Left bundle branch block with acute thrombotic occlusion is associated with increased myocardial jeopardy score and poor clinical outcomes in primary percutaneous coronary intervention activations
- Adam J Brown,
- Stephen P Hoole,
- Liam M McCormick,
- Matt Malone-Lee,
- Paul J Cacciottolo,
- Peter M Schofield,
- Nick E J West
- Correspondence to Dr Nick E J West, Department of Interventional Cardiology, Papworth Hospital, Papworth Everard, Cambridge, Cambridgeshire CB23 3RE, UK;
- Received 15 October 2012
- Revised 21 February 2013
- Accepted 25 February 2012
- Published Online First 20 March 2013
Objective To assess the utility of left bundle branch block (LBBB) as an activation criterion for primary percutaneous coronary intervention (PPCI).
Design Retrospective observational cohort study.
Setting Single UK heart attack centre.
Patients Consecutive patients referred for PPCI September 2008–December 2011 (n=2192).
Interventions Demographic and outcome data were obtained by review of case notes, angiograms and interrogation of local/national databases.
Main outcome measures Angiographic culprit lesion assessment defined appropriate and inappropriate activations. Patients outcomes were assessed by Major adverse cardiac events (MACE), defined as a composite of mortality and unplanned revascularisation at 1-year.
Results LBBB-activation occurred in 120 patients (5.5%), of whom 21 (17.5%) had acute coronary occlusion angiographically, and were adjudicated appropriately. Compared with appropriate activations for ST segment elevation, appropriate LBBB-activations were older (71.0±9.6 vs 64.2±12.4 years, p=0.01) and more likely to be in cardiogenic shock (19.0% vs 4.3%, p<0.01). Extent of disease quantified by the SYNTAX score did not differ (median 21.5, IQR 11.0–27.0 vs 19, 11.0–25.5, p=0.66), but amount of myocardium at-risk was higher in appropriate LBBB-activations (culprit jeopardy score median 4, IQR 2–6 vs 2, 2–4, p=0.02). Final diagnoses for LBBB-activations were acute coronary syndrome (39.2%), non-acute coronary syndrome cardiac chest pain (33.3%) and non-cardiac chest pain (27.5%). In appropriate LBBB-activations 1-year mortality and MACE were higher (23.8% vs 6.6%, p=0.002 and 28.6% vs 10.5%, p=0.007, respectively).
Conclusions Our data suggests that despite its poor specificity for identifying acute coronary occlusion, LBBB should at the present time remain an activation criterion for PPCI and such patients should continue to be transferred to heart attack centres for assessment and treatment.
The activation criterion for primary percutaneous coronary intervention (PPCI) in UK networks requires chest pain of ischaemic aetiology combined with either ST segment elevation (STE) or new-onset left bundle branch block (LBBB) on 12-lead ECG.1 However identification of LBBB on ECG is not specific for myocardial ischaemia and may be associated with hypertension, chronic ischaemic heart disease or underlying structural heart disease including cardiomyopathy.2 Furthermore determining the chronicity of LBBB prehospital is frequently challenging.
Current international guidelines suggest that patients who are thought to have new-onset LBBB, in the context of an acute coronary syndrome (ACS) should be considered for urgent revascularisation.3 ,4 However this recommendation is based on historical studies conducted in the era before PPCI.5 Furthermore, there are concerns that adopting this treatment strategy may expose many patients to unnecessary interventional procedures while incurring additional costs to healthcare providers.6
Our aim was to assess the prevalence of appropriate LBBB referrals within an unselected PPCI population and to examine the outcome of these patients, compared with inappropriate LBBB and appropriate STE PPCI referrals.
Consecutive patients presenting for PPCI at a single tertiary referral centre between September 2008 and December 2011 were included in the analysis (n=2192). Data were obtained from a dedicated cardiology database, CVIS V. R6.1 (Philips Medical Systems, Netherlands), containing information on each PPCI activation including details of prehospital ECG, baseline demographics and procedural details. All data were entered prospectively immediately after angiogram or PPCI procedures by the performing cardiologist, with additional demographics entered into the database following the patient's arrival in the High-Dependency Unit. Case note review was undertaken on all LBBB-activations and on those STE-activations where data was incomplete.
The Sgarbossa criteria were applied as previously described on all LBBB activation ECGs, where STE≥1 mm concordant with the QRS in any lead (5 points), ST depression of ≥1 mm in V1–3 (3 points) or ST elevation of ≥5 mm discordant with QRS in any lead (1 point) were scored.7 A score of ≥3 was considered to be positive for myocardial infarction. Two interventional cardiologists blinded to ECG and patient outcomes reviewed angiographic images to adjudicate if there was an acute thrombotic vessel occlusion and hence appropriate PPCI activation. The SYNTAX score was calculated using SYNTAX Score Calculator 2.11 (http://www.syntaxscore.com) by visual angiographic assessment of each lesion as previously described.8 ,9 The culprit jeopardy score was determined by allocating 2 points for each of six segments (left anterior descending, largest diagonal, largest septal perforator, first obtuse marginal, circumflex and posterior descending artery) adjacent to and downstream of the culprit lesion.10 The maximum achievable score was 12 for a left main stem culprit lesion, in a left dominant circulation. Angiographic data on all appropriate LBBB-activations were compared against a cohort of consecutive appropriate STE-activations that presented from September 2008 until June 2010 (n=674). Outcome data was obtained from our hospital database, which has a direct link to the national spine through the unique National Health Service number. Data is continually updated to the national spine by general practitioners/hospitals, ensuring accurate demographic information. Discharge diagnosis was obtained from case note review. The primary outcome of MACE was defined as a composite of mortality and unplanned revascularisation at 1-year and was completed up to 30 January 2013.
Data was collected as part of a mandatory national cardiac audit. Any fields with patient identifiable data were removed prior to analysis. The project was internally reviewed and received institutional approval.
Continuous variables are presented as mean (SD) and categorical data as counts (%) or median (IQR). Analysis of continuous data was performed using unpaired Student's t test, while the χ2 or Fisher's exact test was used for categorical data where appropriate. Comparison of non-parametric, categorical ranked data was performed using the Mann-Whitney test. Survival was estimated by Kaplan-Meier analysis. A two-sided probability level of <0.05 was considered statistically significant. All calculations were performed using Prism V.5.0a for Apple, GraphPad Software, USA.
There were 120 LBBB-activations (5.5%) in our institution over the study period and 2072 STE-activations (94.5%). Comparing all LBBB-activations with all STE-activations demonstrated that LBBB-activations were older (70.7±12.2 vs 64. 6±13.4, p<0.001), female (76.2% vs 65.8%, p=0.01) and were not current smokers (13.3% vs 23.6%, p=0.01). Other baseline demographics of the whole population are presented in table 1.
Of the total LBBB-activation cohort, 100 patients (83.3%) proceeded to emergent coronary angiography after clinical evaluation. An acute thrombotic lesion was confirmed and treated by PPCI in only 21 patients (17.5%), compared with 1424 (65.0%) in the STE-activation cohort. When a thrombotic culprit lesion was identified the presentation was adjudicated appropriate. Appropriate LBBB-activations were older (71.0±9.6 vs 64.2±12.4, p=0.01), more likely to have sustained a previous myocardial infarction (33.3% vs 11.7%, p<0.01) and to present in cardiogenic shock (19.0% vs 4.3%, p<0.01) when compared with appropriate STE-activations. There were no other significant differences in demographics between those who presented as an inappropriate or appropriate LBBB-activation (table 2). The burden of disease, as quantified by the SYNTAX score, did not differentiate between appropriate LBBB-activations and STE-activations (median 21.5, IQR 11.0–27.0 vs 19, IQR 11.0–25.5, respectively; p=0.66), although the amount of myocardium at risk was significantly higher in the appropriate LBBB-activations as assessed by the culprit jeopardy score (median 4, IQR 2–6 vs 2, IQR 2–4 respectively; p=0.02). Infarct-related artery did not differ between groups (table 3). Analysis of admission ECG determined that when the Sgarbossa score was ≥3, the specificity to predict an acute thrombotic lesion was 88.6% with a sensitivity of 16.7%.
The final adjudicated diagnoses for LBBB-activations was ACS (39.2%), non-ACS cardiac chest pain (33.3%) or non-cardiac chest pain (27.5%) (table 4). In the appropriate LBBB-activations 1-year mortality was numerically greater compared with inappropriate LBBB-activations, although this did not achieve statistical significance (23.8% vs 11.3%, p=0.11). However, there was a significant difference in MACE (28.6% vs 11.3%, p=0.03) (figure 1) driven by increased unplanned revascularisation in the appropriate LBBB-activation cohort. Comparing appropriate LBBB-activations with appropriate STE-activations, there was a significant difference in mortality (23.8% vs 6.6%, p=0.002) and MACE (28.6% vs 10.5%, p=0.007) at 1 year (figure 2).
We demonstrate that the majority of LBBB PPCI activations are inappropriate and do not require coronary intervention for an acutely-occluded coronary artery. However, appropriate LBBB-activations, with a thrombotic lesion confirmed on angiography, are high risk and have a significantly worse outcome compared with STE-activations. Additionally, we found that the Sgarbossa criteria improved the specificity of LBBB to identify an acute thrombotic coronary occlusion, although other clinical factors including patient demographics and burden of coronary disease were not helpful.
Our data agree with the literature: the published incidence of LBBB-activations for PPCI is between 3.5% and 9%11–13 and the incidence of LBBB-activations we observed was similar. Previously published series on LBBB-activations have found that this cohort of patients tends to be older and more likely female14 and our data would again support this. In our cohort, 17.5% of LBBB activations had a thrombotic coronary lesion confirmed at angiography, comparable with the three angiographic studies of LBBB revascularisation with rates between 7% and 14%.11 ,15 ,16 Although outcome data of acute myocardial infarction (MI) presenting with LBBB has been well described,12 ,17 few studies are contemporary.11 ,18 Our data provide the largest experience of patients presenting with LBBB to a dedicated PPCI service.
We have observed no significant differences in the baseline demographics of patients that presented as either an appropriate or inappropriate LBBB-activation, highlighting that identification of appropriate LBBB-activations on patient-specific factors is challenging. However our study did reinforce that the use of the Sgarbossa criteria with a score of ≥3 yielded a specificity of 86% and thus may aid clinical decision making, possibly in conjunction with other non-invasive diagnostic modalities.
Interestingly our data show that patients presenting with appropriate LBBB-activations were more likely to be in cardiogenic shock and exhibit a worse outcome than appropriate STE-activations, despite having a similar burden of coronary disease as quantified by the SYNTAX score and number of diseased vessels. This is likely due to the location of the culprit lesion, as appropriate LBBB-activations are frequently attributable to anteroseptal infarcts.19 Our data showing that appropriate LBBB-activations have a high culprit jeopardy score reinforces this concept, with large volumes of myocardium at risk from ischaemia, leading to cardiogenic shock. Although the SYNTAX scores carries heavy weighting for the anatomical location of a lesion, the overall score is influenced by non-culprit vessel disease which may not have an acute effect on ventricular performance. Our results emphasise that clinicians must use caution in using angiographic tools such as the SYNTAX score acutely for prognostication.20
Although LBBB can occur as a direct result of myocardial infarction, it is more frequently a manifestation of structural or degenerative cardiac pathology.21 Furthermore determining the chronicity of LBBB is challenging. Initial medical contact may err on the side of caution and triage the patient to PPCI when there is doubt about the diagnosis of Acute myocardial infarction (AMI). Although this practice is safe it inevitably leads to unnecessary PPCI activations, incurring additional healthcare costs. Nevertheless, it is clear from our data that the majority of patients presenting with LBBB were diagnosed with a new cardiological condition (72.5%) and are likely to benefit from specialist input. Furthermore the higher mortality for those with appropriate LBBB-activations should only seek to encourage clinicians to refer these patients for the consideration of revascularisation to a specialist centre.
A recent review has proposed a new diagnostic algorithm for assessment of patients suspected of having myocardial infarction with LBBB.22 The authors propose using the Sgarbossa criteria alongside bedside echocardiography and biomarkers to identify those with ST-segment myocardial infarction (STEMI). Although this may improve the accuracy of AMI diagnosis in LBBB in hospitalised patients, translating complex electrographic scoring systems into prehospital care would be challenging and may result in false reassurance. Furthermore the additional diagnostic investigations proposed, if applied to all LBBB-activations, are likely to result in delayed revascularisation for those with a thrombotic culprit lesion. Our data already demonstrates a high incidence of shock and worse outcomes for appropriate LBBB-activations using current PPCI pathways. Potential new algorithms should therefore include a pathway to ensure high-risk cases proceed directly to coronary angiography.
Ultimately whether the proposed algorithm will improve patient care and be adopted by international guidelines can only be answered by future prospective studies. Until this time we recommend that all LBBB-activations be urgently transferred to a heart attack centre where specialist triage methods should be developed to ensure that those most likely to benefit undergo emergent angiography and revascularisation.
Our data were collected retrospectively from a single centre and must be interpreted in this light. Our institution accepts direct referrals from prehospital practitioners and local emergency departments, without prior ECG-triage by a cardiologist. This will alter the cohort of patients with LBBB and could explain the poor specificity of LBBB to diagnose AMI in our study. Twenty LBBB-activations did not proceed to angiography and were classified as inappropriate activations, although this was not corroborated by angiography. Finally, this is an observational study and while we recognise that bedside echocardiography and biomarkers may be useful in improving the care of patients presenting with LBBB and chest pain, we were unable to assess the utility of these diagnostic modalities prospectively.
Less than a fifth of patients presenting with chest pain and LBBB had a thrombotic coronary occlusion requiring PPCI. Nevertheless, appropriate LBBB-activations have a significantly worse prognosis and merit urgent referral. Until methods to reliably identify AMI requiring PPCI in patients presenting with LBBB are available, all such patients should be transferred to heart attack centres for emergent assessment and treatment.
Contributors Planning: AJB, PMS, NEJW. Conduct/Reporting: AJB, SPH, LMM, MML, PJC. Manuscript preparation: AJB, SPH, NEJW. Manuscript review: PMS, NEJW. Guarantor: NEJW.
Competing interests None.
Ethics approval Interventional Cardiology Audit - Ref513.
Provenance and peer review Not commissioned; externally peer reviewed.