Electrocardiographic Criteria to Differentiate Acute Anterior ST Elevation Myocardial Infarction from Left Ventricular Aneurysm Lauren R. Klein MD, Gautam R. Shroff MD, William Beeman MD, Stephen W. Smith MD PII: DOI: Reference:

S0735-6757(15)00190-4 doi: 10.1016/j.ajem.2015.03.044 YAJEM 54887

To appear in:

American Journal of Emergency Medicine

Received date: Revised date: Accepted date:

18 February 2015 19 March 2015 19 March 2015

Please cite this article as: Klein Lauren R., Shroff Gautam R., Beeman William, Smith Stephen W., Electrocardiographic Criteria to Differentiate Acute Anterior ST Elevation Myocardial Infarction from Left Ventricular Aneurysm, American Journal of Emergency Medicine (2015), doi: 10.1016/j.ajem.2015.03.044

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Electrocardiographic Criteria to Differentiate Acute Anterior ST Elevation Myocardial Infarction from Left Ventricular Aneurysm

a.

RI P

T

Lauren R. Klein, MDa, Gautam R. Shroff, MDb, William Beeman, MDa, Stephen W. Smith, MDa Department of Emergency Medicine, Hennepin County Medical Center, 701 Park Avenue, Minneapolis, Minnesota, USA

Division of Cardiology, Department of Medicine, Hennepin County Medical Center, 701 Park Avenue,

SC

b.

MA

NU

Minneapolis, Minnesota, USA

Corresponding author:

[email protected] Hennepin County Medical Center

CE

701 Park Avenue, MC 825

PT

Department of Emergency Medicine

ED

Lauren R. Klein, MD

AC

Minneapolis, Minnesota 55415

Keywords: Left ventricular aneurysm, electrocardiogram, ST elevation

Running Head: ECG of STEMI versus LVA

ACCEPTED MANUSCRIPT Abstract Background: ST elevation (STE) on the electrocardiogram (ECG) may be due to acute myocardial infarction (AMI), or other non-ischemic pathologies such as left ventricular aneurysm (LVA). The objective of this study was to

RI P

T

validate two previously derived ECG rules to distinguish AMI from LVA. The first rule states that if the sum of Twave amplitudes in leads V1-V4 divided by the sum of QRS amplitudes in leads V1-V4 is > 0.22, then acute STEMI is predicted. The second rule states that if any one lead (V1-V4) has a T-wave amplitude to QRS amplitude ratio ≥

SC

0.36, then acute STEMI is predicted.

NU

Methods: This was a retrospective analysis of patients with AMI (n=59) and LVA (n=16) who presented with

MA

ischemic symptoms and STE on the ECG. For each ECG, the T-wave amplitude and QRS amplitude in leads V1V4 were measured. These measurements were applied to the two ECG rules, and sensitivity, specificity, and

ED

accuracy in predicting AMI versus LVA was calculated.

Results: For rule 1 (sum of ratios in V1-V4), sensitivity was 91.5%, specificity was 68.8%, and accuracy was

PT

86.7% in predicting AMI. For rule 2 (maximum ratio in V1-V4), sensitivity was 91.5%, specificity was 81.3%, and

CE

accuracy was 89.3% in predicting AMI.

AC

Conclusions: When patients present to the emergency department with ischemic symptoms and the differential diagnosis for STE on the ECG is AMI versus LVA, these two ECG rules may be helpful in differentiating these two pathologies. Both rules are highly sensitive and accurate in predicting AMI versus LVA.

Keywords: Left ventricular aneurysm, electrocardiogram, ST elevation

2

ACCEPTED MANUSCRIPT 1. Introduction Reperfusion therapy for ST-segment elevation myocardial infarction (STEMI) is indicated for patients with ischemic symptoms and ST elevation (STE) on the electrocardiogram (ECG). Acute myocardial infarction (AMI),

RI P

T

however, is not the only etiology of STE on the ECG, which can significantly complicate the clinician’s reperfusion decision. One etiology of STE that is not AMI is “persistent STE after prior MI.” This is clinically and pathologically manifested as left ventricular aneurysm (LVA), and is found in up to 60% of completed anterior

SC

STEMI when there is no spontaneous or therapeutic reperfusion [1]. When these patients present to the emergency department (ED) with symptoms consistent with ischemia such as chest pain or shortness of breath, this persistent

NU

STE may result in a misinterpretation of the ECG and lead to inappropriate reperfusion therapy [2 – 5].

MA

The association between persistent STE on the ECG after prior MI and LVA has been well documented over the years, confirmed with autopsy studies [6 - 11] as well as echocardiography studies [12 - 15]. LVA is a well-known structural complication of acute myocardial infarction [16]. Anatomical LVA is defined as thinning and

ED

bulging of the myocardial wall. However, in the age of modern cardiac imaging such as echocardiography and ventriculography, the terminology “LVA” has come to encompass diastolic distortion, diastolic dyskinesis, as well

PT

as ventricular wall akinesis, even in the absence of anatomical aneurysm [12 – 15, 17].

CE

The question then arises how to differentiate STE due to AMI versus STE due to LVA. Echocardiography may be utilized but this is not always readily available in the emergency department, and it does not necessarily

AC

differentiate the akinesis of acute MI from prior MI, and may require the use of echocardiographic contrast for accurate diagnosis. There are several features of the ECG of LVA that can help distinguish it from AMI. These findings include the presence of deep Q-waves (usually QS-waves), flattened T-waves (often with some shallow Twave inversion), and a lesser degree of STE [8, 10, 16, 18]. None of these ECG features however are sensitive or specific for LVA. We previously published a derivation of ECG criteria to distinguish acute anterior STEMI from anterior LVA when there is STE on the ECG [19]. This study compared multiple ECG variables, but the variable that best differentiated the two groups was the ratio of T-wave amplitude to QRS amplitude, exploiting the fact that there are larger T-waves in acute STEMI. Based on these findings, we derived two “rules” utilizing ECG measurements to differentiate STE due to AMI from STE due to LVA. These rules were developed based on the median T-wave amplitude to QRS amplitude ratios for each group. The first rule states that if the sum of T-wave amplitudes in leads

3

ACCEPTED MANUSCRIPT V1-V4 divided by the sum of QRS amplitudes in leads V1-V4 is > 0.22, then acute STEMI is predicted. The second rule states that if any one lead (V1 through V4) has a T-wave amplitude to QRS amplitude ratio ≥ 0.36, then acute STEMI is predicted (Figure 1). In the original derivation study, both of these formulas were highly sensitive and

RI P

T

specific for predicting acute STEMI, suggesting good utility for clinicians who need to differentiate AMI from LVA when this is the ECG differential diagnosis. The purpose of the current study was to validate these two ECG rules in

SC

a separate cohort of patients.

NU

2. Methods 2.1. Study design

MA

This was a retrospective cohort analysis comparing ECGs of LVA to ECGs of acute STEMI. The intent was to validate our prior derivation study, which demonstrated that the ratio of T-wave amplitude to QRS amplitude

2.2. Study population and setting

ED

best differentiates STE in AMI from STE in LVA [19]. Approval was obtained from the institutional review board.

PT

The study was conducted in an urban ED, with an annual patient volume of greater than 90,000. The study

CE

cohort included consecutive ECGs of patients with LVA, as diagnosed by echocardiogram. The control group included consecutive ECGs of patients with anterior STEMI and left anterior descending artery (LAD) occlusion, as

AC

later proven by cardiac catheterization. All ECGs utilized were those of patients who presented to the ED with ischemic symptoms such as chest pain or dyspnea.

2.3. Study protocol To identify ECGs for the LVA group, the hospital echocardiography database was searched for the period 2002 – 2009. Using this time period excluded all cases that were included in the derivation study. All echocardiography examinations in our institution are electronically coded. The code diastolic distortion aneurysm is reserved for unequivocal LVA. In addition, many aneurysms are coded as dyskinesis (systolic ventricular distortion). Therefore, because the presence or absence of anatomical aneurysm was not critical for the purposes of the study, we searched for all echocardiograms coded as “diastolic distortion” or “dyskinesis.” These were the same search terms used in the derivation study. Only those with anterior or apical aneurysms were included and those with

4

ACCEPTED MANUSCRIPT inferior or posterior aneurysms were excluded. If a patient had multiple echocardiograms in the database, they were included if at least one of them was interpreted as diastolic distortion or dyskinesis. After these patients were identified via the echocardiography database, the electronic medical records were

RI P

T

reviewed. The most recent ECG recorded for each patient (during an ED visit that was not for a STEMI) was identified. This ECG had to be from a visit that was after the echocardiographic diagnosis of diastolic distortion was made. If no such visit existed, then the patient was excluded. Furthermore, after review of the electronic medical

SC

record, additional cases were excluded if (1) there were no symptoms suggestive of acute MI (chest pain, dyspnea, or arrhythmia), (2) no troponin was measured, (3) the maximum serial troponin I was > 5.0 ng/mL (suggesting this

NU

was possibly a STEMI), (4) the patient had a coronary angiogram at that visit showing acute or recent left anterior

MA

descending artery occlusion, and (5) the evaluating cardiologist retrospectively suspected acute LAD occlusion representing STEMI as the etiology of the symptoms. Any ECG that did not have at least 1 mm of STE in 2 consecutive precordial leads, as measured at the J point and relative to the PR segment, as well as any ECG with

ED

intraventricular conduction delay (QRS duration longer than 120 milliseconds, including any bundle-branch block [BBB]), was also excluded.

PT

For the STEMI group, the catheterization laboratory database was searched for consecutive patients with

CE

anterior STEMI presenting to the ED from October 2007 to March 2009 who underwent primary percutaneous coronary intervention for 100% LAD occlusion. In order to include a substantial number of “subtle”, non-obvious

AC

LAD occlusions, we selected 34 additional ECGs of consecutive subtle acute LAD occlusions from a cohort in a previous study [20]. In that study, ECG exclusion criteria included (1) 5 mm of STE, (2) at least one convex ST segment in V2 – V6, (3) any T-wave inversion, or (4) inferior ST depressions. Exclusion criteria for both AMI cohorts included simultaneous STEMI in another distribution, and those with intraventricular conduction delay (QRS duration longer than 120 milliseconds, including any bundle-branch block [BBB]). A rater who was blinded to the ECG group made all ECG measurements. The rater was not blinded to the purpose of the study. Our previous study showed that intra-class coefficients were very reliable, and therefore only one rater was necessary [19]. The rater measured the QRS in millimeters, from the nadir of any Q-wave, S-wave, or QS-wave to the peak of the R-wave. The T-wave amplitude was measured, which included only the positive portion of the T wave. All measurements were relative to the PR segment and to the nearest 0.5 mm. These data were recorded from every ECG for every one of leads V1 through V4 and recorded directly into an electronic database.

5

ACCEPTED MANUSCRIPT 2.4. Data analysis & statistics The data collected from each ECG was then applied to the two previously derived ECG criteria. The ratio of T-wave amplitude to QRS amplitude for every lead V1-V4 was determined. The maximum ratio in any one lead

RI P

T

was recorded. The ratio of the sum of T-wave amplitudes to the sum of QRS amplitudes in V1-V4 was determined. Statistical analysis was performed using SPSS 22.0 (IBM Corp., 2013, Armonk, NY) after the completion of data collection. For the study and control groups, median ratios for both ECG rules were determined and inter-quartile

SC

ranges were calculated. Sensitivity, specificity, and accuracy were determined for both rules. Comparisons made by Mann-Whitney U Test. Figures 2 and 3 depict examples of ECGs from both cohorts and examples of

MA

NU

measurements.

3. Results 3.1. Study and control groups

ED

There were 155 patients identified with diastolic distortion and 16 of these met inclusion criteria. The rest were excluded due to the previously stated criteria. There were a total of 59 ECGs included for the control group,

PT

found to have an acute STEMI and LAD occlusion. Within the control group, there were 25 in the STEMI group and

CE

34 in the subtle STEMI group. Further analysis regarding the control group going forward was a combination of

AC

these two types of STEMI.

3.2. Comparison of median ratios between study and control groups Table 1 demonstrates the median ratios with interquartile ranges, as applied to both ECG rules. The median ratios were then compared utilizing Mann-Whitney U test. The median ratios for the LVA group and the STEMI group were statistically different for both rules (p < 0.001).

3.3. Sensitivity, specificity, and accuracy of ECG rules Table 2 shows the sensitivity, specificity, accuracy, and misclassification rate of the two ECG rules. Both rules were highly accurate and sensitive for predicting STEMI versus LVA, and rule 2 (using the maximum ratio) was very specific as well. Table 3 depicts the accuracy, sensitivity, and specificity of the two ECG rules as calculated in the present validation study and in the original derivation study.

6

ACCEPTED MANUSCRIPT 4. Discussion When patients present to the emergency department with ischemic symptoms and their ECG manifests ST elevation, it is imperative to discern STE due to AMI from STE due to non-ischemic pathologies, such as left

RI P

T

ventricular aneurysm. In 2005, we published a derivation of two ECG rules for differentiating STE due to AMI from STE due to LVA. Both of these rules were based on the concept of that there are relatively larger T-waves in AMI, and therefore exploited the ratio of T-wave amplitude to QRS amplitude. The current study sought to validate

SC

these two rules within a different cohort of patients. Similar to the derivation study, we found good sensitivity, specificity, and accuracy as well as low rates of misclassification when using these rules to predict acute STEMI

MA

subtle than the ECGs used in the derivation study.

NU

versus LVA. These rules still performed well even when we included ECGs of AMI that were considered more

Clinicians must always consider LVA as an etiology of STE on the ECG in order to avoid inappropriate reperfusion therapy. Unfortunately, studies have shown that LVA is a very difficult diagnosis to make on the ECG,

ED

and may be the most commonly misinterpreted etiology of STE in patients presenting to the emergency department with ischemic symptoms [2, 4, 16, 21]. Brady et al. showed eleven ECGs to 450 emergency physicians (EPs). The

PT

ECG of LVA was misdiagnosed as acute STEMI by 72% of EPs, a higher misclassification rate than any other ECG.

CE

The misclassification would have led to inappropriate thrombolytic therapy by 28% of EPs [2]. In a second study by the same authors, there were 202 ECGs with STE and 12 ECGs were misdiagnosed – five ECGs showed LVA, two

AC

of which were diagnosed as STEMI [4]. Miller et al. studied 100 patients admitted to a cardiac care unit for suspicion of MI. Of these patients, 31 had STE on the ECG; 21 of 21 without prior infarction and 5 of 10 with prior infarction had AMI. All 5 false positive ECGs with STE were in the location of the previous Q-wave infarct (LVA), and the STE did not represent acute injury [3]. In a large, more recent study, Larson et al. showed that 20 out of 123 false catheterization laboratory activations were due to LVA [5]. This frequent misclassification of the ECG of LVA demonstrates the need for specific ECG rules to differentiate STEMI from LVA. Dating back to as early as the 1950’s, authors have sought to describe the ECG of LVA, but these have been descriptive studies, autopsy studies, or clinical-pathological correlation studies, without any comparison or control groups. Findings from these studies have shown that the ECG of LVA may include deep well-formed Q-waves (usually QS-waves), flattened T-waves, T-wave inversions, a lesser degree of STE, or concave STE [8, 10, 16, 18]. Unlike these other studies, our study, along with the original derivation paper, make

7

ACCEPTED MANUSCRIPT direct comparisons of ECG findings between cohorts of LVA and STEMI patients. It is also the first to identify the ratio of T-wave amplitude to QRS amplitude as a means to discriminate the two cohorts. To our knowledge, this is also the most accurate means of differentiating LVA from STEMI. In fact, the commonly misinterpreted ECG cited

RI P

T

in the Brady et al. study, as well as other ECGs published as easily misinterpreted STEMI mimics, would have been unequivocally diagnosed as LVA using our rules [2, 22].

The success of our two ECG rules is based on the concept of “proportionality”, an idea that has received

SC

little attention in the literature. The idea of proportionality refers to the fact that the amplitude of the T-wave has an expected proportion to the amplitude of the QRS. Along these lines, different proportions may suggest different

NU

pathologies. For example, it is well known that tall T-waves (“hyperacute” T waves) are associated with AMI [23 -

MA

25]. In our derivation study, this proportion of the T-wave to the QRS was the most sensitive and specific finding that distinguished the ECG of AMI from the ECG of LVA. There were multiple other ECG measurements examined in the derivation study but they were not as discriminatory. These included (all measurements from the J-

ED

point as well as 60ms after the J-point), degree of ST elevation, summation of ST elevation of V2-V4, highest ratio of ST elevation to QRS amplitude, and ratio of summation of ST elevation to QRS amplitude in V2-4. Other

PT

measurements included height of the maximum T-wave, T-wave sum in V1-V4, maximum Q-wave, Q-wave sum in

CE

V2-V4 and in V1-V6, and absence of inverted T-waves in V1-V6 or V1-V4. The maximum T-wave amplitude (using a cutoff of ≥ 5 mm) and T-wave amplitude sum V1-V4 (using a cutoff of ≥ 11.5 mm) also had good

AC

sensitivity and specificity in differentiating AMI from LVA but did have higher rates of misclassification than the ratio rules (though this was not statistically significant). Even though the two ECG rules evaluated in this study were highly accurate, there were still several ECGs that were misclassified by each rule. One hypothesis regarding the patients that were misclassified is that they had longer duration of symptoms. In acute STEMI, as the MI duration increases, T-wave height diminishes [26 - 29]. As the T-wave amplitudes decreases over time, these two rules would be expected to be less sensitive in predicting AMI, as the ratios are now different. In the derivation study, this hypothesis was evaluated. For Rule 1, there were 2 of 37 AMIs that were incorrectly classified (false negatives), and both patients had > 6 hours of symptoms. Unfortunately, in the present study, data on the time from symptom onset to ECG was not available. Finally, it is important to understand that the utilization of these ECG rules should be taken in their clinical context. There are certain clinical characteristics that may help the EP determine whether the STE is due to AMI or

8

ACCEPTED MANUSCRIPT LVA. First, patient history is essential, particularly when taken by a highly experienced clinician – active chest pain may suggest AMI, while dyspnea or pulmonary edema may suggest LVA (due to poor left ventricular function). A known confirmed diagnosis of LVA greatly increases the likelihood that the STE is due to LVA. Furthermore, a

RI P

T

history of prior anterior AMI will predispose the patient to LVA, also increasing the likelihood of the STE being due to LVA, though one must be aware that a new AMI may be superimposed on an old MI. Presence of Q-waves alone do not necessarily suggest LVA, as QR-waves may be seen within the first hour of ischemic symptoms in up

SC

to 50% of first anterior AMI [30]. However, precordial QS-waves (absence of any R-wave), although also present in cardiomyopathy, left ventricular hypertrophy, and cor pulmonale, should raise suspicion of anterior LVA because

NU

the pathologic Q-waves associated with LVA are characteristically QS waves in leads V2 and V3. QR waves in V2

MA

and V3, unless in the presence of a right bundle branch block, are much less common in LVA. Lastly, ultrasound may identify LVA with the presence of wall thinning or diastolic dyskinesis. A dense wall motion abnormality may

ED

be seen in both AMI and LVA morphologies, and therefore would not distinguish the two entities from one another.

5. Limitations

PT

There are several limitations to the current study. First, the study is limited by its retrospective nature and

CE

subject to accurate coding in the echocardiographic database. Although the rater was blinded to the group from which each ECG came, the rater was not blinded to the study purpose. It is possible that LVA patients who have

AC

persistent STE after previous MI, but who were not in the echocardiography database were missed. It is unlikely that patients with AMI were missed, but it is possible since data was retrospectively obtained. In review of the electronic medical records, few of the LVA patients had emergent angiography, suggesting the suspicion for coronary occlusion was low compared to those with AMI patients (who all had emergency angiography). One must assume that the presenting characteristics of the two populations were sufficiently different that they underwent different management by the treating physician. Review of patient characteristics such as duration of symptoms would have been useful in the analysis of patients who were misclassified by the ECG rules. Unfortunately this data was not available for certain patients therefore this analysis could not be conducted. Finally, the rule only differentiates patients for whom the differential diagnosis is LVA versus AMI. We do not specify prospectively precisely on whom the rule should be applied. It is implied that the ECGs to which this differential applies are those with well-formed Q-waves, especially QS-waves, in the anterior leads.

9

ACCEPTED MANUSCRIPT

6. Conclusion For patients who present to the emergency department with ischemic symptoms and STE on the ECG, and

RI P

T

the ECG differential diagnosis includes STEMI versus LVA, we have derived and validated two rules that rely on a high T-wave to QRS amplitude ratio. These two rules differentiate the two entities with good sensitivity, specificity, and accuracy. These ECG rules taken in clinical context may be helpful in assisting clinicians who are deciding if

AC

CE

PT

ED

MA

NU

SC

reperfusion therapy is appropriate.

10

ACCEPTED MANUSCRIPT FIGURE LEGENDS AND TEXT Figure 1. Two ECG rules for predicting acute STEMI versus LVA Figure 2. ECG of acute STEMI successfully identified by both rules

RI P

T

This ECG with QS-waves could easily be misinterpreted as being due to LVA, or prior anterior MI. In fact, it was an acute STEMI. Measurements annotated on individual lead images (in millimeters). By rule 1, the sum of T-wave amplitudes divided by QRS amplitudes in V1 through V4 = 0.29 (≥

Figure 3. ECG of LVA successfully identified by both rules

SC

0.22  STEMI). By rule 2, the maximum ratio (in this ECG, V4) = 0.8 (≥ 0.36  STEMI).

NU

This ECG could easily be read as acute anteroseptal MI. In fact, it is the ECG of a patient with prior

MA

anterolateral MI and no acute STEMI. There are QS-waves in anterior and lateral leads. Measurements are annotated on individual lead images (in millimeters). By rule 1, the sum of T wave amplitudes divided by QRS amplitudes in V1 through V4 = 0.16 (< 0.22  LVA). By rule 2, the maximum ratio (in this

AC

CE

PT

ED

ECG, V3) = 0.18 (< 0.36  LVA).

11

ACCEPTED MANUSCRIPT REFERENCES [1] Mills RM, Young E, Gorlin R, Lesch M. Natural history of S-T segment elevation after acute myocardial infarction. Am J Cardiol 1975;35:609- 14.

RI P

T

[2] Brady WJ, Perron AD, Chan T. Electrocardiographic ST segment elevation: correct identification of AMI and non-AMI syndromes by emergency physicians. Acad Emerg Med 2001;8:349 - 60.

[3] Miller DH, Kligfield P, Schreiber TL, Borer JS. Relationship of prior myocardial infarction to false-positive

SC

electrocardiographic diagnosis of acute injury in patients with chest pain. Arch Intern Med 1987;147: 257- 61. [4] Brady WJ, Perron A, Ullman E. Errors in emergency physician interpretation of ST-segment elevation in

NU

emergency department chest pain patients. Acad Emerg Med 2000;7:1256 - 60.

MA

[5] Larson DM, Menseen KM, Sharkey SW, Duval S, Schwartz RS, Harris J, et al. “False-positive” cardiac catheterization laboratory activation among patients with suspected ST-segment elevation myocardial infarction. JAMA 2007;298(23):2754-60.

ED

[6] Berger AK, Radford MJ, Wang Y, Krumholz HM. Thrombolytic therapy in older patients. J Am Coll Cardiol 2000;36:366 - 74.

PT

[7] Moyer JB, Hiller GI. Cardiac aneurysm: clinical and electrocardiographic analysis. Am Heart J 1951;41:340- 58.

1016.

CE

[8] Ford RV, Levine HD. The electrocardiographic clue to ventricular aneurysm. Ann Intern Med 1951;34:998-

AC

[9] Schlichter J, Hellerstein HK, Katz LN. Aneurysm of the heart: a comparative study of one hundred and two proved cases. Medicine 1954;33:43-86. [10] Pipilis GA, Wosika PH. Unipolar precordial electrocardiogram in ventricular aneurysm. JAMA 1951;145:14752. [11] Dubnow MH, Burchell HB, Titus JL. Postinfarction ventricular aneurysm. Am Heart J 1965;70:753- 60. [12] Visser CA, Kan G, David GK, Lie KI, Durrer D. Echocardiographic cineangiographi correlation in detecting left ventricular aneurysm: a prospective study of 422 patients. Am J Cardiol 1982;50:337 - 41. [13] Visser CA, Kan G, Meltzer RS, Koolen JJ, Dunning AJ. Incidence, timing and prognostic value of left ventricular aneurysm formation after myocardial infarction: a prospective, serial echocardiographic study of 158 patients. Am J Cardiol 1986;57:729 - 32.

12

ACCEPTED MANUSCRIPT [14] Bhatnagar SK. Observations of the relationship between left ventricular aneurysm and ST segment elevation in patients with a first acute anterior Q wave myocardial infarction. Eur Heart J 1994;15:1500- 4. [15] Arvan S, Varat MA. Persistent ST-segment elevation and left ventricular wall abnormalities: a two-dimensional

RI P

T

echocardiographic study. Am J Cardiol 1984;53:1542-6.

[16] Engel J, Brady WJ, Mattu A, Perron AD. Electrocardiographic ST elevation: left ventricular aneurysm. Am J Emerg Med 2002;20(3):238-42.

SC

[17] Smith SW, Zvosec DL, Henry TD, Sharkey SW. The ECG in acute MI: an evidence-based manual of reperfusion therapy. Philadelphia, Lippincott, Williams, and Wilkins; 2002. p. 320.

NU

[18] Cokkinos DV, Hallman GL, Cooley DA, Zamalloa O, Leachman RD. Left ventricular aneurysm: analysis of

MA

electrocardiographic features and postresection changes. Am Heart J 1971;82:149- 57. [19] Smith SW. T/QRS ratio best distinguishes ventricular aneurysm from anterior myocardial infarction. Am J Emerg Med 2005;23(3):279-87.

ED

[20] Smith SW, Khalil A, Henry TD, Rosas M, Chang RJ, Heller K, et al. Electrocardiographic differentiation of early repolarization from subtle anterior ST segment elevation myocardial infarction. Ann Emerg Med

PT

2012;60(1):45-56.

CE

[21] Deshpande A, Birnbaum Y. ST-segment elevation: Distinguishing ST elevation myocardial infarction ST elevation secondary to nonischemic etiologies. World J Cariol 2014;6(10):1067-79.

AC

[22] Rokos IC, French WJ, Mattu A, Nichol G, Farkouh ME, Reiffel J, et al. Appropriate cardiac cath lab activation: optimizing electrocardiograph interpretation and clinical decision making for acute ST-elevation myocardial infarction. Am Heart J 2010;160(6):995-1003. [23] Smith FM. The ligation of coronary arteries with electrocardiographic study. Arch Intern Med 1918;5:1 - 27. [24] Bayley RH, LaDue JS, York DJ. Electrocardiographic changes (local ventricular ischemia and injury) produced in the dog by temporary occlusion of a coronary artery, showing a new stage in the evolution of myocardial infarction. Am Heart J 1944;27:164 - 9. [25] Pinto IJ, Nanda NC, Biswas AK, Parulkar VG. Tall upright T waves in the precordial leads. Circulation 1967;36:708- 16.

13

ACCEPTED MANUSCRIPT [26] Wilkins ML, Pryor AD, Maynard C, Wagner NB, Elias WJ, Litwin PE, et al. An electrocardiographic acuteness score for quantifying the timing of a myocardial infarction to guide decisions regarding reperfusion therapy. Am J Cardiol 1995;75(8): 617-620.

RI P

T

[27] Corey KE, Maynard C, Pahlm O, Wilkins ML, Anderson ST, Cerqueira MD, et al. Combined historical and electrocardiographic timing of acute anterior and inferior myocardial infarcts for prediction of reperfusion achievable size limitation. Am J Cardiol 1999;83(6): 826-831.

SC

[28] Engblom H, Strauss DG, Heden B, Hedström E, Jovinge S, Götberg M, et al. The evaluation of an

NU

electrocardiographic myocardial ischemia acuteness score to predict the amount of myocardial salvage achieved by early percutaneous coronary intervention: clinical validation with myocardial perfusion single photon emission

MA

computed tomography and cardiac magnetic resonance. J Electrocardiol 2011;44(5):525-532. [29] Hochrein J, Sun F, Pieper KS, Lee KL, Gates KB, Armstrong PW, et al. Higher T-wave amplitude associated with better prognosis in patients receiving thrombolytic therapy for acute myocardial infarction (a GUSTO-1

Am J Cardiol 1998;81(9): 1078-1084.

ED

substudy). Global Utilization of Streptokinase and Tissue plasminogen activator for Occluded Coronary Arteries.

PT

[30] Raitt MH, Maynard C, Wagner GS, Cerqueira MD, Selvester RH, Weaver WD. Appearance of abnormal Q

CE

waves early in the course of acute myocardial infarction: implications for efficacy of thrombolytic therapy. J Am

AC

Coll Cardiol 1995;25(5): 1084-8.

14

ACCEPTED MANUSCRIPT Table 1. Median calculated ratios for the STEMI cohort and LVA cohort Median Ratio for STEMI +/- IQR

Comparison

0.203 (0.127 – 0.236)

0.504 (0.328 – 0.734)

p < 0.001

0.275 (0.177 – 0.341)

0.800 (0.583 – 1.222)

p < 0.001

T

Median Ratio for LVA +/- IQR

RI P

Rule 1: Sum V1-V4 Rule 2: Max ratio

SC

IQR = interquartile range. Comparison by Mann-Whitney U test

Specificity

91.5%

68.8%

91.5%

81.3%

Misclassified

86.7%

13.3%

89.3%

10.7%

ED

Rule 1: Sum V1-V4 Rule 2: Max ratio

Accuracy

MA

Sensitivity

NU

Table 2. Sensitivity, specificity, accuracy, misclassification rates for ECG rules

Specificity – derivation

Specificity – validation

Accuracy – derivation

Accuracy – validation

95%

91.5%

91%

68.8%

93%

86.7%

95%

91.5%

82%

81.3%

90%

89.3%

CE

Sensitivity – validation

AC

Rule 1: Sum V1-V4 Rule 2: Max ratio

Sensitivity – derivation

PT

Table 3. Sensitivity, specificity, and accuracy for ECG rules in the derivation paper and validation paper

15

AC

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

16

AC

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

17

AC

CE

PT

ED

MA

NU

SC

RI P

T

ACCEPTED MANUSCRIPT

18