Clncal Chemstry / EARLY DIAGNOSIS OF ACUTE MYOCARDIAL INFARCTION USING IOCHEMICAL MARKERS Strateges for the Early Dagnoss of Acute Myocardal Infarcton Usng ochemcal Markers Martna Zannotto, Leopoldo Celegon, olsc, SaraAltner, MD,3 andmaro olsc,2mattalachn, Pleban, MD, MD'2 Key Words: Troponn I; Myoglobn; Creatne knase; Creatne knase M; Senstvty; Specfcty; Multple testng; Dagnostc strateges Abstract We evaluated dfferent dagnostc strateges for the early dagnoss of acute myocardal nfarcton, combnng senstvty and specfcty of dfferent markers evaluated sngly and usng combnaton testng n parallel and seral modes. Myoglobn, cardac troponn I (Tnl), creatne knase (CK), and CK-M mass were tested n blood samples from 26 patents wth acute myocardal nfarcton collected at admsson (TQ; mean = 3.3 hours from the onset of chest pan) and 3 and 6 hours later. The comparson group was made up of 70 patents wth renal falure, skeletal muscle dseases, stable angna, unstable angna, and chest pan of nonschemc orgn. Sngle tests showed dfferent senstvtes n relaton to the dfferent release knetcs; myoglobn was the most senstve (69% at TQ) although less specfc (46%), and Tnl showed the hghest specfcty (90%) and a senstvty of 54%. Combnaton testng n a parallel mode usng myoglobn and Tnl or CK-M had the same senstvty and specfcty as myoglobn tested sngly. The best combnaton n a seral mode s myoglobn and Tnl (at TQ senstvty, 54%; specfcty, 98%), as confrmed by the analyss of the postve predctve value, the negatve predctve value, and the accuracy evaluated as a functon of dfferent dsease prevalences. Testng for the dagnoss of acute myocardal nfarcton (AMI) s rapdly changng from the tradtonal enzymatc assay of total creatne knase (CK), CK-M (actvty), and lactate dehydrogenase, to the mass measurement of myoglobn, CKM, and troponn T and troponn I (Tnl)"* to acheve hgh dagnostc senstvty and specfcty wthn a few hours from the onset of tssue necross; the fnal goal s the early and approprate management of patents' condtons.5-7 However, the wde array of mproved analytc procedures for measurng prevously establshed analytes and the ntroducton of new markers forces us to make choces to adopt an effcent and cost-effectve approach assocated wth a ratonal use of laboratory resources, as well as to reduce the number of tests needed n an emergency stuaton.8-0 Therefore, we evaluated dfferent dagnostc strateges for the early dagnoss of AMI usng new and tradtonal bochemcal markers. In partcular, we compared the dagnostc senstvty and specfcty obtaned by measurng dfferent bochemcal markers, some wdely used, such as total CK and ts soenzyme M, and others recently proposed (myoglobn and Tnl), performed as sngle tests and combned n parallel and seral modes.'2 Materals and Methods Patents and lood Collectons We ncluded 26 patents (9 men and 7 women, aged 45 to 83 years) wth AMI dagnosed accordng to World Health Organzaton crtera (characterstc chest pan, unequvocal electrocardographc changes, and seral ncreases n CK and CK-M mass concentraton wth peak values twce the upper lmt of the reference nterval); 8 had Q-wave and 8 non-q-wave nfarctons. The patents were admtted to the coronary care unt wthn 6 hours Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208 Am J Cln Pathol 999; :399-405 3 9 9
Zannotto et al / EARLY DIAGNOSIS OF ACUTE MYOCARDIAL INFARCTIONUSING IOCHEMICAL MARKERS from the onset of symptoms (range, 0.5-6.0 hours; medan, 3 hours); none were recevng thrombolytc treatment. lood samples were collected mmedately on admsson to the coronary care unt (T0; mean = 3.3 hours; range, 30 mnutes to 6 hours from the onset of chest pan) and then 3 (T,), 6 (T2), 9 (T,), and 2 (T4) hours later. All patents gave ther nformed consent for extra blood samples to be obtaned. To evaluate the specfcty of bochemcal markers and ther negatve predctve value n rulng out AMI n a worstcase stuaton, we studed, as a comparson group, 70 patents wth the followng dagnoses accordng to bochemcal and nstrumental data: chest pan of nonschemc orgn, 20; chronc renal falure, 3; chronc skeletal muscle dsease, ; stable angna, ; unstable angna, 9; and multple trauma wthout chest contuson, 6. Only blood sample was obtaned from each of these patents. Methods Total CK Actvty The catalytc concentraton of CK was determned at 37 C (racco, Mlano, Italy) accordng to the method recommended by the Internatonal Federaton of Clncal Chemstry.3 The upper reference lmts were 60 U/L for females and 90 U/L for males. CK-M Mass The mass concentraton of CK-M was measured by usng a commercally avalable fluorometrc enzyme mmunoassay (Stratus CK-M, axter Dade, Mlano, Italy) wth a senstvty of Lg/L. The upper reference lmt was 5 Ltg/L.4 Myoglobn The myoglobn concentraton was measured wth a fluoroenzymometrc method (Stratus Myoglobn, axter Dade) n whch the upper lmt of the reference range was 70 Ltg/L.9 Troponn I The assay of Tnl was performed wth a fluorometrc assay (Stratus Troponn I, axter Dade) wth a senstvty of 0.35 ng/ml (0.35 u.g/l). The upper reference lmt was ng/ml ( u.g/l).5 Data Analyss In the present study, as suggested by other authors,67 combnaton testng n a parallel mode defned as postve any fndngs of values hgher than the upper reference value. Combnaton testng n a seral mode makes use of the frst screenng test, usng the second test only f samples are postve and consderng a result postve f both test results are postve. 400 Am J Cln Pathol 999; :399-405 Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208 The 95% confdence ntervals (CI) were calculated by usng the Confdence Interval Analyss mcrocomputer program,8 to mprove the statstcal nformaton and to provde an explct demonstraton of the magntude of uncertanty of all reported results. Ths statstcal approach was of partcular value n the present study because of the small number of subjects studed. Test senstvty, specfcty, postve and negatve predctve value, and accuracy were determned as descrbed elsewhere.9 Results As prevously demonstrated and as expected from a pathophysologc vewpont, the senstvtes of all bochemcal markers change dependng on the tme of nfarcton, whle specfctes reman unchanged as only blood sample was obtaned from the patents n the comparson group. Fgure II shows the release knetcs of myoglobn, Tnl, and CK-M expressed as percentles (0%, 25%, and 50%) to better compare the frst ncreased concentratons of the markers n the early phases after the onset of chest pan. At admsson, 50% of the myoglobn concentratons are hgher than the upper reference value, whle for CK-M all the percentles evaluated are lower than the upper reference lmt (5 (g/l). At T0 and T Tnl seems to have senstvty values smlar to those of myoglobn, but the analyss of the CI of the medan and the CI of the 25th percentle shows dssmlar behavor; at T0 n fact, the CI of 50% of Tnl values wdely ncludes the upper reference lmt rangng from 0 to.3 ng/ml (0-.3 Ltg/L) and the same occurs for 25% of Tnl values at T, (CI, 0-2.6 ng/ml [0-2.6 tg/l]). In contrast, the CI of myoglobn at T0for 50% and 25% (65.3-250.8 Ltg/L and 95.6-30.0 g/l, respectvely) are sgnfcantly hgher than the upper reference value, supportng the earler dagnostc senstvty of ths assay. The performance of cardac markers evaluated as sngle tests s reported n ITable II. At admsson (T0; mean = 3.3 hours from the onset of chest pan), the senstvty of myoglobn was hgher than that of CK-M and Tnl, reachng the maxmum value (00%) 3 hours later (T,). Furthermore, at T0; 54% of patents wth AMI are correctly dentfed by the Tnl assay, whle CK-M dentfed only 46%. The dagnostc performance of total CK s, on the other hand, extremely poor over the entre clncally mportant nterval for the management of the condtons of patents wth AMI. At the decson thresholds used n the present study, the data showed that the specfcty of Tnl (90%) for the dagnoss of AMI s hgher than that of other markers. In fact, Tnl values of more than ng/ml (> Ltg/L) were found n 7 of 9 patents wth unstable angna (range, 0.6-2.6 ng/ml [0.6-2.6 Ltg/L]).
Clncal Chemstry / ORIGINAL ARTICLE a 340 320 300 280 J 260 ' U 4 220 ~JT 200 3 80 60 40- ^" 20 * 00 60 80-40 - 20 - ' o '- / / S0%\^ N ^ 25% 0% * T, T2 T3 Sample Tmng 4 2 0 4 8 J! 6 C o & 4 ' ^ - 0 * - "- t- Sample Tmng...' ^^50% 25%^ -" ^ "..-"o% I Fgure II Release knetcs of myoglobn, troponn I, and creatne knase (CK)-M represented as 0%, 25% and 50% at >30 4 _, -'' 50% " 25% dfferent ntervals after admsson (T 0, T v T 2, T 3, and T 4 = admsson and 3, 6, 9, and 2 hours later, respectvely). Sold horzontal lnes show the upper reference lmts. _/ / 9 0% J,' ".---"*".-''' Sample Tmng Table II Senstvty and Specfcty Values of ochemcal Markers Performed as a Sngle Test* Senstvty Test Myoglobn Troponn I Creatne knase Creatne knase-m Upper Reference Value T, Specfcty 70ng/L 69(48-86) 00(87-00) 00(87-00) 46(33-60) ng/ml ( ug/l) 54(33-73) 8(6-93) 00(87-00) 90(80-96) Males, 90 U/L; females, 3(4-52) 54(33-73) 88(70-97) 66(52-78) 60 U/L 5 ug/l 46(27-67) 88(70-97) 00(87-00) 78(66-88) " Data are gven as percent (95% confdence nterval). Table 2 gves the results from 2-test parallel combnaton that, when used n the any-test postve mode, has a senstvty hgher than or equal to that of the most senstve test and the specfcty of the less specfc test. Our data thus show that the hghest senstvty s found when myoglobn and Tnl or CK-M are combned n a parallel mode, but the specfcty of ths combnaton s equal to that of the screenng test (myoglobn, Am J Cln Pathol 999; :399-405 40 Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208
Zannotto et al / EARLY DIAGNOSIS OF ACUTE MYOCARDIAL INFARCTION USING IOCHEMICAL MARKERS Table 2 Senstvty and Specfcty Values Usng Two-Test Parallel Combnaton* Senstvty Combnaton Myoglobn >70 ug/l or Tnl > ng/ml (> ug/l) Myoglobn >70 ug/l or CK-M >5 ug/l CK males, >90 U/L; females, >60 U/L orckm>5ug/l T T, 00(87-00) 00 (87-00) 00(87-00) 00(87-00) 46 (33-60) 46 (33-60) 54 (33-73) 88 (70-97) 00(87-00) 66 (52-78) Specfcty Tnl = troponn ; CK = creatne knase. T0, Tp and T2 are admsson and 3 and 6 hours later. "Data are gven as percent (95% confdence nterval). specfcty, 46%). On the other hand, the combnaton of CK wth CK-M gves a specfcty just hgher and a senstvty lower than other parallel combnatons evaluated. Combnaton testng n the seral model ITable 3 provdes the same senstvty of the most specfc test (Tnl or CK-M), mprovng ts specfcty. In partcular, the specfcty of myoglobn and Tnl tested n the seral mode ncreased to 98% because 6 of 7 patents wth unstable angna had Tnl values of more than ng/ml (> ug/l), whle myoglobn concentratons were lower than the upper reference value (range, 39.0-62.3 ug/l), so the combnaton was found to be negatve. When consderng the combned assay of CK and CK-M, the senstvty obtaned was lower than that observed when both tests were evaluated sngly. Ths unexpected behavor depends on the low senstvty of CK when used as a screenng test. The clncal performance of the strategy usng the combnaton n the seral mode, testng the second confrmng test on the T, sample, s reported n ITable 4. Ths strategy, whch provdes the senstvty of the most senstve test and mproves the specfcty of the most specfc test assayed sngly, causes a 3-hour delay n the measurement of the second confrmng test, wth a consequent delay n the fnal dagnoss. IFgure 2 shows the postve predctve value (PPV), the negatve predctve value (NPV), and the accuracy (ACC) calculated at admsson tme for dfferent dsease prevalences. The present retrospectve study does not allow the correct evaluaton of the dsease prevalence n the populaton consdered. Tnl tested sngly performs better than CK-M; ITable 3 Senstvty and Specfcty Values Usng Two-Test Seres Combnaton Assayng oth Tests on the Same Sample* Senstvty Specfcty Combnaton Myoglobn >70 ug/l and Tnl > ng/ml (> ug/l) Myoglobn >70 ug/l and CK-M >5 ug/l CK males, >90 U/L; females, >60 U/L and CK-M >5 ug/l 54(33-73) 46(27-67) 8(6-93) 88(70-97) 00(87-00) 00(87-00) 98(9-00) 83(72-97) 27(2-48) 54(33-73) 88(87-00) 78(66-88) Tnl = troponn I; CK = creatne knase. T, T,, and T2 are admsson and 3 and 6 hours later. *Data are gven as percent (95% confdence nterval). ITable 4 Senstvty and Specfcty Values Usng Two-Test Seres Combnaton Assayng the Screenng Test at the Admsson and the Confrmng Test on the Followng Sample* Senstvty Specfcty Myoglobn >70 ug/l and Tnl > ng/ml (> ug/l) Myoglobn >70 ug/l and CK-M >5 ug/l CK males, >90 U/L; females, >60 U/L and CK-M >5 ug/l 3 (4-52) 98(9-00) 83 (72-97) 78 (66-88) Tnl = troponn I; CK = creatne knase. T0 and T, are admsson and 3 hours later. *Data are gven as percent (95% confdence nterval). 402 Am J Cln Pathol 999; :399-405 Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208
Clncal Chemstry / ORIGINAL ARTICLE CK-M.0- Troponn I,.-a a..-o"' '".,,.EJ'' 0.8- ""e. Val 3 - V. s h> 3 n ra 0 5 0.6- _^-^ ^ 5 ^4C~^~ ^ r ^ - ^. ^ H H"'' 0.8- ^-» '' *"" ^ X P I- U n - ''Q, E m L. o *0) Q- \ 0.2- I *' "' 0.00 0 0.2 o.o- 06 0.8 00 Myoglobn + CK-M Dsease Prevalence I 0.2 0.6 0.8.0 Dsease Prevalence Myoglobn + Troponn I *-......."" '' '"-<K^^ g > ^ - ^ ^ "Q a 0.6- * - ^ ^ '0 '"Q - E a. o t 0) " 0.2-0 0 0.2 Dsease Prevalence ' ' 0.6 ' 0.8. Dsease Prevalence Fgure 2 Performance measures of creatne knase (CK)-M and troponn I assayed as sngle test and n seral combnaton wth myoglobn as a functon of dsease prevalence. = Postve predctve value (PPV), O = Negatve predctve value (NPV), A = Accuracy (ACC). furthermore, the combnaton of each test n the seral mode wth myoglobn showng a hgh NPV (range, -0.930) hardly mproves the performance of CK-M but sgnfcantly ncreases the PPV of Tnl, n partcular when the dsease prevalence s low, as observed n a general emergency department (eg, prevalence 0.; Tnl tested sngly: PPV, 0.375; seral combnaton of myoglobn and Tnl: PPV, 0.750). Interestngly n ths stuaton, the seral combnaton Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208 of myoglobn and Tnl shows the same accuracy (0.58) as that of Tnl tested sngly. Dscusson The present study focuses on the use of new bochemcal markers for the dagnoss of AMI that contnues to be a Am J Cln Pathol 999::399-405 4 0 3
Zannotto et al / EARLY DIAGNOSIS OF ACUTE MYOCARDIAL INFARCTION USING IOCHEMICAL MARKERS dagnostc challenge, partcularly n patents admtted to the emergency department wth uncertan and transent clncal manfestatons and for whom t s mportant to rule out a dagnoss of AMI. In fact, the avalablty of several dfferent bochemcal markers n the emergency department forces laboratorans and clncans to revse and compare dagnostc strateges and dfferent "cardac bochemcal profles," evaluatng whether the new markers are to be used n addton to or nstead of other tradtonal laboratory tests, to adopt a cost-effectve approach n the routne dagnostc settng.20-23 Fndngs of the present study show, as expected, that all bochemcal markers consdered have a tme-dependent senstvty: wthn 9 hours after the onset of chest pan, all markers reach the maxmum value of senstvty (00%), except total CK, whch shows a senstvty of 96% at a mean of 2 hours after the onset of symptoms (data not shown). However, for the early dagnoss of AMI, a sngle serum myoglobn measurement at admsson provdes the hghest senstvty (69%), adequate for dagnosng an AMI, even f the specfcty s very low. On the other hand, Tnl tested sngly shows a senstvty smlar to that of CK-M n the dagnoss of AMI and gves the hghest specfcty,5-24 also n evdencng mnor myocardal damage.25-27 In our experence, combnaton testng n the parallel mode showed no dfference n senstvtes between sngletest myoglobn and 2 tests n any postve combnaton at any tme nterval. However, the overall effect of seres testng s to mprove the specfcty." In fact, whle at admsson all combnaton testng shows a poor senstvty, the hgh specfcty of Tnl ensures that the combned myoglobn and Tnl strategy, performed on the same sample, has the best performance. Thus, a senstvty smlar to that of the most senstve test and a specfcty hgher than that of the most specfc marker s obtaned. Hgher senstvtes were observed when the second confrmng test was performed (f the screenng test result was postve) by usng the followng blood sample, but ths strategy had hgh effcacy only 3 hours later. Fnally, the hgh NPV of myoglobn assocated wth the hgh PPV of Tnl gave the combned measurement n the seral mode a dagnostc accuracy that was, for dfferent prevalences of the dsease, hgher than that provded by testng myoglobn and CK-M n combnaton. Our results suggest the followng dagnostc algorthm based on the combned seral measurement of myoglobn and Tnl n relaton to tme from the onset of symptoms: for patents admtted wthn 4 hours from the onset of chest pan, the myoglobn level s measured frst, and n blood samples obtaned 3 hours later, myoglobn and Tnl are smultaneously evaluated. All postve results confrm AMI, and all negatve results rule out myocardal damage, whle dfferent combnatons of negatve and postve data requre further combned testng untl 9 hours after the onset of chest pan. 404 Am J Cln Pathol 999,:399-405 Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208 In our experence, Tnl shows the maxmum value of senstvty around 9 hours from the onset of symptoms. Therefore, to maxmze the dagnostc effcency of the protocol n patents admtted later (from 4 to 9 hours after the onset of chest pan), myoglobn and Tnl were measured smultaneously n the frst blood sample. Our data, although obtaned from a relatvely small sample of selected patents wth AMI, add to the growng lterature28 confrmng that the practcal approach proposed allows a relable and tmely dagnoss to be made, myocardal necross to be ruled out n patents wth renal falure or wth skeletal muscle dseases,29 and the course of myocardal damage to be montored durng therapy. From the 'Department of Laboratory Medcne, UnverstyHosptal of Padova, Padova, Italy; the 2Center of omedcal Research, Castelfranco Veneto (TV), Italy; and the -'Department of Cardology Hosptal of Castelfranco Veneto (TV), Italy. Address reprnt requests to Dr M. Pleban: Servzo d Medcna d Laboratoro, Azenda Ospedalera d Padova, Va Gustnan 2, 3528 Padova, Italy. References. hayana V, Henderson AR. ochemcal markers of myocardal damage. Cln ochem. 995;28:-29. 2. Keffer JH. Myocardal markers of njury: evoluton and nsghts. Am J Cln Pathol. 996;05:305-320. 3. Mar J. Progress n myocardal damage detecton: new bochemcal markers for clncans. Crt Rev Cln Lab Sc. 997;34:-66. 4. Jaffe AS. More rapd bochemcal dagnoss of myocardal nfarcton: necessary? prudent? cost effectve? [edtoral] Cln Chem. 993;39:567-569. 5. Ryan TJ, Anderson JL, Antman EM, et al. ACC/AHA gudelnesforthe management of patents wth acute myocardal nfarcton: a report of the Amercan College of Cardology/Amercan Heart Assocaton Task Force on Practce Gudelnes (Commttee on Management of Acute Myocardal Infarcton). J Am Coll Cardol. 996;28:328-428. 6. Lndhal, Venge P, Wallentn L. Early dagnoss and excluson of acute myocardal nfarcton usng bochemcal montorng. Coron Artery Ds. 995;6:32-328. 7. Keffer JH. The cardac profle and proposed practce gudelne for acute schemc heart dsease. Am J Cln Pathol. 997;07:398-409. 8. Wu AH, Clve JM. Impact of CK-M testng polces on hosptal length of stay and laboratory costsforpatents wth myocardal nfarcton or chest pan. Cln Chem. 997;43:326-332. 9. Montague C, Krcher T. Myoglobn n the early evaluaton of acute chest pan. Am J Cln Pathol. 995;04:472-476. 0. de Wnter RJ, Koster RW, Sturk A, et al. Value of myoglobn, troponn T, and CK-M mass n rulng out an acute myocardal nfarcton n the emergency room. Crculaton. 995;92:340-3407.
Clncal Chemstry / ORIGINAL ARTICLE. hayana V, Cohoe S, Pellar TG, et al. Combnaton (multple) testng for myocardal nfarcton usng myoglobn, creatne knase-2 (mass), and troponn T. Cln ochem. 994;27:395-406. 2. Levtt MA, Promes S, ullock S, et al. Combned cardac marker approach wth adjuncton two-dmensonal echocardography to dagnose acute myocardal nfarcton n the emergency department. Ann EmergMed. 996;27:-7. 3. Internatonal Federaton of Clncal Chemstry. IFCC methods for the measurement of catalytc concentraton of enzymes, VII: IFCC method for creatne knase (ATP: creatne N-phosphotransferase, EC 2.7.3.2): IFCC recommendaton. Cln Chm Acta. 990;90:S4-S40. 4- Chapelle JP, El Allaf M. Automated quantfcaton of creatne knase M soenzyme n serum by radal partton mmunoassay, wth use of Stratus analyzer. Cln Chem. 990;36:99-0. 5. Zannotto M, Akner S, Lachn M, et al. Fluoroenzymometrc method to measure cardac troponn I n sera of patents wth myocardal nfarcton. Cln Chem. 996;42:460^466. 6. Cebul RD, Hershey JC, Wllams SV. Usng multple tests: seres and parallel approaches. Cln Lab Med. 982;2:87-890. 7. Galen RS, Gambno SR. Combnaton testng-multple testng. In: eyond Normalty: The Predctve Value and Effcency of Medcal Dagnoss. New York, NY: Wley; 975:4-48. 8. Gardner MJ, Gardner S, Wnter PD. Confdence Interval Analyss Mcrocomputer Program. London, England: rtsh Medcal Journal Publshng Group; 989. 9. Ensten AJ, odan CA, Gl J. The relatonshps among performance measures n the selecton of dagnostc tests. Arch Pathol Lab Med. 997;2:0-7. 20. Schreber WE. Laboratory assessment of myocardal damage: whch test s best? [edtoral] Am J Cln Pathol. 997;07:383-3842. Jaffe AS, Landt Y, Parvn CA, et al. Comparatve senstvty of cardac troponn I and lactate dehydrogenase soenzymes for dagnosng acute myocardal nfarcton. Cln Chem. 996;42:770-776. Downloaded from https://academc.oup.com/ajcp/artcle-abstract//3/399/758707 on 20 January 208 22. Martns J, L DJ, askn L, et al. Comparson of cardac troponn I and lactate dehydrogenase soenzymes for the late dagnoss of myocardal njury: demse of the lactate dehydrogenase soenzymes. Am J Cln Pathol. 996;06:705-708. 23. D'Costa M, Flemng E, Patterson MC. Cardac troponn I for the dagnoss of acute myocardal nfarcton n the emergency department. Am J Cln Pathol. 997;08:550-555. 24. McLaurn MD, Apple FS, Voss EM, et al. Cardac troponn I, cardac troponn T, and creatne knase M n dalyss patents wthout schemc heart dsease: evdence of cardac troponn T expresson n skeletal muscle. Cln Chem. 997;43:976-982. 25. Apple FS, Falahat A, Paulsen PR, et al. Improved detecton of mnor schemc myocardal njury wth measurement of serum cardac troponn I. Cln Chem. 997;43:2047-205. 26. Wu AH, Feng YJ, Contos JH, et al. Prognostc value of cardac troponn I n patents wth chest pan. Cln Chem. 996;42:65-652. 27. Antman EM, Tanasjevc MJ, Thompson, et al. Cardacspecfc troponn I levels to predct the rsk of mortalty n patents wth acute coronary syndromes. N Engl J Med. 996;335:342-349. 28. Woo J, Lacbawan FL, Sunhemer R, et al. Is myoglobn useful n the dagnoss of acute myocardal nfarcton n the emergency department settng? Am J Cln Pathol. 995;03:725-729. 29. Trnquer S, Flecheux O, ullenger M, et al. Hghly specfc mmunoassay for cardac troponn I assessed n nonnfarct patents wth chronc renal falure or severe polytrauma. Cln Chem. 995;4:675-676. Am J Cln Pathol 999; :399-405 405