Diagnostic accuracy of heart-type fatty acid–binding protein for the early diagnosis of acute myocardial infarction

American Journal of Emergency Medicine (2012) 30, 267–274

www.elsevier.com/locate/ajem

Original Contribution

Diagnostic accuracy of heart-type fatty acid–binding protein for the early diagnosis of acute myocardial infarction☆,☆☆ C. Geraldine McMahon MD, PhD a , John V. Lamont MSc b , Elizabeth Curtin MSc a , R. Ivan McConnell MSc b , Martin Crockard PhD b , Mary Jo Kurth PhD b,⁎, Peter Crean MD, FRCPI a , S. Peter Fitzgerald DSc b a

Emergency Department and Chest Pain Assessment Unit, St. James's Hospital, Dublin 8, Republic of Ireland Randox Laboratories Ltd., Crumlin BT29 4QY, Northern Ireland, UK

b

Received 15 October 2010; revised 8 November 2010; accepted 14 November 2010

Abstract Objective: The aim of this study was to evaluate the diagnostic efficacy of multiple tests—heart-type fatty acid–binding protein (H-FABP), cardiac troponin I (cTnI), creatine kinase-MB, and myoglobin— for the early detection of acute myocardial infarction among patients who present to the emergency department with chest pain. Methods: A total of 1128 patients provided a total of 2924 venous blood samples. Patients with chest pain were nonselected and treated according to hospital guidelines. Additional cardiac biomarkers were assayed simultaneously at serial time points using the Cardiac Array (Randox Laboratories Ltd, Crumlin, United Kingdom). Results: Heart-type fatty acid–binding protein had the greatest sensitivity at 0 to 3 hours (64.3%) and 3 to 6 hours (85.3%) after chest pain onset. The combination of cTnI measurement with H-FABP increased sensitivity to 71.4% at 3 to 6 hours and 88.2% at 3 to 6 hours. Receiver operating characteristic curves demonstrated that H-FABP had the greatest diagnostic ability with area under the curve at 0 to 3 hours of 0.841 and 3 to 6 hours of 0.894. The specificity was also high for the combination of H-FABP with cTnI at these time points. Heart-type fatty acid–binding protein had the highest negative predictive values of all the individual markers: 0 to 3 hours (93%) and 3 to 6 hours (97%). Again, the combined measurement of cTnI with H-FABP increased the negative predictive values to 94% at 0 to 3 hours, 98% at 3 to 6 hours, and 99% at 6 to 12 hours. Conclusion: Testing both H-FABP and cTnI using the Cardiac Array proved to be both a reliable diagnostic tool for the early diagnosis of myocardial infarction/acute coronary syndrome and also a valuable rule-out test for patients presenting at 3 to 6 hours after chest pain onset. © 2012 Elsevier Inc. All rights reserved.

☆ Ethical approval: this work was approved by the St James's/Adelaide and Meath Incorporating The National Children's Hospital (AMNCH) Research Ethics Committee. ☆☆ This study was supported by funding from Randox Laboratories Ltd, which manufactures the Cardiac Arrays and Evidence instrument. John Lamont, Ivan McConnell, and Drs Crockard, Kurth, and Fitzgerald are employed by Randox Laboratories Ltd. Elizabeth Curtin received grant support from Randox. ⁎ Corresponding author. Tel.: +44 0 28 9442 2413. E-mail address: [email protected] (M.J. Kurth).

0735-6757/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2010.11.022

268

1. Introduction Chest pain is one of the most common complaints in patients presenting to the emergency department (ED). In the United States, this is the second highest reason for admission [1]. Of those admitted, only 10% to 13% have a confirmed acute myocardial infarction (AMI) [1]. In England and Wales, approximately 700 000 present with chest pain to the ED, and two thirds of these are admitted [2]. There are significant challenges with early accurate diagnosis of patients presenting with chest pain of possible cardiac origin because none of the standard diagnostic tests have sufficient diagnostic use to accurately rule out underlying acute coronary syndrome (ACS) in the early stages [3]. Furthermore, there are associated logistic and financial burdens linked to the management of these patients. Inappropriate discharge is associated with a 5-fold increase in mortality and morbidity [4]. The diagnosis of ACS is presently determined by evaluating risk factors, electrocardiographic (ECG) traces, and measurement of cardiac markers. Current biochemical markers used in the diagnosis of ACS include cardiac troponins (cTn), creatine kinase-MB (CK-MB), and myoglobin (MYO). They are limited by either a lack of specificity or a delay of elevation of several hours after symptom onset. Therefore, their clinical use in early diagnosis of AMI is limited. Previous studies have established the success of heart-type fatty acid–binding protein (H-FABP) as an early biochemical marker in the detection of AMI and its use as a prognostic indicator of post-MI recovery [5-9]. Furthermore, both the European Society of Cardiology (ESC) and National Academy of Clinical Biochemistry [10] have suggested a multimarker approach for the detection of AMI. The Cardiac Array from Randox Laboratories Ltd (Crumlin, United Kingdom) contains the early markers H-FABP and MYO and the late markers cardiac troponin I (cTnI) and CK-MB on a ceramic biochip. The biochip is read using biochip array technology that allows simultaneous measurement of all 4 markers from a small sample volume (50 μL). The aim of the this study was to determine the diagnostic efficacy of H-FABP with the additional markers present on the Cardiac Array against standard diagnosis using American College of Cardiology (ACC)/ESC guidelines [11,12] for the early detection of AMI among patients who present to the ED with chest pain.

2. Materials and methods 2.1. Study population This study was undertaken in the ED of an urban university teaching hospital and was approved by St James's/Adelaide and Meath Incorporating the National Children's Hospital

C.G. McMahon et al. Research Ethics Committee. A total of 1128 patients who met the inclusion criteria were enrolled in the study after giving written informed consent. Patients were included in the study if they were 25 years or older and presented to the ED with chest pain of possible cardiac origin. Patients were excluded if they were younger than 25 years, had renal failure, or were not willing to provide informed consent.

2.2. Study design During the period April 2000 to June 2003, eligible patients arriving in the ED with chest pain considered to be of cardiac origin were enrolled in the study. All participants had a 12-lead ECG performed and were diagnosed and treated in accordance with ACC/ESC guidelines for AMI [11,12]. Participating patients were asked to give blood samples on admission then at 2-hour intervals up to 12 hours and at 24 and 48 hours after admission. In a questionnaire, additional information was sought, including the interval between pain onset and admission, medical history, risk factors, and diagnostic and treatment parameters.

2.3. Biochemical measurements Within the hospital, biochemical tests were performed on the patient blood samples, which included when required the cardiac marker cardiac troponin T (cTnT) (Roche electrochemiluminescence immunoassay) on Roche Elecsys 2010 analyzer (Roche, Basel, Switzerland). The analytical sensitivity quoted by Roche was 0.01 μg/L, with 0.03 μg/L of the concentration having a reproducible coefficient of variation (CV) of 10%. Typically, cTnT measurements were taken on admission, followed by repeat tests 6 and 12 hours after admission. Provision also existed for a further test at 48 hours, if this was deemed necessary. Aliquots of serum for analysis with the Cardiac Array were frozen and stored at −80°C.

2.4. Cardiac Array (biochip) Frozen, aliquoted samples were sent to Randox Laboratories Ltd for retrospective analysis on the Cardiac Array using the automated EVIDENCE analyzer (both from Randox Laboratories Ltd) [13,14]. Each biochip contains specific antibodies for the simultaneous, quantitative detection of CK-MB, cTnI, MYO, and H-FABP from a single 50 μL sample. The Randox cTnI assay has an analytical sensitivity of 0.01 μg/L and a reproducible CV of less than 20% at 0.14 μg/L, which Randox quoted as functional sensitivity. Diagnostic cutoffs were determined at the 99th percentile, with CVs of less than 10%, for all analytes from an apparently healthy control group (n = 80; 40 male and 40 female) with no known history of ischemic heart disease. The diagnostic cutoffs used in this study that complied

Use of H-FABP with cTnI for early diagnosis of AMI with the 99th percentile and less than 10% CVs were as follows: cTnI at 0.37 μg/L, CK-MB at 7.18 μg/L, H-FABP at 5.24 μg/L, and MYO at 95.57 μg/L.

2.5. Statistical analysis The sensitivity, specificity, and positive (PPV) and negative (NPV) predictive values were determined for the Cardiac Array markers. Receiver operator characteristic (ROC) curves were constructed to show the diagnostic ability of each marker on the Cardiac Array. All diagnostic indices were determined for each Cardiac Array marker at the following time intervals from onset of chest pain: (a) 0 to 3 hours, (b) 3 to 6 hours, (c) 6 to 12 hours, (d) 12 to 24 hours, (e) 24 to 48 hours, and (f) more than 48 hours. These time points were selected to reflect the release kinetics of the different cardiac markers [15] to determine the most appropriate test window for correct detection of AMI. All statistical procedures were performed using the Statistical Package for the Social Sciences statistical software (SPSS, UK Ltd, Woking, Surrey) version 13.0 for Windows. The results throughout the text, tables, and figures are presented as mean ± SD, unless otherwise stated, and the statistical significance was defined as P b .5.

3. Results All study patients had their samples run on the Cardiac Array as described. Acute MI was confirmed using the predefined criteria [11,12] in 117 (10.4%) of the 1128 patients enrolled in the study. Patient demographics are displayed in Table 1. Patients diagnosed with AMI tended to be older than non–AMI-diagnosed patients. Of the patients with AMI, only 40.2% presented early enough to receive optimum treatment benefit. Twenty-three patients with AMI (19.7%) presented between 6 and 12 hours, and 14 (12%) presented between 12 and 24 hours. Delays of more than 24 hours in presentation were seen in 28.2% of cases with 16 patients (13.7%) presented between 24 and 48 hours, and 17 patients (14.5%) presented more than 48 hours after pain onset. The sensitivity of cTnI was low at the early time points of 0 to 3 hours (50.0%) and 3 to 6 hours (67.6%). Heart-type fatty acid–binding protein had the greatest sensitivity at 0 to 3 hours (64.3%), 3 to 6 hours (85.3%), and 6 to 12 hours (89.9%) after chest pain onset (an average superiority of 13.6% over Tn) (Table 2). The combination of cTnI measurement with H-FABP increased this sensitivity to 71.4% at 0 to 3 hours, 88.2% at 3 to 6 hours, and 92.4% at 6 to 12 hours demonstrating an increase in sensitivity of 20.6% for the combination marker approach at 3 to 6 hours. Troponin, on the other hand, demonstrated a higher specificity than H-FABP at 0 to 3 hours (93.3% compared with 84.2%) and at 3 to 6 hours (94.3% compared with

269 Table 1

Patient demographics AMIs (117 pts; 10.4%)

Non-AMIs (1011 pts; 89.6%)

Male 82 (70.1) 674 (66.6%) Female 35 (29.9) 337 (33.4) n 117 (10.4) 1011 (89.6) Median age and (SD) All patients 64 (13.7) 52 (17.6) Male (y) 62 (13.4) 50 (17.9) Female (y) 71 (13.8) 55 (16.1) Hypertension 51 (43.6) 339 (33.5) Hyperlipidemia 26 (22.2) 226 (22.4) CHD 35 (29.9) 261 (25.8) Fam Hx CHD 53 (45.3) 498 (49.3) Smoking 56 (47.9) 425 (42.0) Diabetes 12 (10.3) 75 (7.4) Obesity 41 (35) 284 (28.1) Diagnosis NSTEMI: 59 (50.4) NC: 618 (61.1) STEMI: 58 (49.6) UA: 169 (16.7) SA: 160 (15.8) Arrh: 18 (1.8) HF: 5 (0.5) Other: 41 (4.1) Total number of patients, n = 1128 (male, 756 [67%]; female, 372 [33%]). Values are given as n (%), unless otherwise indicated. pts indicates patients; y, years; CHD, coronary artery disease; Fam Hx, family history; STEMI, ST-elevation MI; NSTEMI, non–ST-elevation MI; UA, unstable angina; SA, stable angina; Arrh, arrhythmia; HF, heart failure; NC, non–cardiac origin chest pain; other, other conditions.

88.7% for H-FABP) (Table 2). At 6 to 12 hours, the specificity of H-FABP was comparable with that of cTnI (93.5% for H-FABP compared with 94.2% for cTnI). The specificity of cTnI was not increased by the additional measurement of H-FABP; however, the specificity of the dual measurements was still high (Table 2). Additional measurement of a third or fourth marker did not improve sensitivity or specificity. The best PPV (67%-90%) was achieved by CK-MB followed by cTnI, MYO, and H-FABP (Table 3). Although the PPV was lower for H-FABP than cTnI at 3 to 6 hours (55% vs 66%, respectively), the values were equal at the 6- to 12-hour period. The PPV for all markers rose between 0 and 12 hours and continued to rise for cTnI until 24 to 48 hours. Measurement of multiple markers did not improve PPV. Heart-type fatty acid–binding protein as a single marker had the highest NPV between 0 and 12 hours after chest pain onset (93%-98%), closely followed by cTnI (92%-97%). In contrast to the PPV, the combined measurement of cTnI with H-FABP increased the NPV to 94% at 0 to 3 hours, 98% at 3 to 6 hours, and 99% at 6 to 12 hours. The diagnostic ability of each marker on the Cardiac Array for each time point was examined. From the ROC curves, the large area under the curve (AUC) for H-FABP demonstrated that it had the highest accuracy of

270 Table 2

C.G. McMahon et al. Sensitivity and specificity of Cardiac Array markers on EVIDENCE

Time, postpain

% Sensitivity

% Specificity

0-3 h 3-6 h 6-12 h 12-24 h 24-48 h N48 h 0-3 h 3-6 h 6-12 h 12-24 h 24-48 h N48 h Individual markers CK-MB MYO H-FABP cTnI 2-Marker combinations CK-MB + MYO CK-MB + H-FABP CK-MB + cTnI MYO + H-FABP MYO + cTnI H-FABP + cTnI 3-Marker combinations CK-MB + MYO + H-FABP CK-MB + MYO + cTnI CK-MB + H-FABP + cTnI MYO + H-FABP + cTnI 4-Marker combinations CK-MB + MYO + H-FABP + cTnI

39.3 39.3 64.3 50.0

58.8 61.8 85.3 67.6

75.9 65.8 89.9 81.0

87.3 50.7 90.1 95.8

86.7 38.7 62.7 97.3

50.0 21.4 65.5 88.1

95.8 95.8 84.2 93.3

96.2 93.9 88.7 94.3

98.1 96.3 93.5 94.2

97.9 95.8 91.4 94.3

97.9 95.9 90.9 94.7

99.4 96.9 91.0 94.3

53.6 71.4 53.6 64.3 57.1 71.4

76.5 85.3 70.6 85.3 79.4 88.2

87.3 92.4 82.3 89.9 88.6 92.4

91.5 95.8 97.2 90.1 98.6 98.6

86.7 88.0 100 62.7 100 100

53.6 73.8 88.1 65.5 88.1 88.1

93.3 83.6 92.1 84.2 90.3 81.2

92.0 87.7 92.5 87.7 90.1 86.8

96.0 93.5 93.8 93.5 91.9 89.6

94.6 90.5 94.0 91.4 91.1 87.2

94.7 90.0 94.4 90.9 91.4 87.3

96.7 90.9 94.1 90.8 92.0 86.9

71.4 60.7 75.0 71.4

85.3 79.4 88.2 88.2

92.4 88.6 92.4 92.4

95.8 98.6 98.6 98.6

88.0 100 100 100

73.8 88.1 88.1 88.1

83.6 89.7 80.6 81.2

86.8 89.2 85.8 85.8

93.5 91.9 89.6 89.6

90.5 90.8 86.9 87.2

90.0 91.2 87.0 87.3

90.7 92.0 86.9 86.8

75.0

88.2

92.4

98.6

100

88.1

80.6

84.9

89.6

86.9

87.0

86.8

all markers at 0 to 12 hours after pain (Fig. 1 and Table 4). Creatine kinase-MB values were the next highest up to more than 6 hours, after which cTnI values started to increase. Myoglobin had the lowest values, and the pattern overall reflected release kinetics. The AUC values for these

Table 3

4 markers were statistically significant at all time points. Interestingly, the combination of H-FABP and cTnI demonstrated the best diagnostic accuracy by giving the largest AUC over each individual marker up to and including the 24- to 48-hour period (Fig. 1 and Table 4).

PPV and NPV in percent for cardiac markers for MI diagnosis

Time, post pain

% PPV

% NPV

0-3 h 3-6 h 6-12 h 12-24 h 24-48 h N48 h 0-3 h 3-6 h 6-12 h 12-24 h 24-48 h N48 h Individual markers CK-MB MYO H-FABP cTnI 2-Marker combinations CK-MB + MYO CK-MB + H-FABP CK-MB + cTnI MYO + H-FABP MYO + cTnI H-FABP + cTnI 3-Marker combinations CK-MB + MYO + H-FABP CK-MB + MYO + cTnI CK-MB + H-FABP +cTnI MYO + H-FABP + cTnI 4-Marker combinations CK-MB + MYO + H-FABP + cTnI

67 67 43 60

71 62 55 66

87 74 70 70

77 68 56 71

90 67 62 79

89 41 43 61

90 90 93 92

94 94 97 95

96 94 98 97

98 94 99 99

97 88 92 99

95 92 96 99

62 45 57 43 53 41

60 53 60 53 56 52

78 70 68 70 64 59

72 55 67 56 65 51

78 67 79 62 71 64

63 45 60 42 53 41

92 95 92 93 93 94

96 97 95 97 96 98

98 99 97 98 98 99

99 99 100 99 100 100

97 97 100 92 100 100

95 97 99 96 99 99

45 53 42 41

51 54 50 50

70 64 59 59

55 64 51 51

67 71 63 64

45 53 41 40

95 93 95 94

97 96 98 98

99 98 99 99

99 100 100 100

97 100 100 100

97 99 99 99

42

48

59

51

63

40

95

98

99

100

100

99

Use of H-FABP with cTnI for early diagnosis of AMI ROC Curve

1.0

ROC Curve

1.0

0.8

0.8

0.6

0.6

Sensitivity

Sensitivity

271

0.4

0.4

0-3 hrs after Chest Pain

3-6 hrs after Chest Pain

Combination H-FABP and cTnl (AUC=0.868) H-FABP (AUC=0.841)

0.2

CKMB (AUC=0.809)

CKMB (AUC=0.855)

cTnl (AUC=0.762)

MYO (AUC=0.853)

MYO (AUC=0.721)

cTnl (AUC=0.851)

Reference Line

Reference Line

0.0

0.0 0.0

0.2

0.4 0.6 1 - Specificity

0.8

1.0

0.0

ROC Curve

1.0

0.2

0.8

0.8

0.6

0.6

0.4

6-12 hrs after Chest Pain

1.0

0.4 Combination H-FABP and cTnl (AUC=0.979) cTnl (AUC=0.978)

0.2

CKMB (AUC=0.937)

CKMB (AUC=0.976)

cTnl (AUC=0.904)

H-FABP (AUC=0.966)

MYO (AUC=0.896)

MYO (AUC=0.895)

Reference Line

Reference Line

0.0

0.0 0.0

0.2

0.4 0.6 1 - Specificity

0.8

0.0

1.0

ROC Curve

0.2

0.4 0.6 1 - Specificity

0.8

1.0

ROC Curve

1.0

1.0

0.8

0.8

0.6

Sensitivity

Sensitivity

0.8

12-24 hrs after Chest Pain

Combination H-FABP and cTnl (AUC=0.947) H-FABP (AUC=0.938)

0.2

0.4 0.6 1 - Specificity ROC Curve

1.0

Sensitivity

Sensitivity

Combination H-FABP and cTnl (AUC=0.915) H-FABP (AUC=0.894)

0.2

0.4

0.6

0.4

24-48 hrs after Chest Pain Combination H-FABP and cTnl (AUC=0.990) CKMB (AUC=0.987)

0.2

> 48 hrs after Chest Pain cTnl (AUC=0.944) Combination H-FABP and cTnl (AUC=0.937) CKMB (AUC=0.895

0.2

cTnl (AUC=0.976) H-FABP (AUC=0.909)

H-FABP (AUC=0.869)

MYO (AUC=0.760)

MYO (AUC=0.752)

Reference Line

Reference Line

0.0

0.0 0.0

0.2

0.4 0.6 1 - Specificity

0.8

1.0

0.0

0.2

0.4 0.6 1 - Specificity

0.8

1.0

Fig. 1 The ROC curve analysis for H-FABP, cTnI, MYO, CK-MB, and the combination of H-FABP and cTnI for patients presenting within the different periods.

4. Discussion Early accurate diagnosis of ACS in patients presenting to ED with symptoms suggestive of underlying ACS is important in optimizing the care of this large patient cohort. Furthermore, correct identification of patients without AMI is beneficial both for the patient and also financially and logistically for the hospital. Although several diagnostic markers are available,

including Tn, CK-MB, and MYO with relatively high specificities, their reduced sensitivities in the first 6 hours after symptom onset preclude their diagnostic use in the early diagnosis of ACS. Of the 4 biomarkers measured in this study, H-FABP demonstrated the highest sensitivity at the early time points (64.3% at 0-3 hours and 85.3% at 3-6 hours). This sensitivity may be explained by the high concentration of HFABP in the myocardium compared with other tissues; the

272

C.G. McMahon et al.

Table 4 Areas under the ROC curves for cardiac markers split over time brackets (post–pain onset) Test 0-3 h H-FABP cTnI MYO CK-MB Combination of H-FABP and cTnI 3-6 h H-FABP cTnI MYO CK-MB Combination of H-FABP and cTnI 6-12 h H-FABP cTnI MYO CK-MB Combination of H-FABP and cTnI 12-24 h H-FABP cTnI MYO CK-MB Combination of H-FABP and cTnI 24-48 h H-FABP cTnI MYO CK-MB Combination of H-FABP and cTnI N48 h H-FABP cTnI MYO CK-MB Combination of H-FABP and cTnI

Area

P

95% CI

0.841 0.762 0.721 0.809 0.868

b.0001 b.0001 .0004 b.0001 b.0001

0.757-0.924 0.674-0.877 0.591-0.850 0.716-0.902 0.793-0.943

0.894 0.851 0.853 0.855 0.915

b.0001 b.0001 b.0001 b.0001 b.0001

0.815-0.972 0.762-0.940 0.775-0.931 0.761-0.949 0.845-0.984

0.938 0.904 0.896 0.937 0.947

b.0001 b.0001 b.0001 b.0001 b.0001

0.903-0.974 0.854-0.953 0.849-0.943 0.896-0.978 0.915-0.980

0.966 0.978 0.895 0.976 0.979

b.0001 b.0001 b.0001 b.0001 b.001

0.953-0.980 0.959-0.996 0.853-0.936 0.962-0.990 0.969-0.989

0.909 0.944 0.760 0.987 0.990

b.0001 b.0001 b.0001 b.0001 b.0001

0.873-0.944 0.952-1.000 0.697-0.824 0.978-0.996 0.982-0.998

0.869 0.944 0.752 0.895 0.937

b.0001 b.0001 b.0001 b.0001 b.0001

0.827-0.910 0.908-0.981 0.689-0.816 0.851-0.939 0.903-0.972

stability and solubility of H-FABP; its low molecular weight, that is, 15 kDa compared with 18, 80, and 37 kDa for MYO, CK-MB, and cTnT, respectively [16-18]; its rapid release into plasma after myocardial injury, that is, 60 minutes after an ischemic episode [19]; and its relative tissue specificity [20]. Myoglobin is also released early; however, the damage of cardiac vs skeletal muscles cannot be easily distinguished. The myocardial tissue content of H-FABP (0.57 mg/g wet weight) compared with that of MYO (2.7 mg/g wet weight) is 4- to 5fold lower, and the plasma reference concentration is 19-fold lower than that of MYO. This almost 5-fold steeper tissue-toplasma gradient for H-FABP than for MYO results in H-FABP

in plasma rising above its upper reference concentration at an earlier point after AMI onset, which enables an earlier diagnosis of AMI [16,19,21]. The area under the ROC curves demonstrated that H-FABP exhibited the highest diagnostic accuracy of all the markers at 0 to 12 hours post pain, reflecting its early release kinetics [15,22]. The H-FABP cutoff used in this study (5.24 μg/L calculated from 80 healthy individuals) correlates with that obtained by Bathia et al [23], who used samples from 242 patients to obtain cutoff values of 5.3 and 5.8 μg/L for women and men, respectively. Other publications report similar cutoff values for H-FABP [24-28]. Heart-type fatty acid–binding protein demonstrated the highest NPV for AMI detection of the individual markers, up to 12 hours after pain onset, showing it to be an effective ruleout MI marker. The H-FABP results obtained here correlate closely with those of Das et al [29], Glatz et al [5], Kilcullen et al [8], and Viswanathan et al [9], who all found H-FABP to be a valuable marker and recommend it for ACS risk stratification. McCann et al [27] also demonstrated that the sensitivity of H-FABP was superior to that of TnT in patients presenting within 4 hours of symptom onset. Interestingly, Sypniewska et al [28] reported 90.5% sensitivity for H-FABP in all ACS cases. This higher sensitivity may be explained by a very low sample number of 42 patients. Although the results of this study demonstrated that H-FABP had the best sensitivity of all the markers on the Cardiac Array (an average improvement of 13.6% over Tn over the first 3 periods), the specificity of cTnI was always higher than H-FABP even at the early time points. The specificity of H-FABP was improved by combining H-FABP with cTnI measurements. This combination also had increased sensitivity and NPV than each of these 2 markers measured singly at 0 to 3, 3 to 6, and 6 to 12 hours. The combined measurement of cTnI with H-FABP had an NPV of 98% at 3 to 6 hours after chest pain onset. This increased with time to 99% at 6 to 12 hours and 100% thereafter. The combination of H-FABP with cTnI had an AUC of 0.915 at 3 to 6 hours compared with 0.894 for H-FABP and 0.851 for cTnI. This combination demonstrated increasing diagnostic accuracy over H-FABP and cTnI measured singly up to 24 to 48 hours. These results collectively indicate that this combination of markers can be used effectively as a rule-out test to identify those not having AMI at the early time point of 3 to 6 hours after chest pain onset. McCann et al [27] also found that this combination of markers provided a significant improvement in sensitivity for patients presenting within 4 and 12 hours of symptoms. Li et al [30] reported that H-FABP and cTnT are currently the most effective combination for the diagnosis of early AMI within 6 hours of chest pain onset. Interestingly, 40.2% of patients with MI presented within 6 hours of pain onset, which may be too early for a robust cTn assay. This study found that 88.2% of patients would have been correctly diagnosed within 6 hours with the multimarker combination H-FABP and cTnI, representing

Use of H-FABP with cTnI for early diagnosis of AMI a 20% increase compared to the use of cTnI alone. Measuring the combination of H-FABP with cTnI also appears to identify at-risk patients with ACS more effectively than a single cTnI assay. Of the patients without MI identified as positive using the Cardiac Array, 60% were diagnosed as having unstable angina and were clearly patients with ACS rather than negative patients with AMI, as indicated by current diagnostic criteria that use a binary diagnosis cTn test. Elevated H-FABP but negative Tn on admission has been demonstrated to be independently associated with cardiac events in both 6- and 12-month follow-up studies ([31] and [9], respectively). The AMI diagnosis in this study was based on strict criteria described in the ACC/ESC guidelines and did not address ACS conditions and early cardiac damage.

4.1. Limitations of the study The Tn measurements reported in the current study were made using the cTnI on the Cardiac Array and were not compared with another cTn test. The reason for this was that at the time of the study, although patients were diagnosed and treated in accordance with ACC/ESC guidelines 2000 for AMI [11,12], which required a clear peak in cTn or CK-MB and ischemic symptoms, ECG changes (pathologic Q waves, ST elevation, or depression), or coronary artery intervention (angioplasty), TnT levels were not recorded for each patient. However, the cTnI on the Cardiac Array meets 2007 ACC/ESC guidelines of less than 10% at the 99th percentile [32]. In response to these new guidelines, many manufacturers have released new cTn products that now meet the guideline criteria; some of which are described as high-sensitivity assays. In light of the current study, it would be worthwhile comparing the sensitivity of a high-sensitivity Tn assay against that of H-FABP and also against the combination of H-FABP with cTnI on the Cardiac Array. Interestingly, Viswanathan et al [9] have already demonstrated that the prognostic value of elevated H-FABP is additive to cTn (TnI-Ultra) in low- and intermediate-risk patients with suspected ACS and that it is a marker of ischemia even in the absence of necrosis. Although cTn is considered the criterion standard for the diagnosis of AMI, and although serum levels are not raised until 6 to 9 hours after the start of necrosis, there is a lack of standardization among cTn assays. This is currently being addressed by the International Federation of Clinical Chemistry and Laboratory Medicine through their pilot study on standardization of cTnI assays. It has been suggested that renal function might affect serum H-FABP levels because low-molecular-weight proteins such as H-FABP are cleared mostly by the kidneys [33,34], and this clearance would be reduced in patients with renal failure. However, this is also a limitation for cTn. False positives may also occur because of muscle injury, because H-FABP is present in both heart and skeletal

273 muscles, although the concentration of H-FABP is several folds higher in the heart compared with muscles [19]. False positives are also experienced with cTn, and it has been reported that up to 30% of patients presenting with raised cTn do not have typical ACS [35].

5. Conclusion The current study demonstrated that the combination measurement of H-FABP with cTnI using the Cardiac Array was a reliable method for the diagnosis of AMI within 3 to 6 hours of chest pain onset. The combination marker approach provides earlier AMI diagnostic capability, which should in turn be associated with earlier optimum treatment and better survival. Furthermore, the 98% NPV at this time point may improve the accuracy of discharge decisions in this patient cohort benefiting the patient and the ED both logistically and financially.

References [1] Pope HJ, Aufderheide TP, Ruthazer R, et al. Missed diagnoses of acute cardiac ischemia in the emergency department. NEJM 2000; 342(16):1163-70. [2] Goodacre S, Cross E, Arnold J, et al. The health care burden of acute chest pain. Heart 2005;91:229-30. [3] Goldmann L, Kirtane A. Triage of patients with acute chest pain and possible cardiac ischemia: the elusive search for diagnostic perfection. Ann Intern Med 2003;139:987-95. [4] Collinson PO, Premachandram S, Hasehmi K. Prospective audit of incidence of prognostically important myocardial damage in patients discharged from emergency department. BMJ 2000;320: 1702-5. [5] Glatz J, van der Vusse G, Simoons M, et al. Fatty acid–binding protein and the early detection of acute myocardial infarction. Clinica Chimica Acta 1998;272:87-92. [6] O'Donoghue M, deLemos J, Morrow D, et al. Prognostic utility of heart-type FABP in patients with acute coronary syndromes. Circulation 2006;114:550-7. [7] Tanaka T, Hirota Y, Sohmiya K, et al. Serum and urinary human h-FABP in acute myocardial infarction. Clin Biochem 1991;24: 195-201. [8] Kilcullen N, Viswanathan K, Das R, et al. Heart-type fatty acid binding protein predicts long-term mortality after acute coronary syndrome and identifies high-risk patients across the range of troponin values. J Am College Cardiology 2007;50(21):2061-7. [9] Viswanathan K, Kilcullen N, Morrell C, et al. Hearty-type fatty-acid binding-protein (H-FABP) predicts long-term mortality and reinfarction in consecutive patients with suspected acute coronary syndrome who are troponin negative. J Am Coll Cardiol 2010;55(23): 2590-8. [10] Morrow DA, Cannon CP, Jesse RL, et al. National Academy of Clinical Biochemistry Laboratory Medicine Practice Guidelines: clinical characteristics and utilization of biochemical markers in acute coronary syndromes. Clin Chem 2007;53:552-74. [11] Alpert JS, Thygesen K, Antman E, et al. Myocardial infarction redefined—a consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction. J Am Coll Cardiol 2000;36(3): 959-69.

274 [12] Myocardial infarction redefined—a consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction. Eur Heart J 2000;21:1502-13. [13] FitzGerald SP, Lamont JV, Mc Connell RI, et al. Development of a high-throughput automated analyzer using biochip array technology. Clin Chem 2005;51(7):1165-76. [14] Molloy RM, Mc Connell RI, Lamont JV, et al. Automation of biochip array technology for quality results. Clin Chem Lab Med 2005;43(12): 1303-13. [15] Chan CP, Sanderson JE, Glatz JF, et al. A superior early myocardial infarction marker. Human heart-type fatty acid–binding protein. Z Kardiol 2004;93(3):388-97. [16] Glatz JF, Kleine AH, van Nieuwenhoven FA, et al. Fatty acid–binding protein as a plasma marker for the estimation of myocardial infarct size in humans. Br Heart J 1994;71:135-40. [17] Wodzig KW, Kragten JA, Hermens WT, et al. Estimation of myocardial infarct size from plasma myoglobin or fatty acid–binding protein. Influence of renal function. Eur J Clin Chem Clin Biochem 1997;35(3):191-8. [18] Michielsen EC, Diris JH, Kleijnen VW, Wodzig WK, et al. Interpretation of cardiac troponin T behaviour in sizeexclusion chromatography. Clin Chem Lab Med 2006;44(12): 1422-7. [19] Van Nieuwenhoven FA, Kleine AH, Wodzig WH, et al. Discrimination between myocardial and skeletal muscle injury by assessment of the plasma ratio of myoglobin over fatty acid binding protein. Circulation 1995;92:2848-54. [20] Alhadi HA, Fox KA. Do we need additional markers of myocyte necrosis: the potential value of heart fatty-acid–binding protein? QJM 2004;97(4):187-98. [21] Okamoto F, Sohmiya K, Ohkaru Y, et al. Human heart-type cytoplasmic fatty acid–binding protein (H-FABP) for the diagnosis of acute myocardial infarction. Clinical evaluation of H-FABP in comparison with myoglobin and creatine kinase isoenzyme MB. Clin Chem Lab Med 2000;38:231-8. [22] Kleine AH, Glatz JF, van Nieuwenhoven FA, et al. Release of heart fatty acid–binding protein into plasma after acute myocardial infarction in man. Mol Cell Biochem 1992;116:155-62.

C.G. McMahon et al. [23] Bathia DP, Carless DR, Viswanathan K, et al. Serum 99th centile values for two heart-type fatty acid binding protein assays. Ann Clin Biochem 2009;46(6):464-7. [24] Mion MM, Novello E, Altinier S, et al. Analytical and clinical performance of a fully automated cardiac multi-markers strategy based on protein biochip microarray technology. Clin Biochem 2007;40(16-17):1245-51. [25] Zaninotto M, Mion MM, Novello E, et al. Analytical and clinical evaluation of a new heart-type fatty acid–binding protein automated assay. Clin Chem Lab Med 2006;44(11):1383-5. [26] McCann CJ, Glover BM, Menown IBA, et al. Novel biomarkers in early diagnosis of acute myocardial infarction compared with cardiac troponin T. Eur Heart J 2008;29(23):2843-50. [27] McCann CJ, Glover BM, Menown IB, et al. Prognostic value of a multimarker approach for patients presenting to hospital with acute chest pain. Am J Cardiol 2009;103(1):22-8. [28] Sypniewska G, Sawicki M, Krintus M, et al. The use of biochip cardiac array technology for early diagnosis of acute coronary syndromes. JMB 2009;28(4):293-9. [29] Das R, Kilcullen N, Morrell C, et al. The British Cardiac Society Working Group definition of myocardial infarction: implications for practice. Heart 2006;92(1):21-6. [30] Li CJ, Li JQ, Liang XF, et al. Point-of-care test of heart-type fatty acid–binding protein for the diagnosis of early acute myocardial infarction. Acta Pharmacol Sin 2010;31(3):307-12. [31] Ishii J, Ozaki Y, Lu J, et al. Prognostic value of serum concentration of heart-type fatty acid–binding protein relative to cardiac troponin T on admission in the early hours of acute coronary syndrome. Clin Chem 2005;51(8):1397-404. [32] Thygesen K, Alpert JS, White HD. Joint ESC/ACCF/AHA/WHF Task Force for the redefinition of myocardial infarction. J Am Coll Cardiol 2007;50(22):2173-95. [33] Hall CL, Hardwicke J. Low molecular weight proteinuria. Annu Rev Med 1979;30:199-211. [34] Gorski J, Hermens WT, Borawski J, et al. Increased fatty acid–binding protein concentration in plasma of patients with chronic renal failure. Clin Chem 1997;43(1):193-5. [35] Collinson PO, Stubbs PJ. Are troponins confusing? Heart 2003;89: 1285-7.