Timing of ischemic onset estimated from theelectrocardiogram is better than historical timingfor predicting outcome after reperfusion therapy foracute anterior myocardial infarction: A DANish trial inAcute Myocardial Infarction 2 (DANAMI-2) substudyMaria Sejersten, MD,a Rasmus S. Ripa, MD,a Charles Maynard, PhD,b Peer Grande, MD, DMSc,a

Henning Rud Andersen, MD, DMSc,c Galen S. Wagner, MD,d and Peter Clemmensen, MD, DMSc,a

for the DANAMI-2 investigators Copenhagen and Aarhus, Denmark; Seattle, WA; and Durham, NC

Background Acute treatment strategy and subsequently prognosis are influenced by the duration of ischemia inpatients with ST-elevation acute myocardial infarction (AMI). However, timing of ischemia may be difficult to access bypatient history (historical timing) alone. We hypothesized that an electrocardiographic acuteness score is better thanhistorical timing for predicting myocardial salvage and prognosis in patients with anterior AMI treated with fibrinolysis orprimary percutaneous coronary intervention.

Methods One hundred seventy-five patients with anterior infarct without electrocardiogram (ECG) confounding factorswere included. The ECG method for estimating timing of AMI was calculated using core laboratory measurements from theinitial 12-lead ECG. Historical timing was recorded as time from symptom onset to initiation of reperfusion therapy.Myocardial salvage was determined by ECG, using the Aldrich score to determine the initially predicted myocardial infarctsize and the Selvester score to determine the final QRS-estimated myocardial infarct size.

Results The mean amount of myocardium salvage depended on ECG timing (43% [F38%] for bearlyQ vs 1% [F56%] forblateQ; P b .001), whereas myocardial salvage was independent of historical timing ( P = .9). One-year mortality waspredicted from ECG timing ( P = .04).

Conclusions The ECG method of timing was superior to historical timing in predicting myocardial salvage andprognosis after reperfusion therapy. This study suggests that ECG estimated duration of ischemia might provide a better andobjective means to select acute reperfusion therapy rather than the subjective patient history, which could preclude properreperfusion in some patients with salvageable myocardium. (Am Heart J 2007;154:61.e1261.e8.)

In Denmark, approximately 12600 patients suffer

from acute myocardial infarction (AMI) each year.1 It is

well known that time from symptom onset to treatment

is essential in patients with AMI because mortality has

been shown to increase over time.2,3 Current guidelines

state that reperfusion should be initiated within 30

From the aDepartment of Cardiology, The Heart Centre, Rigshospitalet, Copenhagen

University Hospital, Copenhagen, Denmark, bDepartment of Health Services, University

of Washington, Seattle, WA, cDepartment of Cardiology, Skejby University Hospital,

Aarhus, Denmark and dDuke Clinical Research Institute, Durham, NC.

This study was supported by travel grants from the Danish Heart Foundation.

Submitted January 18, 2007; accepted April 1, 2007.

Reprint requests: Dr Maria Sejersten, Department of Cardiology 2142, Rigshospitalet,

Copenhagen University Hospital, Blegdamsvej 9, 2100 Copenhagen a, Denmark.

0002-8703/$ - see front matter

D 2007, Mosby, Inc. All rights reserved.

doi:10.1016/j.ahj.2007.04.003

minutes in patients treated with fibrinolysis and 90

minutes in patients treated with primary percutaneous

coronary intervention (pPCI).4,5 However, these times

are seldom the reality6; and many initiatives have been

taken to reduce time to treatment.7 The information

from patients and bystanders regarding time of symptom

onset is the primary indicator in current clinical use for

determining the duration of the AMI process and

thereby indicating the potential for achieving myocardial

salvage via reperfusion therapy. However, this informa-

tion can be a poor time indicator for the manifestation of

the acute coronary thrombosis because of inaccurate

recollection, nonspecific symptoms, or bsilent ischemia.QBecause the prognosis and the optimal treatment

strategy are influenced by the duration of the ischemia,

it is desirable to have an independent parameter other

than patient history (historical timing) to assist when

Table I. Electrocardiographic criteria for inclusion based onepicardial injury and anterior infarct location

Epicardial injury Anterior AMI location

z1 mm STz in oneor more leads

Max STD is STzin leads V1 to V3

Max STz z max STAMax STz in lead

V2 N STz in lead V1

STz, ST-segment elevation; STA, ST-segment depression; STD, ST-segment deviation;Max, maximal.

Figure 1

A, Electrocardiographic examples of a patient with phase 1A (tallT wave and no abnormal Q wave). B, Electrocardiographicexamples of a patient with phase 2B (positive T wave and anabnormal Q wave).

American Heart Journal

July 200761.e2 Sejersten et al

determining the time interval from thrombotic occlusion

to clinical presentation.

Anderson et al8 and Wilkins et al9 have developed an

electrocardiogram (ECG) acuteness score as a comple-

ment to the historical timing of the evolving infarction

process in the individual patient in the acute situation.

Corey et al10 have documented that the Anderson-

Wilkins (AW) acuteness score provides an equally useful

basis as historical timing for predicting final AMI size

after reperfusion therapy and that the AW acuteness

score and historical timing provide complementary

value. We hypothesized that the AW acuteness score

is better than historical timing for predicting

myocardial salvage and prognosis in patients with

anterior AMI treated with fibrinolysis or primary angio-

plasty (pPCI).

MethodsPatient population

The study design, randomization procedures, and inclusion/

exclusion criteria of the DANAMI-2 trial have been previously

published in detail.11 Patients with ischemic chest discomfort

for V12 hours and an ECG with cumulated ST-segment

elevation z4 mm were eligible for enrollment. Patients were

randomly assigned to fibrinolysis or pPCI. The study was

approved by the National Ethics Committee of Denmark and

complied with the Declaration of Helsinki. All eligible patients

provided written informed consent. The results of the

DANAMI-2 trial have been published.12

The last 500 consecutive patients randomized in the

DANAMI-2 trial were available for consideration in this post

hoc study. Electrocardiographic evidence of epicardial injury

and anterior infarct location, as defined in Table I, was

required. Furthermore, patients were only included if there

were no ECG evidence of bundle-branch or fascicular block,

ventricular hypertrophy, or ventricular paced rhythm. Each

patient had 2 standard resting ECGs (12-lead) analyzed: (1) a

randomization ECG and (2) a follow-up ECG at predischarge or,

if such an ECG was not available, at 1-month follow-up. The

ECGs were analyzed at the ECG core laboratory at Rigshospi-

talet in Copenhagen, Denmark. The ECG core laboratory was

blinded to all patient data. Four experienced ECG readers

analyzed the ECGs. Each patient had both the admission and

the follow-up ECGs analyzed by the same reader. A single

reader did the acuteness scoring on the admission ECG for all

included patients.

Timing indexesThe time of onset of the acute symptoms that resulted in

admission to the hospital was acquired by emergency

department personnel and recorded on the clinical trial

case report form. Historical timing was defined as time

from symptom onset until initiation of thrombolytic therapy

or initial puncture of the femoral artery in the

catheterization laboratory.

The AW acuteness score considered each standard lead

(except aVR) with either z0.1 mV ST elevation or babnormally

tallQ T waves.13 An acuteness phase was designated for each of

these leads based on the presence or absence of a tall T wave

or an abnormal Q wave (criteria previously published)14,15:

phase 1A, tall T wave and no abnormal Q wave; phase 1B,

positive T wave and no abnormal Q wave; phase 2A, tall T

wave and an abnormal Q wave; phase 2B, positive T wave and

an abnormal Q wave. Figure 1, A and B illustrate phase 1A and

phase 2B. The AW acuteness score ranges from 1.0 (late/least

Table II. Demographics

All

Quartiles of acuteness score

P1 Early 2 3 4 Late

n 175 44 27 60 44Age (mean) 61 56 58 63 64 b.001Female sex (%) 22 21 19 25 21 .8Weight (mean kg) 79 81 76 79 79 .6Diabetes4 (%) 7 0 4 10 14 b.01Hypertension Diabetes4 (%) 16 16 7 19 16 .7Hypercholesterol Diabetes4 (%) 4 2 12 3 2 .2Current smoker (%) 82 84 78 83 81 .9Previous angina (%) 26 23 19 30 30 .6Previous MI (%) 9 7 4 12 11 .6Previous PCI (%) 5 2 7 7 2 .5Historical timing 3:36 3:03 3:09 3:50 4:09 .03Initially jeopardized myocardiumy (%) 22 22 24 22 21 .5Randomized for pPCI (%) 45 46 44 42 48 .9

4Receiving treatment.yBy Aldrich score.7

American Heart Journal

Volume 154, Number 1Sejersten et al 61.e3

acute) to 4.0 (early/most acute) and was calculated from the

following formula:

4ð# leads 1AÞþ3ð# leads 1BÞþ2ð# leads 2AÞþ1ð# leads 2BÞTotal # leads with 1A; 1B; 2A; or2B

The original method has been validated for anterior AMI and

changed for inferior AMI.15

Myocardial salvageThe myocardium at risk for infarction was estimated by

summated ST-segment deviation on the admission ECG

according to the Aldrich ST score.16 The Aldrich ST score was

initially developed in a population not receiving reperfusion

therapy, and thus the method predicts the area at risk of

infarction without reperfusion treatment as a percentage of the

left ventricle. The formula has previously been validated for

anterior AMI and changed for inferior AMI.17

The final infarct size was estimated from the predischarge

ECG by the Selvester QRS score.14 The system contains 50

criteria (considering Q and R wave durations and relative Q, R,

and S wave amplitudes) awarding a maximum of 31 points, each

representing approximately 3% infarction of the left ventricle.

This scoring system was originally developed from anatomical

studies of anterior and inferior infarcts and has since been

validated using technetium Tc 99m pyrophosphate single

photon emission computed tomography (SPECT) (r = 0.78)18

and delayed enhancement magnetic resonance imaging.19 The

myocardial salvage index was calculated as follows:20 percent

salvage = 100 [(initially predicted MI size � final QRS-estimated

MI size) / initially predicted MI size]. A prior study21 has found a

moderate relationship between the ECG salvage score and

salvage determined by SPECT.

Clinical end pointThe 2 end points used were death from any cause and

clinical reinfarction at 1 year of follow-up. Detailed definitions

of these end points are available elsewhere.11 An end point

committee that was unaware of the treatment-group assign-

ments reviewed all end point events.22

Statistical analysisAnalyses were performed using SPSS statistical software

(SPSS version 13.0, SPSS Inc, Chicago, IL). By protocol,

historical timing and AW acuteness score were divided into

quartiles to ensure equal-sized groups. The association be-

tween historical time and AW acuteness score was assessed

using Pearson r correlation coefficient. Categorical variables

were compared using the m2 test. Means were compared with

the t test or the analysis of variance test for linearity.

Multivariate linear regression was used to identify predictors

of salvage with special attention to the 2 timing variables.

First, a model including only the 2 timing variables with and

without interaction was used. Second, traditional risk factors

including age, sex, diabetes, and symptom duration were

included into a model with stepwise backward elimination.

The Kaplan-Meier method was used to construct survival

curves, and the log-rank statistic was used to compare survival.

The Cox regression analysis was used to adjust for covariates.

The level of statistical significance was set at P b .05.

ResultsThe study group

Of the last 500 patients randomized in the DANAMI-2

trial, 232 (46%) had an anterior infarct (Table I). Of

these patients, 25% were excluded because of ECG

evidence of left ventricular hypertrophy (n = 16),

complete right bundle-branch block (n = 7), or left

anterior fascicular block (n = 34), resulting in a study

population of 175 patients. Admission ECGs were

available for all 175 patients, whereas predischarge

ECGs were available for 168 patients. Demographics

for all patients and patients grouped according to

Figure 2

Relationship between the AW acuteness score and myocardial salvage. Patients are grouped into quartiles of AW acuteness score based on theinitial ECG. Mean salvage is the percentage of the initially jeopardized myocardium. Numbers above the bars indicate population size. A, Allpatients. B, Patients divided by type of treatment.

Figure 3

Relationship between historical timing and myocardial salvage. Patients are grouped into quartiles of historical timing. Mean salvage is given asthe percentage of the initially jeopardized myocardium. Numbers above the bars indicate population size. A, All patients. B, Patients divided bytype of treatment.

American Heart Journal

July 200761.e4 Sejersten et al

quartiles of AW acuteness score are shown in

Table II. Age, diabetes, and symptom duration were

unevenly distributed among the 4 patient groups.

Patients with long ischemic timing judged by acuteness

score were also presenting later, were older, and

more frequently had diabetes. The correlation between

Figure 4

Mean percentage of myocardial salvage for patients grouped byboth historical timing (V2 vs N2 hours) and AW acuteness score (N3vs V3.0). The numbers of patients included are indicated directlyabove each of the bars.

American Heart Journal

Volume 154, Number 1Sejersten et al 61.e5

historical timing and AW acuteness score was low

(r = �0.2, P = .1).

Myocardial salvageMyocardial salvage could be determined for 168

patients. Figure 2 displays mean myocardial salvage with

patients divided into quartiles based on AW acuteness

score. There was an association between AW acuteness

score and myocardial salvage (Figure 2, A) ( P b .001).

The SDs within the quartiles of acuteness score were

as follows: quartile 1, 38%; quartile 2, 51%; quartile 3,

58%; and quartile 4, 56%. The association was

statistically significant regardless of reperfusion

strategy (Figure 2, B) (thrombolytic therapy, P = .005;

pPCI, P = .007).

Figure 3 demonstrates mean myocardial salvage with

patients grouped according to historical timing. There

was no statistically significant association between

historical timing and myocardial salvage for all patients

( P = .9) (Figure 3, A). The SDs within the 4 timing

quartiles were as follows: quartile 1, 50%; quartile 2,

49%; quartile 3, 64%; and quartile 4, 51%. When analyzed

by treatment group, there was no association between

historical timing and myocardial salvage for patients

randomized to thrombolytic therapy ( P = .9) or pPCI

( P = .6) (Figure 3, B).

There were no significant interactions ( P = .4)

between historical timing and AW acuteness score in a

linear regression model with myocardial salvage index as

dependent variable. The myocardial salvage index was

related to AW acuteness score with a regression

coefficient of 14.4 (95% CI 4.8-24.0), whereas the

equivalent regression coefficient for historical timing

was 1.8 per hour (95% CI 1.3-5.0).

Stepwise linear regression analysis was performed

with age, sex, diabetes, treatment group (fibrinolysis or

pPCI), AW acuteness score, and historical timing as

independent variables. Only low AW acuteness score

(regression coefficient 12.8, 95% CI 3.5-22.2) and male

sex (regression coefficient 21.3, 95% CI 1.7-40.9)

independently predicted poor myocardial salvage. These

results are set into clinical perspective in Figure 4.

Patients are divided into 4 subgroups according to both

median historical timing and median AW acuteness

score. Both patients presenting with short and long

historical timing attained more salvage if they were early

by ECG timing.

Clinical outcomesThe clinical outcomes at 1 year for early and late

timing groups dichotomized by the median value are

shown in Table III and Figures 5 and 6. The relative

difference in the rate of deaths was 84% ( P = .04),

whereas there was no difference in risk of reinfarction

when comparing patients who were early versus late

according to ECG timing. Figure 5, A illustrates that the

difference in mortality was primarily driven by a large

difference within the first 2 weeks after the AMI. When

adjusting for covariates, the difference in mortality was

no longer statistically significant ( P = .14). There were

no differences in mortality or reinfarction when com-

paring patients who were early versus late according to

historical timing. However, Figure 6, A indicates an early

benefit with historical timing b2 hours.

DiscussionThe major finding of this study was a strong relation-

ship between an ECG timing method and myocardial

salvage after acute reperfusion therapy for AMI. This was

followed by a marked difference in mortality during the

first year, although this difference only achieved mar-

ginal statistical significance.

The combination of ECG timing and historical

timing (Figure 4) identified a small group of patients

with short historical timing but late ECG timing. These

patients would be expected to have a large potential for

salvage due to the short historical timing; but converse-

ly, they have no salvage. Corey et al10 studied a similar

patient group using final infarct size (by thallium-201

SPECT) as end point rather than myocardial salvage by

ECG measurements. They found similar results for

patients with long historical timing, whereas final infarct

size was found to be independent of AW acuteness score

Table III. Clinical outcome at 1 year among 175 patients with anterior ST-elevation AMI according to duration of ischemia assessed from theECG or reported historical timing

Early by ECG (n = 71) Late by ECG (n = 104)

P

Early by time (n = 88) Late by time (n = 87)

Pn (%) n (%)

Death 1 (1.4) 9 (8.7) .04 4 (4.5) 6 (6.9) .5Reinfarction 5 (7.0) 8 (7.7) .9 8 (9.1) 5 (5.7) .4

Figure 5

Kaplan-Meier curves showing cumulative event rates for death (A)and clinical reinfarction (B) during 1 year of follow-up. Patients aredivided by AW acuteness score (N3.0 vs V3.0).

Figure 6

Kaplan-Meier curves showing cumulative event rates for death (A)and clinical reinfarction (B) during 1 year of follow-up. Patients aredivided by historical timing (N2.0 vs V2.0 hours).

American Heart Journal

July 200761.e6 Sejersten et al

American Heart Journal

Volume 154, Number 1Sejersten et al 61.e7

if patients had short historical timing. This difference

could be explained by the different end point and by the

fact that all patients in the study by Corey et al received

fibrinolytic therapy.

Previous studies have suggested that the association

between time to treatment and final infarct size/

myocardial salvage depends upon reperfusion strategy;

that is, the association is strong if patients are treated

with fibrinolysis and weak if treated with primary

angioplasty.23 In the present study, historical timing was

unrelated to myocardial salvage irrespective of reperfu-

sion strategy. This could have several explanations. Time

of symptom onset is a subjective measurement, and it

was not a high priority of the DANAMI-2 trial to obtain

the precise time. The exact time of symptom onset can

further sometimes be difficult to determine because of

intermittent spontaneous reperfusion, prodromal angi-

na, and inaccurate recollection. The AW acuteness score

on the other hand is an objective measurement. This

study suggests that the AW acuteness score has similar

value after thrombolysis and pPCI, which is in contrast

to the divergent reports regarding the importance of

time to treatment.23,24

Figure 5 shows that the initial difference in salvage

when stratifying patients by AW acuteness score

resulted in a long-term difference in mortality, whereas

the incidence of reinfarction was independent of the

AW acuteness score. The difference resulted in a relative

mortality reduction of 84% after 1 year driven primarily

by a large difference within the first 2 weeks after

reperfusion therapy. This is the first study addressing

the prognostic value of the AW acuteness score.

The difference between the 2 acuteness groups was

based on only 10 events and should be validated in

larger populations.

This study only included patients with anterior

location of the AMI. Anderson-Wilkins acuteness

scoring is also possible for inferior AMI. However, the

method has not been validated for inferior AMI; and

thus, the scoring was not carried out for patients with

inferior AMI in the DANAMI-2 ECG core laboratory.

Corey et al10 has previously analyzed inferior AMI and

showed that the AW acuteness score did predict final

infarct size, but with less strength than for patients

with anterior AMI. A modification of the AW acuteness

score15 has been published, and it can be expected

that the modified AW acuteness score will predict

salvage equally well for anterior and inferior AMI. This

modification only affects patients with inferior AMI,

and the results presented here will thus remain

unchanged using the modified score.

The AW acuteness score is time consuming to

calculate by hand, and it is unrealistic to incorporate into

the clinical triage decision without automatic calculation

by digitized ECG machines. A study by Ripa et al25

has shown that automatic calculation of the AW

acuteness score is feasible and correlates well with the

score calculated by hand.

LimitationsIt is a limitation of the AW acuteness score that only

patients without ECG confounders such as bundle-

branch block and prior AMI can be considered. The

scoring should however still be applicable in at least half

of all AMI patients and therefore have great clinical

relevance. The AW acuteness phases (1A, 1B, 2A, 2B)

have, in the present form of the equation, equidistance

in their timing of the pathophysiological process; that is,

the coefficients have arbitrarily been chosen as 1 to 4. It

must be assumed that the time phases have different

biological duration, which should be reflected in the

time coefficients. It will be of future interest to identify

the true time coefficients.

Inclusion of all patients randomized in the DANAMI-2

trial would have been preferable, but only the last 500

consecutive patients were available.

ConclusionsThe findings of this study suggest that the AW

acuteness score is superior to historical timing in

predicting the benefit of reperfusion therapy, thereby

potentially changing the proportion of ST-elevation AMI

patients eligible for reperfusion therapy. The informa-

tion contained by the acuteness score is further

substantiated by its long-term prognostic value.

References1. Videbaek J, Madsen M. 2004 HjerteStatistik. 1st. ed. Kbbenhavn,

Denmark7 Hjerteforeningen i samarbejde med Statens Institut forFolkesundhed; 2004.

2. Boersma E, Maas AC, Deckers JW, et al. Early thrombolytictreatment in acute myocardial infarction: reappraisal of the goldenhour. Lancet 1996;348:771-5.

3. De Luca G, Suryapranata H, Ottervanger JP, et al. Time delay totreatment and mortality in primary angioplasty for acute myocardialinfarction: every minute of delay counts. Circulation2004;109:1223 -5.

4. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelinesfor the management of patients with ST-elevation myocardialinfarction: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines(Committee to Revise the 1999 Guidelines for the Management ofPatients with Acute Myocardial Infarction). Circulation2004;110:e82-e292.

5. Van de Werf F, Ardissino D, Betriu A, et al. Management of acutemyocardial infarction in patients presenting with ST-segmentelevation. The Task Force on the Management of Acute MyocardialInfarction of the European Society of Cardiology. Eur Heart J2003;24:28 -66.

6. Nallamothu BK, Bates ER, Herrin J, et al. Times to treatment intransfer patients undergoing primary percutaneous coronary inter-vention in the United States: National Registry of MyocardialInfarction (NRMI)-3/4 analysis. Circulation 2005;111:761 -7.

American Heart Journal

July 200761.e8 Sejersten et al

7. Bradley EH, Herrin J, Wang Y, et al. Strategies for reducing thedoor-to-balloon time in acute myocardial infarction. N Engl J Med2006;355:2308 -20.

8. Anderson ST, Wilkins M, Weaver WD, et al. Electrocardiographicphasing of acute myocardial infarction. J Electrocardiol 1992;25(Suppl):3 -5.

9. Wilkins ML, Pryor AD, Maynard C, et al. An electrocardiographicacuteness score for quantifying the timing of a myocardial infarctionto guide decisions regarding reperfusion therapy. Am J Cardiol1995;75:617 -20.

10. Corey KE, Maynard C, Pahlm O, et al. Combined historical andelectrocardiographic timing of acute anterior and inferior myocar-dial infarcts for prediction of reperfusion achievable size limitation.Am J Cardiol 1999;83:826 -31.

11. Andersen HR, Nielsen TT, Vesterlund T, et al. Danish multicenterrandomized study on fibrinolytic therapy versus acute coronaryangioplasty in acute myocardial infarction: rationale and design ofthe DANish trial in Acute Myocardial Infarction-2 (DANAMI-2). AmHeart J 2003;146:234 -41.

12. Andersen HR, Nielsen TT, Rasmussen K, et al. A comparison ofcoronary angioplasty with fibrinolytic therapy in acute myocardialinfarction. N Engl J Med 2003;349:733 -42.

13. Gambill CL, Wilkins ML, Haisty Jr WK, et al. T wave amplitudesin normal populations. Variation with ECG lead, sex, and ageJ Electrocardiol 1995;28:191 -7.

14. Hindman NB, Schocken DD, Widmann M, et al. Evaluation of aQRS scoring system for estimating myocardial infarct size. V.Specificity and method of application of the complete system. Am JCardiol 1985;55:1485 -90.

15. Heden B, Ripa R, Persson E, et al. A modified Anderson-Wilkinselectrocardiographic acuteness score for anterior or inferior myo-cardial infarction. Am Heart J 2003;146:797 -803.

16. Aldrich HR, Wagner NB, Boswick J, et al. Use of initial ST-segmentdeviation for prediction of final electrocardiographic size of acutemyocardial infarcts. Am J Cardiol 1988;61:749 -53.

17. Clemmensen P, Grande P, Aldrich HR, et al. Evaluation of formulasfor estimating the final size of acute myocardial infarcts from

quantitative ST-segment elevation on the initial standard 12-leadECG. J Electrocardiol 1991;24:77 -83.

18. Juergens CP, Fernandes C, Hasche ET, et al. Electrocardiographicmeasurement of infarct size after thrombolytic therapy. J Am CollCardiol 1996;27:617 -24.

19. Engblom H, Wagner GS, Setser RM, et al. Quantitative clinicalassessment of chronic anterior myocardial infarction with delayedenhancement magnetic resonance imaging and QRS scoring. AmHeart J 2003;146:359 -66.

20. Clemmensen P, Grande P, Saunamaki K, et al. Effect of intravenousstreptokinase on the relation between initial ST-predicted size andfinal QRS-estimated size of acute myocardial infarcts. J Am CollCardiol 1990;16:1252-7.

21. Barbagelata A, Di M, Carli F, Califf RM, et al. Electrocardiographicinfarct size assessment after thrombolysis: insights from the AcuteMyocardial Infarction STudy, ADenosine (AMISTAD) trial. Am HeartJ 2005;150:659 -65.

22. Vejlstrup N, Clemmensen P, Steinmetz E, et al. Blinded end pointadjudication in the Danish multicenter randomized study onfibrinolytic therapy versus acute coronary angioplasty in acutemyocardial infarction (DANAMI-2 trial). Heart Drug 2003;3:127 -33.

23. Schomig A, Ndrepepa G, Mehilli J, et al. Therapy-dependentinfluence of time-to-treatment interval on myocardial salvagein patients with acute myocardial infarction treated withcoronary artery stenting or thrombolysis. Circulation 2003;108:1084 -8.

24. Zijlstra F, Patel A, Jones M, et al. Clinical characteristics andoutcome of patients with early (2 h), intermediate (2-4 h) and late(4 h) presentation treated by primary coronary angioplasty orthrombolytic therapy for acute myocardial infarction. Eur Heart J2002;23:550 -7.

25. Ripa RS, Persson E, Heden B, et al. Comparison betweenhuman and automated electrocardiographic waveform measure-ments for calculating the Anderson-Wilkins acuteness score inpatients with acute myocardial infarction. J Electrocardiol2005;38:96 -9.