• 
• 

Three Part Question

In [patients who present to the Emergency Department with chest pain and have a suspected acute coronary syndrome] does [measurement of plasma heart fatty acid binding protein at the time of presentation] enable [accurate early diagnosis of acute myocardial infarction]?

Clinical Scenario

A 35 year-old man presents to the Emergency Department (ED) 15 minutes after having experienced 30 minutes of central chest pain that resolved spontaneously. He has no previous medical history. Physical examination is unremarkable and the ECG is normal. You know that current rapid rule out protocols have limited sensitivity and are not recommended by the European Society of Cardiology or the International Liaison Committee for Resuscitation. You wonder whether heart fatty acid binding protein (H-FABP), an intracellular constituent of cardiac myocytes that appears in plasma soon after the onset of acute myocardial infarction (AMI), could be used to facilitate rapid exclusion of AMI in the ED.

Search Strategy

Ovid Medline 1950 - 2009 January Week 4
Ovid Embase 1980 - 2009 Week 5
(exp Acute Coronary Syndromes/ OR exp Acute Myocardial Infarction/ OR exp Myocardial Infarction/ OR exp Angina, Unstable/ OR (myocard$ adj (ischem$ OR ischaem$ OR infarct$)).mp. OR exp Chest Pain/) AND (exp Heart Fatty Acid Binding Protein/ OR (HFABP OR H-FABP).mp. OR ((heart) adj2 (fatty acid binding protein)).mp.) limit to humans and English language

Search Outcome

Altogether 63 papers were found in Medline and 71 in Embase. Papers that evaluated the use of H-FABP in patients with suspected cardiac chest pain at the point of presentation to the ED were included. Studies investigating the use of H-FABP in primary care, pre-hospital or among inpatients with confirmed acute coronary syndromes were excluded. Studies that did not investigate the diagnostic value of H-FABP at the point of presentation to the ED were also excluded. Five papers fulfilled the inclusion criteria and are tabulated.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Mad et al 2007 Austria	280 consecutive patients presenting to the ED with chest pain or dyspnoea within 24 hours of symptom onset (median time from symptom onset 3 hours). H-FABP was measured at the time of ED presentation and 4 hours later using a qualitative point of care (POC) assay. Final diagnosis was assigned by a "senior cardiologist" (? blinded to H-FABP) using a troponin T cut-off of >=0.04ng/ml for AMI diagnosis. All patients underwent troponin testing at 0 and 4 hours.	Prospective diagnostic cohort study	H-FABP for diagnosis of AMI	Sensitivity 69% (95% CI 59-77%); specificity 74% (66-80%); LR+ 2.61 (1.97 - 2.46); LR- 0.42 (0.32 - 0.57)	Inadequate reference standard: troponin testing at 0 and 4 hours but not necessarily after 12 hours. This introduces significant verification bias. The treating physician was blinded to H-FABP results but the paper does not mention whether the senior cardiologist who assigned final diagnoses was also blinded. Qualitative assay.
			Performance of H-FABP for diagnosis of AMI stratified by time from symptom onset	Highest sensitivity in patients >6h from symptom onset (sensitivity 65.4% at <2h, 58.8% at 2-6h, 91.3% at >6h)
Alhashemi 2006 Saudi Arabia	64 patients aged >18 years who presented to the ED with suspected cardiac chest pain lasting at least 30 minutes. All patients had blood drawn by finger prick and analysed using a qualitative point of care assay for H-FABP at the point of presentation to the ED. All patients underwent serial troponin I testing for 24 hours. Final diagnosis of AMI was assigned by an independent attending physician who was blinded to H-FABP results, using a troponin I cut-off set at 0.05ng/ml.	Prospective diagnostic cohort study	H-FABP when taken <=4h from symptom onset for diagnosis of AMI	Sensitivity 62.55 (95% CI 35.9 - 83.7%); specificity 100.0% (51.7 - 100.0%); PPV 100.0% (65.5 - 100.0%); NPV 50% (22.3 - 77.7%)	Small numbers - note the wide confidence intervals. Given the small numbers it would have been preferable to pool all of the results for the primary analysis rather than stratify patients by time from symptom onset. No pairwise comparison of AUC or sensitivities and specificities between biomarkers. Qualitative point of care assay, not ideal for ROC analysis.
			H-FABP when taken 4-12h from symptom onset for diagnosis of AMI	Sensitivity 100.0% (73.2 - 100.0%); specificity 80.0% (29.9 - 98.9%); PPV 93.3% (66.0 - 100.0%); NPV 100% (39.6 - 100.0%)
			H-FABP when taken 12-24h from symptom onset for diagnosis of AMI	Sensitivity 100.0% (19.8 - 100.0%); specificity 9.2 - 90.8%); PPV 50.0% (9.2 - 90.8%); NPV 100.0% (19.8 - 100.0%)
			H-FABP when taken >24h from symptom onset for diagnosis of AMI	Sensitivity 77.8% (40.2 - 96.1%); specificity 80.0% (29.9 - 98.9%); PPV 87.5% (46.7 - 99.3%); NPV 66.7% (24.1 - 94.0%)
			Area under the ROC curve (ROC) for each biomarker for patients with symptom onset <=4h	H-FABP 0.813 (95% CI 0.636 - 0.989); TnI 0.650 (0.381 - 0.919); CK 0.588 (0.307 - 0.868)
			Area under the ROC curve (ROC) for each biomarker for patients with symptom onset 4-12h	H-FABP 0.900 (0.685 - 1.000); TnI 0.800 (0.519 - 1.000); CK 0.829 (0.594 - 1.000)
			Area under the ROC curve (ROC) for each biomarker for patients with symptom onset 12-24h	H-FABP 0.750 (0.338 - 1.000); TnI 0.750 (0.338 - 1.000); CK 1.000 (1.000 - 1.000)
			Area under the ROC curve (ROC) for each biomarker for patients with symptom onset >24h	H-FABP 0.938 (0.786 - 1.000); TnI 0.833 (0.489 - 1.000); CK 0.938 (0.786 - 1.000)
Seino et al 2003 Japan	371 consecutive patients who presented to the cardiac Emergency Department at six centres with suspected cardiac chest pain within 36 hours of onset. Patients with STEMI were excluded. AMI diagnosed according to 1979 WHO criteria using CK-MB as the gold standard biomarker. Blood was drawn at presentation. H-FABP was measured by qualitative point of care assay using whole blood. 201 patients also had quantitative measurement of H-FABP in serum. Troponin T was measured using a qualitative assay.	Prospective diagnostic cohort study	Diagnostic performance of qualitative H-FABP test	Sensitivity 95.0%; specificity 48.9%; PPV 63.9%; NPV 91.2%	This was a study in a cardiac ED. The population is probably different to that in standard EDs, as reflected by the high prevalence of AMI (49%). No attempt to select a diagnostic cut-off using the quantitative assay and to report diagnostic performance using that cut-off. Outdated gold standard for AMI; secondary analysis using troponins as gold standard but cut-off not stated.
			Diagnostic performance of qualitative assay at <2h from symptom onset for different biomarkers	H-FABP sensitivity 89% (95% CI 74-97), specificity 96% (87-99). Troponin T sensitivity 22% (10-39); specificity 94% (77-99); myoglobin sensitivity 38% (21-55); specificity 71% (52-65)
			Diagnostic performance of qualitative assay at 2-4h from symptom onset for different biomarkers	H-FABP sensitivity 96% (87-99); specificity 45% (30-60). Troponin T sensitivity 57% (43-71), specificity 70% (55-82); Myoglobin sensitivity 63% (49-76), specificity 64% (49-79)
			Agreement between quantitative and qualitative H-FABP assays	85% concordance.
			Secondary analysis using troponin T as the gold standard biomarker for AMI (quantitative assay) in a subgroup of 201 patients	There were no meaningful differences in the results
			ROC analysis: area under the curve (AUC)	H-FABP 0.79 (0.73 - 0.85); Myoglobin 0.76 (0.70 - 0.82). P=0.02 for comparison. No AUC for troponin T as qualitative assay used
Ghani et al 2000 USA	460 consecutive patients who presented to the ED of one of three hospitals with chest pain. AMI, defined according to 1979 WHO criteria (biomarker criteria not given) was diagnosed in 20.7% of patients. H-FABP was measured in plasma at presentation and every 4h for 16h using an experimental quantitative assay. Levels of H-FABP in patients with AMI were also compared to levels in 39 healthy individuals.	Prospective diagnostic cohort study	Mean H-FABP levels stratified by diagnosis	AMI 59mcg/l, non-cardiac chest pain 14mcg/l, healthy individuals 4.3mcg/l (P<0.001)	Outdated gold standard for AMI and the study protocol did not apparently mandate appropriately timed gold standard biomarker evaluation in order to confirm or exclude a diagnosis of AMI. Sensitivity of myoglobin and cTnI at presentation to the ED not reported for comparison with H-FABP. Optimal cut-off for H-FABP was not explored (the authors only used a fixed specificity of 95%, which is not how we would use the test clinically).
			Diagnostic performance of H-FABP, myoglobin at time of ED presenation	At a specificity of 95% sensitivity was: H-FABP 39%, myoglobin 28%
			AUC at 0-4h	H-FABP 0.80 (0.73 - 0.85); Myoglobin 0.73 (0.65 - 0.79); CK-MB 0.79 (0.72 - 0.85); cTnI 0.91 (0.87 - 0.94)
Valle et al 2008 Spain	419 patients who presented to one of 20 Spanish EDs in May 2004 with suspected ACS within 3 hours of symptom onset. Blood was drawn at the time of presentation for H-FABP and troponin T testing. H-FABP was tested using a qualitative point of care type assay. All patients underwent repeat troponin T testing 6-12h after admission. AMI was diagnosed in 35% of patients. H-FABP levels were considered positive at a cut-off of 7ng/ml.	Prospective diagnostic cohort study	Sensitivity for diagnosis of AMI	H-FABP 60%, troponin I 19% (P<0.05 for comparison)	Qualitative H-FABP assay but interobserver reliability not assessed. Gold standard for AMI not fully defined.
			Specificity for diagnosis of AMI	H-FABP 88%, troponin T 99% (not statistically significant)
			Sensitivity for final diagnosis of ACS	H-FABP 47%, troponin T 12% (P<0.05)
			Specificity for final diagnosis of ACS	H-FABP 94%, troponin T 100% (not statistically significant)
			Negative predictive values (NPV) for AMI and ACS	H-FABP had NPV 80% and 56% for AMI and ACS respectively. Troponin T had NPVs of 69% and 47%.
Nakata et al 2003 Japan	133 patients who presented to the ED with acute chest pain suggestive of ACS. HFABP was measured within 30 minutes of ED presentation and thereafter every six hours for 48 hours. 90/133 (68%) of patients were diagnosed with ACS.	Prospective diagnostic cohort study	Diagnosis of acute myocardial infarction (derived cut-off 9.5ng/ml)	Sensitivity 86.4% (95% CI 72.6 - 94.8), PPV 82%, NPV 54%. AUC 0.907, which was higher than myoglobin (0.860), troponin T (0.838) and CK-MB (0.880)	Outdated gold standard for AMI (CK-MB). Biomarker results were available to clinicians after 24 hours and may have been used to influence final diagnosis. High prevalence of ACS in the study population suggests that there was selection bias.
			Diagnosis of acute coronary syndromes	Sensitivity 84.5% (95% CI 72.6 - 92.6%), specificity 86.7% (76.8 - 93.4%).

Comment(s)

H-FABP is a cytoplasmic protein that is abundantly expressed in human myocardial cells. It is also found in much lower concentrations in skeletal muscle, kidney and brain tissue (Glatz et al. 2002). Experimental data first suggested that H-FABP may be a potential novel biomarker of AMI as early as 1988 (Glatz et al. 1988). In 1991 Tanaka et al reported elevated H-FABP levels in patients with AMI, with levels peaking earlier than CK-MB (Tanaka et al. 1991). Despite interest in H-FABP as an early marker of AMI for many years it has never gained widespread acceptance for use in clinical practice. Indeed the 2007 universal definition of myocardial infarction and the 2007 AHA/ACC and ESC guidelines for the management of NSTE-ACS do not even mention H-FABP, although they do discuss the use of other markers of myocardial necrosis including CK-MB, myoglobin and troponin I as well as a number of novel biomarkers. Several studies have investigated the utility of H-FABP for assisting in the diagnosis of AMI and for prognostic stratification. We identified six studies that investigated the diagnostic utility of H-FABP when used for the diagnosis of AMI at the time of presentation to the ED. Four of these studies utilised qualitative assays that are available as point of care tests. All six studies had significant weaknesses, with most studies employing now outdated gold standards for AMI diagnosis and being subject to significant verification bias. The data reporting in the small study by Alhashemi et al precludes calculation of total sensitivity and specificity. If the remainder of the results are pooled this would give H-FABP a total sensitivity of 71.6% (95% CI 67.8 – 75.3%) and a total specificity of 82.0% (79.6 – 84.3%). Excluding the studies by Ghani et al and Nakata et al, in which quantitative assays were used, the pooled sensitivity is 76.8% (72.6 – 80.5%), pooled specificity 72.5% (68.9 – 75.8%), pooled PPV 65.8% (61.6 – 69.8%) and pooled NPV 82.0% (78.6 – 85.0%). The positive likelihood ratio would be 2.79 and negative likelihood ratio 0.32. As such, if we were to apply the test in a typical United Kingdom ED population with suspected cardiac chest pain who have a prevalence of AMI of approximately 18% (Body R et al. 2008;Carley et al. 2005) the post-test probability of AMI given a normal H-FABP test would be 6.6%. This provides similar predictive value to a normal ECG in this cohort (Panju AA et al. 1998) but is still far from excluding the diagnosis.

Clinical Bottom Line

H-FABP is a promising biomarker for early exclusion of AMI in the ED but cannot be used alone to rule out the diagnosis. Multivariate analysis is likely to be necessary in order to evaluate any potential role for H-FABP in the evaluation of the ED patient with suspected cardiac chest pain.

Level of Evidence

Level 2 - Studies considered were neither 1 or 3.

References

1. Mad P; Domanovits H; Fazelnia C; Stiassny K; Rassmuller G; Cseh A; Sodeck G; Binder T; Christ G; Szekeres T; Laggner A; Herkner H Human heart-type fatty-acid-binding protein as a point-of-care test in the early diagnosis of acute myocardial infarction QJM 2007; 100: 203-210
2. Alhashemi JA Diagnostic accuracy of a bedside qualitative immunochromatographic test for acute myocardial infarction American Journal of Emergency Medicine 2006; 24(2): 149-155
3. Seino Y; Ogata K; Takano T; Ishil J; Hishida H; Morita H; Takeshita H; Takagi Y; Sugiyama H; Tanaka T; Kitaura Y Use of a whole blood rapid panel test for heart-type fatty acid-binding protein in patients with acute chest pain: comparison with rapid troponin T and myoglobin tests American Journal of Medicine 2003; 115(3): 185-190
4. Ghani F; Wu AH; Graff L; Petry C; Armstrong G; Prigent F; Brown M Role of heart-type fatty acid-binding protein in early detection of acute myocardial infarction Clinical Chemistry 2000; 46(5): 718-719
5. Valle HA; Riesgo LGC; Bel MS; Gonazalo FE; Sanchez MS; Oliva LI Clinical assessment of heart-type fatty acid binding protein in early diagnosis of acute coronary syndrome European Journal of Emergency Medicine 2008; 15(3): 140-144
6. Nakata T; Hashimoto A; Hase M; Tschuchihashi K; Shimamoto K Human heart-type fatty acid-binding protein as an early diagnostic and prognostic marker in acute coronary syndrome [Abstr] Cardiology 2003; 99(2): 96-104
7. Body R, McDowell G, Carley S, & Mackway-Jones K Do risk factors for chronic coronary heart disease help diagnose acute myocardial infarction in the Emergency Department? Resuscitation 2008; 79: 41-45
8. Glatz JFC, van der Voort D, Hermens WT Fatty acid-binding protein as the earliest available plasma marker of acute myocardial injury Journal of Clinical Ligand Assay 2002; 25(2):167-177
9. Glatz JFC, van Bilsen M, Paulussen RJA, Veerkamp JH, van der Vusse GJ, Reneman RS Release of fatty acid-binding protein from isolated rat heart subjected to ischemia and reperfusion or to the calcium paradox Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism 1988; 961(1):148-152
10. Tanaka T, Hirota Y, Sohmiya K, Nishimura S, Kawamura K and urinary human heart fatty acid-binding protein in acute myocardial infarction Clinical Biochemistry 1991; 24(2): 195-201
11. Panju AA, Hemmelgam BR, Guyatt GH, Simei DL The rational clinical examination: Is this patient having a myocardial infarction JAMA 1998; 280(14): 1256-1263
12. Carley SD; Jenkins M; Mackway-Jones K Body surface mapping versus the standard 12 lead ECG in the detection of myocardial infarction amongst Emergency Department patients: a Bayesian approach Resuscitation 2005; 64: 309-14