increasingly sensitive assays for cardiac troponins

CLINICIANS CORNERGRAND ROUNDSAT THE UNIVERSITY OF TEXASSOUTHWESTERN MEDICAL CENTER

Increasingly Sensitive Assaysfor Cardiac TroponinsA ReviewJames A. de Lemos, MD

CASE PRESENTATIONA 62-year-old man presented to theemergency department following a3-hour plane flight from Guatemalawith symptoms of intermittent chestdiscomfort and shortness of breath.He described pressure in his chest andshoulders that was not worsened byexertion and had no obvious precipi-tating cause. The patient had a historyof diabetes, hypertension, stage 2chronic kidney disease (CKD), gastro-esophageal reflux disease, chronicheart failure with a left ventricularejection fraction of 30%, and paroxys-mal atrial flutter. The results of coro-nary angiography, performed 3 yearspreviously at an outside hospital,were reported by the patient to havebeen normal.On examination, his heart rate was

60 beats/min; his blood pressure was118/70 mm Hg. Pulses were equal inall extremities and no edema, ascites,or jugular venous distension wasnoted. His 12-lead electrocardiogram(ECG) showed sinus rhythm with lowlimb-lead voltages and nonspecificST-segment and T-wave abnormalitiesand was unchanged from his previousECG. His cardiac troponin T (cTnT)level at presentation was 20 ng/L

(upper limit of normal, 10 ng/L);during an admission 6 monthspreviously, his cTnT level wasless than 10 ng/L.The patient was admitted with a di-

agnosis of nonST-segment elevationmyocardial infarction (MI). His warfa-rin was temporarily discontinued, andhe received aspirin, clopidogrel, enox-aparin, simvastatin, lisinopril, andcarvedilol. Cardiac catheterization wasdeferred because of an elevated inter-national normalized ratio, but when se-rial measurements of cTnT levels dem-onstrated unchanged levels of 20 ng/Lover the subsequent 2 days, noninva-sive risk stratification was pursued. Avasodilator perfusion scan revealed

a left ventricular ejection fractionof 32% and mild, fixed inferiorand inferolateral perfusion defects de-scribed as patchy and more consis-tent with nonischemic cardiomyo-pathy than prior infarction.This case highlights challenges in in-

terpreting small elevations of tro-ponin levels in patients with clinicalpresentations that are not specific for

CME available online atwww.jamanetworkcme.comand questions on p 2276.

Author Affiliation: Division of Cardiology, Depart-ment of Medicine, University of Texas SouthwesternMedical Center, Dallas.Corresponding Author: James A. de Lemos,MD, Uni-versity of Texas Southwestern Medical Center, 5323Harry Hines Blvd, Dallas, TX 75390-8830 ([email protected]).Grand Rounds Section Editor: Mary McGraeMcDermott, MD, Contributing Editor, JAMA.

Cardiac troponins are the preferred biomarkers for diagnosis of myocardialinfarction because of their high sensitivity and specificity for myocardial in-jury. However, acute and chronic conditions distinct from acute coronary syn-dromes (ACS) commonly lead to small elevations in troponin levels, withfew data available regardingmanagement of care for patients with such con-ditions. Recently developed highly sensitive troponin assays will likely leadto a substantial increase in the proportion of detectable troponin levels at-tributable to non-ACS conditions. Novel algorithms with highly sensitive as-says, incorporating baseline troponin values and changes in values over 1to 2 hours, may allow rapid exclusion of myocardial infarction and help toaddress specificity concerns butmust be validated in appropriate target popu-lations. Enhanced detection of very low troponin levels with highly sensi-tive assays has made feasible several potential new indications for troponintesting, including in the ambulatory setting, where assessment for low-level chronic myocardial injury may enhance risk stratification for heart fail-ure and cardiac death.JAMA. 2013;309(21):2262-2269 www.jama.com

2262 JAMA, June 5, 2013Vol 309, No. 21 2013 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a Saint Joseph's University User on 08/22/2013

acute coronary syndromes (ACS). Thispatient had multiple potential non-ACS etiologies for his elevated tro-ponin levels. Application of ACS treat-ment guidelines1,2 to patients withelevated troponin levels not attribut-able to ACS may expose such patientsto unnecessary testing and risk.

Current Status of TroponinMeasurement to Rule Out MI

Evaluation for chest pain accounts formore than 5 million emergencydepartment visits each year in theUnited States.3 Troponins T and I arecurrently the preferred biochemicalmarkers of myocardial necrosis inpatients with suspected ACS, becauseof these markers high sensitivity andalmost complete cardiac specificity.4

The regulatory troponin complex iscomposed of 3 protein subunits, tro-ponins T, I, and C. Troponins T and Ihave unique cardiac isoforms,whereas cardiac and skeletal muscleshare troponin C isoforms, renderingthis protein unsuitable for diagnosticuse. Cardiac troponins are complexedwith actin in cardiac myofibrils, witha smaller fraction (3%-6%) soluble inthe cytoplasm.5 Ischemia is thought toalter cell membrane integrity, causingrapid depletion of the soluble cyto-plasmic pool, followed by larger andmore sustained release of troponininto the circulation as the contractileapparatus breaks down (FIGURE 1).6

Cardiac troponins T and I providelargely identical information, and se-lection between them is typically in-fluenced by equipment and vendor se-lection in the central laboratory. ThecTnT assay is produced by a singlemanufacturer. However, multiplemanufacturers make cardiac troponinI (cTnI) assays and eachmanufactureruses different antibody pairs, so as-says are not interchangeable. Medi-care reimbursement for measurementof troponin levels is approximately $13to $18.It is important to recognize that tro-

ponin is a laboratory measure of myo-cardial necrosis, whereas MI is a clini-cal diagnosis. The Universal Definition

of MI, a consensus document repre-sentingmultiple international cardiol-ogy societies, requires an elevated tro-ponin level combined with symptomsor signs of ischemia for the diagnosisof MI (BOX 1).4 In addition to requir-ing at least 1 troponin value above theMI diagnostic threshold, an increase ordecrease in troponin levels over serialmeasurement is necessary to meet theMI definition. The decision limit forMIis set by consensus at the 99th percen-tile value of a healthy normal popula-tion and varies with each troponinassay.The universal definition recognizes

the distinction between classic (type1) MI, caused by ACS with atheroscle-rotic plaque rupture or erosion and

coronary thrombus, and secondary(type 2) MI, in which a conditionother than coronary artery disease(CAD) contributes to increased myo-cardial oxygen demand or decreasedoxygen supply, resulting in myocar-dial necrosis. Plaque disruption is notthe cause of type 2 MI, but underlyingCAD may be present. Myocardialinfarction in the setting of atrialarrhythmia, hypertensive emergency,and critical medical or surgical illnessmay represent type 2 MI.With conventional clinical assays, se-

rialmeasurements of troponin levels areneeded to excludeMI, because a singlemeasurement has only approximately70% to 85% sensitivity.7,8 Although theperiod of monitoring is typically 6 to

Figure 1. Release of Cardiac Troponins Following Ischemic Cardiac Injury

Cytosolic troponins

Myofibrillar troponins

Actin

cTnTcTnI TnC

cTnT

cTnI

Release ofcardiac troponins into the bloodstream

B L O O DV E S S E L

Myosin

Tropomyosin

B Ischemia-induced cardiomyocyte damage

A Structure of cardiac troponins

Cardiomyocytes

Ischemic cardiomyocytes

TnC indicates troponin C; cTnI, cardiac troponin I; cTnT, cardiac troponin T.

INCREASINGLY SENSITIVE ASSAYS FOR CARDIAC TROPONINS

2013 American Medical Association. All rights reserved. JAMA, June 5, 2013Vol 309, No. 21 2263


9 hours, shorter intervals are appropri-ate among individuals with a low prob-ability of MI who are also at low riskfor death or ischemic complications. Aninnovative approach proposed by Thanet al9 incorporated assessment of pre-test probability forMI and risk for com-plications using the Thrombolysis inMyocardial Infarction (TIMI) risk score,together with ECG findings and serialmeasurement of cTnI levels at 0 and 2hours. Among 1975 patients present-ingwith suspectedACS, 20%were clas-sified as at low risk, defined prospec-tively as a TIMI risk score of 0, theabsence of ischemic ECG changes, andcTnI level below the MI detectionthreshold at both timepoints. Thenega-tive predictive value (NPV)was greaterthan 99% in the low-risk group, sug-gesting that this expedited algorithm,or a similar one, may allow rapid dis-charge of low-risk individuals with-out additional testing.

Risk Stratification and TherapeuticDecision Making in Suspected ACS

In patients with suspected ACS, evenminimally elevated troponin levels iden-

tify patients at increased risk for deathand recurrent ischemic events.10 Pa-tients with suspected ACS and el-evated troponin levels have more se-vere and complex CAD, with a largerthrombus burden and greater micro-vascular injury than similar patientswith normal troponin levels.11,12 Thesepathophysiological observations alsohave implications for therapeutic de-cision making, because clinical trialshave generally demonstrated signifi-cant benefit from intensive antiplate-let and antithrombotic regimens, aswellas percutaneous coronary interven-tions, in patients with ACS and el-evated troponin levels, with less favor-able risk-benefit and cost-benefit ratiosfor these therapies among patients withnormal troponin levels.13-17

The Specificity Problem

Although troponins are generallyspecific for myocardial injury, theyare not specific for a particularmechanism of myocardial injury.Any conditionacute or chronicthat injures cardiomyocytes may leadto measurable increases in levels ofcirculating troponins. Acute condi-tions that can elevate troponin levelsinclude heart failure, sepsis, and pul-monary embolism, although multipleother potential etiologies exist(BOX 2).18,19 When sensitive assays areused, elevated troponin levels are com-mon after endurance athletics, such asmarathon running, but these eleva-tions are transient and are not associ-ated with cardiac abnormalities orrisk.20,21

Persistent low-level elevations in lev-els of cardiac troponins are commonlyseen in patients with chronic cardiacconditions, including patients stabi-lized months after an ACS event andthose with chronic ambulatory heartfailure.22,23 Less frequently, elevatedcTnT or cTnI levels are seen among in-dividuals from the general popula-tion, in whom chronic structural heartdisease, ormajor determinants of struc-tural heart disease such as diabetes andCKD, rather than coronary atheroscle-rosis, appear to explainmost cases.24-26

Alcalai et al,27 adjudicating all el-evated troponin levelsmeasured in theirhospital over a 1-year period, deter-mined that only about half of the el-evated levels were attributable to ACS.In older patients withCKD, the propor-tionof elevated levels attributable toACSdecreased to 37%. Of note, among pa-tients with elevated troponin levels,2-year mortality was more than doublein thenon-ACSgroupcomparedwith theACS group.27 Thus, although cliniciansfrequently trivialize elevated troponinlevelsnot attributable toACSusing termssuch as troponinemia or troponino-sis, these elevations are of importantprognostic significance.

Approach to the Patient WithNon-ACS Elevated Troponin LevelsNot Attributable to ACS

Few data are available to guide evalu-a t i on and manag emen t whenelevated troponin levels are notattributable to ACS. Because elevatedtroponin levels (identified usingassays in current clinical use in theUnited States) are uncommon in theabsence of underlying cardiovascularor renal disease and are associatedwith a poor prognosis, an attemptshould be made to explain the causeof all elevated troponin levels. How-ever, care must be taken to imple-ment a prudent strategy that mini-mizes unnecessary treatment andresource utilization. Evaluationshould include a detailed history,physical examination, assessment ofrenal function, and an ECG. Giventhe strong association betweenelevated troponin levels and cardiacstructural and functional abnormali-ties, echocardiography should beconsidered for most unexplainedelevations in troponin levels. How-ever, there is no evidence that per-forming echocardiography in allp a t i en t s w i th e l e v a t ed l e v e l simproves outcomes. Using the infor-mation gathered, elevated levelsshould be classified as attributable totype 1 or type 2 MI, to a non-ACSacute condition, or to a non-ACSchronic condition (FIGURE 2).19,28

Box 1. Third UniversalDefinition of MyocardialInfarctiona

Detection of increase, decrease, orboth of levels of cardiac biomarkers(preferably cardiac troponin) with atleast 1 value above the 99th percen-tile of the upper reference limit andwith at least 1 of the following:

Symptoms of ischemia

New significant ST-segment andT-wave changes or new left bundle-branch block

Development of pathologicalQ waves

Cardiac imaging findings consis-tent with ischemia or infarction(ie, new wall motion abnormality)

Identification of an intracoronarythrombus by angiography

aAdapted from Thygesen et al.4




Although evidence-based guide-lines exist for managing type 1 MI,1,2

little is known about the best strate-gies for treating patientswith type 2MI,which is caused by imbalance in myo-cardial oxygen supply and demand inthe absence of plaque rupture. A moreindividualized approach is appropri-ate for type 2 MI, focusing on treatingprecipitating medical or surgical con-ditions and reducing themismatch be-tweenmyocardial oxygen demand andsupply.Hypertension and tachyarrhyth-mias should be controlled, usually withregimens that include -blockers, andsevere anemia should be corrected. As-pirin should be considered unless con-traindications are present, but antico-agulants and invasive therapies may bebest reserved for patients with recur-rent ischemia after treating the precipi-tating events. However, these treat-ment strategies have not been tested inrandomized controlled trials for pa-tients with type 2 MI.Data on elevated troponin levels in

conditions such as heart failure, pul-monary embolism, sepsis, and renal fail-ure are still evolving. Although prog-nosis is worse in these conditionswhentroponin levels are elevated,29-32 it is notyet clear whether or how this informa-tion should be used to alter the diag-nostic or therapeutic approach. In thesepatients, clinicians should be alert to thepossibility ofmisdiagnosis of ACS, lead-ing to inappropriate therapy for ACSand delays in treating other causes ofelevated troponin levels.

Do Highly Sensitive TroponinAssays Improve Diagnosis of MI?

Highly sensitive assays for cTnT andcTnI are available and are widely usedin many parts of the world, althoughthey are not approved for use in theUnited States. These assays can detectlevels of troponin well below 10 ng/L,identifying a wide range of values be-low theMI detection threshold (see theTABLE for currently recommended 99thpercentile cut points).Several European studies compared

highly sensitive assays with standardtroponin assays for MI detection.7,8,33

These studies included patients with ahigh overall rates of MI (17%-23%).Compared with standard troponin as-says, the highly sensitive assays im-proved both sensitivity and discrimina-tion for MI, particularly in the first 3 to6 hours after symptomonset (Table).7,33

These advantageswerepartially offset bya decrease in specificity forMI,7,33 a find-ing that raises concerns regarding broadapplication of these tests to popula-tions with a lower MI prevalence. Forexample, the probability ofMI in a typi-cal US chest pain observation unit is lessthan 5%.34,35

Although the high NPV and lowfalse-negative rate of the highly sensi-tive assays are clear, the decrement inpositive predictive value (PPV) hastroubling implications. Among indi-viduals who rule in with a highly sen-

sitive assay but who would have beenruled out with a standard assay, false-positive MI diagnoses will be morelikely than true-positive diagnoses, par-ticularly when the a priori probabilityof MI is low.36 Moreover, because thetarget population in whom the test willbe performed commonly has eithermultiple cardiac risk factors or exist-ing cardiac disease, many patients willhave chronic elevation of troponin lev-els above the MI detection threshold.In one study in a US chest pain popu-lation, more than 15% of patients hadcTnT levels above the MI detectionthreshold with the highly sensitive as-say, but only 2% were determined tohaveMI.37 Thus, problems related to in-terpretation and management of non-ACS elevations of troponin levels, in-cluding increased downstream costs

Box 2. Acute Causes of Elevated Troponin Levelsa

Acute Myocardial Infarction

ACS (type 1 MI)

ST-segment elevation myocardialinfarction

NonST-segment elevation myo-cardial infarction

Supply/demand mismatch (type 2 MI)

Severe hypertension

Tachyarrhythmia

Severe anemia

Decreased supply (non-ACS)

Spasm

Embolism

Drugs

Cocaine

Methamphetamines

Procedure-related

Percutaneous coronaryintervention

CABG surgery

Nonischemic Acute Myocardial Injury

Direct Myocardial Damage

Congestive heart failure

Infection

Viral myocarditis

Endocarditis

Inflammation

Myocarditis

Pericarditis

Malignancy

Cancer chemotherapy

Trauma

Electrical shock

Ablation procedures

Infiltrative diseases

Stress cardiomyopathy (Takosubo)

Other Causes

Pulmonary embolism

Sepsis

Renal failure

Stroke

Subarachnoid hemorrhage

Abbreviations: ACS, acute coronary syndromes; CABG, coronary artery bypass graft;MI, myocardial infarction.aAdapted from Newby et al.19




and anxiety for patients and clini-cians, are likely to be greater with thehighly sensitive assays.

It is also important to consider theresource implications of troponin lev-els that are measurable but below the

MI detection threshold. Based on re-cent population-based studies, 50% ormore of patients in a typical chest painobservation unit would be expected tohave a low but detectable cTnT levelwith the highly sensitive assay,38,39 withsimilar or even higher proportions de-tectable with emerging highly sensi-tive cTnI assays.40 If such individualsare routinely referred for cardiologyconsultation, additional testing, or both,the indirect costs and potential harmsfrom excess testingmay be substantial.

Serial Troponin Measurementsto Improve Specificity

Although theUniversalDefinition ofMIrequires dynamic changes in troponinconcentrations to meet the MI defini-tion, it does not provide guidance re-garding the magnitude of change re-quired.4 Existing recommendationsfrom the National Academy of Clini-cal Biochemistry suggest that relative

Figure 2. Proposed Algorithm for Classifying Elevated Troponin Levels

Cardiac injury presentTroponin level >99th percentile

Acute troponin elevationRise and/or fall in troponin level over serial measurements

Chronic troponin elevationTroponin level elevated but no rise and/or fall over serial measurements

Yes No

Clinical presentation consistent with atherosclerotic plaque rupture

Type 1 MI

Treat underlying disease

Recognize elevated troponin level as poor prognostic marker

Possible structural heartdisease, chronic renal disease

Consider echocardiogramFollow treatment guidelines for acute MI

Precipitant other than coronary artery diseaseAnemia, tachyarrhythmia, severe hypertension, etca

Type 2 MI

Ischemic mechanisim present?Are the patient history, ECG, or cardiac imaging consistent with myocardial ischemia?

Correct precipitant

Consider aspirin and -blockertreatment

Nonischemic acute myocardial injuryPulmonary embolism, acute heart failure, etca

This algorithm is based on available literature19,28 but has not been separately validated. ECG indicates electrocardiogram; MI, myocardial infarction.aSee Box 2 for more detail.

Table. Comparison of Test Performance Characteristics for Detection of Myocardial InfarctionBetween Standard and High-Sensitivity Assays for Troponins T and I at the Time of Presentationa

Characteristic

cTnTa cTnIb

Fourth-GenerationHighly

SensitiveContemporary

SensitiveHighly

Sensitive

AUC 0.90 (0.86-0.94) 0.96 (0.94-0.98) 0.92 (0.90-0.94) 0.96 (0.95-0.97)

99th Percentilecut point, ng/L

10 14 32 30

10% CV threshold,ng/L

35 13 32 5.2

Sensitivity, % 83 (76-90) 95 (90-98) 79 (75-84) 82 (77-87)

Specificity, % 93 (91-95) 80 (77-83) 95 (93-96) 92 (90-94)

Positive predictivevalue, %

72 (64-79) 50 (43-56) 81 (76-85) 75 (70-80)

Negative predictivevalue, %

97 (95-98) 99 (97-100) 94 (92-95) 95 (93-96)

Abbreviations: AUC, area under the receiver operating characteristic curve; cTnI, cardiac troponin I; cTnT, cardiac troponinT; CV, coefficient of variation.

aData for the comparisons of the Roche fourth-generation cTnT assay vs Roche highly sensitive cTnT assay are ab-stracted from Reichlin et al.7

bData for comparisons of the contemporary sensitive (Abbott Architect STAT cTnI) assay vs highly sensitive (Abbott Ar-chitect hs-cTnI) assay are abstracted from Keller et al.33




changes 20% greater than baseline overserial follow-up are sufficient to docu-ment rising troponin levels.41 The 20%change value was selected empiricallyas a value that would exceed analyti-cal variation alone. However, withhighly sensitive assays, substantial rela-tive increases in troponin levels arecommon even among individuals with-out evidence of ischemia or infarc-tion,42,43 and a large relative change introponin levelmay occur despite a verysmall absolute increase.44 Such small ab-solute changes do not demonstratespecificity for MI.The APACE (Advantageous Predic-

tors ofAcuteCoronary SyndromeEvalu-ation) investigators recently comparedabsolute and relative changes over se-rialmeasurements at 1 and 2hourswithhighly sensitive troponin assays. Usingreceiver operating characteristic curveoptimized cut points for changes over2hours, they reported aPPVofonly35%for a relative change in cTnT level of 30%or greater comparedwith a PPV of 64%for an absolute change of 7 ng/L orgreater.44 This observation was sup-ported by a study of a highly sensitivecTnI assay in a similar population,whichidentified a large relative changeof 250%or greater over 3 hours as the optimalthreshold, yielding a PPV of 65%.33 Ex-clusive focus on such large change val-ues, however, would reduce sensitivityfor MI detection, because MIs can oc-cur with much smaller changes in tro-ponin levels. Thus, optimized thresh-olds for these assays may differ in theemergency department, where sensitiv-ity and high NPV are critical, com-pared with the inpatient floor, wherespecificity and high PPV are needed toguide decision making.28

Rapid MI Diagnostic ProtocolsWith Highly Sensitive Assays

One strategy to maximize advantagesof the lower detection range of thehighly sensitive assays is to focus noton high-end values above the MI diag-nosis threshold but rather on the lowend of the assay range, at the limit ofdetection. In one study of 703 pa-tients, 28% had a cTnT value at pre-

sentation below the 3 ng/L detectionthreshold. Although MI occurred in19% of the cohort, none of the indi-viduals with undetectable cTnT at pre-sentation subsequently ruled in (NPV,100%).45 In the study by Keller et al,33

which used a highly sensitive cTnI as-say, a similar proportion of individu-als had cTnI levels below the detec-tion limit of 3.4 ng/L at presentation(27%), and the NPV was also 100%.These preliminary findings suggest thatserialmonitoring of troponin levelsmaynot be necessary among patients withtroponin levels undetectable using ahighly sensitive assay, with the likelyexception of those presenting very early(1-2 hours) after symptom onset.Reichlin et al46 derived an algo-

rithm incorporating baseline values and1-hour absolute changes in cTnT lev-els among 436 randomly selected pa-tients with chest pain from the APACEstudy and validated the algorithm in theremaining 436 patients. The algo-rithm assigned individuals with both abaseline cTnT level lower than 12 ng/Land a change of less than 3 ng/L overthe first hour to the rule-out group(n=259 [60%]), those with a baselinecTnT level of 52 ng/L or greater or a1-hour change of 5 ng/L or greater tothe rule-in group (n=76 [17%]), andthe remaining 101 patients (23%) to anobservational zone group. Preva-lence of MI in the observational zonewas 8%. Sensitivity andNPVwere both100% in the rule-out group, with a 30-day mortality of only 0.2%. In therule-in group, specificity was 97% andPPV was 84%. If validated in addi-tional data sets, this algorithm, whichallowed safe rule out or relatively ac-curate rule in for more than three-fourths of patients within 1 hour, mayhelp to address some of the importantchallenges limiting application and in-terpretation of the highly sensitive tro-ponin assays in the United States.Clearly, the transition to increasinglysensitive troponin assays for MI diag-nosis will require clinicians to movefrom largely qualitative interpretationto amore quantitative assessment of tro-ponin levels.

Risk Assessment in the OutpatientSetting: A Better Use of the HighlySensitive Assays?The development of highly sensitive as-says has expanded the possible appli-cations of troponin testing, includingpotentially to the outpatient office visit.With the highly sensitive assay, cTnTcan be detected at very low concentra-tions in more than 90% of outpatientswith stable chronic heart failure47 orchronic CAD.48 In these chronic con-ditions, dose-dependent associationshave been observed between both cTnTand cTnI and the subsequent risk fordeath and heart failure, at levels wellbelow the detection threshold of stan-dard assays.47-49

More recently, population screen-ing with highly sensitive cTnT assayswas assessed in 3 large epidemiologiccohorts totaling more than 17 500 in-dividuals. The prevalence of detect-able cTnT (3 ng/L) with the highlysensitive assay ranged from 25% in theDallas Heart Study (ages 30-65 years)50

to 66.5% in the ARIC (AtherosclerosisRisk in Communities) study (ages54-74 years)39 and 66.2% in the CHS(Cardiovascular Health Study) (ages65 years).38 Concentrations of cTnTincreased with age and were higher inmen, African Americans, and individu-als withCKD.Cardiac troponin T dem-onstrated strong associations withpathological cardiac remodeling, in thatlevels increased in parallel with thepresence and severity of left ventricu-lar hypertrophy and left ventricular sys-tolic dysfunction.50 Of interest is thatprior MI, angina, and coronary cal-cium were not associated with detect-able levels of cTnT.In each population-based study,

higher cTnT level was associated withall-cause and cardiovascular mortalityand with incident heart failure,38,39,50,51

findings that remained significant af-ter adjustment for traditional risk fac-tors, renal function, and concentra-tions ofN-terminal pro-brainnatriureticpeptide (NT-proBNP) and high-sensitivity C-reactive protein (CRP).When added to traditional risk assess-ment models, cTnT improved metrics




of discrimination and risk classifica-tion, performing at least as well as NT-proBNP and outperforming high-sensitivity CRP.38,39,50 Although cTnTwas associatedwith coronary heart dis-ease events in the ARIC study, themag-nitude of association was weaker thanfor death and heart failure events.39

In a recent analysis from theFraming-ham Offspring Study incorporating anovelhighly sensitivecTnIassay, thepro-portion of adults with detectable cTnIlevelswas81%; in that study,higher cTnIlevels were associated with higher ratesdeath andheart failurebutnotMI,52 find-ings similar to those reported above forhighly sensitive cTnT assays.In the CHS, approximately two-

thirds of individuals underwent a fol-low-up cTnTmeasurement 2 to 3 yearsafter the baseline measurement. Inanalyses adjusting for baseline cTnTlevels as well as other risk predictors,increases in cTnT level of 50% ormoreduring follow-up identified individu-als at increased subsequent risk fordeath and heart failure, with de-creases in cTnT level associated withlower risk.38 This observation sug-gests that risk reflected by higher tro-ponin concentrations may be modifi-able.However, to date, only limited dataare available regarding the factors thatmight favorably influence troponin lev-els in the population. An exploratoryanalysis from the CHS demonstratedthat higher levels of baseline physicalactivity were associated with a lowerprobability of increased cTnT levelsover follow-up.53 In the ARIC study,higher hemoglobin A1c levels (eventhose below the diagnostic range fordiabetes) were associated with highercTnT levels measured years later.54

Muchwork still needs to be done be-fore highly sensitive troponin assays canbe considered in the outpatient set-ting. Additional research is needed toidentify treatments that prevent addi-tional cardiac injury andmodify risk as-sociated with elevated troponin lev-els. Although structural heart diseaseand CKD explain some of the varia-tion in troponin levels, many otherknown and unknown factors contrib-

ute as well. Troponins are agnosticmarkers of cardiac damage thatmay re-flect multiple different sources of in-jury. The effects and costs of noninva-sive cardiovascular imaging to identifythe source of chronic cardiac injury andidentify targets of therapy must be de-lineated prior to implementation of anyscreening strategy.55

CONCLUSIONSThe new highly sensitive troponin as-says will present myriad challenges forclinicians evaluating patientswith chestpain. The adoption of highly sensitiveassays for MI detection should be ac-companied by implementation of algo-rithms to enhance specificity and PPVand to shorten the period of observa-tion. In addition, recommendations foradditional testing and referral for pa-tients with increased cardiac troponinlevels and a low clinical suspicion forACS need to be developed.The application of highly sensitive

troponin assays for risk assessment inthe ambulatory setting is a potentiallyattractive approach tomaximize the ad-vantages of increased assay sensitiv-ity. In this setting, the detection of verylow troponin concentrations identi-fies risk not captured with other tools.Future studies should clarify the clini-cal value and identify the best ap-proach to incorporating these assays forpopulation screening.

Conflict of Interest Disclosures: The author has com-pleted and submitted the ICMJE Form for Disclosureof Potential Conflicts of Interest and reported servingas a consultant for Janssen Pharmaceuticals, Dia-dexus Inc, AstraZeneca, Bristol-Myers Squibbsanofi-aventis, Tethys Bioscience, St Jude Medical, and Dai-ichi Sankyo; receiving grants or grants pending fromRoche Diagnostics and Abbott Diagnostics; and re-ceiving payment for lectures from Bristol-MyersSquibbsanofi-aventis and AstraZeneca.Funding/Support:Dr de Lemos has received grant sup-port as the principal investigator for investigator-initiated grants from Roche Diagnostics and AbbottDiagnostics.Role of the Sponsors: Roche Diagnostics and AbbottDiagnostics had no role in the preparation, review, orapproval of the manuscript or the decision to submitthe manuscript for publication.

REFERENCES

1. Anderson JL, Adams CD, Antman EM, et al; Ameri-can College of Cardiology; American Heart Associa-tion Task Force on PracticeGuidelines. ACC/AHA2007guidelines for the management of patients with un-stable angina/non-ST-elevation myocardial infarc-

tion: a report of the American College of Cardiology/American Heart Association Task Force on PracticeGuidelines. J Am Coll Cardiol. 2007;50(7):e1-e157.2. OGara PT, Kushner FG, Ascheim DD, et al. 2013ACCF/AHA guideline for the management of ST-elevationmyocardial infarction: a report of the Ameri-can College of Cardiology Foundation/AmericanHeartAssociation Task Force on Practice Guidelines. J AmColl Cardiol. 2013;61(4):e78-e140.3. Bhuiya FA, Pitts SR, McCaig LF. Emergency de-partment visits for chest pain and abdominal pain:United States, 1999-2008. NCHS Data Brief. 2010;(43):1-8.4. Thygesen K, Alpert JS, Jaffe AS, et al; Joint ESC/ACCF/AHA/WHF Task Force for the Universal Defi-nition of Myocardial Infarction. Third universal defi-nition of myocardial infarction. Circulation. 2012;126(16):2020-2035.5. Apple FS. Tissue specificity of cardiac troponin I,cardiac troponin T and creatine kinase-MB.Clin ChimActa. 1999;284(2):151-159.6. Remppis A, Scheffold T, Greten J, et al. Intracel-lular compartmentation of troponin T: release kinet-ics after global ischemia and calcium paradox in theisolated perfused rat heart. J Mol Cell Cardiol. 1995;27(2):793-803.7. Reichlin T, Hochholzer W, Bassetti S, et al. Earlydiagnosis of myocardial infarction with sensitive car-diac troponin assays. N Engl J Med. 2009;361(9):858-867.8. Keller T, Zeller T, Peetz D, et al. Sensitive troponinI assay in early diagnosis of acutemyocardial infarction.N Engl J Med. 2009;361(9):868-877.9. Than M, Cullen L, Aldous S, et al. 2-Hour accel-erated diagnostic protocol to assess patients with chestpain symptoms using contemporary troponins as theonly biomarker: the ADAPT trial. J Am Coll Cardiol.2012;59(23):2091-2098.10. Bonaca M, Scirica B, Sabatine M, et al. Prospec-tive evaluation of the prognostic implications of im-proved assay performance with a sensitive assay forcardiac troponin I. J Am Coll Cardiol. 2010;55(19):2118-2124.11. Lindahl B, Diderholm E, Lagerqvist B, Venge P,Wallentin L; FRISC II (Fast Revascularization duringInStability in CAD) Investigators. Mechanisms be-hind the prognostic value of troponin T in unstablecoronary artery disease: a FRISC II substudy. J AmCollCardiol. 2001;38(4):979-986.12. Wong GC, Morrow DA, Murphy S, et al. Eleva-tions in troponin T and I are associated with abnor-mal tissue-level perfusion: a TACTICS-TIMI 18substudy. Circulation. 2002;106(2):202-207.13. HammCW,Heeschen C, Goldmann B, et al; c7E3Fab Antiplatelet Therapy in Unstable Refractory An-gina (CAPTURE) Study Investigators. Benefit of ab-ciximab in patients with refractory unstable angina inrelation to serum troponin T levels.NEngl JMed. 1999;340(21):1623-1629.14. Heeschen C, Hamm CW, Goldmann B, Deu A,Langenbrink L,White HD; PRISM Study Investigators.Troponin concentrations for stratification of patientswith acute coronary syndromes in relation to thera-peutic efficacy of tirofiban. Lancet. 1999;354(9192):1757-1762.15. Lindahl B, Venge P, Wallentin L; Fragmin in Un-stable Coronary Artery Disease (FRISC) Study Group.Troponin T identifies patients with unstable coronaryartery disease who benefit from long-term antithrom-botic protection. J Am Coll Cardiol. 1997;29(1):43-48.16. MorrowDA,CannonCP,RifaiN,etal;TACTICS-TIMI18 Investigators. Ability of minor elevations of tro-ponins I and T to predict benefit from an early invasivestrategy in patients with unstable angina andnon-ST elevation myocardial infarction: results froma randomized trial. JAMA. 2001;286(19):2405-2412.




17. Bavry AA, Kumbhani DJ, Quiroz R, RamchandaniSR, Kenchaiah S, Antman EM. Invasive therapy alongwith glycoprotein IIb/IIIa inhibitors and intracoro-nary stents improves survival in nonST-segment el-evation acute coronary syndromes: ameta-analysis andreview of the literature. Am J Cardiol. 2004;93(7):830-835.18. Jeremias A, Gibson CM. Narrative review: alter-native causes for elevated cardiac troponin levels whenacute coronary syndromes are excluded. Ann InternMed. 2005;142(9):786-791.19. Newby LK, Jesse RL, Babb JD, et al. ACCF 2012expert consensus document on practical clinical con-siderations in the interpretation of troponin eleva-tions: a report of the American College of CardiologyFoundation task force on Clinical Expert ConsensusDocuments. J Am Coll Cardiol. 2012;60(23):2427-2463.20. Mingels A, Jacobs L, Michielsen E, SwaanenburgJ, Wodzig W, van Dieijen-Visser M. Reference popu-lation and marathon runner sera assessed by highlysensitive cardiac troponin T and commercial cardiactroponin T and I assays. Clin Chem. 2009;55(1):101-108.21. Baker P, Davies SL, Larkin J, et al. Changes tothe cardiac biomarkers of non-elite athletes complet-ing the 2009 London Marathon [published onlineMarch 19, 2013]. Emerg Med J. 2013. doi:10.1136/emermed-2012-201465.22. Eggers KM, Lagerqvist B, Oldgren J, Venge P,Wallentin L, Lindahl B. Pathophysiologic mecha-nisms of persistent cardiac troponin I elevation in sta-bilized patients after an episode of acute coronarysyndrome. Am Heart J. 2008;156(3):588-594.23. Sato Y, Yamada T, Taniguchi R, et al. Persis-tently increased serum concentrations of cardiac tro-ponin T in patients with idiopathic dilated cardiomy-opathy are predictive of adverse outcomes.Circulation.2001;103(3):369-374.24. Wallace TW, Abdullah SM, Drazner MH, et al.Prevalence and determinants of troponin T elevationin the general population. Circulation. 2006;113(16):1958-1965.25. Eggers KM, Lind L, Ahlstrom H, et al. Prevalenceand pathophysiological mechanisms of elevated car-diac troponin I levels in a population-based sample ofelderly subjects. Eur Heart J. 2008;29(18):2252-2258.26. Leistner DM, Klotsche J, Pieper L, et al; DETECTStudy Group. Circulating troponin as measured by asensitive assay for cardiovascular risk assessment in pri-mary prevention. Clin Chem. 2012;58(1):200-208.27. Alcalai R, Planer D, Culhaoglu A, Osman A, PollakA, Lotan C. Acute coronary syndrome vs nonspecifictroponin elevation: clinical predictors and survivalanalysis. Arch Intern Med. 2007;167(3):276-281.28. Korley FK, Jaffe AS. Preparing the United Statesfor high-sensitivity cardiac troponin assays. J Am CollCardiol. 2013;61(17):1753-1758.29. PeacockWF IV, DeMarco T, Fonarow GC, et al;ADHERE Investigators. Cardiac troponin and out-

come in acute heart failure. N Engl J Med. 2008;358(20):2117-2126.30. Giannitsis E,Muller-BardorffM, Kurowski V, et al.Independent prognostic value of cardiac troponin Tin patients with confirmed pulmonary embolism.Circulation. 2000;102(2):211-217.31. Ammann P, Maggiorini M, Bertel O, et al. Tro-ponin as a risk factor for mortality in critically ill pa-tients without acute coronary syndromes. J Am CollCardiol. 2003;41(11):2004-2009.32. Apple FS,MurakamiMM, Pearce LA, Herzog CA.Predictive value of cardiac troponin I and T for sub-sequent death in end-stage renal disease.Circulation.2002;106(23):2941-2945.33. Keller T, Zeller T, Ojeda F, et al. Serial changes inhighly sensitive troponin I assay and early diagnosisof myocardial infarction. JAMA. 2011;306(24):2684-2693.34. Napoli AM, Arrighi JA, Siket MS, Gibbs FJ. Phy-sician discretion is safe and may lower stress test uti-lization in emergency department chest pain unitpatients. Crit Pathw Cardiol. 2012;11(1):26-31.35. Mahler SA, Hiestand BC, Goff DC Jr, HoekstraJW, Miller CD. Can the HEART score safely reducestress testing and cardiac imaging in patients at lowrisk for major adverse cardiac events? Crit PathwCardiol. 2011;10(3):128-133.36. de Lemos JA, Morrow DA, deFilippi CR. Highlysensitive troponin assays and the cardiology commu-nity: a love/hate relationship? Clin Chem. 2011;57(6):826-829.37. Januzzi JL Jr, Bamberg F, Lee H, et al. High-sensitivity troponin T concentrations in acute chest painpatients evaluatedwith cardiac computed tomography.Circulation. 2010;121(10):1227-1234.38. deFilippi CR, de Lemos JA, Christenson RH, et al.Association of serial measures of cardiac troponin Tusing a sensitive assay with incident heart failure andcardiovascular mortality in older adults. JAMA. 2010;304(22):2494-2502.39. Saunders JT, Nambi V, de Lemos JA, et al. Car-diac troponin T measured by a highly sensitive assaypredicts coronary heart disease, heart failure, andmor-tality in the Atherosclerosis Risk in Communities Study.Circulation. 2011;123(13):1367-1376.40. Apple FS, Ler R, Murakami MM. Determinationof 19 cardiac troponin I and T assay 99th percentilevalues froma commonpresumably healthy population.Clin Chem. 2012;58(11):1574-1581.41. Morrow DA, Cannon CP, Jesse RL, et al; Na-tional AcademyofClinical Biochemistry. National Acad-emy of Clinical Biochemistry LaboratoryMedicine Prac-tice Guidelines: clinical characteristics and utilizationof biochemical markers in acute coronary syndromes.Clin Chem. 2007;53(4):552-574.42. Turer AT, Addo TA, Martin JL, et al. Myocardialischemia induced by rapid atrial pacing causes tro-ponin T release detectable by a highly sensitive as-say: insights from a coronary sinus sampling study.J Am Coll Cardiol. 2011;57(24):2398-2405.43. Sabatine MS, Morrow DA, de Lemos JA, JarolimP, Braunwald E. Detection of acute changes in circu-

lating troponin in the setting of transient stresstestinducedmyocardial ischaemia using an ultrasen-sitive assay: results from TIMI 35. Eur Heart J. 2009;30(2):162-169.44. Reichlin T, Irfan A, Twerenbold R, et al. Utility ofabsolute and relative changes in cardiac troponin con-centrations in the early diagnosis of acute myocardialinfarction. Circulation. 2011;124(2):136-145.45. Body R, Carley S, McDowell G, et al. Rapid ex-clusion of acute myocardial infarction in patients withundetectable troponin using a high-sensitivity assay.J Am Coll Cardiol. 2011;58(13):1332-1339.46. Reichlin T, Schindler C, Drexler B, et al. One-hour rule-out and rule-in of acute myocardial infarc-tion using high-sensitivity cardiac troponin T.Arch In-tern Med. 2012;172(16):1211-1218.47. Latini R, Masson S, Anand IS, et al; Val-HeFTInvestigators. Prognostic value of very low plasma con-centrations of troponin T in patients with stable chronicheart failure. Circulation. 2007;116(11):1242-1249.48. Omland T, de Lemos JA, Sabatine MS, et al; Pre-vention of Events with Angiotensin Converting En-zyme Inhibition (PEACE) Trial Investigators. A sensi-tive cardiac troponin T assay in stable coronary arterydisease. N Engl J Med. 2009;361(26):2538-2547.49. Omland T, PfefferMA, Solomon SD, et al; PEACEInvestigators. Prognostic value of cardiac troponin Imeasuredwith a highly sensitive assay in patients withstable coronary artery disease. J AmColl Cardiol. 2013;61(12):1240-1249.50. de Lemos JA, Drazner MH, Omland T, et al. As-sociation of troponin T detected with a highly sensi-tive assay and cardiac structure and mortality risk inthe general population. JAMA. 2010;304(22):2503-2512.51. Neeland IJ, DraznerMH, Berry JD, et al. Biomark-ers of chronic cardiac injury and hemodynamic stressidentify a malignant phenotype of left ventricular hy-pertrophy in the general population. J AmColl Cardiol.2013;61(2):187-195.52. Wang TJ, Wollert KC, Larson MG, et al. Prog-nostic utility of novel biomarkers of cardiovascularstress: the FraminghamHeart Study.Circulation. 2012;126(13):1596-1604.53. deFilippi CR, de Lemos JA, TkaczukAT, et al. Physi-cal activity, change in biomarkers of myocardialstress and injury, and subsequent heart failure risk inolder adults. J Am Coll Cardiol. 2012;60(24):2539-2547.54. Rubin J,Matsushita K, Ballantyne CM,HoogeveenR, Coresh J, Selvin E. Chronic hyperglycemia and sub-clinical myocardial injury. J Am Coll Cardiol. 2012;59(5):484-489.55. HlatkyMA, Greenland P, Arnett DK, et al; Ameri-can Heart Association Expert Panel on Subclinical Ath-erosclerotic Diseases and Emerging Risk Factors andthe Stroke Council. Criteria for evaluation of novelmarkers of cardiovascular risk: a scientific statementfrom the American Heart Association [published cor-rection appears in Circulation. 2009;119(25):e606].Circulation. 2009;119(17):2408-2416.


