Outcomes in patients with chronicity of left bundle- branch block with possible acute Michael C. Kontos, MD, FACC, a,b Hammad A. Aziz, MD, a Vinh Q. Chau, BS, a Charlotte S. Roberts, MSN, a Joseph P. Ornato, MD, FACC, b and George W. Vetrovec, MD, FACC a Richmond, VA

Introduction Guidelines derived from patients in clinical trials indicate that emergency department patients with likely myocardial infarction (MI) who have new left bundle-branch block (LBBB) should undergo rapid reperfusion therapy. Whether this pertains to lower risk emergency department patients with LBBB is unclear. Methods A total of 401 consecutive patients with LBBB undergoing an MI rule-out protocol were included. Left bundle- branch blocks were classified as chronic; new; or, if no prior electrocardiogram (ECG) was available, as presumably new. Left bundle-branch blocks were considered concordant if there was ≥1 mm concordant ST elevation or depression. Rates of MI, peak MB values in MI patients, and 30-day mortality were compared across groups. Results A majority of patients (64%) had new (37%) or presumably new LBBB (27%). A total of 116 patients (29%) had MI, with no significant difference in prevalence or size of MI among the 3 ECG groups. Myocardial infarction was diagnosed in 86% of patients with concordant ECG changes versus 27% of patients without concordant ECG changes (P b .01). Peak MB was N5× normal in 50% who had concordant ST changes compared to none of those who did not. Concordant ST changes were the most important predictor of MI (odds ratio 17, 95% CI 3.4-81, P b .001) and an independent predictor of mortality (odds ratio 4.3, 95% CI 1.3-15, P b .001); new or presumably new LBBB was neither. Conclusions Most patients with possible MI with new or presumably new LBBB do not have MI. Concordant ECG changes were an important predictor of MI and death. Current guidelines regarding early reperfusion therapy for patients with LBBB should be reconsidered. (Am J 2011;161:698-704.)

Patients with left bundle-branch block (LBBB) represent Methods an important minority of patients with suspected acute The chest pain protocol used at our institution has been coronary syndrome. Current recommendations from the described in detail previously.12 Patients with possible myocardial American College of /American Heart Associ- undergo an evaluation by emergency department (ED) ation1 and European Society of Cardiology2 indicate that physicians, with further treatment and evaluation dictated based on patients with new or presumed new LBBB should the initial risk stratification.12 Patients were included in this analysis undergo early reperfusion therapy (fibrinolytics or if they presented with possible myocardial ischemia, underwent primary percutaneous coronary intervention [PCI]), serial marker assessment, and had an LBBB on the initial ED ECG. although this occurs in only a minority.3-5 However, Data were collected prospectively as part of an ongoing quality early reperfusion treatment may not be appropriate for all assurance process using a standardized data collection form with retrospective analysis of patient's records to supply mission patients with new LBBB because only a minority is information. This study was approved by the Committee on the diagnosed with myocardial infarction (MI).6-8 In contrast, 9 Conduct on Human Research at Virginia Commonwealth Univer- objective electrocardiographic (ECG) criteria may iden- sity. Of patients included in the current study, 182 were included in tify those who have larger MIs in whom treatment is more apriorpublication6 and have been combined with an additional 10,11 likely to be beneficial. 219 patients to form the current analysis. Therefore, we assessed the effect of the chronicity of LBBB on outcomes in patients with possible acute MI. Electrocardiographic interpretation Left bundle-branch block wasdefinedasthepresenceofaQRS

a ≥ From the Division of Cardiology, Department of Internal Medicine, Virginia Common- duration of 120 milliseconds, a QS or rS complex in lead V1,and wealth University, Richmond, VA, and bPauley Heart Center, and Department of an R wave peak time of at least 60 milliseconds in the absence of Q

Emergency Medicine, Virginia Commonwealth University, Richmond, VA. waves in leads I, V5,orV6. Concordant ECG changes were Submitted October 6, 2010; accepted January 8, 2011. considered present if 1 of 2 previously defined criteria for acute Reprint requests: Michael C. Kontos, MD, PO Box 980051, Richmond, VA 23298-0051. myocardial infarction (MI) were present9:ST-segmentelevation≥1 E-mail: [email protected] 0002-8703/$ - see front matter mm concordant with the QRS complex or ST-segment depression ≥ © 2011, Mosby, Inc. All rights reserved. 1mminleadsV1,V2,orV3. In addition, a third criterion was also 13 doi:10.1016/j.ahj.2011.01.008 analyzed: the ratio of the ST segment to the S wave ≥25%. All American Heart Journal Kontos et al 699 Volume 161, Number 4

Table I. Demographics and risk factors based on chronicity of the LBBB

New or unknown All patients, Chronic LBBB, New LBBB, Unknown duration, duration, N = 401 n = 145 n = 147 n = 109 n = 256

Age, y 66 ± 14 (66) 67 ± 14 (67) 65 ± 14 (64) 67 ± 16 (68) 66 ± 15 (66) Age ≥65 y 193 (48) 71 (49) 60 (41) 62 (56) 122 (48) ⁎ Male 158 (39) 46 (32) 63 (43) 45 (45) 112 (44) Hypertension 312 (78) 118 (81) 120 (83) 73 (66) 193 (75) Diabetes 157 (39) 52 (36) 66 (46) 38 (35) 104 (41) CHF 95 (24) 42 (29) 33 (23) 30 (27) 63 (25) EF, % 36 ± 16 (33) 35 ± 16 (32) 37 ± 18 (35) 34 ± 15 (30) 36 ± 17 (33) EF b40% 219 (61) 80 (62) 78 (58) 60 (64) 138 (60) † Prior MI 109 (27) 50 (34) 39 (27) 20 (18) 59 (23) † Prior revascularization 103 (27) 50 (34) 41 (28) 17 (15) 58 (23) CrCl, mL/min 59 ± 31 (57) 57 ± 33 (55) 59 ± 30 (60) 62 ± 32 (56) 60 ± 32 (56) CrCl b60 mL/min 201 (53) 76 (56) 66 (49) 58 (55) 124 (51) TnI MI 116 (30) 46 (32) 35 (24) 35 (32) 70 (27) CK-MB MI 93 (23) 32 (22) 32 (22) 29 (27) 61 (24) Peak CK-MB, ng/mL 18 (11, 45) 17 (12, 37) 19 (10, 48) 23 (10, 54) 19 (10, 52) Peak CK, U/L 350 (190, 590) 250 (180, 450) 470 (230, 750) 390 (170, 680) 400 (200, 710)

Categorical data are presented as n (%). Continuous data are presented as mean ± SD (median). Peak CK and CK-MB values were calculated in the subgroup of patients who had MI and are presented as median (25th, 75th percentiles). CrCl, Creatinine clearance. ⁎ P b .05 compared to patients with new LBBB. † P b .01 compared to unknown duration LBBB.

ECGs were read independently by 2 interpreters unaware of the initial creatinine. Significant coronary disease was defined as a clinical variables and patient outcome. Disagreements were stenosis ≥50% in a coronary artery or its branches or in a resolved by a third interpreter (3% of cases). The chronicity of coronary bypass graft. the LBBB was determined by comparison with the most recent previous ECG available in our hospital's computerized ECG records. An LBBB was considered new if the most recent, available ECG was Statistical analysis not LBBB. If no prior ECG was available for comparison, patients Comparisons were made among patients with chronic LBBB, were classified as having presumed new LBBB. Patients who had new LBBB, and presumed new LBBB. Because patients with new LBBB on the most recent prior ECG were classified as having or presumed new LBBB are considered candidates for early chronic LBBB. reperfusion treatment, they were combined for several analyses. Student t test or χ2 analysis was used for categorical and Definitions proportional variables, respectively. P =.05 was considered statistically significant. Significant univariate variables at P b .10 All patients underwent serial sampling of total creatine kinase were included in a multivariate analysis to identify independent (CK)–MB (CK-MB) and troponin I (TnI). Diagnosis of MI predictors of acute MI and 30-day mortality. Statistical analysis required at least 1 TnI value that exceeded the reference was preformed with a standard statistical software package (SAS range (coefficient of variation b10% for that assay) with a 6.11, SAS, Cary, NC). serial rise and fall. Five different assays for CK-MB and TnI were used: the Opus Magnum Analyzer (Boston, MA) (diagnos- The authors are solely responsible for the design and conduct tic value 1.0 ng/mL, June 1996-May 1998), the Bayer Immuno of this study, all study analyses, and the drafting and editing of One assay (Tarrytown, NY) (diagnostic value 0.3 ng/mL, May the paper and its final contents. 1998-April 2000), the DPC assay (Los Angeles, CA) (diagnostic value 1.0 ng/mL, April 2000-December 2002), the Bayer TnI assay (Tarrytown, NY) (diagnostic value 0.5 ng/mL, December 2002-April 2007), and the Bayer TnI Ultra Assay (Tarrytown, Results NY) (diagnostic value 0.1 ng/mL, April 2007-end of the study). Clinical characteristics Because different TnI assays were used, infarct size was From 1994 to 2009, 401 consecutive patients who estimated using peak CK-MB values obtained during the initial underwent evaluation for possible acute coronary syn- 24 to 36 hours after admission. The number of patients who had drome had LBBB on the initial ECG and were included in N CK-MB MI (CK-MB 6 ng/mL) in conjunction with an elevated this analysis. In general, the patient population was older TnI was also determined. Size of MI was described based on (median age 66 years), with frequent comorbidities multiples of peak CK and CK-MB. To compare relative MI size, patients with LBBB and MI were compared with 547 patients (Table I). Ejection fraction (EF) was available in 360 b with ST-elevation MI (STEMI) and 2,467 patients with non- patients (90%), of whom 76% had an EF 50% and with STEMI admitted during the same period. 61%, b40%. Renal failure was defined as a creatinine clearance b60 mL/ A majority of patients (64%) had either a new LBBB min, estimated using the Cockcroft-Gault equation,14 using the (36%) or presumed new LBBB (28%). Clinical American Heart Journal 700 Kontos et al April 2011

characteristics of patients with chronic LBBB were similar Figure 1 to those with new or presumably new LBBB, except patients with chronic LBBB were less likely to be male A and have prior MI or prior revascularization (Table I). 100%

Myocardial infarction 80% Myocardial infarction was diagnosed in 116 patients (29%), of whom 23 (6%) had TnI elevations alone without 60% >10 x >5-10x CK-MB elevations, whereas 93 had both TnI and CK-MB 1 to 5 x elevations (23%). The frequency of MI and was not 40% <1 significantly different in patients with chronic, new, or presumed new LBBB (Table I). In the subgroup of patients 20% who had MI, the size of MI based on peak CK-MB and CK values was relatively similar among the 3 groups, although 0% patients with known LBBB were more likely to have MI Old New Unknown New+Unk without CK-MB elevations (Figure 1A and B). B Univariate predictors of MI included older age, 100% concordant ECG changes, congestive (CHF) at admission, prior MI, prior revascularization, 80% and abnormal renal function (Table II). Multivariate predictors included concordant ECG changes (P b .001, 60% >10 x odds ratio [OR] 16, 95% CI 3.4-81) and age ≥65 years >5-10x (P = .02, OR 1.8, 95% CI 1.1-2.9). 1 to 5 x 40% <1

Electrocardiographic criteria 20% Myocardial infarction was diagnosed in 12 (86%) of 14 patients who had concordant ST changes compared to 0% 109 (28%) of 388 without ECG criteria (P b .001). Old New Unknown New+Unk Concordant ST depression was present in 4 patients, and Comparison of multiples of peak CK A, and peak CK-MB B, based on concordant ST elevation was present in 10, of whom chronicity of LBBB. Unk indicates unknown. 8 had MI (Table III). Of the 2 patients with concordant ST elevation who did not have MI, one had 1 mm of elevation in leads I and L that was new compared to a Univariate predictors of MI are shown in Table II. prior ECG, and the other had 1 mm of elevation isolated Multivariate predictors included only concordant ST to lead V ; no prior ECG was available. 4 changes (OR 17, 95% CI 3.4-81, P b .001) and age Patients who had concordant ST changes and MI were ≥65 years (OR 1.7, 95% CI 1.1-2.7, P = .02). more likely to have a peak CK-MB and CK N10× normal (both P b .05) and were less likely to have a peak CK-MB or CK b1× normal (both P b .05) compared to patients Coronary angiographic outcomes without concordant ST changes. The distribution of Coronary angiography was performed in 152 patients. peak CK (1,820 [680, 6,020] U/L vs 2,700 [580, 3,460] This included 59 (51%) of the TnI (+) patients and 93 − U/L) and CK-MB values (99 [48, 322] ng/mL vs 200 (33%) of the TnI ( ) patients. Significant disease was [44, 290] ng/mL) were similar for patients with concor- present in 43 (73%) of the TnI (+) patients, of whom 21 − dant ST changes and the control group of patients who (49%) had revascularization, and 37 (40%) of the TnI ( ) had STEMI evaluated during the same period (Figure 2A patients, of whom 16 (37%) had revascularization. and B). In patients with MI and LBBB who did not have Sixteen patients underwent emergent treatment: 1 patient concordant ECG changes, the distribution of peak CK was treated with fibrinolytics(hadMI),and15underwent (240 [120, 490] U/L vs 780 [130, 590] U/L) and CK-MB emergent coronary angiography. Significant disease was values (13 [6.8, 29] ng/mL vs 42 [4.9, 38] ng/mL) was presentin9,ofwhom7hadPCI,1hadcoronaryartery similar to the control group of patients with non-STEMI bypass surgery (CABG), and 1 was managed medically. Six (Figure 2). patientshadnosignificantdisease. Six patients had an ST/T ratio N25%, of whom 4 had MI. Peak MB and CK levels tended to be low; peak MB Thirty-day mortality values were 4.1, 16, 20, and 58 ng/mL, and peak CK Thirty-day mortality was 11% overall; 6.7% in patients values were 89, 187, 372, and 493 U/L. without and 22% in patients with MI. In patients with American Heart Journal Kontos et al 701 Volume 161, Number 4

Table II. Distribution of variables in patients with and without MI Figure 2 MI (n = 116) No MI (n = 285) P A Chronic LBBB 46 (40) 99 (35) .35 100% New LBBB 35 (30) 110 (39) .11 Unknown duration 35 (29) 76 (27) .48 80% New or unknown 70 (60) 186 (66) .33 Age, y 70 ± 13 (70) 65 ± 15 (64) .002 Concordant ST-T changes 12 (10) 2 (0.7) b.001 60% >10 x Age ≥65 y 71 (61) 122 (43) b.001 5-10x Male 46 (40) 112 (39) .95 1 to 5 x 40% Hypertension 96 (83) 216 (76) .13 <1 Diabetes 52 (45) 106 (37) .16 CHF 40 (35) 66 (23) .02 20% Prior MI 43 (37) 66 (23) .005 Prior revascularization 39 (34) 69 (24) .06 CrCl, mL/min 50 ± 27 (49) 63 ± 32 (61) .15 0% LBBB (+) ST STEMI LBBB (-) ST NSTEMI CrCl b60 mL/min 74 (64) 127 (49) b.001 30-day mortality 27 (23) 19 (6.7) b.001 B 100% Categorical data are presented as n (%), and continuous data are presented as mean ± SD (median). 80%

60% >10 x >5-10x Table III. Demographics and risk factors based on the presence 1 to 5 x of concordant ST changes in patients with MI 40% <1 ST-T changes No ST-T changes (n = 14) (n = 107) P 20%

Age, y 69 ± 9 (69) 70 ± 14 (70) .58 0% Age ≥ 65 10 (71) 64 (60) .65 LBBB(+) ST STEMI LBBB (-) ST NSTEMI Male 7 (50) 41 (38) .40 Hypertension 10 (71) 90 (84) .24 Comparison of peak CK A, and peak CK-MB B, based on whether Diabetes 4 (29) 48 (45) .25 Prior MI 5 (36) 23 (21) .24 the ECG showed LBBB with concordant changes (LBBB [+] ST), STEMI, − CHF 3 (21) 37 (35) .32 LBBB without concordant changes (LBBB [ ] ST), or NSTEMI. NSTEMI EF 34 ± 16 (35) 32 ± 15 (28) .72 indicates non–ST-elevation MI. EF b40% 8 (62) 68 (67) .64 Prior revascularization 6 (43) 35 (33) .45 CrCl, mL/min 45 ± 27 (45) 51 ± 27 (51) .44 changes had a high likelihood of MI. In addition, CrCl b60 mL/min 10 (71) 66 (62) .48 concordant ECG changes were the most important TnI MI 12 (86) 107 (100) b.001 independent predictor of MI as well as an independent CK-MB MI 12 (86) 81 (76) b.24 predictor of 30-day mortality. 30-day mortality 6 (50) 21 (20) .02 Administration of fibrinolytic therapy to patients with Categorical data are presented as n (%). Continuous data are presented as mean ± AMI and ST-segment elevation decreases mortality.15 The SD (median). results of the Fibrinolytic Collaborative Group, which included patients with bundle-branch block and pre- sumed AMI, demonstrated a 25% reduction in mortality MI, 30-day mortality was 50% in those with and 20% with treatment with fibrinolytics15 and thus formed the in patients without concordant ECG changes (Table IV). basis for current treatment recommendations.1,2 How- Troponin MI was the most important univariate ever, important limitations of this analysis were that and multivariate predictor of mortality. Other multiva- patients with right bundle-branch block and LBBB riate predictors included concordant ECG changes and were not differentiated nor was the chronicity of the male gender. LBBB determined.

Prevalence of LBBB and MI Discussion Clinical trials that have included patients with LBBB We found that there was no difference in the incidence have found that the rate of MI confirmation is significantly of MI based on the chronicity of the LBBB on the lower in patients with LBBB than those who have ST presenting ECG. Patients who had concordant ST elevation.10,11 In the ASSENT 2 and 3 trials, despite American Heart Journal 702 Kontos et al April 2011

Table IV. 30-day mortality

30-day No 30-day mortality mortality Multivariate (n = 47) (n = 354) P P value OR (95% CI)

ST-T changes 7 (15) 7 (2) b.001 .02 4.3 (1.3-15) Age, y 68 ± 17 (70) 66 ± 14 (66) .52 Chronic 15 (32) 130 (37) .52 New 18 (38) 127 (36) .74 Unknown 14 (30) 97 (27) .73 New + unknown 32 (68) 224 (63) .52 Age ≥65 y 29 (62) 164 (46) b.05 .54 1.3 (0.61-2.6) Male 27 (57) 131 (37) b.01 .005 2.6 (1.3-5.1) Hypertension 32 (68) 280 (79) .09 Diabetes 17 (36) 140 (40) .66 CHF 17 (36) 89 (25) .11 Prior MI 15 (32) 94 (27) .44 Prior revascularization 10 (21) 98 (28) .35 CrCl b60 mL/min 33 (70) 168 (47) .004 .03 2.3 (1.1-4.8) TnI MI 28 (60) 91 (26) b.001 .001 3.3 (1.7-6.7)

Categorical data are presented as n (%). Continuous data are presented as mean ± SD (median). rigorous entry criteria, 38% of those enrolled with LBBB Increasing data indicate that patients with LBBB and did not have MI confirmed.11 Similarly, in the HERO Trial, MI presenting to the ED represent a heterogenous patients with LBBB were less likely to have MI.10 In more group of patients: a minority who have STEMI equivalent, heterogeneous ED patients, the prevalence of AMI who are likely to benefit from acute reperfusion among patients who have LBBB is even lower.7,8 treatment; patients with non-STEMI equivalent, who have myocardial necrosis but without complete vessel New LBBB occlusion; and, with increased use of more sensitive We found that the frequency as well as the size of MI assays, patients with acute CHF exacerbation with minor was similar among patients with chronic, new, or LBBB of myocardial necrosis who may be mistakenly classified as unknown duration. Although the presence of a new (or having MI. Although all may benefit from invasive presumably new) LBBB is thought to identify patients evaluation, it is only the first group in whom emergent who have AMI and emergent reperfusion is recom- reperfusion treatment is beneficial. Therefore, accurate mended, data to support these criteria are limited. No identification of patients with LBBB who are likely to have references are provided in current guidelines.1,2 Some an occluded vessel is important. studies,7 but not all,8 have shown that a new LBBB is associated with an increased rate of MI. However, even in Diagnostic ECG changes patients with new LBBB, MI is diagnosed in only a Early diagnosis of MI in patients with LBBB can be minority. For example, Chang et al8 found MI in 7.2%, improved by applying specific diagnostic ECG criteria. and Li et al7 found MI in 16% of patients with new or Sgarbossa et al9 reported that concordant ST elevation or LBBB of unknown duration. Therefore, this criterion ST depression ≥1 mm was highly specific, although not appears insufficient for selecting patients for early sensitive, for identifying MI in patients with LBBB. We reperfusion therapy. found similar results; concordant ST changes were The low diagnostic accuracy of the ECG and increased specific for MI and were the most important independent numbers of atypical presentations coupled with the high predictor of MI. In addition, it was an independent rate of comorbidities3,4 in patients with LBBB can result predictor of 30-day mortality; in contrast, a new LBBB in diagnostic confusion and likely contribute to the predicted neither MI nor mortality. lower rate3,4 and increased time3-5,11 to emergent Our results are concordant with other studies. In a reperfusion in these patients. In an analysis from the combined analysis of patients with LBBB enrolled in the National Registry of Acute MI, patients who had LBBB ASSENT 2 and 3 Trials that included 253 patients with were significantly less likely to undergo early reperfusion LBBB, MI was diagnosed in only 13% of patients with treatment3,4; in those who did, door-to-drug time was LBBB without concordant ECG changes compared to longer than those with STEMI. Similarly, more recent data 49% of those with concordant ECG changes.11 Addi- in patients undergoing primary PCI from the ACTION- tional analyses that have included patients enrolled in GWTG Registry found that primary PCI rates were clinical trials10 or patients admitted from the ED have significantly lower and were associated with greater confirmed the high specificity of these criteria. In a delay than that of STEMI patients.5 meta-analysis of 10 studies that included 1,614 patients, American Heart Journal Kontos et al 703 Volume 161, Number 4

the presence of concordant ECG changes had a Limitations specificity of 98%, with a positive likelihood ratio of 16 Although patient data from an extended period were 7.9 for predicting MI. used, the criteria for emergent reperfusion in patients We also found that patients with LBBB who had MI and with LBBB have been present because guidelines were concordant ECG changes had significantly larger infarcts, published in 1996. We used CK and CK-MB to estimate likely contributing to the higher mortality. Peak CK and peak MI size because of the multiple different TnI assays CK-MB values were similar to that of STEMI patients used during the study period. In addition, use of CK and and significantly higher than that in non-STEMI patients. CK-MB allows comparison across other studies, a Similar results were reported in the HERO Trial; as peak technique used in clinical trials. CK values were more than twice to those with con- Despite these limitations, our study has several cordant ECG changes, mortality was similar to the STEMI strengths. These include the large consecutive cohort of patients and almost twice as high as LBBB patients with- 10 patients with LBBB, the ability to assess the chronicity of out concordant ST changes. the LBBB, the ability to determine 30-day rather than just Requiring diagnostic ECG changes could induce in-hospital mortality, and use of standardized diagnostic complexity into the decision making. However, a high criteria for MI, including use of troponins, instead of CK agreement between cardiologists and emergency medi- or CK-MB. cine physicians for diagnosing MI using these criteria has been reported,17 indicating that emergency physicians could be able to accurately use these criteria. Conclusions In the absence of concordant ST elevation or depres- Most patients who present to the ED with possible sion, the ability to rapidly identify MI is difficult. Standard AMI who have an underlying LBBB do not have an MI. In tools, such as cardiac markers or emergent imaging, are those who do, MI size is relatively small and therefore limited. Body-surface potential ECG mapping18,19 has are unlikely to benefit from emergent reperfusion treat- shown promise but has not gained wide-spread use. ment. The diagnostic use of a new LBBB was low. Specific Therefore, use of specific ECG criteria appears to be the ECG criteria, concordant ST depression or elevation, had best available tool for rapid selection of patients who are a high specificity and identified patients who had large most likely to benefit from emergent reperfusion. MIs. The use of emergent reperfusion treatment in patients with new LBBB as a performance indicator Performance measures should be reconsidered. Only a subset of guidelines should be considered performance measures—those specifically intended for public reporting, external comparisons, and pay-for- Disclosures performance programs.20 They should be limited to Funded, in part, from research funds from the Pauley those that meet the highest level of evidence, with Heart Center. substantial data that the treatment is beneficial and that failure to treat appropriate patients is associated with worse outcomes.21 However, we and others7,8 found that References MI is diagnosed in only a minority of patients with new 1. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines LBBB. The size of MI was relatively small in most patients for the management of patients with ST-elevation myocardial — who had MI and therefore less likely to benefit from infarction executive summary. A report of the American College of emergent reperfusion, especially if treated with fibrino- Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to revise the 1999 guidelines for the lytics.15 In addition, attempts to meet the 90-minute door- management of patients with acute myocardial infarction). J Am Coll to-balloon time could lead to inadvertent patient harm. Cardiol 2004;44:671-719. There is a high false-positive rate in patients with LBBB 2. Van De WF, Bax J, Betriu A, et al. Management of acute myocardial undergoing emergent coronary angiography for pre- infarction in patients presenting with persistent ST-segment elevation: 22 sumed STEMI. The risk of emergent angiography may the Task Force on the Management of ST-Segment Elevation Acute outweigh the benefits given the high rates of underlying Myocardial Infarction of the European Society of Cardiology. Eur CHF, renal insufficiency, and other comorbidities that Heart J 2008;29:2909-45. increase the risk of contrast nephropathy.23 Use of 3. Go AS, Barron HV, Rundle AC, et al. Bundle-branch block and in- troponins may inadvertently classify patients with CHF hospital mortality in acute myocardial infarction. National Registry of Myocardial Infarction 2 Investigators. Ann Intern Med 1998;129: as having MI and therefore subject to this performance 690-7. measure. Because most patients with a new LBBB do 4. Shlipak MG, Go AS, Frederick PD, et al. Treatment and outcomes not have MI, clinical judgment and use of objective of left bundle-branch block patients with myocardial infarction ECG criteria offer the best way to determine the need for who present without chest pain. J Am Coll Cardiol 2000;36: early reperfusion. 706-12. American Heart Journal 704 Kontos et al April 2011

5. Yeo KK, Li S, Wang TY, et al. New or presumed new left bundle from all randomised trials of more than 1000 patients. Lancet 1994; branch block in patients with acute myocardial infarction: findings 343:311-22. From ACTION Registry-GWTG. J Am Coll Cardiol 2010;55:A107. 16. Tabas JA, Rodriguez RM, Seligman HK, et al. Electrocardiographic 6. Kontos MC, McQueen MJ, Jesse RL, et al. Can myocardial infarction criteria for detecting acute myocardial infarction in patients with left be rapidly identified in emergency department patients who have left : a meta-analysis. Ann Emerg Med 2008;52: bundle-branch block? Ann Emerg Med 2001;37:431-8. 329-36. 7. Li SF, Walden PL, Marcilla O, et al. Electrocardiographic diagnosis 17. Sokolove PE, Sgarbossa EB, Amsterdam EA, et al. Interobserver of myocardial infarction in patients with left bundle branch block. agreement in the electrocardiographic diagnosis of acute myocardial Ann Emerg Med 2000;36:561-5. infarction in patients with left bundle branch block. Ann Emerg Med 8. Chang AM, Shofer FS, Tabas JA, et al. Lack of association between 2000;36:566-71. left bundle-branch block and acute myocardial infarction in 18. Maynard SJ, Menown IB, Manoharan G, et al. Body surface mapping symptomatic ED patients. Am J Emerg Med 2009;27:916-21. improves early diagnosis of acute myocardial infarction in patients 9. Sgarbossa EB, Pinski SL, Barbagelata A, et al. Electrocardiographic with chest pain and left bundle branch block. Heart 2003;89: diagnosis of evolving acute myocardial infarction in the presence of 998-1002. left bundle-branch block. GUSTO-1 (Global Utilization of Streptoki- 19. Owens C, McClelland A, Walsh S, et al. Comparison of value of nase and Tissue Plasminogen Activator for Occluded Coronary leads from body surface maps to 12-lead electrocardiogram for Arteries) Investigators. N Engl J Med 1996;334:481-7. diagnosis of acute myocardial infarction. Am J Cardiol 2008;102: 10. Wong CK, French JK, Aylward PE, et al. Patients with prolonged 257-65. ischemic chest pain and presumed-new left bundle branch block have 20. Krumholz HM, Keenan PS, Brush Jr JE, et al. Standards for measures heterogeneous outcomes depending on the presence of ST-segment used for public reporting of efficiency in health care: a scientific changes. J Am Coll Cardiol 2005;46:29-38. statement from the American Heart Association Interdisciplinary 11. Al Faleh H, Fu Y, Wagner G, et al. Unraveling the spectrum of left Council on Quality of Care and Outcomes research and the bundle branch block in acute myocardial infarction: insights from the American College of Cardiology Foundation. J Am Coll Cardiol Assessment of the Safety and Efficacy of a New Thrombolytic 2008;52:1518-26. (ASSENT 2 and 3) trials. Am Heart J 2006;151:10-5. 21. Krumholz HM, Anderson JL, Brooks NH, et al. ACC/AHA 12. Tatum JL, Jesse RL, Kontos MC, et al. Comprehensive strategy for the clinical performance measures for adults with ST-elevation and evaluation and triage of the chest pain patient. Ann Emerg Med non-ST-elevation myocardial infarction: a report of the 1997;29:116-23. American College of Cardiology/American Heart Association 13. Smith SW, Heegaard W, Bachour FA, et al. Acute myocardial Task Force on Performance Measures. J Am Coll Cardiol 2006; infarction with left bundle-branch block: disproportional anterior ST 47:236-65. elevation due to right ventricular myocardial infarction in the 22. Larson DM, Menssen KM, Sharkey SW, et al. “False-positive” cardiac presence of left bundle-branch block. Am J Emerg Med 2008;26: catheterization laboratory activation among patients with suspected 342-7. ST-segment elevation myocardial infarction. JAMA 2007;298: 14. Cockcroft DW, Gault MH. Prediction of creatinine clearance from 2754-60. serum creatinine. Nephron 1976;16:31-41. 23. Mehran R, Aymong ED, Nikolsky E, et al. A simple risk score for 15. Fibrinolytic Therapy Trialists' (FTT) Collaborative Group. Indications prediction of contrast-induced nephropathy after percutaneous for fibrinolytic therapy in suspected acute myocardial infarction: coronary intervention: development and initial validation. J Am Coll collaborative overview of early mortality and major morbidity results Cardiol 2004;44:1393-9.