- Email: [email protected]
Pre-hospital detection of acute myocardial infarction with ultra-rapid human fatty acid-binding protein (H-FABP) immunoassay
International Journal of Cardiology 119 (2007) 349 – 354 www.elsevier.com/locate/ijcard
Pre-hospital detection of acute myocardial infarction with ultra-rapid human fatty acid-binding protein (H-FABP) immunoassay Patrick Ecollan a , Jean-Philippe Collet b , Guillaume Boon a , Marie-Laure Tanguy c , Marie-Laurence Fievet a , Rosita Haas a , Nicolas Bertho a , Shidash Siami a , Jean-Christophe Hubert a , Pierre Coriat a , Gilles Montalescot b,⁎ a SAMU, Pitié-Salpêtrière University Hospital, Paris, France Institut de Cardiologie, Pitié-Salpêtrière University Hospital, Paris, France Department of Biostatistics, Pitié-Salpêtrière University Hospital, Paris, France b
c
Received 4 January 2006; received in revised form 8 September 2006; accepted 13 September 2006 Available online 13 November 2006
Abstract Aims: To evaluate the diagnostic performance of a portable semi-quantitative whole blood immunoassay measuring soluble human fatty acidbinding protein [H-FABP] (CardioDetect®) for the pre-hospital detection of myocardial infarction (MI). Methods and results: We enrolled prospectively 108 consecutive patients with acute ischemic type chest pain in whom the first medical care was delivered by a mobile intensive care unit (MICU). CTnI, myoglobin, CK-MB and CardioDetect® were first assessed in the MICU before hospital admission using point-of-care assays and then cTnI was serially measured during the hospital stay. MI was defined as a positive cTnI in any sample over the first 24 h. The vast majority of the patients (77/108) were admitted b 3 h of symptoms onset. Pre-hospital cTnI sample was normal in 96 patients (88.9%) of whom 43 had subsequent cTnI elevation. A positive H-FABP using the CardioDetect® assay had a significantly better sensitivity than cTnI, myoglobin and CK-MB (87.3% vs 21.8%, 64.2% and 41.5%, respectively) to identify MI. The significant better sensitivity of the CardioDetect® assay was also observed among patients who presented b 3 h of symptom onset. The specificity of the CardioDetect® assay was significantly better than that of myoglobin, irrespective of the time delay from symptom onset to measurement. In patients with normal pre-hospital cTnI and no STelevation (n = 63), assessment based only on a positive H-FABP had 83.3% sensitivity, 93.3% specificity for the diagnosis of an evolving MI. Conclusion: Early assessment of H-FABP in patients presenting with chest pain improves the diagnosis of ongoing MI. © 2006 Published by Elsevier Ireland Ltd. Keywords: Myocardial infarction; Coronary disease; Biomarkers
1. Introduction Early detection of acute myocardial infarction (AMI) remains suboptimal. Although quite specific [1], STelevation has only a 50 to 60% sensitivity for diagnosis of ⁎ Corresponding author. Institut de Cardiologie, Hôpital Pitié-Salpêtrière (AP-HP) Bureau 2-236, 47, boulevard de l'Hôpital, 75013 Paris, France. Tel.: +33 1 42 16 30 06; fax: +33 1 42 16 29 31. E-mail address: [email protected] (G. Montalescot). 0167-5273/$ - see front matter © 2006 Published by Elsevier Ireland Ltd. doi:10.1016/j.ijcard.2006.09.003
AMI [2]. Cardiac troponin I (cTnI) does not reliably exclude AMI without repeated negative measurements over 4–6 h and myoglobin is limited by poor specificity [1]. This is an important issue given the fact that 40% of out-of-hospital patients with an evolving AMI are getting medical care within the first 3 h [3]. There is a need for a fast and reliable test to facilitate triage, diagnosis and adequate treatment strategies. This is particularly important in patients presenting with atypical symptoms and/or non-contributive ECG. The human heart-type fatty acid-binding protein (H-FABP) is 15 kDA soluble protein which is a powerful regulator of the
350
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
mitochondrial beta-oxidative system [4]. It represents 10% of the whole cardiac myocytes cytosolic proteins [5]. It is undetectable in normal conditions (b5 μg/L), but is released from the myocardium under various types of injury, including myocardial ischemia [5]. It has been shown to be a more effective biological marker than myoglobin for the very early detection of AMI (b 3 h) [2,6,7]. However, its use has been restricted to clinical research because of the lack of a fast and easy-to-use test. A novel one-step qualitative assay for the detection of H-FABP has recently been developed, the CardioDetect® assay, allowing diagnosis of MI within 30 min of chest pain [8]. Our primary objective was to compare the diagnostic accuracy of the point-of-care (POC) assay CardioDetect® to that of pre-hospital troponin, myoglobin and CK-MB in the early detection of MI in patients obtaining a first medical contact quickly after onset of chest pain. 2. Methods 2.1. Study population We prospectively enrolled 108 consecutive out-hospital patients during the period from February 2004 to September 2004, who called the central emergency phone number “15” for chest pain. The decision of sending a mobile intensive care unit (MICU) to the patients' home was made at the dispatching centre by a senior emergency physician. Within 30 min, each patient obtained at home a first physical examination and an 18-lead ECG by an emergency physician. Patients with cardiogenic shock or those with any evidence of a recent chest trauma were excluded. Then according to the diagnosis of NSTEMI or STEMI, the patients were admitted directly to the CCU for further evaluation and treatment or directly to the catheterization laboratory for primary percutaneous coronary intervention. A single MICU participated to this pilot study including four medical doctors who were trained for the use of the CardioDetect® assay.
test zone to which the capture antibody is immobilized, the analyte–detector reagent complex is subsequently bound and the excess detector reagent is trapped by the control zone. Two red bands at the test and control zones are developed with no further addition of reagent. If the sample contains HFABP with concentration below the detection limit, only one band at the control zone is visualized. The whole blood assay uses a blood separator instead of a sample pad and is based on an ‘immuno-threshold’ principle for minimizing falsepositive results. After application of three drops of blood onto the test-strip, the result is available within 15 min (Fig. 1). Interpretation of the tests was performed at the hospital in the same conditions of lighting. The analytical performance of this new assay was obtained from two cohorts of 39 and 218 patients respectively [8,9] and defined according to the National Committee for Clinical Laboratory Standards [10]. The lower limit of detection (LLD) was 2.8 ng/mL corresponding to a signal 3 SD above the mean for a sample in which H-FABP was absent [8,9]. The intraassay coefficient variation was 8–10%. The cut-off for the semi-quantitative CardioDetect® assay was set up at 7 ng/mL according to the manufacturer [8]. The Biosite Triage cTnI cut-off was set up at 0.4 ng/mL according to previous findings [11–13] and the requirements of ESC or ACC/AHA guidelines [14,15]. The LLD of this
2.2. Biological assays All patients underwent blood testing by the MICU physician before reaching the hospital. Simultaneous determinations of cTnI, myoglobin and CK-MB were performed with the Triage® Meter Plus laboratory kit (Biosite SAS, Buc, France). All patients were also tested with the CardioDetect® assay (Rennesens, Berlin, Germany) during pre-hospital management. It is a novel one-step qualitative FABP test which is a rapid chromatographic immunoassay designed for qualitative determination of soluble H-FABP in whole blood samples. This test is derived from the serum lateral-flow test in which a sample applied to a pad containing a specific capture antibody rapidly wets through a conjugate pad containing a detector antibody for H-FABP (Fig. 1). As the sample passes over the
Fig. 1. Principle of the lateral-flow assay (top) and images of a negative and of a positive test (bottom).
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
pre-hospital point-of-care assay is 0.05 ng/mL. The Biosite Triage Myoglobin cut-off was set up at 150 ng/mL according to previous published data with a CVof 11% [12,13]. Finally, the Biosite Triage CK-MB was considered to be positive when above and 10 ng/mL. All patients underwent serial measurements of cTnI every 6 h over the first 24 h after admission to the hospital. All inhospital cTnI measurements were performed with a centrallaboratory assay (SCS of Dade Behring) which met all the requirements of ESC or ACC/AHA guidelines. The cut-off used for this assay was set up at a value of 0.07 ng/mL, which corresponded to the 99th percentile for this assay. We also performed a correlation analysis between the Biosite Triage Meter Plus laboratory kit used for pre-hospital cTnI determination and the central-laboratory assay SCS de Dade Behring to assess whether pre-hospital cTnI determination lack of sensitivity. CTnI was measured with the two assays on the same blood sample in 60 additional patients with suspected acute ischemic chest pain. 2.3. Study end point and statistical analysis Our primary objective was to compare the diagnostic accuracy of the CardioDetect® assay to that of pre-hospital troponin, myoglobin and CK-MB in the early detection of MI in patients presenting with chest pain. The final diagnosis of MI was based on a positive cTnI test at anytime point over the first 24 h, of which one at least had to be determined with the central-laboratory assay. The sensitivity, specificity, negative predictive value (NPV), and diagnostic accuracy (defined as the sum of true positives and true negatives divided by the total number of patients) were determined for all four biological POC assays. These analyses were performed in the whole population and in two prespecified subgroups including patients presenting b 3 hours after symptom onset and those with no ST-elevation as a negative pre-hospital cTnI. The 95% confidence intervals of sensibilities, specificities and accuracies were based on binomial distribution. Comparisons of sensibilities, specificities and accuracies were made by exact Mac Nemar's tests. Statistical significance was defined as p b 0.05. Analyses were performed with the SAS software package V8.2 (SAS Institute, Cary, NC). The sample size was determined according to our primary objective (comparison of sensitivity) and to previous studies which has evaluated the diagnostic performance of the CardioDetect® assay as compared to conventional biomarkers of myonecrosis [16,17]. 3. Results 3.1. Patients' characteristics Patient's characteristics are presented in Table 1. As expected in this unselected population, there was a high proportion of elderly and patients with a prior history of cardiovascular events. Chest pain was considered typical in
351
68% of patients. The MICU met the patients within the first 3 h of symptom onset in 79% of cases. There was no difference for the delay from symptom onset to first medical contact between patients with or without ST-elevation. However, the elderly (N 80 yrs) and females tended to call later and to receive first medical care later than younger patients and males (150 ± 27 vs 134 ± 16 min, for elderly vs younger patients, p = 0.62 and 161 ± 26 vs 126 ± 15 min for females vs males, p = 0.23). 3.2. Accuracy of the pre-hospital POC troponin assay There was a strong correlation between pre-hospital cTnI measured with the Biosite Triage Meter Plus laboratory kit and cTnI measured on the same sample with the centrallaboratory assay (r = 0.98, p b 0.0001). However, there was a significant difference between the two cTnI assays in
Table 1 Patients baseline characteristics
Demographics Age (yrs) (mean ± std) ≥80 yrs Female Risk factors Current smoker % Hypertension % Diabetes (%) Prior history of cardiovascular events b Time from symptom onset to first medical contact (min) (median (min–max)) Initial medical presentation Systolic blood pressure (mm Hg) (mean ± std) Heart rate (/min) (mean ± std) Killip ≥2 (%) ST-segment depression
n = 108
STEMI n = 33
Non-STEMI No MI n = 22 n = 53
68 ± 16 29.6% 37%
69 ± 15 27.3% 51.5% a
71 ± 12 36.4% 18.2%
66 ± 17 28.3% 35.9%
34.0% 41.0% 11.0% 64.0%
42.4% 36.4% 6.1% 40.6% c
31.8% 36.4% 18.2% 72.7%
30.2% 45.3% 11.3% 73.6%
139 153 109 (15–660) (30–600) (15–480)
141 (15–660)
140 ± 30
138 ± 30
154 ± 30 d
133 ± 27
82 ± 21
83 ± 19
89 ± 18
79 ± 23
16.0% 19.0% (n = 21) ST-segment elevation or 38.0% recent LBB (n = 41) Positive cTnI 11.0% (MICU measurement) (n = 12) In-hospital mortality (%) 3.7% (n = 4) Percutaneous coronary 40.9% intervention (%) (n = 43)
27.3%
13.6% 50.0%
9.4% 19.0% 15.0%
24.2%c
18.2%d
0%
9.4% (n = 3) 84.9%a,c
0%
2.0%
52.4%d
7.8%
Adjustments for multiple comparisons were performed with the Tukey– Kramer method for quantitative variables and the Bonferroni method for qualitative variables. LBB = Left Bundle Block. a Indicates a significant difference between STEMI and NSTEMI. b Includes prior myocardial infarction, prior stroke/TIA, prior coronary revascularization, chronic oral antiplatelet therapy. c Indicates a significant difference between STEMI and no MI. d Indicates a significant difference between NSTEMI and no MI.
352
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
Fig. 2. Flow chart.
identifying MI according to the cut-off of the present study. Indeed, only 60.4% of patients with MI according to the central-laboratory assay were recognized to have MI with the Biosite Triage Meter Plus laboratory kit. 3.3. Diagnostic performance of the CardioDetect® assay in the whole population This analysis was conducted irrespective of pre-hospital ECG findings and irrespective of pre-hospital cTnI value. In the pre-hospital setting, 51% of patient had a positive CardioDetect® assay whereas 48%, 24% and 11% of patients had an elevated myoglobin, CK-MB and troponin plasma level, respectively. Fig. 2 summarizes patients' presentation according to initial ECG and biological markers and diagnostic performance of CardioDetect® are summarized
in Table 2. Overall, the semi-quantitative CardioDetect® assay was found to have a significant better sensitivity than all the other POC assays and a better specificity except for that of CK-MB and of cTnI (that is 100% by definition). In addition, the negative predictive value of the CardioDetect® assay was 87.7% [76.3%–94.9%] as compared to 53.3% [44.1%–68.1%], 60.4% [45.3%–74.2%] and 55.2% [44.7%–65.4%] for CK-MB, myoglobin and pre-hospital cTNI, respectively. The diagnostic accuracy of the semiquantitative CardioDetect® assay was found to have a significant better accuracy than all the other POC assays in the whole population but also among the prespecified subgroups including patients presenting less than 3 h after symptom onset and those with a negative pre-hospital cTNI. A total of 77/108 patients with an initial negative cTnI were enrolled in the study within the first 3 h of symptoms
Table 2 Sensitivity, specificity and diagnostic accuracy of all point-of care assays in the whole population and in two prespecified subgroup of patients Whole population (n = 108) CardioDetect® Positive test (n)% 51 Sensitivity 87.3% [75.7%–94.7%] Specificity 94.3% [84.3%–98.8%] Diagnostic accuracy 90.7% [83.6%–95.5%] b3 h (n = 77) Sensitivity 85.7% [71.5%–94.6%] Specificity 92.7% [80.1%–98.5%] Diagnostic accuracy 89.2% [80.4%–94.9%] Negative pre-hospital cTnI and no ST-elevation (n = 63) Sensitivity 83.3% [58.6%–96.4%] Specificity 93.3% [81.7%–98.6%] Diagnostic accuracy 90.5% [80.4%–96.4%] NA=Not Applicable. a Indicates a significant difference with H-FABP.
CK-MB
Myoglobin
Pre-hospital cTnI
25 41.5%a [28.4%–55.9%] 93.0% [80.9%–98.5%] 64.6%a [54.2%–74.1%]
48 64.2%a [40.9%–66.6%] 67.4%a [51.4%–80.9%] 65.6%a [55.2%–75.02%]
12 21.8%a [11.8%–35%] 100% 60.2%a [50.3%–69.5%]
37.5%a [22.7%–54.2%] 93.8%[79.2%–99.2%] 62.5%a [50.3%–73.6%]
57.5%a [40.9%–72.9%] 71.9%a [53.3%–86.3%] 63.9a [51.7 %– 74.9%]
14.3%a [5.4%–28.5%] 100% 56.6%a [45.3%–67.5%]
29.4a [10.3%–56.0%] 91.9 [78.1%–98.3%] 72.2%a [58.4%–83.5%]
70.6 [44%–89.7%] 70.3a [53%–84.1%] 70.4%a [56.4%–82.02%]
NA NA NA
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
onset, of whom 36/77 (46.8%) had a final diagnosis of MI. Time from symptom onset to first medical contact had no impact on the diagnostic performance of the CardioDetect® assay (Table 2). Similarly, the negative predictive value of the CardioDetect® assay (86.4% [72.6%–92.8%]) remained better than any other pre-hospital biomarker (54.6% [40.1%–68%], 57.5% [40.9%–73%], 53.3 [41.5%–64.7%] for CK-MB, myoglobin and pre-hospital cTNI, respectively) in patients admitted b3 h after symptom onset. 3.4. Patients presenting with ST-elevation on the initial ECG The vast majority of patients presenting with ST-elevation on the first ECG had a final diagnosis of myocardial infarction (80.5%). Of interest, these patients had more frequently a positive H-FABP than a positive cTnI (70.7% vs 19.5%, p b 0.0001) in the MICU. In addition, none of them with a negative cTnI over the first 24 hours had a positive CardioDetect® assay. On the other hand, a total of 4 patients out of 33 presenting with STEMI had a negative a positive CardioDetect® assay. 3.5. Patients without ST-elevation and with a negative prehospital POC troponin assay Most patients without ST-elevation on the presenting ECG had a negative cTnI in the ambulance (63/67). In this subgroup of patients, 28.6% (n = 18/63) had a final diagnosis of myocardial infarction of which 83.3% (n = 15/18) were found to have a positive CardioDetect® assay. Of interest was that 18 out of 22 patients with NSTEMI had a positive CardioDetect® assay indicating that 18% of the NSTEMI patients would have been missed when measuring H-FABP and ECG instead of cTnI. In addition, 3 patients without NSTEMI had a positive a positive CardioDetect® assay. The negative predictive value of the CardioDetect® assay (93.3% [81.7%–98.6%]) remained higher than the other prehospital biomarkers (73.9 [58.9%–85.7%], 83.9 [66.3%– 94.6%] for CK-MB and myoglobin, respectively). 4. Discussion Our investigation demonstrates that early determination of H-FABP using a semi-quantitative assay in patients calling for acute chest pain has a better diagnostic performance than any other biological point-of-care assays. This is especially true in patients who were screened within the first 3 h of symptoms onset, which represented the vast majority of patients with a pre-hospital management. Because of complicated assay procedures, most of the immunochemical assays for H-FABP have been of limited use in clinical practice. The newly developed lateral-flow assay was designed to facilitate and to speed up the whole biological procedure and was shown to correlate perfectly with conventional ELISA. The one-step CardioDetect® assay was designed to allow H-FABP determination in
353
whole blood at bed side. It is derived from serum lateral-flow assay with a great specificity, a great sensitivity, and a negative predictive value of 100% in patients presenting to a primary care unit for suspect acute ischemic chest pain. It was claimed that 100% of non-AMI patients could be excluded without false-negative results [8,9]. To our knowledge, our study provides for the first time, data on the accuracy of this new one-step, easy-to-use assay, in a population of all comers with high risk real life characteristics and managed early thanks to the pre-hospital MICU system. The diagnostic performance of the CardioDetect® assay that we report is similar to that of in-hospital patients admitted for chest pain [16]. In addition, we found H-FABP to remain very accurate in patients who had a first medical contact within the first 3 h of symptom onset. These patients represent 2/3 of our population, which is far above the proportion usually reported in registries and in previous studies dedicated to the CardioDetect® assay [16]. This is due to the recruitment through our pre-hospital system of care [3]. The use of the CardioDetect® assay appears to be of unique value in the early hours of acute MI. Indeed, the specificity of the CardioDetect® in the MICU was found to be similar to that of repeated cTnI measurements but with a much greater sensitivity in the first 3 h of AMI. Better and earlier detection of AMI should improve care delivery to these patients. Patients without ST-elevation on the initial ECG and without positive troponin but with suspected acute ischemic chest pain are the most difficult to handle in terms of diagnosis. Indeed, they represent 2/3 of the whole population and 1/3 of them have a final diagnosis of NSTEMI. H-FABP was found to be very efficient to exclude non-ST-elevation MI in this group of patients. This test may be particularly valuable to manage patients with atypical symptoms or silent MI or in those with non-contributive ECG (LBBB, pacemaker) and in whom the need for urgent reperfusion is uncertain. Indeed, the CardioDetect® assay allowed us to exclude 93.3% non-AMI patients, with only three falsenegative values. Our data are very similar to that previously reported for in-hospital patients [16]. Rapid onset of positivity appears to be a major advantage over troponin. Our study shows a faster onset of positivity of H-FABP but does not demonstrate that H-FABP can replace cTnI, which remains the gold standard to identify MI. We also acknowledge that the usefulness of biomarkers is not established in symptomatic patients with persistent ST-segment elevation or new LBBB. This is even true within the early phase of symptom onset. It represents a potential limitation of the present investigation to evaluate the accuracy of this biomarker together with the limited number of patients with non-ST-segment elevation MI. In particular, negative pre-hospital biomarkers should never delay the decision to initiate effective reperfusion therapies, although it is disappointing to that only 80.5% of all patients of our study with ST-segment elevation had confirmed acute ST-elevation MI. It is however reassuring to report that none of the patients
354
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
with ST-segment elevation and without a final diagnosis of MI had a positive H-FABP. Obviously, additional information regarding analytical performance of pre-hospital biomarkers is warranted to meet the requirements of ESC or ACC/AHA guidelines [15]. This is especially true for the Biosite Triage cTnI, which lacks sensitivity as compared to central-laboratory assay [15]. In addition, scant data are available on the inter-observer reliability of the whole blood CardioDetect® assay. Indeed, colour interpretation may be subject to great variations [16]. The inability to discriminate between moderate and high HFABP elevations could be major drawbacks of this test. Finally, comparison of the accuracy of the one-step CardioDetect® assay with that of the quantitative determination of H-FABP is warranted since there is now evidence that this former assay is as much sensitive and as much specific than a central-laboratory cTnI assay, especially within the first 3 h of chest pain onset [9]. This study is exploratory but these preliminary results are encouraging and consistent with recent reports [16]. Whether this easy-to-use and fast assay is a correlate of clinical outcome needs to be further addressed. In addition, whether having such a test available in a mobile unit may add a benefit to patient classification and therefore patient care remains to be established in the setting of suspected acute ischemic chest pain without ST-segment elevation. References [1] Bertrand ME, Simoons ML, Fox KA, et al. Management of acute coronary syndromes: acute coronary syndromes without persistent ST segment elevation; recommendations of the Task Force of the European Society of Cardiology. Eur Heart J 2002;21:1809–40. [2] Adams JE, Trent R, Rawles J. Earliest electrocardiographic evidence of myocardial infarction: implications for thrombolytic treatment. Br Med J 1993(307):409–13. [3] Danchin N, Blanchard D, Sauval PG, et al, for the USIC 2000 Investigators. Impact of prehospital thrombolysis for acute myocardial infarction on 1-year outcome: results from the French Nationwide USIC 2000 Registry. Circulation 2004;110:1909–15.
[4] Fournier NC, Richard MA. Role of fatty acid-binding protein in cardiac fatty acid oxidation. Mol Cell Biochem 1990;98:149–59. [5] Pelsers MM, Hermens WT, Glatz JF. Fatty acid-binding proteins as plasma markers of tissue injury. Clin Chim Acta 2005;352:15–35. [6] Ishii J, Wang JH, Naruse H, et al. Serum concentrations of myoglobin vs. human heart-type cytoplasmic fatty acid-binding protein in early detection of acute myocardial infarction. Clin Chem 1997;43:1372–8. [7] Kleine AH, Glatz JFC, van Nieuwenhoven FA, van der Vusse GJ. Release of heart fatty acid-binding protein into plasma after acute myocardial infarction in man. Cell Biochem 1992;116:155–62. [8] Chan CP, Sum KW, Cheung KY, et al. Development of a quantitative lateral-flow assay for rapid detection of fatty acid-binding protein. J Immunol Methods 2003;279:91–100. [9] Chan CP, Sanderson JE, Glatz JF, Cheng WS, Hempel A, Renneberg R. A superior early myocardial infarction marker. Human heart-type fatty acid-binding protein. Z Kardiol 2004;93:388–97. [10] National Committee for Clinical Laboratory Standards. Evaluation of precision performance of clinical chemistry devices—tentative guideline. NCCLS document EP5-T2. 2nd ed. 1-56238-145-8; 1992. [11] Apple FS, Christenson RH, Valdes R, et al. Simultaneous rapid measurement of whole blood myoglobin, creatine kinase MB, and cardiac troponin I by the triage cardiac panel for detection of myocardial infarction. Clin Chem 1999;45:199–205. [12] McCord J, Nowak RM, McCullough PA, et al. Ninety-minute exclusion of acute myocardial infarction by use of quantitative pointof-care testing of myoglobin and troponin I. Circulation 2001;104: 1483–8. [13] Ming S, Krishnaswamy P, Morissey R, Clopton P, Fitzgerald R, Maisel AS. Ninety-minute accelerated critical pathway for chest pain evaluation. Am J Cardiol 2001;88:611–7. [14] 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: 1331–6. [15] Wu AHB, Smith A, Christenson RH, Murakami MAM, Apple FS. Evaluation of a point-of-care assay for cardiac markers for patients suspected of acute myocardial infarction. Clin Chim Acta 2004;346: 211–9. [16] Alhashemi JA. Diagnostic accuracy of a bedside qualitative immunochromatographic test for acute myocardial infarction. Am J Emerg Med 2006;24:149–55. [17] Chan CPY, Cheng WS, Glatz JFC, et al. Early diagnosis of acute myocardial infarction using immunosensors and immunotests. Anal Lett 2003;36:1987–2004.
Pre-hospital detection of acute myocardial infarction with ultra-rapid human fatty acid-binding protein (H-FABP) immunoassay Patrick Ecollan a , Jean-Philippe Collet b , Guillaume Boon a , Marie-Laure Tanguy c , Marie-Laurence Fievet a , Rosita Haas a , Nicolas Bertho a , Shidash Siami a , Jean-Christophe Hubert a , Pierre Coriat a , Gilles Montalescot b,⁎ a SAMU, Pitié-Salpêtrière University Hospital, Paris, France Institut de Cardiologie, Pitié-Salpêtrière University Hospital, Paris, France Department of Biostatistics, Pitié-Salpêtrière University Hospital, Paris, France b
c
Received 4 January 2006; received in revised form 8 September 2006; accepted 13 September 2006 Available online 13 November 2006
Abstract Aims: To evaluate the diagnostic performance of a portable semi-quantitative whole blood immunoassay measuring soluble human fatty acidbinding protein [H-FABP] (CardioDetect®) for the pre-hospital detection of myocardial infarction (MI). Methods and results: We enrolled prospectively 108 consecutive patients with acute ischemic type chest pain in whom the first medical care was delivered by a mobile intensive care unit (MICU). CTnI, myoglobin, CK-MB and CardioDetect® were first assessed in the MICU before hospital admission using point-of-care assays and then cTnI was serially measured during the hospital stay. MI was defined as a positive cTnI in any sample over the first 24 h. The vast majority of the patients (77/108) were admitted b 3 h of symptoms onset. Pre-hospital cTnI sample was normal in 96 patients (88.9%) of whom 43 had subsequent cTnI elevation. A positive H-FABP using the CardioDetect® assay had a significantly better sensitivity than cTnI, myoglobin and CK-MB (87.3% vs 21.8%, 64.2% and 41.5%, respectively) to identify MI. The significant better sensitivity of the CardioDetect® assay was also observed among patients who presented b 3 h of symptom onset. The specificity of the CardioDetect® assay was significantly better than that of myoglobin, irrespective of the time delay from symptom onset to measurement. In patients with normal pre-hospital cTnI and no STelevation (n = 63), assessment based only on a positive H-FABP had 83.3% sensitivity, 93.3% specificity for the diagnosis of an evolving MI. Conclusion: Early assessment of H-FABP in patients presenting with chest pain improves the diagnosis of ongoing MI. © 2006 Published by Elsevier Ireland Ltd. Keywords: Myocardial infarction; Coronary disease; Biomarkers
1. Introduction Early detection of acute myocardial infarction (AMI) remains suboptimal. Although quite specific [1], STelevation has only a 50 to 60% sensitivity for diagnosis of ⁎ Corresponding author. Institut de Cardiologie, Hôpital Pitié-Salpêtrière (AP-HP) Bureau 2-236, 47, boulevard de l'Hôpital, 75013 Paris, France. Tel.: +33 1 42 16 30 06; fax: +33 1 42 16 29 31. E-mail address: [email protected] (G. Montalescot). 0167-5273/$ - see front matter © 2006 Published by Elsevier Ireland Ltd. doi:10.1016/j.ijcard.2006.09.003
AMI [2]. Cardiac troponin I (cTnI) does not reliably exclude AMI without repeated negative measurements over 4–6 h and myoglobin is limited by poor specificity [1]. This is an important issue given the fact that 40% of out-of-hospital patients with an evolving AMI are getting medical care within the first 3 h [3]. There is a need for a fast and reliable test to facilitate triage, diagnosis and adequate treatment strategies. This is particularly important in patients presenting with atypical symptoms and/or non-contributive ECG. The human heart-type fatty acid-binding protein (H-FABP) is 15 kDA soluble protein which is a powerful regulator of the
350
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
mitochondrial beta-oxidative system [4]. It represents 10% of the whole cardiac myocytes cytosolic proteins [5]. It is undetectable in normal conditions (b5 μg/L), but is released from the myocardium under various types of injury, including myocardial ischemia [5]. It has been shown to be a more effective biological marker than myoglobin for the very early detection of AMI (b 3 h) [2,6,7]. However, its use has been restricted to clinical research because of the lack of a fast and easy-to-use test. A novel one-step qualitative assay for the detection of H-FABP has recently been developed, the CardioDetect® assay, allowing diagnosis of MI within 30 min of chest pain [8]. Our primary objective was to compare the diagnostic accuracy of the point-of-care (POC) assay CardioDetect® to that of pre-hospital troponin, myoglobin and CK-MB in the early detection of MI in patients obtaining a first medical contact quickly after onset of chest pain. 2. Methods 2.1. Study population We prospectively enrolled 108 consecutive out-hospital patients during the period from February 2004 to September 2004, who called the central emergency phone number “15” for chest pain. The decision of sending a mobile intensive care unit (MICU) to the patients' home was made at the dispatching centre by a senior emergency physician. Within 30 min, each patient obtained at home a first physical examination and an 18-lead ECG by an emergency physician. Patients with cardiogenic shock or those with any evidence of a recent chest trauma were excluded. Then according to the diagnosis of NSTEMI or STEMI, the patients were admitted directly to the CCU for further evaluation and treatment or directly to the catheterization laboratory for primary percutaneous coronary intervention. A single MICU participated to this pilot study including four medical doctors who were trained for the use of the CardioDetect® assay.
test zone to which the capture antibody is immobilized, the analyte–detector reagent complex is subsequently bound and the excess detector reagent is trapped by the control zone. Two red bands at the test and control zones are developed with no further addition of reagent. If the sample contains HFABP with concentration below the detection limit, only one band at the control zone is visualized. The whole blood assay uses a blood separator instead of a sample pad and is based on an ‘immuno-threshold’ principle for minimizing falsepositive results. After application of three drops of blood onto the test-strip, the result is available within 15 min (Fig. 1). Interpretation of the tests was performed at the hospital in the same conditions of lighting. The analytical performance of this new assay was obtained from two cohorts of 39 and 218 patients respectively [8,9] and defined according to the National Committee for Clinical Laboratory Standards [10]. The lower limit of detection (LLD) was 2.8 ng/mL corresponding to a signal 3 SD above the mean for a sample in which H-FABP was absent [8,9]. The intraassay coefficient variation was 8–10%. The cut-off for the semi-quantitative CardioDetect® assay was set up at 7 ng/mL according to the manufacturer [8]. The Biosite Triage cTnI cut-off was set up at 0.4 ng/mL according to previous findings [11–13] and the requirements of ESC or ACC/AHA guidelines [14,15]. The LLD of this
2.2. Biological assays All patients underwent blood testing by the MICU physician before reaching the hospital. Simultaneous determinations of cTnI, myoglobin and CK-MB were performed with the Triage® Meter Plus laboratory kit (Biosite SAS, Buc, France). All patients were also tested with the CardioDetect® assay (Rennesens, Berlin, Germany) during pre-hospital management. It is a novel one-step qualitative FABP test which is a rapid chromatographic immunoassay designed for qualitative determination of soluble H-FABP in whole blood samples. This test is derived from the serum lateral-flow test in which a sample applied to a pad containing a specific capture antibody rapidly wets through a conjugate pad containing a detector antibody for H-FABP (Fig. 1). As the sample passes over the
Fig. 1. Principle of the lateral-flow assay (top) and images of a negative and of a positive test (bottom).
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
pre-hospital point-of-care assay is 0.05 ng/mL. The Biosite Triage Myoglobin cut-off was set up at 150 ng/mL according to previous published data with a CVof 11% [12,13]. Finally, the Biosite Triage CK-MB was considered to be positive when above and 10 ng/mL. All patients underwent serial measurements of cTnI every 6 h over the first 24 h after admission to the hospital. All inhospital cTnI measurements were performed with a centrallaboratory assay (SCS of Dade Behring) which met all the requirements of ESC or ACC/AHA guidelines. The cut-off used for this assay was set up at a value of 0.07 ng/mL, which corresponded to the 99th percentile for this assay. We also performed a correlation analysis between the Biosite Triage Meter Plus laboratory kit used for pre-hospital cTnI determination and the central-laboratory assay SCS de Dade Behring to assess whether pre-hospital cTnI determination lack of sensitivity. CTnI was measured with the two assays on the same blood sample in 60 additional patients with suspected acute ischemic chest pain. 2.3. Study end point and statistical analysis Our primary objective was to compare the diagnostic accuracy of the CardioDetect® assay to that of pre-hospital troponin, myoglobin and CK-MB in the early detection of MI in patients presenting with chest pain. The final diagnosis of MI was based on a positive cTnI test at anytime point over the first 24 h, of which one at least had to be determined with the central-laboratory assay. The sensitivity, specificity, negative predictive value (NPV), and diagnostic accuracy (defined as the sum of true positives and true negatives divided by the total number of patients) were determined for all four biological POC assays. These analyses were performed in the whole population and in two prespecified subgroups including patients presenting b 3 hours after symptom onset and those with no ST-elevation as a negative pre-hospital cTnI. The 95% confidence intervals of sensibilities, specificities and accuracies were based on binomial distribution. Comparisons of sensibilities, specificities and accuracies were made by exact Mac Nemar's tests. Statistical significance was defined as p b 0.05. Analyses were performed with the SAS software package V8.2 (SAS Institute, Cary, NC). The sample size was determined according to our primary objective (comparison of sensitivity) and to previous studies which has evaluated the diagnostic performance of the CardioDetect® assay as compared to conventional biomarkers of myonecrosis [16,17]. 3. Results 3.1. Patients' characteristics Patient's characteristics are presented in Table 1. As expected in this unselected population, there was a high proportion of elderly and patients with a prior history of cardiovascular events. Chest pain was considered typical in
351
68% of patients. The MICU met the patients within the first 3 h of symptom onset in 79% of cases. There was no difference for the delay from symptom onset to first medical contact between patients with or without ST-elevation. However, the elderly (N 80 yrs) and females tended to call later and to receive first medical care later than younger patients and males (150 ± 27 vs 134 ± 16 min, for elderly vs younger patients, p = 0.62 and 161 ± 26 vs 126 ± 15 min for females vs males, p = 0.23). 3.2. Accuracy of the pre-hospital POC troponin assay There was a strong correlation between pre-hospital cTnI measured with the Biosite Triage Meter Plus laboratory kit and cTnI measured on the same sample with the centrallaboratory assay (r = 0.98, p b 0.0001). However, there was a significant difference between the two cTnI assays in
Table 1 Patients baseline characteristics
Demographics Age (yrs) (mean ± std) ≥80 yrs Female Risk factors Current smoker % Hypertension % Diabetes (%) Prior history of cardiovascular events b Time from symptom onset to first medical contact (min) (median (min–max)) Initial medical presentation Systolic blood pressure (mm Hg) (mean ± std) Heart rate (/min) (mean ± std) Killip ≥2 (%) ST-segment depression
n = 108
STEMI n = 33
Non-STEMI No MI n = 22 n = 53
68 ± 16 29.6% 37%
69 ± 15 27.3% 51.5% a
71 ± 12 36.4% 18.2%
66 ± 17 28.3% 35.9%
34.0% 41.0% 11.0% 64.0%
42.4% 36.4% 6.1% 40.6% c
31.8% 36.4% 18.2% 72.7%
30.2% 45.3% 11.3% 73.6%
139 153 109 (15–660) (30–600) (15–480)
141 (15–660)
140 ± 30
138 ± 30
154 ± 30 d
133 ± 27
82 ± 21
83 ± 19
89 ± 18
79 ± 23
16.0% 19.0% (n = 21) ST-segment elevation or 38.0% recent LBB (n = 41) Positive cTnI 11.0% (MICU measurement) (n = 12) In-hospital mortality (%) 3.7% (n = 4) Percutaneous coronary 40.9% intervention (%) (n = 43)
27.3%
13.6% 50.0%
9.4% 19.0% 15.0%
24.2%c
18.2%d
0%
9.4% (n = 3) 84.9%a,c
0%
2.0%
52.4%d
7.8%
Adjustments for multiple comparisons were performed with the Tukey– Kramer method for quantitative variables and the Bonferroni method for qualitative variables. LBB = Left Bundle Block. a Indicates a significant difference between STEMI and NSTEMI. b Includes prior myocardial infarction, prior stroke/TIA, prior coronary revascularization, chronic oral antiplatelet therapy. c Indicates a significant difference between STEMI and no MI. d Indicates a significant difference between NSTEMI and no MI.
352
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
Fig. 2. Flow chart.
identifying MI according to the cut-off of the present study. Indeed, only 60.4% of patients with MI according to the central-laboratory assay were recognized to have MI with the Biosite Triage Meter Plus laboratory kit. 3.3. Diagnostic performance of the CardioDetect® assay in the whole population This analysis was conducted irrespective of pre-hospital ECG findings and irrespective of pre-hospital cTnI value. In the pre-hospital setting, 51% of patient had a positive CardioDetect® assay whereas 48%, 24% and 11% of patients had an elevated myoglobin, CK-MB and troponin plasma level, respectively. Fig. 2 summarizes patients' presentation according to initial ECG and biological markers and diagnostic performance of CardioDetect® are summarized
in Table 2. Overall, the semi-quantitative CardioDetect® assay was found to have a significant better sensitivity than all the other POC assays and a better specificity except for that of CK-MB and of cTnI (that is 100% by definition). In addition, the negative predictive value of the CardioDetect® assay was 87.7% [76.3%–94.9%] as compared to 53.3% [44.1%–68.1%], 60.4% [45.3%–74.2%] and 55.2% [44.7%–65.4%] for CK-MB, myoglobin and pre-hospital cTNI, respectively. The diagnostic accuracy of the semiquantitative CardioDetect® assay was found to have a significant better accuracy than all the other POC assays in the whole population but also among the prespecified subgroups including patients presenting less than 3 h after symptom onset and those with a negative pre-hospital cTNI. A total of 77/108 patients with an initial negative cTnI were enrolled in the study within the first 3 h of symptoms
Table 2 Sensitivity, specificity and diagnostic accuracy of all point-of care assays in the whole population and in two prespecified subgroup of patients Whole population (n = 108) CardioDetect® Positive test (n)% 51 Sensitivity 87.3% [75.7%–94.7%] Specificity 94.3% [84.3%–98.8%] Diagnostic accuracy 90.7% [83.6%–95.5%] b3 h (n = 77) Sensitivity 85.7% [71.5%–94.6%] Specificity 92.7% [80.1%–98.5%] Diagnostic accuracy 89.2% [80.4%–94.9%] Negative pre-hospital cTnI and no ST-elevation (n = 63) Sensitivity 83.3% [58.6%–96.4%] Specificity 93.3% [81.7%–98.6%] Diagnostic accuracy 90.5% [80.4%–96.4%] NA=Not Applicable. a Indicates a significant difference with H-FABP.
CK-MB
Myoglobin
Pre-hospital cTnI
25 41.5%a [28.4%–55.9%] 93.0% [80.9%–98.5%] 64.6%a [54.2%–74.1%]
48 64.2%a [40.9%–66.6%] 67.4%a [51.4%–80.9%] 65.6%a [55.2%–75.02%]
12 21.8%a [11.8%–35%] 100% 60.2%a [50.3%–69.5%]
37.5%a [22.7%–54.2%] 93.8%[79.2%–99.2%] 62.5%a [50.3%–73.6%]
57.5%a [40.9%–72.9%] 71.9%a [53.3%–86.3%] 63.9a [51.7 %– 74.9%]
14.3%a [5.4%–28.5%] 100% 56.6%a [45.3%–67.5%]
29.4a [10.3%–56.0%] 91.9 [78.1%–98.3%] 72.2%a [58.4%–83.5%]
70.6 [44%–89.7%] 70.3a [53%–84.1%] 70.4%a [56.4%–82.02%]
NA NA NA
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
onset, of whom 36/77 (46.8%) had a final diagnosis of MI. Time from symptom onset to first medical contact had no impact on the diagnostic performance of the CardioDetect® assay (Table 2). Similarly, the negative predictive value of the CardioDetect® assay (86.4% [72.6%–92.8%]) remained better than any other pre-hospital biomarker (54.6% [40.1%–68%], 57.5% [40.9%–73%], 53.3 [41.5%–64.7%] for CK-MB, myoglobin and pre-hospital cTNI, respectively) in patients admitted b3 h after symptom onset. 3.4. Patients presenting with ST-elevation on the initial ECG The vast majority of patients presenting with ST-elevation on the first ECG had a final diagnosis of myocardial infarction (80.5%). Of interest, these patients had more frequently a positive H-FABP than a positive cTnI (70.7% vs 19.5%, p b 0.0001) in the MICU. In addition, none of them with a negative cTnI over the first 24 hours had a positive CardioDetect® assay. On the other hand, a total of 4 patients out of 33 presenting with STEMI had a negative a positive CardioDetect® assay. 3.5. Patients without ST-elevation and with a negative prehospital POC troponin assay Most patients without ST-elevation on the presenting ECG had a negative cTnI in the ambulance (63/67). In this subgroup of patients, 28.6% (n = 18/63) had a final diagnosis of myocardial infarction of which 83.3% (n = 15/18) were found to have a positive CardioDetect® assay. Of interest was that 18 out of 22 patients with NSTEMI had a positive CardioDetect® assay indicating that 18% of the NSTEMI patients would have been missed when measuring H-FABP and ECG instead of cTnI. In addition, 3 patients without NSTEMI had a positive a positive CardioDetect® assay. The negative predictive value of the CardioDetect® assay (93.3% [81.7%–98.6%]) remained higher than the other prehospital biomarkers (73.9 [58.9%–85.7%], 83.9 [66.3%– 94.6%] for CK-MB and myoglobin, respectively). 4. Discussion Our investigation demonstrates that early determination of H-FABP using a semi-quantitative assay in patients calling for acute chest pain has a better diagnostic performance than any other biological point-of-care assays. This is especially true in patients who were screened within the first 3 h of symptoms onset, which represented the vast majority of patients with a pre-hospital management. Because of complicated assay procedures, most of the immunochemical assays for H-FABP have been of limited use in clinical practice. The newly developed lateral-flow assay was designed to facilitate and to speed up the whole biological procedure and was shown to correlate perfectly with conventional ELISA. The one-step CardioDetect® assay was designed to allow H-FABP determination in
353
whole blood at bed side. It is derived from serum lateral-flow assay with a great specificity, a great sensitivity, and a negative predictive value of 100% in patients presenting to a primary care unit for suspect acute ischemic chest pain. It was claimed that 100% of non-AMI patients could be excluded without false-negative results [8,9]. To our knowledge, our study provides for the first time, data on the accuracy of this new one-step, easy-to-use assay, in a population of all comers with high risk real life characteristics and managed early thanks to the pre-hospital MICU system. The diagnostic performance of the CardioDetect® assay that we report is similar to that of in-hospital patients admitted for chest pain [16]. In addition, we found H-FABP to remain very accurate in patients who had a first medical contact within the first 3 h of symptom onset. These patients represent 2/3 of our population, which is far above the proportion usually reported in registries and in previous studies dedicated to the CardioDetect® assay [16]. This is due to the recruitment through our pre-hospital system of care [3]. The use of the CardioDetect® assay appears to be of unique value in the early hours of acute MI. Indeed, the specificity of the CardioDetect® in the MICU was found to be similar to that of repeated cTnI measurements but with a much greater sensitivity in the first 3 h of AMI. Better and earlier detection of AMI should improve care delivery to these patients. Patients without ST-elevation on the initial ECG and without positive troponin but with suspected acute ischemic chest pain are the most difficult to handle in terms of diagnosis. Indeed, they represent 2/3 of the whole population and 1/3 of them have a final diagnosis of NSTEMI. H-FABP was found to be very efficient to exclude non-ST-elevation MI in this group of patients. This test may be particularly valuable to manage patients with atypical symptoms or silent MI or in those with non-contributive ECG (LBBB, pacemaker) and in whom the need for urgent reperfusion is uncertain. Indeed, the CardioDetect® assay allowed us to exclude 93.3% non-AMI patients, with only three falsenegative values. Our data are very similar to that previously reported for in-hospital patients [16]. Rapid onset of positivity appears to be a major advantage over troponin. Our study shows a faster onset of positivity of H-FABP but does not demonstrate that H-FABP can replace cTnI, which remains the gold standard to identify MI. We also acknowledge that the usefulness of biomarkers is not established in symptomatic patients with persistent ST-segment elevation or new LBBB. This is even true within the early phase of symptom onset. It represents a potential limitation of the present investigation to evaluate the accuracy of this biomarker together with the limited number of patients with non-ST-segment elevation MI. In particular, negative pre-hospital biomarkers should never delay the decision to initiate effective reperfusion therapies, although it is disappointing to that only 80.5% of all patients of our study with ST-segment elevation had confirmed acute ST-elevation MI. It is however reassuring to report that none of the patients
354
P. Ecollan et al. / International Journal of Cardiology 119 (2007) 349–354
with ST-segment elevation and without a final diagnosis of MI had a positive H-FABP. Obviously, additional information regarding analytical performance of pre-hospital biomarkers is warranted to meet the requirements of ESC or ACC/AHA guidelines [15]. This is especially true for the Biosite Triage cTnI, which lacks sensitivity as compared to central-laboratory assay [15]. In addition, scant data are available on the inter-observer reliability of the whole blood CardioDetect® assay. Indeed, colour interpretation may be subject to great variations [16]. The inability to discriminate between moderate and high HFABP elevations could be major drawbacks of this test. Finally, comparison of the accuracy of the one-step CardioDetect® assay with that of the quantitative determination of H-FABP is warranted since there is now evidence that this former assay is as much sensitive and as much specific than a central-laboratory cTnI assay, especially within the first 3 h of chest pain onset [9]. This study is exploratory but these preliminary results are encouraging and consistent with recent reports [16]. Whether this easy-to-use and fast assay is a correlate of clinical outcome needs to be further addressed. In addition, whether having such a test available in a mobile unit may add a benefit to patient classification and therefore patient care remains to be established in the setting of suspected acute ischemic chest pain without ST-segment elevation. References [1] Bertrand ME, Simoons ML, Fox KA, et al. Management of acute coronary syndromes: acute coronary syndromes without persistent ST segment elevation; recommendations of the Task Force of the European Society of Cardiology. Eur Heart J 2002;21:1809–40. [2] Adams JE, Trent R, Rawles J. Earliest electrocardiographic evidence of myocardial infarction: implications for thrombolytic treatment. Br Med J 1993(307):409–13. [3] Danchin N, Blanchard D, Sauval PG, et al, for the USIC 2000 Investigators. Impact of prehospital thrombolysis for acute myocardial infarction on 1-year outcome: results from the French Nationwide USIC 2000 Registry. Circulation 2004;110:1909–15.
[4] Fournier NC, Richard MA. Role of fatty acid-binding protein in cardiac fatty acid oxidation. Mol Cell Biochem 1990;98:149–59. [5] Pelsers MM, Hermens WT, Glatz JF. Fatty acid-binding proteins as plasma markers of tissue injury. Clin Chim Acta 2005;352:15–35. [6] Ishii J, Wang JH, Naruse H, et al. Serum concentrations of myoglobin vs. human heart-type cytoplasmic fatty acid-binding protein in early detection of acute myocardial infarction. Clin Chem 1997;43:1372–8. [7] Kleine AH, Glatz JFC, van Nieuwenhoven FA, van der Vusse GJ. Release of heart fatty acid-binding protein into plasma after acute myocardial infarction in man. Cell Biochem 1992;116:155–62. [8] Chan CP, Sum KW, Cheung KY, et al. Development of a quantitative lateral-flow assay for rapid detection of fatty acid-binding protein. J Immunol Methods 2003;279:91–100. [9] Chan CP, Sanderson JE, Glatz JF, Cheng WS, Hempel A, Renneberg R. A superior early myocardial infarction marker. Human heart-type fatty acid-binding protein. Z Kardiol 2004;93:388–97. [10] National Committee for Clinical Laboratory Standards. Evaluation of precision performance of clinical chemistry devices—tentative guideline. NCCLS document EP5-T2. 2nd ed. 1-56238-145-8; 1992. [11] Apple FS, Christenson RH, Valdes R, et al. Simultaneous rapid measurement of whole blood myoglobin, creatine kinase MB, and cardiac troponin I by the triage cardiac panel for detection of myocardial infarction. Clin Chem 1999;45:199–205. [12] McCord J, Nowak RM, McCullough PA, et al. Ninety-minute exclusion of acute myocardial infarction by use of quantitative pointof-care testing of myoglobin and troponin I. Circulation 2001;104: 1483–8. [13] Ming S, Krishnaswamy P, Morissey R, Clopton P, Fitzgerald R, Maisel AS. Ninety-minute accelerated critical pathway for chest pain evaluation. Am J Cardiol 2001;88:611–7. [14] 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: 1331–6. [15] Wu AHB, Smith A, Christenson RH, Murakami MAM, Apple FS. Evaluation of a point-of-care assay for cardiac markers for patients suspected of acute myocardial infarction. Clin Chim Acta 2004;346: 211–9. [16] Alhashemi JA. Diagnostic accuracy of a bedside qualitative immunochromatographic test for acute myocardial infarction. Am J Emerg Med 2006;24:149–55. [17] Chan CPY, Cheng WS, Glatz JFC, et al. Early diagnosis of acute myocardial infarction using immunosensors and immunotests. Anal Lett 2003;36:1987–2004.