Prognostic Value of Serum Concentration of Heart-TypeFatty Acid-Binding Protein

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 SyndromeJunnichi Ishii,1*Yukio Ozaki,2Jingchao Lu,2Fumihiko Kitagawa,3Takahiro Kuno,3 Tadashi Nakano,2Yuu Nakamura,2Hiroyuki Naruse,2Yoshihisa Mori,2Shigeru Matsui,2Hisaji Oshima,3Masanori Nomura,2Kouji Ezaki,2andHitoshi Hishida2Background:Heart-type fatty acid–binding protein(H-FABP)is proposed as an early biomarker for acute myocardial infarction(AMI),but its prognostic value is unclear in acute coronary syndrome(ACS).We evaluated the prognostic value of the H-FABP concentration relative to cardiac troponin T(cTnT)in the early hours of ACS. Methods:Serum concentrations of H-FABP and cTnT were measured on admission in328consecutive patients hospitalized for ACS within6h after the onset of chest pain[AMI,241(73.5%)patients;ST-segment elevation myocardial infarction,154(47.0%)patients;and emer-gent coronary angiography within24h after admission, 287(87.5%)patients].Cardiac events,which were de-fined as cardiac death or subsequent nonfatal AMI, were monitored for6months after admission. Results:During the6-month follow-up period,there were 25cardiac events,including15cardiac deaths and10 subsequent nonfatal AMIs.Stepwise multivariate analy-ses including clinical,electrocardiographic,and biochem-ical variables revealed that increased H-FABP(above the median of9.8␮g/L),but not increased cTnT(above the median of0.02␮g/L),was independently associated with cardiac events in all patients[relative risk(RR)؍8.96;P؍0.0004],the subgroup of patients with ST-segment eleva-tion myocardial infarction(RR؍11.3;P؍0.02),and the subgroup of patients with unstable angina and non-ST-segment elevation myocardial infarction(RR؍8.31;P؍0.007).The area under the ROC curve was higher for H-FABP than for cTnT(0.711vs0.578;P؍0.08),suggesting that H-FABP concentrations have a greater predictive ca-pacity for cardiac events than cTnT.Conclusion:Serum H-FABP is a potential independent predictor of cardiac events within6months of patient admission and may provide prognostic information su-perior to cTnT in the early hours of ACS.©2005American Association for Clinical ChemistryHeart-type fatty acid–binding protein(H-FABP)4is a low–molecular-mass(14–15kDa)cytoplasmic protein that,along with myoglobin,is among the earliest markers released into circulation after myocardial damage.The H-FABP content of skeletal muscle is only10%–20%of that found in cardiac muscle,whereas the skeletal muscle content of myoglobin is approximately twice that of cardiac muscle(1–3).Along with other groups,we have shown that quantitative measurement of the serum H-FABP concentration is more sensitive and specific than myoglobin for the early diagnosis of acute myocardial infarction(AMI)(4,5).The value of serum H-FABP1Division of Critical Care,Fujita Health University Graduate School of Health Sciences,Toyoake,Japan.2Department of Internal Medicine,Fujita Health University School of Medicine,Toyoake,Japan.3Department of Joint Research Laboratory of Clinical Medicine,Fujita Health University Hospital,Toyoake,Japan.*Address correspondence to this author at:Division of Critical Care,Fujita Health University Graduate School of Health Sciences,1-98Dengakugakubo, Kutsukake-cho,Toyoake470-1192,Japan.Fax81-562-93-2315;e-mail jishii@fujita-hu.ac.jp.Received December31,2004;accepted May3,2005.Previously published online at DOI:10.1373/clinchem.2004.0476624Nonstandard abbreviations:H-FABP,heart-type fatty acid–binding pro-tein;AMI,acute myocardial infarction;ACS,acute coronary syndrome;cTnT, cardiac troponin T;CK-MB,creatine kinase MB isoenzyme;CI,confidence interval;STEMI,ST-segment elevation myocardial infarction;and UA/ NSTEMI,unstable angina and non-ST-segment elevation myocardial infarc-tion.Clinical Chemistry51:81397–1404(2005)Proteomics andProtein Markers1397measurements in the early diagnosis of AMI was high-lighted recently by a prospective multicenter trial in Japan for a newly developed qualitative whole-blood rapid panel test for H-FABP that positively detects serum H-FABP concentrationsՆ6.2␮g/L(6).The prognostic value of serum H-FABP concentrations has not been established for acute coronary syndrome(ACS).The present study was designed to prospectively as-sess the prognostic value of serum concentrations of H-FABP relative to cardiac troponin T(cTnT)determined on admission for adverse cardiac outcomes within6 months in patients presenting with ACS in the first6h after the onset of chest pain.Materials and Methodsstudy populationBetween March2000and December2001,a total of328 consecutive patients admitted to the coronary care unit of Fujita Health University for ACS within6h after the onset of chest pain were enrolled in this study.Patients were eligible for inclusion if they had,within the6h before admission,an episode of resting anginal pain lastingϾ10 min and at least one of the following:ST-segment eleva-tion of at least0.05mV,ST-segment depression of at least 0.05mV,T-wave inversion of at least0.1mV in at least2 contiguous leads,or presumed new left-bundle branch block.Patients with a history of coronary revasculariza-tion within the preceding6months or a serum creatinine concentrationϾ15mg/L were rmed consent was obtained from all patients before participation,and the protocol was approved by the local ethics committee.All patients were treated according to routine clinical protocols.Early invasive management,including emergent coronary angiography within24h of admission followed by percutaneous coronary intervention,was routinely per-formed in our institution.Demographics and clinical data, including age,sex,diagnosis,revascularization,coronary risk factors,and Killip class(7)on admission,were collected from hospital medical records and patient interviews. laboratory analysesVenous whole blood samples were drawn and centri-fuged within30min of collection,and the serum obtained was stored atϪ70°C.H-FABP was measured by2-step direct sandwich ELISA(Market-M H-FABP;Dainippon Pharmaceutical).The calibrators for the ELISA covered the range0–250␮g/L.The lower limit of detection for the H-FABP assay was1.25␮g/L,and a value of6.2␮g/L has been used for an AMI cutoff(5,8).The assay was linear within the range0–220␮g/L.Myoglobin(200mg/L)and myosin(84mg/L)did not cross-react in the H-FABP assay. The total imprecision for the H-FABP assay was8.2%at6.2␮g/L,7.9%at12.0␮g/L,and5.6%at33.7␮g/L.cTnT and creatine kinase MB(CK-MB)were deter-mined routinely at admission,as well as16or24h after admission.For the cTnT measurements(Elecsys2010 analyzer;Roche Diagnostics),the manufacturer-stated de-tection limit and cutoff for AMI were0.01and0.1␮g/L, respectively.The total imprecision for the cTnT assay was14%at0.012␮g/L and8.7%at0.024␮g/L.For the CK-MB measurements(Dade Dimension RxL Max analyzer;Di-mension Flex reagent cartridge MMB;Dade Behring),thedetection limit and upper limit of the reference interval were0.5and3.6␮g/L,respectively.The total imprecision for the CK-MB assay was4.6%at2.9␮g/L and2.3%at9.1␮g/L. study end point and follow-upCardiac events were defined as cardiac death or subse-quent nonfatal AMI.Cardiac death was defined as anydeath for which there was no clearly documented noncar-diac cause.Sudden death occurring outside the hospitalfor which no other cause was assigned was also regardedas cardiac death.Patients were defined as having admis-sion AMI when serum cTnT concentrations wereՆ0.1␮g/L in blood samples collected within24h of admission. Only a new AMI occurringϾ24h after admission wasconsidered a subsequent AMI.Patients were defined ashaving a subsequent AMI if they presented with symp-toms suggestive of ACS and typical increased concentra-tions of biochemical markers with CK-MB greater thanthe upper reference limit(and a CK-MB increaseՆ50%from the previous concentration in the case of an AMIoccurring soon after the admission AMI).The study endpoint was a cardiac event at6monthsafter admission.A review of medical records and fol-low-up telephone interviews were conducted to surveyfor patient cardiac events.The review of case notes forfollow-up was performed by2of the authors(J.L.andH.N.),who were blinded to H-FABP concentrations. statistical analysisContinuous variables were analyzed(Ver.11.0,StatisticalPackage for the Social Sciences)with the Mann–WhitneyU-test,Wilcoxon paired sign-rank test,or linear regres-sion analysis,and data were expressed as mean(SD)ormedian(25th–75th percentiles).Values below the detec-tion limit of the assay were defined as zero.Categoricalvariables were compared by the␹2test.The relative risk with the95%confidence interval(CI)is presented.Uni-variate and stepwise multivariate Cox regression analyses were used to evaluate the prognostic value of variables. The log-rank test was performed for the Kaplan–Meier probability estimates.The ROC curve was used to assess the discriminatory ability of biomarkers(9).A univariate Z-test was used to compare the area under the ROC curve of biomarkers,as described by Hanley and McNeil(10).P Ͻ0.05was considered significant.ResultsOf the328patients(Table1),241(73.5%)were diagnosed with AMI at admission.On the basis of electrocardio-graphic findings on admission,154patients who had a persistent(Ͼ30min)ST-segment elevation of at least0.05 mV in at least2contiguous leads or presumed new1398Ishii et al.:FABP in ACSleft-bundle branch block were assigned to the ST-segmentelevation myocardial infarction(STEMI)group,and theother174patients were placed in the unstable angina andnon-ST-segment elevation myocardial infarction(UA/NSTEMI)group.Mean(SD)serum concentrations of bothH-FABP[63.8(92.4)␮g/L vs25.2(46.4)␮g/L;PϽ0.0001] and cTnT[0.552(1.77)␮g/L vs0.093(0.211)␮g/L;P Ͻ0.0001]on admission in the STEMI group were signifi-cantly higher than those in the UA/NSTEMI group.Ofthe patients with STEMI,139(90.3%)underwent emer-gent coronary angiography within24h after admission[median of0.8(0.4–1.1)h after admission],with123(79.9%)of these subsequently undergoing coronaryrevascularization.Of the patients with UA/NSTEMI,148(85.1%)underwent emergent coronary angiographywithin24h after admission[median of3.2(1.0–17.5)h],with76(43.7%)of these subsequently undergoing coro-nary revascularization.The serum concentration of H-FABP on admission was Ͻ1.25␮g/L in3patients(0.9%), 1.25–6.2␮g/L in 113(34.5%),andՆ6.2␮g/L in212(64.6%).The cTnT concentration on admission wasϽ0.01␮g/L in136pa-tients(41.5%),0.01–0.1␮g/L in95(29.0%),andՆ0.1␮g/L in97(29.6%).A significantly higher percentage of patients had H-FABP concentrationsՆ6.2␮g/L(64.6%) than cTnT concentrationsՆ0.01␮g/L(58.5%;Pϭ0.02)or Ն0.1␮g/L(29.6%;PϽ0.0001),suggesting that H-FABP concentration is a more sensitive indication than cTnT of myocardial damage in patients with ACS within6h after the onset of chest pain.The time from the onset of chest pain to admission correlated significantly with H-FABP concentration(rϭ0.197;Pϭ0.0003),but not with cTnT concentration.There was a weak but significant correlation between H-FABP and cTnT concentrations at admission(rϭ0.17;Pϭ0.002).During the6months after admission,there was 1(0.3%)noncardiac death(gastric cancer)and25(7.6%) cardiac events,including15(4.6%)cardiac deaths and 10(3.0%)subsequent nonfatal AMIs.The causes of cardiac death were fatal AMI in10patients,heart failure in3 patients,and sudden death occurring outside hospital in2 patients.A comparison of patients who had subsequent cardiac events with those who did not revealed that cardiac events were associated with patient age(mean age,69.8vs 64.5years;Pϭ0.02),higher serum concentrations of H-FABP(119vs46.6␮g/L;Pϭ0.0004)and cTnT(1.72vs 0.45␮g/L;Pϭ0.01),AMI(96%vs71.6%;Pϭ0.008)or anterior AMI(56%vs30%;Pϭ0.01),hypertension(72% vs53.5%;Pϭ0.09),and a tendency not to receive emergent coronary angiography within24h after admis-sion(64%vs89.4%;Pϭ0.001;Table1).A comparison of patients who experienced a cardiac event and those whoparison of demographics and clinical characteristics for328ACS patients who did or did not have acardiac event.All patients(n؍328)No cardiac event(n؍303)Cardiac event(n؍25)P Mean(SD)age,years64.9(10.4)64.5(10.4)69.8(9.9)0.02 Men,n(%)264(80.5)246(81.2)18(72.0)NS a Mean(SD)time from onset,h 3.2(1.6) 3.2(1.6) 3.3(1.6)NS AMI,n(%)241(73.5)217(71.6)24(96.0)0.008 STEMI,n(%)154(47.0)139(45.9)15(60.0)NS AdmissionMean(SD)heart rate,beats/min76(17)75(16)82(19)NS Mean(SD)systolic BP,a mmHg138(28)138(28)141(27)NS Mean(SD)diastolic BP,mmHg73(17)73(17)76(15)NS Mean(SD)serum H-FABP,␮g/L52.1(99.2)46.6(86.5)119(187)0.0004 Mean(SD)serum cTnT,␮g/L0.54(3.22)0.45(2.5) 1.72(7.83)0.01 Mean(SD)serum creatinine,mg/L7.9(3.2)7.8(3.1)8.6(3.6)NS Killip classϾ1,n(%)10(3.0)8(2.6)2(8)NS Anterior AMI,n(%)105(32.2)91(30)14(56)0.01 Previous historyMyocardial infarction,n(%)58(17.6)51(16.8)7(28)NS Heart failure,n(%)3(0.9)3(1)0(0)NS Coronary risk factorsDiabetes,n(%)105(32)96(31.7)9(36)NS Hypertension,n(%)180(54.9)162(53.5)18(72)0.09 Hyperlipidemia,n(%)145(44.2)133(43.9)12(48)NS Current smoking,n(%)178(54.3)165(54.5)13(52)NS Emergent coronary angiography,n(%)287(87.5)271(89.4)16(64)0.001 PCI,n(%)182(55.5)168(55.4)14(56)NS CABG,n(%)17(5.2)16(5.3)1(4)NSa NS,not significant;BP,blood pressure;PCI,percutaneous coronary intervention;CABG,coronary artery bypass graft surgery.Clinical Chemistry51,No.8,20051399did not revealed that there was no difference in the number of patients who had diabetes or hyperlipidemia,currently smoked,or had received early coronary revas-cularization.Kaplan–Meier analyses showed that patients with H-FABP concentrations above the median value of 9.8␮g/L on admission had a significantly higher risk of cardiac mortality (P ϭ0.0006)and cardiac events (P Ͻ0.0001)within 6months after admission compared with those who did not (Fig.1).Although weaker (P ϭ0.018and 0.012,respectively),a similar relationship was observed for H-FABP Ն6.2␮g/L;we therefore used the median H-FABP value of 9.8␮g/L as the cutoff for predicting cardiac death and cardiac events.Patients with increased H-FABP (above the median value of 9.8␮g/L)on admission had a significantly higher risk of cardiac death and cardiac events within 30days (P ϭ0.002and 0.006,respectively)and 6months (P ϭ0.0008and P Ͻ0.0001,respectively)after admission com-pared with those who did not (Table 2).Although weaker,a similar relationship was observed for increased cTnT (above the median value of 0.02␮g/L;Table 3).The relationship between increased H-FABP and car-diac death or cardiac events within 6months after admis-sion in patients with or without increased cTnT is shown in Fig.2.Increased H-FABP was significantly associated with an increased risk of cardiac events within 6months after admission in patients with or without increased cTnT concentrations (P ϭ0.0001and 0.003,respectively).There was a greater tendency for cardiac death within 6months after admission among patients with increased cTnT and H-FABP concentrations compared with those patients with only increased cTnT concentrations (P ϭ0.07).In the subgroup of patients with STEMI or UA/NSTEMI,patients with increased H-FABP also had a higher risk of cardiac events within 6months after admis-sion compared with those who did not (P ϭ0.04and 0.005,respectively;Table 2).There was no significantdifference in the tendency for cardiac event risk within 6months after admission between patients with and with-out increased cTnT (Table 3).The relationship between early invasive management and incidence of cardiac events within 6months after admission for UA/NSTEMI patients with or without increased H-FABP or cTnT is shown in Fig.3.Patients with increased H-FABP who received early invasive man-agement had a significantly lower risk of cardiac events within 6months after admission compared with those who did not (P ϭ0.015).However,the risk of cardiac events within 6months after admission did notdifferFig.1.Kaplan–Meier curves showing incidence of cardiac death (A )and cardiac events (B )within 6months after admission according to H-FABP (above vs at or below the median value of 9.8␮g/L).Table 2.Risk stratification of ACS patients based on increased H-FABP (above the median value of 9.8␮g/L).H-FABP,␮g/L P>9.8<9.8All patients (n ϭ328)n16416430-Day cardiac mortality,n (%)13(7.9)1(0.6)0.00230-Day cardiac event,n (%)15(9.1)3(1.8)0.0066-Month cardiac mortality,n (%)14(8.5)1(0.7)0.00086-Month cardiac event,n (%)22(13.4)3(1.8)Ͻ0.0001STEMI (n ϭ154)n10648Coronary revascularization,n (%)82(77.3)41(85.4)NS a 30-Day cardiac mortality,n (%)11(10.4)1(2.1)0.130-Day cardiac event,n (%)12(11.3)1(2.1)0.066-Month cardiac mortality,n (%)12(11.3)1(2.1)0.066-Month cardiac event,n (%)14(13.2)1(2.1)0.04UA/NSTEMI (n ϭ174)n58116Coronary revascularization,n (%)39(67.2)37(31.9)Ͻ0.000130-Day cardiac mortality,n (%)2(3.4)0(0)NS 30-Day cardiac event,n (%)3(5.2)2(1.7)NS 6-Month cardiac mortality,n (%)2(3.4)0(0)NS 6-Month cardiac event,n (%)8(13.8)2(1.7)0.005aNS,not significant.1400Ishii et al.:FABP in ACSsignificantly between patients without increased H-FABP who received early invasive management and those who did not.Similar results were observed for increased cTnT.Univariate Cox proportional regression analyses re-vealed that increased H-FABP and cTnT were signifi-cantly associated with cardiac death (P ϭ0.009and 0.01,respectively)and cardiac events (P ϭ0.0009and 0.03,respectively;Table 4).Stepwise multivariate analyses in-cluding age,gender,time from the onset of chest pain,STEMI,increased H-FABP,increased cTnT,creatinineconcentration,Killip class Ͼ1,anterior AMI,and previous history of myocardial infarction revealed that increased H-FABP,but not increased cTnT,was independently associated with cardiac death (P ϭ0.04)and cardiac events (P ϭ0.0004)within 6months after admission (Table 5).Similar to all patients,for patients with STEMI stepwise multivariate analyses including age,gender,time from the onset of chest pain,increased H-FABP,increased cTnT,creatinine concentration,Killip class Ͼ1,anterior AMI,and previous history of myocardial infarc-tion revealed that increased H-FABP (P ϭ0.02),age (P ϭ0.004),and a history of myocardial infarction (P ϭ0.006)were independent predictors of cardiac events within 6months after admission (Table 5).For patients with UA/NSTEMI,only increased H-FABP was independently as-sociated with cardiac events within 6months after admis-sion (P ϭ0.007).Comparison of the areas under the ROC curves for H-FABP and cTnT revealed that H-FABP (0.711;95%CI,0.810–0.612)was a better predictor of cardiac events within 6months after admission than cTnT (0.578;95%CI,0.695–0.462;P ϭ0.08;Fig.4A).The area under the H-FABP ROC curve for patients with STEMI was 0.684(95%CI,0.814–0.553),and the H-FABP concentra-tion associated with maximum sensitivity (60%)and specificity (65%)was 35␮g/L (Fig.4B).For patients with UA/NSTEMI,the area under the H-FABP ROC curve was 0.730(95%CI,0.893–0.567),and the H-FABP concentra-tion associated with maximum sensitivity (80.0%)and specificity (71%)was 10.4␮g/L.DiscussionThis prospective study demonstrated that measurement of serum H-FABP concentrations on admission provides important and independent information for risk stratifi-cation of patients presenting with ACS in the first 6h afterTable 3.Risk stratification of ACS patients based on increased cTnT (above the median value of 0.02␮g/L).cTnT,␮g/L P>0.02<0.02All patients (n ϭ328)n16416430-Day cardiac mortality,n (%)12(7.3)2(1.2)0.0130-Day cardiac event,n (%)13(7.9)5(3.0)0.096-Month cardiac mortality,n (%)13(7.9)2(1.2)0.0066-Month cardiac event,n (%)18(11.0)7(4.3)0.04STEMI (n ϭ154)n9658Coronary revascularization,n (%)74(77.1)49(84.5)NS a 30-Day cardiac mortality,n (%)10(10.4)2(3.4)NS 30-Day cardiac event,n (%)10(10.4)3(5.2)NS 6-Month cardiac mortality,n (%)11(11.5)2(3.4)NS 6-Month cardiac event,n (%)12(12.5)3(5.2)NSUA/NSTEMI (n ϭ174)n68106Coronary revascularization,n (%)41(60.3)35(33.0)0.000530-Day cardiac mortality,n (%)2(2.9)0(0)NS 30-Day cardiac event,n (%)3(4.4)2(1.9)NS 6-Month cardiac mortality,n (%)2(2.9)0(0)NS 6-Month cardiac event,n (%)6(8.8)4(3.8)NSaNS,notsignificant.Fig.2.Association of increased H-FABP concentra-tions (above the median value of 9.8␮g/L)with cardiac death (A )and cardiac events (B )within 6months after admission in patients with or without increased cTnT (above the median value of 0.02␮g/L).Clinical Chemistry 51,No.8,20051401the onset of chest pain.To the best of our knowledge,this is the first time that an increase in the concentration of serum H-FABP early after symptomatic onset has been shown to have a profound and independent impact on adverse outcome within 6months of admission for pa-tients with STEMI or UA/NSTEMI.Such findings may have important implications for immediate management of high-risk patients in the early hours of ACS.Although there was a relationship between increased serum cTnT concentrations and adverse outcomes within 6months after admission,the predictive ability of H-FABP concentration was higher.Increased H-FABP con-centrations,but not cTnT concentrations,retained inde-pendent significance based on a stepwise Cox regression analysis.Supporting these findings,ROC curve analyses revealed that H-FABP concentrations tended to have a higher predictive ability than cTnT concentrations.Fur-thermore,increased H-FABP was associated with an increased risk of cardiac events within 6months after admission independent of increased cTnT (Fig.2).These findings show that serum H-FABP concentrations provide prognostic information superior to that provided by cTnT concentrations for patients presenting with ACS in the first 6h after the onset of chest pain.The superiority of H-FABP over cTnT may result from its higher sensitivity for the detection of myocardial damage in the early hours of ACS.A recent report on 134patients with suspected ACS,including 74(55%)patients presenting within 6h after symptom onset,showed that H-FABP Ն6.2␮g/L had higher sensitivities than cTnT Ն0.1␮g/L for identifying patients requiring emergency hospitalization (64%vs 43%;P Ͻ0.05),emergency angiography (63%vs 41%;P Ͻ0.05),and early coronary intervention within 7days after admission (78%vs 58%;P Ͻ0.05)(11).This study,however,showed no correlation of H-FABP and subsequent AMI or death because the authors reported no subsequent AMI and death during the 1-month follow-up period.These differences between the 2studies might result from differ-ences in population sizes (larger in our study)and the severity of patient condition (more severe in our study).Given the diagnostic and prognostic properties of cTnT for patients with ACS during the first 6h after the onset of chest symptoms (12–14),our findings clearly demonstrate the need to consider either serum H-FABP or cTnT for diagnosis of AMI and risk stratification,depending on the ischemic interval at presentation.Our findings indicate that if patients present within 6h after the onset of chest pain,measurement of serum H-FABP concentrationspro-Fig.3.Association of early invasive management with cardiac events within 6months after admission in UA/NSTEMI patients with or without increased H-FABP (above the median value of 9.8␮g/L;A )or increased cTnT (above the median value of 0.02␮g/L;B ).Table 4.Univariate predictors of cardiac death and cardiac events within 6months after admission in 328ACS patients.VariablesCardiac deathCardiac event RR a95%CIPRR95%CIPAge (years)1.09 1.03–1.160.004 1.05 1.01–1.100.02Gender (male ϭ1) 3.75 1.36–10.40.01 1.660.70–3.98NS Time from onset (h)0.950.67–1.32NS 1.040.82–1.32NS STEMI (yes ϭ1)7.58 1.71–33.60.008 1.750.79–3.89NS H-FABP Ͼ9.8␮g/L (yes ϭ1)14.5 1.91–110.50.0097.70 2.30–25.70.0009cTnT Ͼ0.02␮g/L (yes ϭ1) 6.69 1.51–29.60.01 2.65 1.11–6.340.03Serum creatinine (mg/L) 1.110.98–1.250.10 1.060.96–1.18NS Killip class Ͼ1(yes ϭ1) 5.19 1.17–23.00.03 2.990.70–12.7NS Anterior AMI (yes ϭ1)1.880.68–5.18NS2.78 1.26–6.120.01History of myocardial infarction (yes ϭ1)1.170.33–4.13NS 1.850.77–4.43NSaRR,relative risk;NS,not significant.1402Ishii et al.:FABP in ACSvides a more appropriate platform for diagnosis and prognosis than measurement of cTnT concentrations;con-versely,if patients present later than the first 6h after the onset of chest pain,measurement of serum cTnT concen-trations is appropriate.Because accurately predicting the ischemic interval in ACS is difficult (15,16),a combina-tion of serum H-FABP and cTnT measurements would provide a better early risk assessment for patients with ACS than would measurement of either of these markers rger prospective studies would allow the effec-tiveness of combined H-FABP and cTnT serum measure-ments for patients presenting with ACS to be ascertained.Recent multicenter studies demonstrated the use of cTnT concentrations to identify UA/NSTEMI patients,who par-ticularly benefit from early invasive management (17,18).By comparing the serum H-FABP concentrations of UA/NSTEMI patients,the present study showed that early invasive management reduced cardiac event rates within 6months after admission in patients with increased serum H-FABP on admission but had no significant effect when H-FABP concentrations were not increased (Fig.3).These results provide evidence that measurement of serum H-FABP concentrations on admission would allow identifica-tion of high-risk patients presenting in the early hours of ACS,who would benefit greatly from an early invasive strategy.A disadvantage of this study is that because the population of patients who did not receive early invasive management was small,the treatments were not rger prospective studies would confirm the results from this study and clarify the utility of H-FABP for identi-fying appropriate therapeutic options.This study highlights the advantages of using serum H-FABP concentrations for risk assessment of patients only with a high clinical probability of ACS.Further investigations are required to explore the application of this approach to a broad spectrum of patients presenting with chest pain and a low to intermediate probability of ACS.Because H-FABP has a very low molecular mass,its elimination rate is partly determined by the patient’s kidney function (19).It would be interesting to determine whether our findings might apply to patients with chronic renal failure.To minimize the effects of renal clearance,however,we excluded patients with serum creatinine concentrations Ͼ15mg/L.In addition,it would be useful to calculate the H-FABP/creatinine concentration ratio in serum.Further studies are needed to clarify the useful-ness of this ratio in patients with ACS.A further limitation of this study is the potential increase in H-FABP concen-trations in patients with skeletal muscle damage (3,4)or heart failure (20).Recent studies have also suggested that H-FABP concentrations may be increased in patients with minor brain injuries (21)or neurodegenerative disease (22).It is important to ascertain whether the prognosticTable 5.Multivariate predictors of cardiac death and cardiac events within 6months after admission in 328ACS patients.VariablesRR a95%CIPCardiac death within 6months All patients (n ϭ328)H-FABP Ͼ9.8␮g/L (yes ϭ1)8.92 1.15–69.40.04Age (years)1.09 1.03–1.150.002STEMI (yes ϭ1)5.22 1.16–23.50.03Cardiac event within 6months All patients (n ϭ328)H-FABP Ͼ9.8␮g/L (yes ϭ1)8.96 2.64–30.40.0004Age (years)1.05 1.00–1.090.03History of myocardial infarction (yes ϭ1)2.52 1.03–6.160.04STEMI (n ϭ154)H-FABP Ͼ9.8␮g/L (yes ϭ1)11.3 1.41–90.60.02Age (years)1.08 1.02–1.140.004History of myocardial infarction (yes ϭ1)5.41 1.64–17.80.006UA/NSTEMI (n ϭ174)H-FABP Ͼ9.8␮g/L (yes ϭ1)8.31 1.76–39.10.007aRR,relativerisk.Fig. 4.ROC curves for cardiac events within 6months after admission based on increased H-FABP and cTnT concentrations in the study population (A )and for increased H-FABP in patients with STEMI or UA/NSTEMI (B ).Clinical Chemistry 51,No.8,20051403。