Diagnosis of ST-Elevation Myocardial Infarction in the Presence of Left Bundle Branch Block With the ST-Elevation to S-Wave Ratio in a Modified Sgarbossa Rule

Study objective

Sgarbossa's rule, proposed for the diagnosis of acute myocardial infarction in the presence of left bundle branch block, has had suboptimal diagnostic utility. We hypothesize that a revised rule, in which the third Sgarbossa component (excessively discordant ST-segment elevation as defined by ≥5 mm of ST-segment elevation in the setting of a negative QRS) is replaced by one defined proportionally by ST-segment elevation to S-wave depth (ST/S ratio), will have better diagnostic utility for ST-segment elevation myocardial infarction (STEMI) equivalent, using documented coronary occlusion on angiography as reference standard.

Methods

We collected admission ECGs for all patients with an acutely occluded coronary artery and left bundle branch block at 3 institutions. The ECGs of emergency department patients with chest pain or dyspnea and left bundle branch block, but without coronary occlusion, were used as controls. The R or S wave, whichever was most prominent, and ST segments, relative to the PR segment, were measured to the nearest 0.5 mm. The ST/S ratio was calculated for each lead that has both discordant ST deviation of greater than or equal to 1 mm and an R or S wave of opposite polarity; others were set to 0. The cut point for the most negative ST/S ratio with at least 90% specificity was determined. The revised rule is unweighted, requiring just 1 of 3 criteria. Diagnostic utilities of the original and revised Sgarbossa rules were computed and compared. McNemar's test was used to compare sensitivities and specificities.

Results

The study and control groups included 33 and 129 ECGs, respectively. The cut point selected for relative discordant ST-segment elevation was −0.25. Excessive absolute discordant ST-segment elevation of 5 mm was present in at least one lead in 30% of ECGs in patients with confirmed coronary occlusion versus 9% of the control group, whereas excessive relative discordant ST-segment elevation less than −0.25 was present in 79% versus 9%. Sensitivity of the revised rule in which ST-segment elevation with an ST/S ratio less than or equal to −0.25 replaces ST-segment elevation greater than or equal to 5 mm was significantly greater than either the weighted (P<.001) or unweighted (P=.008) Sgarbossa rule: 91% (95% confidence interval [CI] 76% to 98%) versus 52% (95% CI 34% to 69%) versus 67% (95% CI 48% to 82%). Specificity of the revised rule was lower than that of the weighted rule (P=.002) and similar to that of the unweighted rule (P=1.0): 90% (95% CI 83% to 95%) versus 98% (95% CI 93% to 100%) versus 90% (95% CI 83% to 95%). Positive and negative likelihood ratios for the revised rule were 9.0 (95% CI 8.0 to 10) and 0.1 (95% CI 0.03 to 0.3). The revised rule was significantly more accurate than both the weighted (16% difference; 95% CI 5% to 27%) and unweighted (12% difference; 95% CI 2% to 22%) Sgarbossa rules.

Conclusion

Replacement of the absolute ST-elevation measurement of greater than or equal to 5 mm in the third component of the Sgarbossa rule with an ST/S ratio less than −0.25 greatly improves diagnostic utility of the rule for STEMI. An unweighted rule using this criterion resulted in excellent prediction for acute coronary occlusion.

Introduction

Timely and accurate identification of acute coronary occlusion in the setting of ischemic symptoms is critical to initiating urgent angiography and appropriate reperfusion therapy. Although the increase or decrease of cardiac biomarker levels is essential to the diagnosis of acute myocardial infarction, positive biomarker results alone do not differentiate ST-elevation myocardial infarction (STEMI) from non-STEMI. ST elevation on the ECG is the primary indication for emergency reperfusion therapy; however, identification of STEMI in the setting of left bundle branch block remains challenging.¹

In the setting of left bundle branch block, ST-segment elevation or ST-segment depression commonly occurs in the absence of acute myocardial infarction and is predictable in that the ST-segment and T-wave abnormalities are normally “discordant” to (in the opposite direction of) the majority of the QRS (Figure 1). “Concordant” is the term used when the ST segment or T wave is in the same direction as the QRS and is not normally observed in baseline (normal) left bundle branch block.

Sgarbossa et al proposed requiring at least 3 points from the following criteria for the diagnosis of acute myocardial infarction in the presence of left bundle branch block: (1) concordant ST-segment elevation of 1 mm (0.1 mV) in at least 1 lead (5 points), (2) concordant ST-segment depression of at least 1 mm in leads V1 to V3 (3 points), or (3) excessively discordant ST-segment elevation, defined as greater than or equal to 5 mm of ST-segment elevation when the QRS result is negative (2 points)² (Figure 2). There have been many evaluations of Sgarbossa's criteria, with variable methodologies and patient populations.3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 In a systematic review, although specificity for greater than or equal to 3 Sgarbossa points was high (98%), sensitivity was only 20%.¹⁷ For a score greater than or equal to 2 (ie, the unweighted rule), the sensitivities in the studies ranged from 20% to 79%, and specificities ranged from 61% to 100%.

Two main issues may contribute to the low sensitivity of Sgarbossa's rule. First, all validating studies cited above used a reference standard of creatine kinase (CK) (with or without MB fraction) for acute myocardial infarction, not coronary occlusion by angiography, meaning non-STEMI (emergency reperfusion therapy unnecessary) and STEMI (emergency reperfusion required) were included in the acute myocardial infarction group. Second, anterior STEMI is most often diagnosed by ST-segment elevation in leads V1 to V4; however, in left bundle branch block, these leads normally already have discordant ST-segment elevation. Therefore, some means of assessment of excessive anterior ST-segment elevation is necessary to diagnose most anterior STEMI. Specifically, Sgarbossa's rule uses an absolute 5-mm cutoff for discordant ST-segment elevation when an ST-segment elevation proportional to the preceding QRS or S wave may be more useful. We sought to evaluate the performance of the Sgarbossa rule in patients with left bundle branch block and angiographic evidence of coronary occlusion. We hypothesized that changing the third component of the Sgarbossa rule to a proportional rule would improve its sensitivity and specificity.