Electrocardiogram Differentiation of Benign Early Repolarization Versus Acute Myocardial Infarction by Emergency Physicians and Cardiologists

Samuel D. Turnipseed, MD, Aaron E. Bair, MD, J. Douglas Kirk, MD, Deborah B. Diercks, MD, Poroshat Tabar, DO, Ezra A. Amsterdam, MD

Abstract
Objectives: ST-segment elevation (STE) related to benign early repolarization (BER), a common normal variant, can be difcult to distinguish from acute myocardial infarction (AMI). The authors compared the electrocardiogram (ECG) interpretations of these two entities by emergency physicians (EPs) and cardiologists. Methods: Twenty-ve cases (13 BER, 12 AMI) of patients presenting to the emergency department with chest pain were identied. Criteria for BER required four of the following: 1) widespread STE (precordial greater than limb leads), 2) J-point elevation, 3) concavity of initial up-sloping portion of ST segment, 4) notching or irregular contour of J point, and 5) prominent, concordant T waves. Additional BER criteria were 1) stable ECG pattern, 2) negative cardiac injury markers, and 3) normal cardiac stress test or angiography. AMI criteria were 1) regional STE, 2) positive cardiac injury markers, and 3) identication of culprit coronary artery by angiography in less than eight hours of presentation. The 25 ECGs were distributed to 12 EPs and 12 cardiologists (four in academic medicine, four in community practice, and four in community academics [health maintenance organization] in each physician group). The physicians were informed of the patients age, gender, and race, and they then interpreted the ECGs as BER or AMI. Undercalls (AMI misdiagnosed as BER) and overcalls (BER misdiagnosed as AMI) were calculated for each physician group. Results: Cardiologists correctly interpreted 90% of ECGs, and EPs correctly interpreted 81% of ECGs. The proportion of undercalls (missed AMI/total AMI) was 2.8% for cardiologists (95% condence interval [CI] = 0.09% to 5.5%) compared with 9.7% for EPs (95% CI = 4.8% to 14.6%) (p = 0.02). The proportion of overcalls (missed BER/total BER) was 17.3% for cardiologists (95% CI = 11.4% to 23.3%) versus 27.6% for EPs (95% CI = 20.6% to 34.6%) (p = 0.03). The mean number of years in practice was 19.8 for cardiologists (95% CI = 19 to 20.5) and 11 years for EPs (95% CI = 10.5 to 12.0) (p < 0.001). Conclusions: Although correct interpretation was high in both groups, cardiologists, who had signicantly more years of practice, had fewer misinterpretations than EPs in distinguishing BER from AMI electrocardiographically. ACADEMIC EMERGENCY MEDICINE 2006; 13:961967 2006 by the Society for Academic Emergency Medicine Keywords: benign early repolarization, acute myocardial infarction, electrocardiogram

From the Department of Emergency Medicine (SDT, AEB, JDK, DBD, PT) and Department of Internal Medicine, Division of Cardiovascular Medicine (EAA), University of California, Davis, Medical Center, Sacramento, CA. Received November 3, 2005; revisions received January 24, 2006, and March 4, 2006; accepted April 8, 2006. Address for correspondence and reprints: Samuel D. Turnipseed, MD, Department of Emergency Medicine, University of California, Davis, Medical Center, 4150 V Street, Suite 2100, Sacramento, CA 95817. Fax: 916-734-7950; e-mail: sdturnipseed@ ucdavis.edu.

he proven benet of urgent coronary reperfusion for patients with acute myocardial infarction (AMI) necessitates rapid recognition of ST-segment elevation myocardial infarction (STEMI) by emergency physicians (EP). However, AMI is the etiology of ST-segment elevation (STE) in a minority of patients presenting with chest pain.1,2 Among the multiple causes of noninfarction electrocardiographic (ECG) patterns that may mimic STEMI are benign early repolarization (BER), left bundle branch block, left ventricular aneurysm, left ventricular hypertrophy, paced ventricular rhythms, and

2006 by the Society for Academic Emergency Medicine doi: 10.1197/j.aem.2006.04.014

ISSN 1069-6563 PII ISSN 1069-6563583

961

962

Turnipseed et al.

BENIGN EARLY REPOLARIZATION VS AMI

pericarditis. The potential for these conditions to be confused with STEMI is reected in the report of Sharkey et al., in which 11% of patients receiving a thrombolytic agent were subsequently found not to have AMI.3 Benign early repolarization, a normal variant, is a relatively frequent ECG nding and can be particularly difcult to distinguish from AMI in patients presenting to the emergency department (ED) with chest pain. It is found in approximately 1% of the population; the majority of these individuals are younger than 50 years of age and are rarely older than 70 years.4 African American men between the ages of 20 and 40 years comprise a large proportion of this group.5,6 In patients presenting to the ED with STE, Brady et al. found that BER was almost as common as AMI (15% vs. 13%).1 In the report by Sharkey et al. citing an 11% frequency of erroneous administration of a thrombolytic agent, almost one third of patients without AMI had BER.3 It is essential to recognize BER to avoid subjecting patients to unnecessary medications or procedures. The risks of thrombolytic therapy include an 8% rate of major bleeding, of which 1%2% are intracerebral hemorrhages usually resulting in devastating stroke.7 Unnecessary emergency percutaneous coronary intervention is also associated with an unfavorable risk/benet prole. On the other hand, misinterpretation of AMI as BER deprives patients of potential lifesaving therapy. Because of the continuing challenge of distinguishing BER from STEMI, we compared the ECG interpretations of these two entities by EPs and cardiologists. METHODS Study Design This was a retrospective study that compared ECG interpretations by EPs and cardiologists in terms of AMI vs. BER. The study was approved by our institutional review board. Study Setting and Population The initial ECGs of each patient were distributed to 12 EPs and 12 cardiologists who were blinded to the nal

Table 1 Benign Early Repolarization Criteria At least four of the following: 1. Widespread ST-segment elevation (precordial greater than limb leads) 2. J-point elevation 3. Concavity of initial up-sloping portion of ST segment 4. Notching or irregular contour of J point 5. Prominent, concordant T waves (large amplitude)
Adapted from Brady and Chan.8

diagnosis. To achieve diversity among the ECG readers, we chose physicians from three unique practice environments. Of these 12 physicians in each of the two groups, four in each group practiced at our academic training center, four practiced at a large local health maintenance organization (teaching) community hospital, and four were in community practice. All cardiologists involved in the study routinely evaluated and admitted patients from the ED. The number of years practiced after completion of training was recorded for each physician. Study Protocol We selected 25 ECGs of patients who had presented to the ED with the chief complaint of chest pain and who had complete cardiac evaluations. The ECGs included 13 with BER and 12 with STEMI. BER ECGs were chosen from our Chest Pain Evaluation Unit database under the heading ECG Interpretation: BER. Criteria for BER have been previously established8 and are listed in Table 1. In our study, the diagnosis of BER required four of the ve criteria (Figure 1). Additional criteria of BER used in this study included the following: 1) ECGs demonstrated a stable pattern, 2) patients had at least three consecutive negative cardiac injury markers (creatine phosphokinase-MB and/or troponin I), and 3) patients had a normal noninvasive cardiac stress test or normal coronary angiography. STEMI ECGs were chosen from a computergenerated list using the International Classication of Disease billing code for AMI. The criteria for AMI were

Figure 1. Typical example of an electrocardiogram demonstrating benign early repolarization in a 43-year-old African American man. All ve criteria for benign early repolarization (Table 1) are present.

ACAD EMERG MED

September 2006, Vol. 13, No. 9

www.aemj.org

963

as follows: 1) regional STE, 2) positive serial cardiac injury markers, and 3) coronary angiography within eight hours of the patients arrival in the ED that identied the culprit vessel. Chest pain was the chief complaint in all 25 cases. ECG readers were informed of the age, gender, and race of the patient for each ECG. No other clinical data were provided. Readers were asked to interpret each ECG as either BER or AMI. It was also clearly stated that if STEMI was selected, the reader should assume the patient would receive thrombolytic therapy, not intervention in a catheterization laboratory. ECG interpretations were then collected by a single researcher. Physician interpretations were classied as correct, undercall, and overcall. Undercalls were dened as AMI misdiagnosed as BER, and overcalls were dened as BER misdiagnosed as AMI. Data Analysis Physician groups continuous data were expressed using mean or median and then were compared using Students t-test where appropriate. Proportion and frequency data related to ECG interpretations were compared using Fishers exact test. To account for clustering of repeated measures, a generalized linear model was used for linear regression analysis. When appropriate, 95% condence intervals (CIs) are presented. All tests were based on two-tailed alternatives. We performed the statistical analyses using Stata 7.0 (Stata Corp., College Station, TX) for Windows (Microsoft Corp., Redmond, WA). RESULTS The number of interpretations for each specialty group was calculated by multiplying the number of physicians in each group (n = 12) by the number of ECGs (n = 25), yielding a total of 300 interpretations for each group. The cardiologists correctly interpreted 90% (269/300;

Table 2 Cardiologist and Emergency Physician Interpretations Cardiologists Emergency Physicians pvalue

Total correct ECGs 90% (269/300) 81% (243/300) Average no. correct 22 20 ECGs per physician Proportion undercalls 2.8% (4/144) 9.7% (14/144) 0.02 Proportion overcalls 17.3% (27/156) 27.6% (43/156) 0.03 Average years 19.8 11.0 <0.001 practiced
Undercalls = acute myocardial infarction misdiagnosed as benign early repolarization; overcalls = benign early repolarization misdiagnosed as acute myocardial infarction.

95% CI = 86% to 93%) of the ECGs, and the EPs interpretations were correct in 81% (243/300; 95% CI = 76% to 85%) of cases (Table 2). The average number of correct interpretations for the individual cardiologists was 22 (range, 2024) and for each EP was 20 (range, 1523). The proportion of undercalls (AMI misdiagnosed as BER) was different between specialty groups. Overall, the proportion of undercalls (missed AMI/total AMI) was 2.8% for cardiologists (95% CI = 0.09% to 5.5%) compared with 9.7% for EPs (95% CI = 4.8% to 14.6%) (p = 0.02). The proportion of overcalls (BER misdiagnosed as AMI) was also different between specialty groups. The proportion of overcalls (missed BER/total BER) was 17.3% for cardiologists (95% CI = 11.4% to 23.3%) versus 27.6% for EPs (95% CI = 20.6% to 34.6%) (p = 0.03). The percentage of correct ECG interpretations by specialty and type of ECG is shown in Table 3. A commonly missed BER ECG is shown in Figure 2, and a commonly missed AMI ECG is shown in Figure 3. There was a disparity of years in practice between specialty groups. Among the cardiology group, the mean number of years in practice was 19.8 (95% CI = 19 to

Table 3 Percentage of Correct ECG Interpretations by Physician Specialty and Type of ECG Condition BER BER BER BER BER BER BER BER BER BER BER AMI AMI AMI AMI AMI AMI AMI ECG No. 1 3 5 6 12 14 18 19 22 23 25 4 7 8 13 15 17 20 Patient Age (yr) 54 52 56 60 41 46 49 46 57 36 43 52 71 55 52 38 42 67 Patient Gender Male Male Female Male Male Male Male Male Male Male Male Male Male Male Male Male Male Male Patient Race African American African American African American African American White African American African American African American White African American African American White White Hispanic Hispanic White White White % Emergency Physicians Correct (n = 12) 58 83 66 75 75 58 75 75 66 100 50 92 75 83 92 92 92 66 % Cardiologists Correct (n = 12) 92 50 75 58 50 58 100 100 92 92 75 100 92 92 100 100 100 83

All ECGs that were correctly interpreted by all physicians of both groups were excluded from the table. BER = benign early repolarization; AMI = acute myocardial infarction.

964

Turnipseed et al.

BENIGN EARLY REPOLARIZATION VS AMI

Figure 2. A commonly misdiagnosed electrocardiogram with benign early repolarization. The incorrect diagnosis of acute myocardial infarction was made by 38% of all physician readers in the study.

20.5). The EPs, however, had only been in practice for a mean of 11 years (95% CI = 10.5 to 12.0) (p < 0.001). ECG misdiagnoses (both overcalls and undercalls) correlated inversely with years in practice for all 24 physicians considered as a single group (Figure 4). Among misdiagnosed ECGs (overcalls and undercalls in both physician groups), the mean number of years in practice was 13 (95% CI = 12 to 15). In contrast, the mean number of years of practice reported among those who correctly diagnosed the ECGs was 16 (95% CI = 15 to 16.5) (p = 0.001). After adjusting for covariates (specialty training, years in practice, practice location, and diagnosis) and clustering of physician responses, the most signicant variable contributing to a correct diagnosis was the presence of AMI (b coefcient, 0.16; 95% CI = 0.08 to 0.24; p < 0.001). DISCUSSION Our study demonstrates the difculty in differentiating BER from STEMI. The range of correct answers for the

25 ECGs among the cardiologists and EPs was 1524. The cardiologists had fewer undercalls and overcalls compared with the EPs. However, the cardiologists had also practiced signicantly longer than the EPs. It is noteworthy that after adjusting for covariates (specialty training, years in practice, practice location, and diagnosis) and clustering of physician responses, the most signicant variable associated with a correct diagnosis was the presence of AMI. This nding suggests that BER and STEMI can be distinguished electrocardiographically in a majority of patients by experienced clinicians but that frequent errors of overcalls and undercalls will still unavoidably exist. Benign early repolarization is a pseudoinfarction ECG pattern frequently encountered in the ED. Of 171 patients presenting to an ED with STE, Brady et al. found that BER was the cause in 12%.9 This nding has been associated with inappropriate administration of brinolytic therapy to patients with BER.3 In patients admitted to the hospital with ECG ndings of STE, Brady et al. compared

Figure 3. A commonly misdiagnosed electrocardiogram with acute myocardial infarction. The incorrect diagnosis of benign early repolarization was made by 17% of all physician readers in the study.

ACAD EMERG MED

September 2006, Vol. 13, No. 9

www.aemj.org

965

Figure 4. Physician years of experience versus number of missed diagnoses (benign early repolarization + acute myocardial infarction) of the 25 study electrocardiograms.

the interpretation of STE by EPs with the cardiologists nal interpretation that was supported by clinical testing. The rate of ECG STE misinterpretation was 6% by the EPs.10 A survey involving hypothetical patients with chest pain and ECGs that demonstrated different etiologies of STE such as AMI, left ventricular hypertrophy, and BER was administered to 458 attending and resident physicians. These physicians were asked to decide whether brinolytic therapy should be given (assuming no contraindications), and it was selected appropriately in 83% (1,525/1,832) of cases. BER was the diagnosis in 23% of cases in which brinolytic therapy was given inappropriately.11 Despite the clinical importance of differentiating BER from STEMI, the literature pertaining to this problem is limited. Regional STE denes STEMI in a patient with compatible symptoms until proven otherwise and reciprocal ST depression further supports the diagnosis.12 Although both AMI and non-AMI causes of STE may demonstrate a similar number of leads with STE,2 STE in BER is less likely to be isolated to one ECG region than is AMI.13 Additionally, the use of current computerized ECG retrieval allows the clinician to readily access previous ECGs in cases where they are in fact available. Comparative ECGs are invaluable in distinguishing BER from AMI. Patients whose ECGs demonstrate BER should be encouraged to carry a wallet-sized ECG.14 However, when only the ED ECG is available, a number of steps in interpretation are helpful in arriving at the correct diagnosis. Serial ECGs that show no change suggest BER. In a report by Sharkey et al. of patients presenting with AMI, 16% did not exhibit STE until the second or third ECG.3 In addition, the presence of at least four of the ve criteria in Table 1 supports BER. It should be noted, however, that although a notch in the J point (usually in the precordial leads) has been considered indicative of

BER, this nding has not been systematically studied.15 The value of upward concavity of elevated ST segments was assessed by Brady et al. In their study, nonconcave ST elevation had a sensitivity and specicity for AMI of 77% and 97%, respectively, and a positive and negative predictive value for AMI of 94% and 88%, respectively.9 If these steps are performed and the diagnosis remains unclear, a review of the ECG, faxed to the on-call cardiologist, may be of benet. Immediate coronary angiography affords a denitive diagnosis in patients with chest pain and STE of uncertain etiology. In the absence of a cardiac catheterization laboratory, echocardiography may be helpful in detecting wall motion abnormalities in patients presenting with AMI. However, emergency echocardiography is limited by its inability to dene the age of a wall motion abnormality. When the diagnosis remains unclear and neither angiography nor echocardiography is available, consideration can be given to transfer a stable patient to an institution with these capabilities. While several diagnostic tools may be available to distinguish the etiology of STE, ECG interpretation remains fundamental. This study demonstrates that both cardiologists and EPs have difculty differentiating BER from AMI electrocardiographically. As shown in Table 3, only seven ECGs were correctly diagnosed by all physicians in both groups. Thus, physicians must have heightened awareness that one of these entities may masquerade as the other and thus attempt to optimize the use of available resources to establish the correct diagnosis. LIMITATIONS A limitation of our study was the use of a hypothetical situation wherein the ECG reader did not have the advantage of a history and examination of the patient, serial or previous ECGs, or the ability to obtain consultation. However, it was our specic purpose to isolate and evaluate the success rate of the ECG interpretation among groups of experienced clinicians. It is clear from prior literature that the miss rate of AMI is lower when the overall clinical picture is available.16 A further limitation is the relatively small number of readers. However, to our knowledge, this is the rst study to compare ECG interpretation of STE between physicians of different specialties in different practice settings. Finally, our ndings may not apply to other institutions. CONCLUSIONS In conclusion, 1) BER, a common ECG nding, can be difcult to distinguish from AMI, 2) cardiologists who had signicantly more years of clinical practice had fewer undercalls and overcalls than EPs in distinguishing BER from AMI electrocardiographically, and 3) clinical experience may be an important factor in differentiating BER from AMI on ECG. References 1. Brady WJ, Perron AD, Martin ML, Beagle C, Aufderheide TP. Cause of ST segment abnormality in ED chest pain patients. Am J Emerg Med. 2001; 19:258.

966

Turnipseed et al.

BENIGN EARLY REPOLARIZATION VS AMI

2. Brady WJ, Perron AD, Ullman EA, et al. Electrocardiographic ST segment elevation: a comparison of AMI and non-AMI ECG syndromes. Am J Emerg Med. 2002; 20:60912. 3. Sharkey SW, Berger CR, Brunette DD, Henry TD. Impact of the electrocardiogram on the delivery of thrombolytic therapy for acute myocardial infarction. Am J Cardiol. 1994; 73:5503. 4. Mehta MC, Jain AC. Early repolarization on scalar electrocardiogram. Am J Med Sci. 1995; 309: 30511. 5. Thomas J, Harris E, Lassiter G. Observations on the T wave and S-T segment changes in the precordial electrocardiogram of 320 young Negro adults. Am J Cardiol. 1960; 5:46874. 6. Klatsky AL, Oehm R, Cooper RA, Udaltsova N, Armstrong MA. The early repolarization normal variant electrocardiogram: correlates and consequences. Am J Med. 2003; 115:1717. 7. Weaver WD, Simes RJ, Betriu A, et al. Comparison of primary coronary angioplasty and intravenous thrombolytic therapy for acute myocardial infarction. JAMA. 1997; 278:20938. 8. Brady WJ, Chan TC. Electrocardiographic manifestations: benign early repolarization. J Emerg Med. 1999; 17:4738. 9. Brady WJ, Syverud SA, Beagle C, et al. Electrocardiographic ST-segment elevation: the diagnosis of acute

myocardial infarction by morphologic analysis of the ST segment. Acad Emerg Med. 2001; 8:9617. 10. Brady WJ, Perron A, Ullman E. Errors in emergency physician interpretation of ST-segment elevation in emergency department chest pain patients. Acad Emerg Med. 2000; 7:125660. 11. Brady WJ, Perron AD, Chan T. Electrocardiographic ST-segment elevation: correct identication of acute myocardial infarction (AMI) and non-AMI syndromes by emergency physicians. Acad Emerg Med. 2001; 8: 34960. 12. Brady WJ, Peron AD, Syverud SA, et al. Reciprocal ST segment depression: impact on the electrocardiographic diagnosis of ST segment elevation acute myocardial infarction. Am J Emerg Med. 2002; 20:358. 13. Brady WJ. Benign early repolarization: electrocardiographic manifestations and differentiation from other ST segment elevation syndromes. Am J Emerg Med. 1998; 16:5927. 14. Smith SW. ST-elevation acute myocardial infarction: a critical but difcult electrocardiographic diagnosis. Acad Emerg Med. 2001; 8:3825. 15. Mehta M, Jain AC, Mehta A. Early repolarization. Clin Cardiol Rev. 1999; 22(2):5965. 16. McCarthy BD, Beshansky JR, DAgostino RB, Selker HP. Missed diagnoses of acute myocardial infarction in the emergency department: results from a multicenter study. Ann Emerg Med. 1993; 22:57982.

Please see accompanying video Data Supplement available at www.aemj.org.

Colles Fracture Reduction Using Ultrasound

Case: A 28-year-old intoxicated woman presented to the emergency department with right wrist pain and deformity after a fall during an altercation. We conrmed a Colles fracture with x-rays (Figure 1). The fracture was also visualized using a 7.5MHz ultrasound probe (Figure 2). After adequate sedation, the fracture was reduced and visualized at the bedside using ultrasound, which revealed persistent, slight malalignment (Figure 3). Further reduction and utilization of repeat ultrasound visualization efciently achieved optimal alignment (Figure 4), which was conrmed by standard post-reduction x-rays (Figure 5). The four-minute video available as a Data Supplement (http:// www.aemj.org/cgi/content/full/j.aem.2006.07.013/DC1) explains in detail the use of ultrasound during Colles fracture reduction. Nate Unkefer, MD Scott Joing, MD Rob Reardon, MD Hennepin County Medical Center Minneapolis, MN

Figure 1. X-ray showing the initial diagnosis of the Colles fracture.