J 2006;23(7):537-9. [16] Falkensammer J, Stojakovic T, Huber K, et al. Serum levels of ischemiamodified albumin in healthy volunteers after exercise-induced calfmuscle ischemia. Clin Chem Lab Med 2007;45(4):535-40. [17] Roy D, Quiles J, Sharma R, et al. Ischemia-modified albumin concentrations in patients with peripheral vascular disease and exerciseinduced skeletal muscle ischemia. Clin Chem 2004;50(9):1656-60. [18] Roy D, Quiles J, Sinha M, et al. Effect of radiofrequency catheter ablation on the biochemical marker ischemia modified albumin. Am J Cardiol 2004;94(2):234-6. [19] Roy D, Quiles J, Sinha M, Aldama G, Gaze D, Kaski JC. Effect of direct-current cardioversion on ischemia-modified albumin levels in patients with atrial fibrillation. Am J Cardiol 2004;93(3):366-8. [20] Borderie D, Allanore Y, Meune C, Devaux JY, Ekindjian OG, Kahan A. High ischemia-modified albumin concentration reflects oxidative stress but not myocardial involvement in systemic sclerosis. Clin Chem 2004;50(11):2190-3. [21] Roche Diagnostics. Troponin T STAT (Short Turn Around Time), CARDIAC T 04660307 190 2007-07. 2007. [22] Apple FS, Quist HE, Doyle PJ, Otto AP, Murakami MM. Plasma 99th percentile reference limits for cardiac troponin and creatine kinase MB mass for use with European Society of Cardiology/American College of Cardiology consensus recommendations. Clin Chem 2003;49(8):1331-6. [23] Pelsers MM, Chapelle JP, Knapen M, et al. Influence of age and sex and day-to-day and within-day biological variation on plasma concentrations of fatty acid-binding protein and myoglobin in healthy subjects. Clin Chem 1999;45(3):441-3. [24] Lippi G, Montagnana M, Salvagno GL, Guidi GC. Standardization of ischemia-modified albumin testing: adjustment for serum albumin. Clin Chem Lab Med 2007;45(2):261-2. [25] Worster A, Devereaux PJ, Heels-Ansdell D, et al. Capability of ischemia-modified albumin to predict serious cardiac outcomes in the short term among patients with potential acute coronary syndrome. CMAJ 2005;172(13):1685-90.
www.elsevier.com/locate/ajem
Original Contribution
Diagnostic value of ischemia-modified albumin in patients with suspected acute coronary syndrome☆ Søren Hjortshøj MD a,⁎, Søren Risom Kristensen MD, DMSc b , Jan Ravkilde MD, DMSc a a
Department of Cardiology, Cardiovascular Research Centre, Aalborg Hospital, Aarhus University Hospital, DK-9000 Alborg, Denmark b Department of Clinical Biochemistry, Cardiovascular Research Centre, Aalborg Hospital, Aarhus University Hospital, DK-9000 Alborg, Denmark Received 17 September 2008; accepted 26 October 2008
Abstract Introduction: Ischemia-modified albumin (IMA) has been proposed as a useful rule-out marker for the diagnosis of acute coronary syndrome (ACS) in the emergency department. This study evaluated the ability of IMA to predict the acute myocardial infarction (AMI) diagnosis in a population of chest pain patients. Methods: The study population comprised 107 subjects (men, 62%; women, 38%) admitted with suspected ACS. None of the patients had ST-segment elevations that qualified for immediate revascularization. Ischemia-modified albumin was determined from serum with albumin cobalt binding test (Inverness Medical Innovations Inc, Stirling, UK). Furthermore, cardiac troponin T, creatinine kinase MB mass, myoglobin, and heart-type fatty acid binding protein (H-FABP) were determined on arrival, after 6 to 9 hours, and after 12 to 24 hours. All patients had at least 2 blood samples taken to exclude/verify the AMI. AMI was defined by a cardiac troponin T level greater than 0.03 μg/L. Results: Thirty-three percent of the patients (n = 35) had a final diagnosis of AMI. The sensitivity of admission IMA for a final diagnosis of ACS was 0.86 (95% confidence interval [95% CI], 0.69-0.95). Specificity was 0.49 (95% CI, 0.36-0.60). Negative predictive value was 0.88 (95% CI, 0.72-0.95). The optimal cutoff threshold derived from the receiver operating characteristics (ROC) curve (ROC analysis) was determined as 91 U/mL. The area under the ROC curve was 0.73. Ischemia-modified albumin did not, at any time, provide superior sensitivity or specificity compared with other biomarkers.We do not find the data supportive of IMA as a standard marker in the emergency department. © 2010 Elsevier Inc. All rights reserved.
1. Introduction Biochemical diagnosis of acute coronary syndrome (ACS) and myocardial infarction (MI) relies heavily on biomarkers of myocardial necrosis, that is, cardiac troponins ☆ Grants and support: This study was supported with assays, reagents, and technical support by Inverness Medical Inc, Stirling, UK. ⁎ Corresponding author. Tel.: +45 9932 2178 or +45 2293 9778. E-mail address: [email protected] (S. Hjortshøj).
0735-6757/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2008.10.038
and creatine kinase MB (CKMB) fraction [1]. The routine use of these markers has been shown to improve risk stratification, therapy, and outcomes [2,3]. Ischemia-modified albumin is a new ischemia marker and has been proposed as a useful marker for the diagnosis of severe ischemic heart disease and ACS. Ischemia-modified albumin represents a new class of markers that depend on the ischemia preceding myocardial necrosis. The goal is to have a biomarker challenging the electrocardiogram (ECG) as the fastest diagnostic tool in cardiac ischemia.
Ischemia modified albumin in acute coronary syndrome
171
Ischemia-modified albumin is considered to be a fraction of human serum albumin (HSA) where the Nterminal end has undergone chemical changes mediated by ischemia. This, in turn, decreases the binding capacity of HSA for exogenous cobalt. Initial studies suggested that damage by reactive ischemia-dependent oxidative species was responsible for damage to the N-terminal of HSA [4]. Although recent studies have investigated the underlying mechanisms in depth, the formation and clearance of IMA still remain largely unexplained [5]. The albumin cobalt binding test (ACB test) is the first commercially available assay of IMA and has been approved by the US Food and Drug Administration for the detection of cardiac ischemia. Ischemia-modified albumin has been investigated in different study populations. In chest pain patients with suspected ACS, IMA has been shown to be a useful adjunct to electrocardiography and troponins, increasing sensitivity for the diagnosis when used in combination [6-12]. A high negative predictive value (NPV) found in some studies has led to speculations that IMA may be a useful rule-out marker in the emergency department [8,13]. Others, however, have found IMA to be of limited practical value, with a number of false-negatives that could undermine the safety of the test [14]. Furthermore, there have been reports that increases in IMA may also reflect conditions other than cardiac ischemia [15-20]. The aim of this study was to evaluate the ability of IMA to predict the AMI diagnosis as well as its NPV in a population of subjects presenting with chest pain to the emergency department.
onset of symptoms as accurately as possible (ie, description of chest pain, pulmonary edema, severe dyspnea, and syncope), a general patient history, clinical examination, 12-lead ECG, and the laboratory tests described below.
2. Material and methods 2.1. Patient specimens The study population comprised 107 subjects who were admitted with chest pain and suspected of ACS during a 4-month period (February to May 2007). The study population included 66 men (62%) and 41 women (38%). Patients were referred by a general practitioner or the ambulance service. They were subsequently admitted and included by the physician on duty. Patients admitted more than once during the investigation period were included only on their first admission. 12-Lead ECG was performed in all patients. None of the patients had ST-segment elevations that qualified for immediate angiography and primary percutaneous coronary intervention (PCI) or thrombolytic therapy.
2.2. Routine diagnostic procedures At the emergency department and in the coronary care unit (CCU), routine procedures comprise establishing time of
2.3. Electrocardiogram An ECG was taken on admission to the emergency department, on arrival at the CCU, and at least once daily thereafter. Patients with ECGs readily diagnostic for STsegment elevation MI went to primary PCI and were excluded. Electrocardiogram changes were classified as progressive, nonprogressive, or uncodable. Electrocardiogram changes were considered transient if new specific STsegment (ST depression or ST elevation N0.1 mV) and Twave changes would subside in later recordings. Changes were considered progressive if ST-segment changes progressed in 2 or more leads from earlier recordings. Nonprogressive changes comprise ST-segment and T-wave changes that did not change in appearance in serial recordings. Electrocardiograms were considered uncodable in the presence of left bundle branch block, pacemaker artifacts, or complete atrioventricular block.
2.4. Medical therapy All patients were transferred to the CCU with a standard regimen for patients suspected of having ACS. On arrival, 300 mg of aspirin was administered to all patients (thereafter 75 mg OD). If the ECG showed ST depressions, T-wave inversions, or other signs that increased the likelihood of ACS, the physician on duty administered low-molecularweight heparin (enoxaparin 1 mg/kg body weight BD) and clopidogrel (300 mg bolus, 75 mg OD). Otherwise, results of troponin T analysis were awaited. If ACS was diagnosed by increased cardiac troponin T (cTnT) or if the patient was otherwise considered unstable, treatment with low-molecular-weight heparin, clopidogrel, aspirin, and, in many cases, statins and β-blockers was instituted.
2.5. Cardiac biomarkers Ischemia-modified albumin was determined from serum with the ACB test (Inverness Medical Innovations Inc, Stirling, UK). The principle of the assay is described elsewhere [5]. Analyses were performed on a Cobas MIRA Plus instrument (Roche Diagnostics, Mannheim, Germany). The within-series coefficient of variation (CV) was 8.3%, and the between-day imprecision was 11%. A reference level was previously determined from a control reference population of 258 healthy blood donors. Analyses from the reference population were performed on a Hitachi 911 instrument (Roche Diagnostics, Mannheim, Germany). Blood samples were handled and analyses were performed according to the manufacturer's protocols.
172 Cardiac troponin T, CKMB mass (CKMBmass), and myoglobin were determined from serum at 37°C on Elecsys 2010 (Roche Diagnostics, Mannheim, Switzerland). For cTnT (fourth-generation assay), the lowest concentration exhibiting an imprecision CV of less than 10% was 0.03 μg/L [21]. The upper reference value for CKMBmass was 4.0 μg/L for women and 7.0 μg/L for men [22], and that for myoglobin was 51 μg/L for women and 72 μg/L for men [23]. Heart-type fatty acid binding protein (H-FABP) was analyzed from dipotassium ethylenediamine tetraacetic acid (K-EDTA) plasma using ELISA technique (HyCult Biotechnology, Uden, The Netherlands); upper reference value is 6.0 μg/L with a CV of 10% or less [23].
2.6. Blood sampling Blood sampling was scheduled according to the department's routine procedures, where blood samples for cTnT and CKMBmass were obtained on arrival, after 6 to 9 hours,
S. Hjortshøj et al. and after 12 to 24 hours. All patients had at least 2 blood samples taken to exclude/verify the AMI diagnosis. On arrival, patients also had blood samples drawn for the evaluation of electrolytes, renal parameters, and lipid metabolism. Blood samples for determination of IMA and H-FABP were obtained together with the above-mentioned samples. Blood samples for determination of cTnT, CKMBmass, and IMA were drawn into dry tubes without anticoagulants. Samples for determination of H-FABP were drawn into K-EDTA tubes. After coagulation, samples were centrifuged at 4000g for 10 min. The serum was stored at −80°C until analysis. The total processing time from obtainment of samples to storage did not exceed 2.5 hours as recommended by the manufacturer. Frozen samples were mixed after thawing and recentrifuged before analysis. Samples did not undergo repeat freeze-thaw cycles. To evaluate sensitivity of the assays, relative concentrations were calculated as measured concentration divided by the 99th percentile given a CV of 10% or less. For IMA, relative
Fig. 1 Control reference population. A, IMA in control reference population followed a normal distribution. B, F test showing normal distribution. C, IMA levels were not influenced by age. D, An inverse relationship was seen with albumin levels. However, this did not influence other results.
Ischemia modified albumin in acute coronary syndrome Table 1
Baseline characteristics and levels of biomarkers
No. of patients admitted Sex, male:female, n Age, median (range), y Medical history before admission, n (%) Angina pectoris MI Coronary artery bypass grafting (CABG) Diabetes mellitus Biochemical markers, median (range) Peak IMA (U/mL) Peak cTnT (μg/L) Peak CKMBmass (μg/L) Peak H-FABP (μg/L) Electrocardiogram, no. (%) Q waves Progressive ST-segment changes Nonprogressive ST-segment changes Normal Uncodable
Chest pain with AMI (cTnT N0.03 μg/L)
Chest pain without AMI Peak IMA N88.2 U/mL
Peak IMA b88.2 U/mL
35 23:12 62 (25-92)
34 20:14 60 (30-88)
38 27:11 63 (34-84)
35 (100) 8 (23) 4 (11) 10 (29)
9 (26) 12 (35) 3 (9) 3 (9)
7 6 1 6
96.3 (58.5-117) 0.52 (0.03-2.82) 38 (1.7-387) 48.1 (0.1-151.6)
97.0 (88.5-110) 0.001 (0-0.02) 17 (1-128) 1.6 (0.1-9)
75.5 (63-88) 0.002 (0-0.02) 18 (1-91) 1.35 (0.4-6.8)
3 9 6 9 8
4 (11) 0 (0) 8 (24) 19 (56) 3 (9)
3 (8) 0 (0) 7 (18) 23 (61) 5 (13)
(9) (26) (16) (26) (23)
concentrations were calculated as measured concentrations divided by 95th percentile.
2.7. Definition of acute MI Patients were classified as having acute MI (AMI) according to the new universal definition of AMI [1]. These included detection of rise and/or fall in cTnT above the 99th percentile of the upper reference limit (N 0.03 μg/L) together with signs indicative of the following: • • • •
173
Relevant clinical symptoms of ischemia New ST-segment changes in the ECG Development of new Q waves in the ECG Echocardiographic signs of new regional motion defects.
(18) (16) (3) (16)
2.8. Exclusion criteria Patients were excluded, if they had a documented MI within the last week before admission. Patients who were on hemodialysis or had jaundice were also excluded from the study.
2.9. Statistics Results from the control reference population were evaluated with the F-test method for a normal distribution fit. Comparison between groups was performed by using Fischer's exact test. Medians and ranges were used for descriptive purposes. Receiver operating characteristics (ROC) analysis was performed by the method of Hanley and McNeil. Wilcoxon rank sum test was used for comparison of patients who were positive and negative for cTnT. The level of significance chosen was .05. The statistical analyses were performed using STATA statistical software package (StataCorp, College Station, Tex).
2.10. Ethics The study was approved by the ethical committee of North Jutland and Viborg Counties. Patients gave informed consent.
3. Results 3.1. Reference population Fig. 2 IMA in patients with normal and increased cTnT on admission. Cutoff for cTnT is 0.03 μg/L. There was significant difference between the 2 groups (P b .0001).
The ACB test values for the reference population were normally distributed (Fig. 1). Values for the control
174
S. Hjortshøj et al.
reference population were 49 to 136.5 U/mL (mean, 74.1 U/mL; median, 73.5 U/mL). The age distribution was 19 to 65 years (mean, 43 years; median, 43.5 years). The assay exhibited a CV of 10% or less at both the upper 95th and 99th percentile (88.2 and 111.8 U/mL, respectively). The 95th percentile has been used in other studies [8], and 88.2 U/mL was therefore defined as upper limit of normal. Because of a slight inverse relationship between IMA and levels of HSA, correction for albumin level was performed as proposed by Lippi et al [24] [(individual serum albumin concentration/median albumin concentration of the population) × IMA value = corrected IMA value]. This did not, however, lead to significant changes at all, and results are therefore given without correction.
3.2. Chest pain patients
Fig. 3 ROC curve for the ability of admission IMA to predict a cTnT level of 0.03 μg/L or greater (from admission to 24 hours).
The baseline characteristics of enrolled patients are shown in Table 1. Thirty-three percent of the patients (n = 35) had a final diagnosis of AMI with a cTnT level greater than 0.03 μg/L. Fig. 2 shows levels of admission IMA in patients with cTnT level greater than 0.03 μg/L and 0.03 μg/L or less. The sensitivity of admission IMA for a final diagnosis of AMI was 0.86 (95% confidence interval [95% CI], 0.69-0.95) (Table 2). Specificity was 0.49 (95% CI, 0.360.60). The NPV was 0.88 (95% CI, 0.72-0.95). Figs. 3 and 4 show the ROC curves (ROC analysis). The area under the ROC curve for admission IMA was 0.73. The optimal cutoff threshold derived from the ROC analysis was determined to be 91 U/mL. This corresponds to approximately the 97.25th percentile in the control reference material. Sensitivity, specificity, and NPV for a cutoff of 91 U/mL are also shown in Table 2. Sensitivity was decreased whereas specificity increased at this cutoff. Table 2
For comparison of different biomarkers, Table 2 shows the sensitivity, specificity, NPV, and area under the ROC curve for IMA, CKMBmass, myoglobin, H-FABP, and ECG.
4. Discussion This study was intended to evaluate IMA in a population of chest pain patients. The main results show that the admission sample of IMA identified only 86% (sensitivity) of patients who were otherwise diagnosed as having AMI according to current recommendations from ESC/ACC [1]. Likewise, the test had a low specificity of 49% and an NPV of 88%. Also, the ROC curve (1 − specificity vs sensitivity) showed an area under the curve of 0.73 (Fig. 3). Compared
Overview of IMA diagnostic performance of admission samples in different markers
Marker Admission IMA cutoff 88.2 U/mL IMA cutoff 91 U/mL CKMBmass Myoglobin H-FABP IMA at 6 to 9 h or peak value IMA 6-9 h, 88.2 U/mL Peak IMA, 88 U/mL Combinations with ECG ECG on admission IMA + ECG CKMB + ECG Myoglobin + ECG H-FABP + ECG
Sensitivity
Specificity
NPV
ROC analysis (area under curve)
0.86 (0.69-0.95) 0.71 (0.53-0.85) 0.85 (0.67-0.94) 0.64 (0.48-0.78) 0.83 (0.65-0.93)
0.49 (0.36-0.60) 0.65 (0.53-0.75) 0.67 (0.81-0.96) 0.88 (0.77-0.94) 0.92 (0.83-0.97)
0.88 (0.72-0.95) 0.82 (0.69-0.91) 0.93 (0.84-0.97) 0.79 (0.68-0.88) 0.92 (0.83-0.97)
0.73 (0.62-0.83) — 0.93 (0.87-0.99) 0.90 (0.82-0.98) 0.91 (0.83-0.99)
0.66 (0.48-0.80) 0.86 (0.69-0.95)
0.65 (0.53-0.76) 0.51 (0.39-0.63)
0.80 (0.67-0.89) 0.88 (0.74-0.96)
0.73 (63-0.83) 0.78 (0.69-0.87)
0.31 (0.21-0.43) 0.91 (0.76-0.98) 0.94 (0.79-0.99) 0.91 (0.71-0.98) 0.89 (0.72-0.96)
0.49 (0.32-0.67) 0.46 (0.32-0.55) 0.32 (0.22-0.44) 0.41 (0.27-0.58) 0.67 (0.54-0.77)
0.27 (0.16-0.38) 0.91 (0.79-0.98) 0.92 (0.72-0.98) 0.90 (0.79-0.98) 0.92 (0.81-0.98)
0.60 (0.50-0.70) 0.67 (0.58-0.76) 0.81 (0.74-0.89) 0.77 (0.72-0.86) 0.78 (0.70-0.85)
Data shown with 95% confidence intervals in parentheses. AMI is defined by the presence of a cTnT level greater than 0.03 μg/L according to the ESC/ACC consensus report.
Ischemia modified albumin in acute coronary syndrome
175
Fig. 4 ROC curves for the ability of IMA to predict the following: A, AMI; B, CKMBmass; C, myoglobin; and D, and H-FABP. AMI is defined by the presence of cTnT level greater than 0.03 mg/L according to the ESC/ACC consensus report.
with all other early biomarkers of myocardial injury, IMA exhibited lower values for diagnostic performance (Table 2). There have been confusing results regarding IMA and its value in the diagnosis of ACS. Some studies have found the test of possible use, and in particular, a high NPV has attracted attention [8,13]. However, we found an NPV of only 0.88 (95% CI, 0.72-0.95), which means that one cannot safely rely on a negative IMA as a single test for excluding IMA. To explore further into this particular patient population, we determined an optimal cutoff value of 91 U/mL from the ROC analysis. This value is between the 95th and the 97.5th percentile of our control reference material. The use of 91 U/mL led to a decreased sensitivity (0.71; 95% CI, 0.530.85) and increased specificity (0.65; 95% CI, 0.53-0.75). The NPV at this cutoff was even lower (0.82; 95% CI, 0.690.91). This patient cohort was enrolled at a tertiary hospital. One has to assume that some form of selection has occurred and that the prevalence of cTnT-positive patients would be larger than in many other settings in the health service. When applying classic Bayesian thinking, the prevalence of AMI in the population where the ACB test is introduced becomes very important for the test's ability to distinguish between
diseased and nondiseased individuals. The mediocre sensitivity, specificity, and NPV of the test make the ACB test particularly vulnerable to changes in the prevalence of AMI. An example illustrating this can be calculated. In a setting of 1000 patients with an AMI prevalence of 33%, for example, a CCU, a sensitivity of 0.91 and a specificity of 0.46 for the ACB test (+ECG) will lead to a positive predictive value of 0.45 and an NPVof 0.91. The test will thus lead to 30 patients being false-negative, whereas 362 patients would be falsepositive. In another setting, for example, an emergency department, with a prevalence of AMI of 5%, the test would generate a positive predictive value of 8% and an NPV of 99%. This translates into only 4 patients being false-negative but an overwhelming 513 patients who are false-positive and thus need further testing. It is therefore not easy balancing the scales and deciding where to apply the ACB test. No consensus on exact cutoff levels for IMA has yet been achieved from the studies that have been performed. Some studies have used 80 U/mL [25], whereas others have found even lower values of 75 U/mL [8,12]. Other groups have used the manufacturer's recommendations of 85 U/mL [9-11]. Although the cardiology and biochemistry communities have
176 decided that the optimal cutoff for other cardiac biomarkers (cTnT and CKMBmass) should be at the 99th percentile with a CV of 10% or less, there seems to be no consensus on this important point regarding IMA. Although our data from the control reference population indicate that an acceptable CV% could be achieved at the 99th percentile, the ROC analysis clearly shows that a cutoff at this level would decrease the specificity of the assay to an unacceptable degree. The limitations of this study are, first and foremost, the size of the patient group. On the other hand, our control reference population is the largest published normal material to this date and therefore provides a solid basis. In this study, we test the diagnostic performance of IMA in the setting of possible ACS. This may be an unfair comparison because no standardized test apart from the ECG at present measures ischemia. However, as a proposed ischemia marker, IMA must be able to detect the severest forms of ischemic heart disease, namely, AMI. Several issues have slowed the progress of IMA as a standard marker in coronary care. First, the general concept of marker of ischemia and its possible flaws has been regarded with suspicion within the cardiology community. Second, complicated issues regarding handling and storage of samples have limited widespread use in automated laboratories and clinical trials [14]. In conclusion, in our population of chest pain patients, we do not find the data supportive of IMA as a standard marker in the emergency department.
Acknowledgments The staff at the Departments of Cardiology and Clinical Chemistry, Aalborg Hospital, Aarhus University Hospital, is greatly thanked for their help and support during sampling and analysis of samples. The study was financially supported by The Danish Heart Foundation and the County of North Jutland Research Foundation. Also, Inverness Medical Innovations, Inc, is thanked for providing assays and technical support.
References [1] Thygesen K, Alpert JS, White HD, et al. Universal definition of myocardial infarction. Circulation 2007;116(22):2634-53. [2] Ravkilde J. Risk stratification in acute coronary syndrome using cardiac troponin I. Clin Chem 2000;46(4):443-4. [3] Wallentin L, Lagerqvist B, Husted S, Kontny F, Stahle E, Swahn E. Outcome at 1 year after an invasive compared with a non-invasive strategy in unstable coronary-artery disease: the FRISC II invasive randomised trial. FRISC II Investigators. Fast Revascularisation During Instability in Coronary Artery Disease. Lancet 2000;356 (9223):9-16. [4] Bar-Or D, Curtis G, Rao N, Bampos N, Lau E. Characterization of the Co(2+) and Ni(2+) binding amino-acid residues of the N-terminus of human albumin. An insight into the mechanism of a new assay for myocardial ischemia. Eur J Biochem 2001;268(1):42-7. [5] Bar-Or D, Rael LT, Bar-Or R, et al. The cobalt-albumin binding assay: insights into its mode of action. Clin Chim Acta 2008;387(1-2):120-7.
S. Hjortshøj et al. [6] Abadie JM, Blassingame CL, Bankson DD. Albumin cobalt binding assay to rule out acute coronary syndrome. Ann Clin Lab Sci 2005;35 (1):66-72. [7] Bar-Or D, Lau E, Winkler JV. A novel assay for cobalt-albumin binding and its potential as a marker for myocardial ischemia—a preliminary report. J Emerg Med 2000;19(4):311-5. [8] Christenson RH, Duh SH, Sanhai WR, et al. Characteristics of an albumin cobalt binding test for assessment of acute coronary syndrome patients: a multicenter study. Clin Chem 2001;47(3):464-70. [9] Collinson PO, Gaze DC, Bainbridge K, et al. Utility of admission cardiac troponin and “ischemia modified albumin” measurements for rapid evaluation and rule out of suspected acute myocardial infarction in the emergency department. Emerg Med J 2006;23(4):256-61. [10] Roy D, Quiles J, Aldama G, et al. Ischemia modified albumin for the assessment of patients presenting to the emergency department with acute chest pain but normal or non-diagnostic 12-lead electrocardiograms and negative cardiac troponin T. Int J Cardiol 2004;97(2):297-301. [11] Sinha MK, Roy D, Gaze DC, Collinson PO, Kaski JC. Role of “ischemia modified albumin”, a new biochemical marker of myocardial ischaemia, in the early diagnosis of acute coronary syndromes. Emerg Med J 2004;21(1):29-34. [12] Wu AH, Morris DL, Fletcher DR, Apple FS, Christenson RH, Painter PC. Analysis of the albumin cobalt binding (ACB) test as an adjunct to cardiac troponin I for the early detection of acute myocardial infarction. Cardiovasc Toxicol 2001;1(2):147-51. [13] Peacock F, Morris DL, Anwaruddin S, et al. Meta-analysis of ischemia-modified albumin to rule out acute coronary syndromes in the emergency department. Am Heart J 2006;152(2):253-62. [14] Keating L, Benger JR, Beetham R, et al. The PRIMA study: presentation ischaemia-modified albumin in the emergency department. Emerg Med J 2006;23(10):764-8. [15] Can M, Demirtas S, Polat O, Yildiz A. Evaluation of effects of ischaemia on the albumin cobalt binding (ACB) assay in patients exposed to trauma. Emerg Med J 2006;23(7):537-9. [16] Falkensammer J, Stojakovic T, Huber K, et al. Serum levels of ischemiamodified albumin in healthy volunteers after exercise-induced calfmuscle ischemia. Clin Chem Lab Med 2007;45(4):535-40. [17] Roy D, Quiles J, Sharma R, et al. Ischemia-modified albumin concentrations in patients with peripheral vascular disease and exerciseinduced skeletal muscle ischemia. Clin Chem 2004;50(9):1656-60. [18] Roy D, Quiles J, Sinha M, et al. Effect of radiofrequency catheter ablation on the biochemical marker ischemia modified albumin. Am J Cardiol 2004;94(2):234-6. [19] Roy D, Quiles J, Sinha M, Aldama G, Gaze D, Kaski JC. Effect of direct-current cardioversion on ischemia-modified albumin levels in patients with atrial fibrillation. Am J Cardiol 2004;93(3):366-8. [20] Borderie D, Allanore Y, Meune C, Devaux JY, Ekindjian OG, Kahan A. High ischemia-modified albumin concentration reflects oxidative stress but not myocardial involvement in systemic sclerosis. Clin Chem 2004;50(11):2190-3. [21] Roche Diagnostics. Troponin T STAT (Short Turn Around Time), CARDIAC T 04660307 190 2007-07. 2007. [22] Apple FS, Quist HE, Doyle PJ, Otto AP, Murakami MM. Plasma 99th percentile reference limits for cardiac troponin and creatine kinase MB mass for use with European Society of Cardiology/American College of Cardiology consensus recommendations. Clin Chem 2003;49(8):1331-6. [23] Pelsers MM, Chapelle JP, Knapen M, et al. Influence of age and sex and day-to-day and within-day biological variation on plasma concentrations of fatty acid-binding protein and myoglobin in healthy subjects. Clin Chem 1999;45(3):441-3. [24] Lippi G, Montagnana M, Salvagno GL, Guidi GC. Standardization of ischemia-modified albumin testing: adjustment for serum albumin. Clin Chem Lab Med 2007;45(2):261-2. [25] Worster A, Devereaux PJ, Heels-Ansdell D, et al. Capability of ischemia-modified albumin to predict serious cardiac outcomes in the short term among patients with potential acute coronary syndrome. CMAJ 2005;172(13):1685-90.