Sgarbossa criteria for myocardial infarction in left bundle branch block

Clinical background

Left bundle branch block shifts repolarisation and means that the usual ST elevation criteria for acute myocardial infarction cannot be applied. Historically, "new or presumed new left bundle branch block" with chest pain was treated as a STEMI equivalent, but a systematic review and meta-analysis of 51 studies shows that the age of the bundle branch block has no diagnostic value: the likelihood ratio for new versus old LBBB was 1.30 (95% CI 0.75 to 1.85), with a confidence interval crossing 1.0 [5]. The decision therefore cannot be based on chronology but must rest on specific ECG patterns.

The Sgarbossa criteria were developed to meet this need: to identify, with high specificity, patients with left bundle branch block (or a ventricular paced rhythm) who have an acute coronary occlusion and who should therefore proceed to emergency reperfusion therapy. The score is not a screening tool but a rule set for distinguishing occlusion from the background changes that the bundle branch block itself produces.

Applying the Sgarbossa criteria

Three ECG criteria are assessed, each with an assigned point value:

$$\text{Score} = 5 \cdot [\text{concordant STE} \geq 1,\text{mm}] + 3 \cdot [\text{concordant STD} \geq 1,\text{mm in V1--V3}] + 2 \cdot [\text{discordant STE} \geq 5,\text{mm}]$$

The square brackets denote an indicator variable: 1 if the criterion is met, 0 if not. Concordant means that the ST deviation is in the same direction as the QRS complex. Discordant means that the ST deviation is in the opposite direction. A total score ≥3 is considered diagnostic of acute myocardial infarction.

The derivation study drew on patients in the GUSTO-1 trial, a large thrombolysis trial with 26,003 North American patients. Of these, 131 (0.5%) had both acute myocardial infarction and left bundle branch block. Their ECGs were compared blindly with ECGs from control patients with chronic coronary artery disease and left bundle branch block. The three criteria were identified by multivariable analysis and were then validated in an independent cohort of patients with acute chest pain and left bundle branch block [1]. The scoring system (0 to 10 points) gave a high specificity for the diagnosis of acute myocardial infarction.

Interpretation in practice

The low sensitivity of the original criteria (see below) means that a negative result cannot be used to exclude occlusion reliably. A positive result, on the other hand, is highly specific and warrants immediate action.

Validation and performance

In several external validations, the original Sgarbossa criteria have shown high specificity but low sensitivity. In a retrospective case-control study at two American emergency departments and a regional referral centre, with a total of 45 cases of acute coronary occlusion and 249 controls, the original weighted criteria had a sensitivity of 49% and a specificity of 100% [2]. The unweighted criteria (in which each criterion counts as positive irrespective of its point value) gave a sensitivity of 56% at an unchanged specificity of 94% [2].

To improve sensitivity, Smith and colleagues proposed a modification of the third criterion: instead of an absolute cut-off of 5 mm discordant ST elevation, a ratio is used in which the ST elevation must constitute at least 25% of the amplitude of the S wave (ST elevation/S wave ≤ −0.25). This modified rule increased sensitivity to 80% in the same validation cohort without significant loss of specificity (99%) [2]. A systematic review and meta-analysis of 51 studies, published in 2025, summarised the performance of the modified Sgarbossa criteria for angiographically confirmed occlusion myocardial infarction as a sensitivity of 83.6% (95% CI 55.4 to 95.5%) and a specificity of 92.6% (95% CI 78.9 to 97.7%), with a positive likelihood ratio of 11.34 [5].

A Swedish study from 2024 included 623 patients with chest pain and left bundle branch block or ventricular paced rhythm at five emergency departments [4]. Of these, 15 (2.4%) had an occlusion myocardial infarction. In this low-prevalence population the original unweighted Sgarbossa criteria performed with a sensitivity of 26.7% and a specificity of 86.2%, whereas the modified criteria reached a sensitivity of 60.0% at a specificity of 86.0% [4]. The positive predictive value was low (4.6% and 9.4% respectively), reflecting the low prevalence of occlusion in a general emergency department population, while the negative predictive value exceeded 98% for all criteria [4].

In a larger retrospective case-control study of the OMI (occlusion myocardial infarction) paradigm, structured ECG interpretation was compared with the usual STEMI criteria in 808 patients with suspected acute coronary syndrome [3]. The modified Sgarbossa criteria were included as one of eight predefined ECG findings for patients with left bundle branch block or paced rhythm. Overall OMI interpretation had a sensitivity of 86% and a specificity of 91%, compared with 41% and 94% for the STEMI criteria [3]. Patients with occlusion who did not meet the STEMI criteria had a similar infarct size and mortality to STEMI-positive patients, but a longer delay to angiography [3].

Limitations

The original Sgarbossa criteria have low sensitivity, particularly in populations with a low prevalence of occlusion. A negative result does not exclude acute coronary occlusion, and patients with persistent symptoms and clinical suspicion should be managed irrespective of the score.

The criteria were derived for left bundle branch block but are also applied to ventricular paced rhythm, in which the pattern of concordant and discordant ST deviation resembles that in bundle branch block. Their validity is, however, less well documented for paced rhythm than for left bundle branch block.

The absolute 5 mm threshold for discordant ST elevation is the weakest part of the original criteria. It misses cases in which the ST elevation is substantial relative to the QRS amplitude but falls below 5 mm, particularly in patients with a low R or S amplitude. The modified rule using a ratio of ≥0.25 (ST elevation/S wave) addresses this and is often recommended as a complement to, or a replacement for, the third criterion [2, 5].

In a general emergency department population, where the prevalence of occlusion myocardial infarction in left bundle branch block is low (around 2 to 3%), the positive predictive value is low even at high specificity [4]. This means that a positive result should be confirmed with prompt angiography or biomarkers, but should not delay immediate referral to the catheterisation laboratory when the score is diagnostic and the clinical picture supports the suspicion.

Prospective validation in a low-prevalence population

The Swedish validation study by Lindow and colleagues, with data from five emergency departments, is to date the largest prospective evaluation of the Sgarbossa criteria in a Nordic low-prevalence population [4]. The results confirm that ECG criteria alone are not sufficient to predict occlusion reliably in a general emergency department population, and that both the original and the modified criteria should be combined with clinical assessment, troponins and repeat ECG recording. The study also compared the newer Barcelona criteria and the Selvester criteria, without any method clearly outperforming the others in this setting [4].

References

1. Sgarbossa EB, Pinski SL, Barbagelata A et al. Electrocardiographic diagnosis of evolving acute myocardial infarction in the presence of left bundle-branch block. GUSTO-1 Investigators. N Engl J Med 1996. PMID: 8559200
2. Meyers HP, Limkakeng AT Jr, Jaffa EJ et al. Validation of the modified Sgarbossa criteria for acute coronary occlusion in the setting of left bundle branch block: A retrospective case-control study. Am Heart J 2015. PMID: 26678648
3. Pendell Meyers H, Bracey A, Lee D et al. Accuracy of OMI ECG findings versus STEMI criteria for diagnosis of acute coronary occlusion myocardial infarction. Int J Cardiol Heart Vasc 2021. PMID: 33912650
4. Lindow T, Mokhtari A, Nyström A et al. Comparison of diagnostic accuracy of current left bundle branch block and ventricular pacing ECG criteria for detection of occlusion myocardial infarction. Int J Cardiol 2024. PMID: 37931659
5. Alencar JN, Lima GWF, Geraldo HACS et al. Accuracy of Left Bundle Branch Block Chronology and Electrocardiography Criteria for Acute Myocardial Infarction Diagnosis: A Systematic Review and Meta-analysis. Arq Bras Cardiol 2025. PMID: 41191702