Journal of the American College of Cardiology © 2005 by the American College of Cardiology Foundation Published by Elsevier Inc.

Vol. 46, No. 1, 2005 ISSN 0735-1097/05/$30.00 doi:10.1016/j.jacc.2005.02.084

Patients With Prolonged Ischemic Chest Pain and Presumed-New Left Bundle Branch Block Have Heterogeneous Outcomes Depending on the Presence of ST-Segment Changes Cheuk-Kit Wong, MD, FACC,*† John K. French, MB, PHD, FACC,* Philip E. G. Aylward, MD, FACC,‡ Ralph A. H. Stewart, MD, FACC,* Wanzhen Gao, PHD,* Paul W. Armstrong, MD, FACC,§ Frans J. J. Van De Werf, MD, FACC,储 R. John Simes, MD,¶ O. Christopher Raffel, MB,* Christopher B. Granger, MD, FACC,# Robert M. Califf, MD, FACC,# Harvey D. White, DSC, FACC,* for the HERO-2 Trial Investigators Auckland and Dunedin, New Zealand; Adelaide, South Australia; Edmonton, Canada; Leuven, Belgium; Sydney, Australia; and Durham, North Carolina The purpose of this research was to examine the prognostic value of ST-segment changes (concordant ST-segment elevation and/or precordial V1 to V3 ST-segment depression) during presumed-new left bundle branch block (LBBB) in patients receiving fibrinolytic therapy. BACKGROUND These patients are often considered high-risk, but their outcome is not well-defined. METHODS The Hirulog and Early Reperfusion or Occlusion (HERO)-2 trial compared bivalirudin with heparin in patients receiving streptokinase for ST-segment elevation or presumed-new LBBB. Each patient with LBBB was matched with a control (with normal intraventricular conduction) for age, gender, pulse rate, systolic blood pressure, Killip class, and region. RESULTS A total of 300 patients had LBBB (92 with and 208 without ST-segment changes) and 15,340 had normal conduction. Acute myocardial infarction (AMI) occurred in 80.7% of LBBB patients and 88.7% of controls (p ⫽ 0.006). ST-segment changes were specific (96.6%) but not sensitive (37.8%) for enzymatic diagnosis of AMI. Mortality at 30 days was similar in LBBB patients with ST-segment changes (21.7%) and controls (25.0%, p ⫽ 0.563), but lower in LBBB patients without ST-segment changes than in controls (13.5% vs. 21.6%, p ⫽ 0.022). In the whole HERO-2 cohort, the LBBB patients with ST-segment changes had higher mortality than patients with normal conduction (odds ratio [OR] 1.37, 95% confidence interval [CI] 0.78 to 2.42). The LBBB patients without ST-segment changes had lower mortality than patients with normal conduction (OR 0.52, 95% CI 0.33 to 0.80). CONCLUSIONS ST-segment changes during LBBB are specific for the diagnosis of AMI and predict 30-day mortality; LBBB patients without ST-segment changes have lower adjusted 30-day mortality than those with normal conduction. Trials are required to determine the best treatment for high-risk and low-risk patients with LBBB. (J Am Coll Cardiol 2005;46:29 –38) © 2005 by the American College of Cardiology Foundation OBJECTIVES

Repolarization changes occurring with left bundle branch block (LBBB) can obscure the classical electrocardiographic changes of ST-segment elevation in patients presenting See page 39 with acute myocardial infarction (AMI). Treatment guidelines recommend reperfusion therapy for patients presenting with LBBB and a history suggestive of AMI, regardless of From the *Cardiovascular Research Unit, Green Lane Hospital, Auckland, New Zealand; †Department of Medical and Surgical Sciences, University of Otago, Dunedin, New Zealand; ‡Cardiology Department, Flinders Medical Centre, Adelaide, South Australia; §Department of Medicine, University of Alberta, Edmonton, Canada; 储Department of Cardiology, Gasthuisberg University Hospital, Leuven, Belgium; ¶National Health and Medical Research Council Clinical Trials Centre, University of Sydney, Sydney, Australia; and #Duke Clinical Research Institute, Durham, North Carolina. Honorary Professor White received partial salary support from the Green Lane Research and Educational Fund Board (Auckland, New Zealand). Manuscript received August 18, 2004; revised manuscript received February 2, 2005, accepted February 8, 2005.

associated ST-segment changes (1,2). Although these patients have generally been considered to be at higher risk, it is uncertain if LBBB, in itself, predicts an adverse outcome independently of other prognostic factors. When LBBB is present, delayed left ventricular activation occurs due to electrical activity spreading rapidly from the terminations of the right bundle branch. Normally the ST segment is depressed or elevated in the opposite direction to the main QRS deflection due to secondary repolarization changes (3). Manifestation of concordant ST-segment elevation (with a positive QRS complex) or ST-segment depression in leads V1 to V3 (with a negative QRS complex) during LBBB requires large ST-segment shifts, which may reflect severe transmural myocardial ischemia (4). Using a raised creatine kinase-myocardial band (CKMB) level as the “gold standard” for the diagnosis of AMI, Sgarbossa et al. (4) studied a cohort of patients from the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO)-I

30

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

Abbreviations and Acronyms AMI ⫽ acute myocardial infarction CI ⫽ confidence interval CK ⫽ creatine kinase GUSTO ⫽ Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries HERO ⫽ Hirulog and Early Reperfusion or Occlusion LBBB ⫽ left bundle branch block OR ⫽ odds ratio RBBB ⫽ right bundle branch block ULN ⫽ upper limit of normal

trial, and found that two features occurring during LBBB had independent diagnostic value for AMI. ST-segment elevation measuring ⱖ1 mm concordant with (i.e., in the same direction as) the QRS complex in any lead had the highest diagnostic value, followed by ST-segment depression measuring ⱖ1 mm in any lead from V1 to V3. ST-segment elevation measuring ⱖ5 mm discordant with the QRS complex was not, in itself, diagnostic of AMI. These criteria have recently been tested in community and emergency room studies, and were found to be specific, but not sensitive, for the diagnosis of AMI (5,6). In the large, international Hirulog and Early Reperfusion or Occlusion (HERO)-2 trial (7), electrocardiograms were recorded at randomization and repeated 60 min after commencing streptokinase, allowing identification of serial changes during LBBB. In this substudy, we examined the prognostic value (for 30-day mortality) of concordant ST-segment elevation or ST-segment depression in leads V1 to V3 during LBBB.

METHODS The HERO-2 trial has been described previously (7). Patients presenting with ⬎30 min of ischemic chest discomfort and either ST-segment elevation or presumednew-onset LBBB (i.e., with no previous record of LBBB) within 6 h of symptom onset were randomized to receive either bivalirudin or unfractionated heparin as adjunctive therapy with streptokinase and aspirin. All patients gave written, informed consent. The primary end point was 30-day mortality, which did not differ between the two treatment groups. The electrocardiograms (ECG) recorded at randomization (i.e., before fibrinolytic therapy) and 60 min after commencement of fibrinolytic therapy were sent to a core laboratory at Green Lane Hospital for analysis by experienced readers. Enzymatic confirmation of AMI was defined as a CK level exceeding twice the upper limit of normal (ULN) or a CK-MB level exceeding the ULN. Electrocardiographic analysis of LBBB. The diagnostic criteria for LBBB were a QRS duration of ⱖ0.120 s in the presence of sinus or supraventricular rhythm, a QS or rS complex in lead V1, and an R-wave peak time of ⱖ0.06 s in leads I, V5, or V6 associated with absence of a Q-wave in the

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

same lead (4). The following measurements were made by a cardiologist blinded to the patients’ treatment and outcome, with ST-segment deviation being measured at the J point and compared with the TP segment isoelectric line: Maximum magnitude of ST-segment elevation in leads with a positive QRS complex (concordant ST-segment elevation). Maximum magnitude of ST-segment depression in leads V1, V2, or V3. Maximum magnitude of ST-segment elevation in leads with a negative QRS complex (discordant ST-segment elevation). Change in net QRS polarity (direction) over 60 min in the lead showing a positive QRS complex and maximum concordant ST-segment elevation at randomization. Statistical analysis. Patients with LBBB were compared with: 1) a matched control group with normal intraventricular conduction, and 2) all patients with normal intraventricular conduction. Each patient with LBBB at randomization was matched (for age, gender, pulse rate, systolic blood pressure, and Killip class at randomization) with a control patient chosen from those in HERO-2 who had normal intraventricular conduction at both randomization and 60 min. As this substudy was performed in 46 countries, the patients were also matched for their recruitment region (Russia, Eastern Europe, Western countries, Latin America, and Asia). After matching for these six variables, further matching was done for diabetes and the time from symptom onset to randomization. The data are presented as proportions or median values with interquartile ranges where appropriate. Comparisons between groups were done using the chi-square test for categorical variables and the Mann-Whitney U test for continuous variables. Comparisons between the LBBB patients and matched control patients were done using McNemar’s test (for two levels) or Bowker’s test (for more than two levels) for categorical variables, and the Wilcoxon signed rank test for continuous variables. McNemar’s test was also used to compare the incidence of enzymatically confirmed AMI in patients randomized with LBBB and the matched control patients. Also tested were the sensitivity, specificity, positive predictive accuracy, and negative predictive accuracy (for diagnosing AMI) of the three different criteria for ST-segment deviation during LBBB (i.e., concordant ST-segment elevation measuring ⱖ1 mm, STsegment depression measuring ⱖ1 mm in leads V1 to V3, and discordant ST-segment elevation measuring ⱖ5 mm). Logistic regression analysis was used to establish whether any of these three criteria independently predicted 30-day mortality. Logistic regression analysis was also done including all patients with normal intraventricular conduction in HERO-2 and incorporating the baseline variables described above, including the time from symptom onset. The model calculations were repeated using the GUSTO prognostic

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

risk score, which takes into account the heart rate, systolic blood pressure, AMI location, previous AMI, and the interaction between age and Killip class (8). Unadjusted and adjusted 30-day survival curves were generated using Cox regression analysis to visualize survival rates for three groups of patients: LBBB with ST-segment changes, LBBB without ST-segment changes, and normal intraventricular conduction. In these models, the factor “ST-segment changes” was treated as a stratifying factor rather than as a covariate. The GUSTO score (continuous variable) and region (five levels) were entered as covariates in the adjusted model. The p values were generated by fitting extra Cox models, in which ST-segment change was entered as a covariate rather than as a stratum, as well as the GUSTO score and region. The assumption was that hazards were proportional to the baseline hazard, and this was tested graphically and statistically (data not shown). Cox regression analysis was done using S-Plus 4.5 software (MathSoft Inc., Cambridge, Massachusetts), and the rest of the analysis was done using SAS 8.01 software (SAS Institute Inc., Cary, North Carolina).

RESULTS Of the 17,073 patients randomized into HERO-2, 300 (1.76%) had LBBB at randomization. Of these 300, 41 had resolution of LBBB within 60 min (36 patients changed to normal intraventricular conduction and 5 changed to right bundle branch block [RBBB]). A total of 15,340 patients had normal intraventricular conduction at both randomization and 60 min. Another 26 patients (25 with normal intraventricular conduction and 1 with RBBB at randomization) developed new LBBB within 60 min. Table 1 describes the baseline characteristics, in-hospital interventions, and 30-day mortality of the 300 patients with LBBB at randomization, the 300 matched control patients, and the 15,340 patients with normal intraventricular conduction at both time points. Of the 300 patients with LBBB, 148 received bivalirudin and 152 received heparin, and their 30-day mortality rates were 16.2% and 15.8%, respectively (p ⫽ 0.92). When compared with the 15,340 patients with normal intraventricular conduction, patients with LBBB were older and had a greater prevalence of previous AMI and other cardiovascular risk factors, a higher pulse rate, and a higher Killip classification at randomization, as reflected by a higher GUSTO risk score. The 30-day mortality rates were 16% in patients with LBBB and 9.1% in those with normal intraventricular conduction (p ⬍ 0.001), and approximately 40% of these deaths occurred within the first 24 h (Table 1). However, patients with LBBB had a lower 30-day mortality rate than matched control patients (16% vs. 22.7%); in 67% of matched pairs, both the LBBB patient and the control patient survived; in 5.7% of matched pairs, both patients died; in 17% of matched pairs, only the control patient died;

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

31

and in 10.3% of matched pairs, only the LBBB patient died (p ⫽ 0.027). Incidence of enzymatically confirmed AMI in patients randomized with LBBB. Patients with LBBB had a lower incidence of enzymatically confirmed AMI than matched control patients (80.7% vs. 88.7%, p ⫽ 0.006) or unselected patients with normal intraventricular conduction (80.7% vs. 92.1%, p ⬍ 0.001; Table 2). These differences persisted when deaths within 24 h were assumed to be due to AMI and when a lower CK cut-point (⬎1.5 times the ULN) was used as the criterion for detection of myocyte necrosis (Table 2). Peak cardiac enzyme levels were lower in patients with LBBB than in matched control patients, and this finding was also present when only patients with enzymatically confirmed AMI were included in the analysis (Table 2). Usefulness of concordant ST-segment elevation or lead V1 to V3 ST-segment depression during LBBB for diagnosis of enzymatically confirmed AMI. The criterion of concordant ST-segment elevation measuring ⱖ1 mm on the randomization ECG had high specificity (98.3%) but low sensitivity (33.5%) for the diagnosis of enzymatically confirmed AMI. The criterion of ST-segment depression measuring ⱖ1 mm in any of the V1 to V3 leads had similarly high specificity, but only 14.1% sensitivity. Lowering the cut-point for ST-segment changes to ⱖ0.5 mm for each of the two criteria did not improve sensitivity. When both criteria were combined (i.e., concordant ST-segment elevation or lead V1 to V3 ST-segment depression), the specificity for detection of enzymatically confirmed AMI was 96.6%, and the sensitivity was 37.2% (Table 3). The criterion of discordant ST-segment elevation measuring ⱖ5 mm had 58.6% specificity and 29.3% sensitivity. Conduction and ST-segment changes between baseline and 60 min. Of the 41 patients with baseline LBBB that resolved within 60 min after commencing fibrinolytic therapy, 26 had ST-segment elevation at 60 min, and 15 did not. Together with the 26 patients who had ST-segment elevation at randomization and subsequently developed new LBBB within 60 min, a total of 52 patients had LBBB on one ECG and a pattern of ST-segment elevation without LBBB on the other ECG (50 had enzymatically confirmed AMI, and 1 without enzymatic confirmation of AMI died within 24 h). Using these 52 cases as having a confirmed “ST-segment elevation acute coronary syndrome,” we evaluated the diagnostic sensitivity of ST-segment changes during LBBB for enzymatically confirmed AMI. The sensitivity of concordant ST-segment elevation measuring ⱖ1 mm during LBBB was 44.2% (53.9% in the 26 patients developing new LBBB within 60 min), while the sensitivity of ST-segment depression measuring ⱖ1 mm during LBBB in any of the V1 to V3 leads was 23.1% (30.8% in the 26 patients developing new LBBB). In 14 of the 82 patients (17%) with concordant STsegment elevation measuring ⱖ1 mm during LBBB at randomization, ST-segment elevation in the lead showing

32

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

Table 1. Baseline Demographics, In-Hospital Interventions, and Mortality in Patients With LBBB Versus Matched Control Patients, and Versus All Patients With Normal Intraventricular Conduction

Demographics Median age, yrs (IQR) Male gender, % History of cardiovascular disease Angina, % AMI, % Stroke, % Transient ischemic attack, % Percutaneous coronary intervention, % Coronary bypass graft, % Risk factors Smoking history Never smoked, % Past smoker, % Current smoker, % Hypertension, % Diabetes, % Recruitment region Russia, % Eastern Europe, % Western countries, % Latin America, % Asia, % Time from symptom onset to randomization ⱕ2 h, % ⬎2 but ⱕ4 h, % ⬎4 h, % Median time, h (IQR) Anthropometrics Median weight, kg (IQR) Median height, cm (IQR) Median body mass index, kg/m2 (IQR) Hemodynamics Median systolic blood pressure, mm Hg (IQR) Median diastolic blood pressure, mm Hg (IQR) Median heart rate, beats/min (IQR) Killip class I, % II, % III or IV, % Median GUSTO risk score (IQR) Randomization to bivalirudin, % In-hospital interventions Percutaneous coronary intervention, % Coronary bypass graft, % Mortality Within 24 h, % Within 30 days, %

LBBB (n ⴝ 300)

Controls (n ⴝ 300)

p Value vs. LBBB*

Normal Intraventricular Conduction (n ⴝ 15,340)

p Value vs. LBBB†

68.5 (61–75) 57

68 (60–74) 57.7

0.013 0.593

61 (51–70) 71.8

⬍0.001 ⬍0.001

62.7 30.7 0.7 2 0.3 0

55 24.7 0.7 1 1 1.3

0.032 0.009 1 0.317 0.317 N/A

46.4 14.5 0.4 0.9 1/2 0.8

⬍0.001 ⬍0.001 0.321 0.058 0.277 0.286

51.3 27.3 21.3 68.7 17

53 30.3 16.7 62.7 19.3

36 42.3 9.7 9 3

36 42.3 9.7 9 3

16 50 34 3.4 (2.4–4.4)

15 52.3 32.7 3.3 (2.4–4.3)

75 (65–84) 168 (162–174) 26.1 (23.8–29.2)

⬍0.001

0.182

0.089 0.463 N/A

38.3 44.7 17 51.4 13.7

⬍0.001 0.101 0.012

35.6 34.5 15.1 10.2 4.5 0.951

0.032

0.974

21.1 50.5 28.4 3.1 (2.2–4.2)

0.016

75 (67–83) 168 (160–173) 26.4 (24.2–29.0)

0.947 0.133 0.83

75 (68–85) 170 (164–175) 26.4 (24.2–29.3)

0.064 0.035 0.963

140 (120–160) 80 (70–90) 88 (75–101)

140 (120–156) 81 (75–90) 86 (74–100)

0.18 0.111 0.044 N/A

135 (120–150) 80 (70–90) 76 (65–88)

54.3 26.7 19 74 (63–84) 49.3

54.3 26.7 19 72 (61–83) 45.0

0.006 0.287

80.1 16.7 3.2 61 (50–70) 50.0

2.7 1.3

3.0 0.7

0.763 0.414

5.7 0.7

0.026 0.177

6.7 16

8.7 22.7

0.289 0.027

3.5 9.1

0.003 ⬍0.001

0.057 0.514 ⬍0.001 ⬍0.001

⬍0.001 0.828

*p value for comparison between patients with left bundle branch block (LBBB) and matched control patients, as generated from paired comparisons; †p value for comparison between patients with LBBB and those with normal intraventricular conduction. AMI ⫽ acute myocardial infarction; GUSTO ⫽ Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries; IQR ⫽ interquartile range.

maximum concordant ST-segment elevation became discordant within 60 min after the start of fibrinolytic therapy because the net QRS deflection became negative (Fig. 1). ST-segment changes during LBBB and 30-day mortality. A total of 92 patients had concordant ST-segment elevation measuring ⱖ1 mm or lead V1 to V3 ST-segment depression measuring ⱖ1 mm during LBBB, and 208

patients did not (Table 4). Significantly, peak cardiac enzyme levels were more than twice as high in patients with ST-segment changes during LBBB than in those without, and this finding persisted when the analysis excluded patients without enzymatically confirmed AMI (Table 4). Thirty-day mortality tended to be higher in patients with ST-segment changes during LBBB than in those without

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

33

Table 2. Proportion of Patients With Enzymatically Confirmed AMI or Myocardial Necrosis Among Those With LBBB at Randomization

Enzymatically confirmed AMI (CK level ⬎2 ⫻ ULN or CK-MB level ⬎ULN), % Enzymatically confirmed AMI or early death within 24 h, % Myocardial necrosis (CK level ⬎1.5 ⫻ ULN or CK-MB level ⬎ULN), % Myocardial necrosis or early death within 24 h, % Peak enzyme level expressed as multiple of ULN‡ In all patients (IQR) In matched pairs of patients who both had enzymatically confirmed AMI (IQR)

LBBB (n ⴝ 300)

Controls (n ⴝ 300)

p Value vs. LBBB*

Normal Intraventricular Conduction and ST-Segment Elevation (n ⴝ 15,340)

80.7

88.7

0.006

92.1

⬍0.001

82.0

92.7

⬍0.001

93.7

⬍0.001

84.0

91.3

0.01

93.9

⬍0.001

85.3

95.0

⬍0.001

95.2

⬍0.001

4.1 (2.0–9.1) 5.6 (3.0– 10.4)

7.2 (3.1–14.0) 8.1 (4.0–14.9)

⬍0.0001 ⬍0.0001

7.5 (3.4–14.1) 8.2 (4.2–14.9)

⬍0.0001 ⬍0.0001

p Value vs. LBBB†

*p value for comparison between patients with left bundle branch block (LBBB) and matched control patients, as generated from paired comparisons; †p value for comparison between patients with LBBB and those with normal intraventricular conduction; ‡the peak enzyme level was based on the creatine kinase (CK)-MB level where available, or on the CK level if the CK-MB level was not available. AMI ⫽ acute myocardial infarction; IQR ⫽ interquartile range; ULN ⫽ upper limit of normal.

(21.7% vs. 13.5%, p ⫽ 0.067). There was no difference in mortality between the 92 LBBB patients with these STsegment changes and matched control patients (21.7% vs. 25%, p ⫽ 0.563), but mortality was lower in the 208 LBBB patients without these ST-segment changes than in matched control patients (13.5% vs. 21.6%, p ⫽ 0.022; Fig. 2). Predictors of 30-day mortality. On univariable analysis of the 300 patients with LBBB, the odds ratio (OR) for 30-day mortality in those with versus those without ST-segment changes (i.e., concordant ST-segment elevation measuring ⱖ1 mm or lead V1 to V3 ST-segment depression measuring ⱖ1 mm) was 1.79 (95% confidence interval [CI] 0.95 to 3.37, p ⫽ 0.074). The OR for 30-day mortality per 10-U increase in the GUSTO risk score was 1.89 (95% CI 1.49 to 2.41, p ⬍ 0.0001). On multivariable analysis, including the 300 patients with LBBB and the 15,340 patients with normal intraventricular conduction, there was a significant association between the presence of ST-segment changes during LBBB and 30-day mortality (p ⫽ 0.007). The ORs for 30-day mortality were 2.65 (95% CI 1.31 to 5.38) in patients with versus patients without ST-segment changes during LBBB, 1.37 (95% CI 0.78 to 2.42) in patients with ST-segment changes during LBBB versus those with normal intraventricular conduction, and 0.52 (95% CI 0.33 to 0.80) in patients without

ST-segment changes during LBBB versus those with normal intraventricular conduction. The GUSTO risk score was also a predictor of 30-day mortality (OR 2.18 per 10-U increase, 95% CI 2.09 to 2.28, p ⬍ 0.001; Table 5). When the analysis was repeated excluding patients who had normal enzyme levels and survived for ⬎24 h, the ORs were 2.25 (95% CI 1.08 to 4.67) for LBBB with ST-segment changes versus LBBB without ST-segment changes, 1.36 (95% CI 0.77 to 2.41) for LBBB with ST-segment changes versus normal intraventricular conduction, and 0.61 (95% CI 0.38 to 0.97) for LBBB without ST-segment changes versus normal intraventricular conduction. Figures 3 and 4 show the unadjusted and adjusted 30-day survival curves for the 92 LBBB patients with concordant ST-segment elevation measuring ⱖ1 mm or lead V1 to V3 ST-segment depression measuring ⱖ1 mm, the 208 LBBB patients without these changes, and the 15,340 patients with normal intraventricular conduction.

DISCUSSION This study has two major new findings. First, in patients presenting with LBBB and a history of ⬎30 min of ischemic chest discomfort, the presence of concordant ST-segment elevation or lead V1 to V3 ST-segment depression independently predicted higher 30-day mortality. Sec-

Table 3. Application of ST-Segment Criteria for the Diagnosis of AMI in the 300 Patients With LBBB at Randomization

Concordant ST-segment elevation ⱖ1 mm Lead V1 to V3 ST-segment depression ⱖ1 mm Concordant ST-segment elevation ⱖ1 mm or lead V1 to V3 ST-segment depression ⱖ1 mm

n

Sensitivity (%)

Specificity (%)

Positive Predictive Value (%)

Negative Predictive Value (%)

82 35 92

33.5 (27.6–39.8) 14.1 (10.1–19.2) 37.2 (31.1–43.6)

98.3 (89.5–99.9) 98.4 (89.5–99.9) 96.6 (87.0–99.4)

98.8 (92.5–99.9) 97.1 (83.4–99.9) 97.8 (91.6–99.6)

26.1 (20.6–32.6) 21.5 (16.8–27.0) 26.9 (21.1–33.6)

AMI ⫽ acute myocardial infarction; LBBB ⫽ left bundle branch block.

34

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

Figure 1. (A) Randomization and (B) 60-min electrocardiograms in a patient in whom concordant ST-segment elevation in lead V5 became discordant ST-segment elevation because the net QRS deflection in lead V5 became negative by 60 min.

ond, patients with LBBB had more high-risk features (e.g., older age, previous AMI, and a higher GUSTO risk score) than patients with normal intraventricular conduction. After adjustment for these baseline risk factors, patients with LBBB did not have a higher 30-day mortality rate. In fact, the absence of concordant ST-segment elevation or lead V1 to V3 ST-segment depression during LBBB independently predicted a lower 30-day mortality rate than that of patients with normal intraventricular conduction. Our analysis focused on risk assessment at the time when patients were first

seen and treatment decisions needed to be made, before the results of cardiac enzyme tests were available. When multivariable analysis was performed only on patients who were subsequently confirmed as having abnormal enzyme levels, the findings were identical. Over 80% of HERO-2 trial patients with LBBB had enzymatically confirmed AMI. As in previous studies (4 – 6), concordant ST-segment elevation measuring ⱖ1 mm and lead V1 to V3 ST-segment depression measuring ⱖ1 mm within the LBBB morphology were both highly specific

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

35

Table 4. Enzymatically Confirmed AMI and Mortality in Patients With and Patients Without ST-Segment Changes* During LBBB LBBB and ST-Segment Changes* (n ⴝ 92) Enzymatically confirmed AMI Peak enzyme level expressed as multiple of ULN All patients (IQR) Patients with enzymatically confirmed AMI (IQR) 30-day mortality Timing of death 24-h mortality Patients with enzymatically confirmed AMI Causes of death†

Mortality between 24 h and 30 days Patients with enzymatically confirmed AMI Causes of death†

LBBB and No ST-Segment Changes* (n ⴝ 208)

p Value

90 (97.8%)

152 (73.1%)

p ⬍ 0.0001

8.5 (4.0–17.5) 9.4 (4.2–18.0) 20 (21.7%)

3.2 (1.5–6.2) 4.4 (2.9–7.5) 28 (13.5%)

p ⬍ 0.0001 p ⬍ 0.0001 p ⫽ 0.067

10 10 All cardiac

10 6 9 cardiac and 1 from other causes; 4 of the 9 cardiac deaths had no enzymatic confirmation of AMI 18 16 16 cardiac, 2 of whom had no enzymatic confirmation of AMI; 1 from stroke on day 16; 1 from other causes on day 4

10 10 8 cardiac; 2 out-of-hospital deaths not defined

*ST-segment changes were defined as concordant ST-segment elevation measuring ⱖ1 mm or ST-segment depression measuring ⱖ1 mm in electrocardiographic leads V1 to V3 during left bundle branch block (LBBB); †the causes of in-hospital death were classified as cardiac, stroke, bleeding, anaphylaxis, or other causes. AMI ⫽ acute myocardial infarction; IQR ⫽ interquartile range; ULN ⫽ upper limit of normal.

for the diagnosis of enzymatically confirmed AMI. Concordant ST-segment elevation had modest sensitivity of 33.5%, and lead V1 to V3 ST-segment depression had only 14.1% sensitivity. The use of a lower ST-segment cut-point of ⱖ0.5 mm did not improve diagnostic sensitivity. Discordant ST-segment elevation measuring ⱖ5 mm was neither sensitive (29.3%) nor specific (58.6%). In the 45-patient “validation cohort” described in the original report by Sgarbossa et al. (4), the sensitivity of either concordant ST-segment elevation or lead V1 to V3 ST-segment depression for the diagnosis of enzymatically confirmed AMI was 36%, which was consistent with our findings. These criteria had lower sensitivity (⬍20%) in emergency room and community studies (5,6). There are several explanations for the relatively infrequent occurrence of concordant ST-segment elevation during LBBB in patients with enzymatically confirmed AMI. First, some patients with enzymatically confirmed AMI and LBBB may have the “equivalent” of non–ST-segment elevation AMI. From serial electrocardiographic recordings, we identified 52 patients who had LBBB on one ECG and ST-segment elevation without LBBB on the other ECG. In these 52 patients with an “ST-segment elevation acute coronary syndrome,” the diagnostic sensitivity of concordant ST-segment elevation measuring ⱖ1 mm during LBBB was modest (44.2%). Second, the ST segment may have risen from below the original baseline during AMI, but not enough to be detectable as concordant ST-segment elevation during LBBB. The secondary repolarization changes during LBBB are opposite in direction to those of the main QRS deflection (3), leading to a negative ST-segment baseline, and several millimeters of ST-segment elevation during AMI from this negative baseline may be insufficient to raise the ST segment above the isoelectric line. Third, the QRS vector changes as AMI progresses over

time. In the 17% of our patients who exhibited concordant ST-segment elevation during LBBB at randomization, concordant ST-segment elevation became discordant within 60 min as the net QRS complex became negative. Thus, in some LBBB patients who presented with a more evolved AMI, the concordant ST-segment elevation pattern may have evolved into a discordant ST-segment elevation pattern. In the American College of Cardiology/American Heart Association guidelines (1), the combination of bundle branch block and a history suggestive of AMI is listed as a class 1 indication for fibrinolytic therapy. This recommendation is based on the findings of placebo-controlled fibrinolytic trials (9 –11), which showed that patients with bundle branch block on the randomization ECG had high mortality and benefited significantly from fibrinolytic therapy, with a 21% reduction in mortality. However, these trials did not specify whether the ECG showed RBBB or

Figure 2. Thirty-day mortality in 92 patients with ST-segment changes during left bundle branch block (LBBB) versus matched control patients, and in 208 patients without ST-segment changes during LBBB versus matched control patients.

36

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

Table 5. Multivariable Logistic Regression Models for Predictors of 30-Day Mortality Predictors

OR

95% CI

p Value

GUSTO risk score (every 10-U rise) ST-segment changes* during LBBB (A) LBBB with ST-segment changes versus LBBB without ST-segment changes (B) LBBB with ST-segment changes versus normal intraventricular conduction (C) LBBB without ST-segment changes versus normal intraventricular conduction

2.18

2.09–2.28

⬍0.001 0.007

2.65

1.31–5.38

1.37

0.78–2.42

0.52

0.33–0.80

The association between the three groups (ST-segment changes* present during left bundle branch block (LBBB), ST-segment changes* absent during LBBB, and normal intraventricular conduction) was similar when the recruitment region and the time from symptom onset were also entered into the model: (A) odds ratio (OR) 2.70, 95% confidence interval (CI) 1.33–5.49; (B) OR 1.38, 95% CI 0.79 –2.44; (C) OR 0.51, 95% CI 0.33– 0.80. The results were also similar when the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO) risk score in the model was replaced by age, sex, pulse, systolic blood pressure, and Killip class: (A) OR 2.73, 95% CI 1.33–5.62; (B) OR 1.44, 95% CI 0.81–2.54; (C) OR 0.53, 95% CI 0.33– 0.83. The results were also similar when the analysis was repeated excluding patients who had a normal enzyme level and survived for ⬎24 h: (A) OR 2.25, 95% CI 1.08 – 4.67; (B) OR 1.36, 95% CI 0.77–2.41; (C) OR 0.61, 95% CI 0.38 – 0.97. *ST-segment changes were defined as concordant ST-segment elevation measuring ⱖ1 mm or ST-segment depression measuring ⱖ1 mm in electrocardiographic leads V1 to V3 during LBBB.

LBBB, whether the conduction abnormalities were new, or whether there were associated ST-segment changes. STsegment elevation can be detected in patients with RBBB, but in those with LBBB, ST-segment changes reflecting transmural ischemia may be obscured by repolarization changes. There are, therefore, no data available from randomized clinical trials that specifically demonstrate a mortality reduction with fibrinolytic therapy in patients with LBBB. In the HERO-2 trial, which specified prolonged ischemic chest discomfort lasting ⬎30 min and presumed-new LBBB as inclusion criteria, enzymatically confirmed cases of AMI were about 10% less frequent in patients with LBBB than in matched control patients. Some of these patients with LBBB who did not have enzymatically confirmed

AMI might have had nonischemic chest discomfort or prolonged ischemia without AMI. When patients have concordant ST-segment elevation or lead V1 to V3 ST-segment depression during LBBB, these findings are highly specific (97%) for the diagnosis of enzymatically confirmed AMI. The use of early biomarkers such as myoglobin (12,13) and heart-type fatty-acid binding protein (14) in patients without ST-segment changes may identify a group of patients in whom subsequent confirmation of AMI by cardiac markers is unlikely and in whom fibrinolytic therapy may not prove beneficial. In contrast with unselected HERO-2 patients with normal intraventricular conduction, the patients with LBBB were older and had a greater prevalence of previous AMI and other cardiovascular risk factors, a higher pulse rate, and

Figure 3. Nonadjusted 30-day survival curves in 92 patients with either concordant ST-segment elevation or lead V1 to V3 ST-segment depression during left bundle branch block (LBBB), 208 patients without concordant ST-segment elevation or lead V1 to V3 ST-segment depression during LBBB, and 15,340 patients with normal intraventricular conduction (p ⫽ 0.029 for normal conduction vs. LBBB without ST-segment changes; p ⬍ 0.001 for normal conduction vs. LBBB with ST-segment changes; p ⫽ 0.07 for LBBB without ST-segment changes vs. LBBB with ST-segment changes).

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

37

Figure 4. Adjusted 30-day survival curves in 92 patients with and 208 patients without concordant ST-segment elevation or lead V1 to V3 ST-segment depression during left bundle branch block (LBBB), and in 15,340 patients with normal intraventricular conduction (p ⫽ 0.004 for normal conduction vs. LBBB without ST-segment changes; p ⫽ 0.08 for normal conduction vs. LBBB with ST-segment changes; p ⫽ 0.002 for LBBB without ST-segment changes vs. LBBB with ST-segment changes).

a higher Killip classification at randomization. Although their 30-day mortality rate was higher than that of unselected patients with ST-segment elevation and normal conduction (16% vs. 9.1%, p ⬍ 0.001), it was lower than that of matched control patients (16% vs. 22.7%, p ⫽ 0.027). In patients with LBBB, concordant ST-segment elevation or lead V1 to V3 ST-segment depression during LBBB independently predicted the risk of 30-day mortality, which was 2.65 times higher in patients with associated ST-segment changes than in those without these features, after adjustment for other prognostic factors (Table 5). These ST-segment changes may have represented large shifts from the original baseline ST-segment level during LBBB (3,15,16), reflecting more severe ischemia. Although some patients with LBBB and no STsegment changes did not have enzymatic confirmation of AMI, this cannot fully explain their better outcome because the results were similar when multivariable analysis was repeated after excluding patients with normal enzyme levels who survived for ⬎24 h. The better 30-day prognosis in these patients with LBBB and no ST-segment changes may be at least partly explained by smaller infarct sizes as reflected by lower peak enzyme levels, and/or by the possibility that their infarction was more characteristic of non–ST-segment elevation AMI than ST-segment elevation AMI. In the North American cohort of patients with LBBB randomized into the GUSTO-I trial, 30-day mortality was identical in the 131 patients with LBBB and control patients matched for age and Killip class (17), but further subgroup analysis of the LBBB patients was not performed. Both LBBB and AMI are more common in elderly patients, and, with increasing longevity, the diagnostic and

management challenges posed by patients presenting with LBBB accompanying prolonged ischemic chest discomfort will arise more frequently. In the HERO-2 trial, LBBB at the time of randomization was presumed to be new unless the patient had a previous record of LBBB. Our findings suggest that the presence of LBBB (even when judged clinically to be new) in a patient with symptoms compatible with AMI should not in itself be regarded as an independent marker of higher risk. Instead, patients with LBBB may be risk-stratified by the presence or absence of concordant ST-segment elevation or lead V1 to V3 ST-segment depression. Although patients with ST-segment changes during LBBB had a higher 30-day mortality rate than those without ST-segment changes during LBBB, the latter subgroup of LBBB patients actually had a lower adjusted 30-day mortality rate than the majority of patients with ST-segment elevation AMI and normal intraventricular conduction. These lower-risk patients with LBBB and no associated ST-segment changes may have had no AMI, a smaller AMI, or an event more characteristic of non–STsegment elevation AMI with normal intraventricular conduction. Further studies are needed to test different management strategies in patients presenting with symptoms compatible with AMI accompanied by LBBB with or without ST-segment changes.

Acknowledgments The authors gratefully acknowledge the statistical assistance of Teena West, MSc, and the editorial assistance of Anna Breckon, ELS.

38

Wong et al. ST-Segment Changes During LBBB Predict Outcomes

Reprint requests and correspondence: Honorary Professor Harvey White, Cardiology Department, Green Lane Cardiovascular Service, Auckland City Hospital, Private Bag 92024, Auckland 1030, New Zealand. E-mail: [email protected].

REFERENCES 1. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST-segment elevation myocardial infarction— executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of Patients With Acute Myocardial Infarction). Circulation 2004;110:588 – 636. 2. Van de Werf F, Ardissino D, Betriu A, et al. Management of acute myocardial infarction in patients presenting with ST-segment elevation: the Task Force on the Management of Acute Myocardial Infarction of the European Society of Cardiology. Eur Heart J 2003;24:28 – 66. 3. Schamroth L. Myocardial infarction associated with left bundle branch block. In: Schamroth L, editor. The 12-Lead Electrocardiogram. Cambridge, MA: Blackwell Scientific, 1989:193–201. 4. Sgarbossa EB, Pinski SL, Barbagelata A, et al. Electrocardiographic diagnosis of evolving acute myocardial infarction in the presence of left bundle-branch block. N Engl J Med 1996;334:481–7. 5. Kontos MC, McQueen RH, Jesse RL, Tatum JL, Ornato JP. Can myocardial infarction be rapidly identified in emergency department patients who have left bundle-branch block? Ann Emerg Med 2001; 37:431– 8. 6. Gula LJ, Dick A, Massel D. Diagnosing acute myocardial infarction in the setting of left bundle branch block: prevalence and observer variability from a large community study. Coron Artery Dis 2003;14:387–93. 7. The Hirulog and Early Reperfusion or Occlusion (HERO)-2 Trial Investigators. Thrombin-specific anticoagulation with bivalirudin versus heparin in patients receiving fibrinolytic therapy for acute myocardial infarction: the HERO-2 randomised trial. Lancet 2001;358: 1855– 63.

JACC Vol. 46, No. 1, 2005 July 5, 2005:29–38 8. Califf RM, Woodlief LH, Harrell FE, Jr., et al. Selection of thrombolytic therapy for individual patients: development of a clinical model. Am Heart J 1997;133:630 –9. 9. Gruppo Italiano per lo Studio della Streptochinasi nell’Infarto Miocardico (GISSI) Investigators. Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986;i:397– 402. 10. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988;ii:349 – 60. 11. Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet 1994;343:311–22. 12. White HD, Van de Werf FJJ. Thrombolysis for acute myocardial infarction. Circulation 1998;97:1632– 46. 13. Newby LK, Storrow AB, Gibler WB, et al. Bedside multimarker testing for risk stratification in chest pain units: the Chest Pain Evaluation by Creatine Kinase-MB, Myoglobin, and Troponin I (CHECKMATE) study. Circulation 2001;103:1832–7. 14. Okamoto F, Sohmiya K, Ohkaru Y, et al. Human heart-type cytoplasmic fatty acid-binding protein (H-FABP) for the diagnosis of acute myocardial infarction: clinical evaluation of H-FABP in comparison with myoglobin and creatine kinase isoenzyme MB. Clin Chem Lab Med 2000;38:231– 8. 15. Stark KS, Krucoff MW, Schryver B, Kent KM. Quantification of ST-segment changes during coronary angioplasty in patients with left bundle branch block. Am J Cardiol 1991;67:1219 –22. 16. Cannon A, Freedman SB, Bailey BP, Bernstein L. ST-segment changes during transmural myocardial ischemia in chronic left bundle branch block. Am J Cardiol 1989;64:1216 –7. 17. Sgarbossa EB, Pinski SL, Topol EJ, et al. Acute myocardial infarction and complete bundle branch block at hospital admission: clinical characteristics and outcome in the thrombolytic era. J Am Coll Cardiol 1998;31:105–10.