A total of 23 patients that underwent CA (40%) had no ECG confounders (ECG is "otherwise unremarkable”). LMCA disease was found in only 3 of them (13%), LMEQ in 2 (9%) and three vessel disease in 5 (22%). The prevalence of three vessel disease, LMCA or LMEQ was 43%, not different from that of the whole group. A total of 16 patients (70%) underwent revascularization with 30% undergoing PCI and 43% CABG.

Most of the patients that underwent CA had positive T waves (53 patients). Among the 4 patients with negative T waves, one had three vessel disease, one LMCA and two nonsignificant coronary artery disease (Table 4).

Among the 57 patients that underwent CA, 25 (44%) had dynamic ECG changes and in 17 patients (30%) the pattern of diffuse ST segment depression with ST elevation in aVR was a chronic pattern. Previous ECG was unavailable in 15 patients (26%). There was no difference in the prevalence of LMCA (p=0.11), LMEQ (p=0.50) or LMCA/LMEQ/three vessel disease (p=0.41) among the groups. However, LMCA or LMEQ (p=0.18) tended to occur more often in the group with dynamic ECG changes (Table 5). Patients who had dynamic ECG changes underwent more often revascularization than patients with stable or unknown ECG pattern (p=0.09), with more patients in with dynamic ECG changes needed CABG (p=0.134).

Discussion

The LMCA supplies blood to a majority of the anterolateral and septal aspects of the left ventricle as it branches into the LAD and left circumflex arteries. A decrease in blood flow after LMCA narrowing affects a significant region of the heart, producing electrical disturbances visible by ECG due to myocardial ischemia/injury. Isolated ST segment elevation in lead aVR as well as diffuse ST segment depression with/without ST segment elevation in aVR has been described previously as patterns consistent with either LMCA or triple vessel stenosis.^{1, 7 and 8}

Isolated ST elevation in aVR and V₁

Gorgels et al were the first to report that ST elevation in lead aVR, together with ST depression in leads I, II and V4–V6 were predictive of LMCA or three vessel disease, especially when the total magnitude of ST changes exceeded 12 mm.⁹ Yamaji et al. reported that ST elevation in lead aVR (> 0.05 mV) that was greater than ST elevation in lead V₁ "distinguished” LMCA obstruction from LAD and right coronary artery obstruction with 81% and 88% sensitivity, 80% and 92% specificity and 81% and 90% accuracy, respectively.> 0.15 mV ST elevation in lead aVR predicted mortality with 75% sensitivity, 75% specificity and 75% accuracy in patients with ACS.¹⁰ The positive end of lead aVR points toward the patient's right shoulder, so if the LMCA is obstructed, the septal branches of the LAD would also be affected. It is believed that LMCA obstruction leads to ischemia/injury of the basal septum with a resultant right superior pointing injury vector causing ST segment elevation in lead aVR.¹⁰ Since lead aVR faces into the left ventricular cavity, the same electrical vector that causes ST segment depression in leads V₅ and V₆ (that are oriented towards the left ventricular apex) will be recorded by lead aVR as ST segment elevation.¹¹ These reciprocal changes may be seen in situations where the left ventricular end diastolic pressure is elevated. Gorgels et al. suggested that ischemia of the basal interventricular septum should cause ST elevation in leads aVR and V_1.¹² Thus, both proximal LAD obstruction and LMCA occlusion should cause ST segment elevation in leads aVR and V₁.¹²Kosuge et al. reported that ST elevation in aVR has a sensitivity of 78%, specificity of 86%, positive predictive value (PPV) of 57%, negative predictive value (NPV) of 95% and predictive accuracy of 84% for predicting LMCA or 3VD in a retrospective study of 310 patients with NSTE-ACS.⁴

Diffuse ST depression

Nikus et al. describe that the ECG pattern of widespread ST-segment depression and inverted T waves, maximally in leads V_4–5 is associated with LMCA disease, LMEQ coronary artery disease, or severe there vessel disease.¹³ In 50 patients with NSTE-ACS this pattern had a sensitivity of 100%, specificity of 57%, PPV of 24% and NPV of 100% for severe 3VD. While a sensitivity of 91%, specificity of 79%, PPV of 76% and NPV of 92% were reported for LMCA or LMEQ coronary artery disease. This pattern had a sensitivity of 93%, specificity of 100%, PPV of 100% and NPV of 92% for severe 3VD, LMCA or LMEQ coronary artery disease.¹³ In another study, Nikus and Sclarovsky describe "transient ST depression with negative T waves seen maximally in leads V₄ and V₅” as a pattern associated with significant LMCA/3VD. The pattern had a PPV of 100% and an NPV of 92% in a prospective study of 25 patients. ST elevation in aVR was explained as reciprocal changes to the ST depression in leads V₄ and V₅, which are electrically opposite to aVR.¹³Taglieri et al. describe the predictive value of isolated diffuse ST depression compared to ST depression plus ST elevation in lead aVR in identifying LMCA disease and in-hospital as well as one-year cardiovascular mortality in patients with NSTE-ACS. They concluded that only diffuse ST depression plus ST elevation in lead aVR was useful in identifying LMCA disease and mortality. Multivariable analysis showed an odds ratio (4.72; 95% CI 2.31 to 9.64; p < 0.001) for this combined pattern to predict LMCA obstruction.¹⁴ The same analysis was completed for prediction of mortality, providing a hazard ratio (1.52; 95% CI 1.44 to 3.64; p < 0.001), favoring the combined pattern over isolated diffuse ST deviation as well.¹⁴ Hanna and Glancy describe that during an episode of ischemic chest pain, a pattern of diffuse ST depression in eight or more leads combined with ST elevation in aVR and V₁ has a 75% predictive accuracy for LMCA or 3-VD.¹⁵

However, as abovementioned, diffuse ST depression may not even be the most common ECG pattern seen in patients presenting with LMCA obstruction. Atie et al reported that only 60% of the patients with NSTE-ACS due to LMCA obstruction had ST elevation in lead aVR and 67% had maximal ST depression in lead V₄.¹⁶ The average number of leads with ST-T changes was only 6.4, suggesting that in many patients there were less than 6 leads with ST depression.¹⁶ More recently, Taglieri et al reported that only 24 out of 57 patients (42%) with NSTE-ACS due to LMCA stenosis had ST elevation in lead aVR with ST depression in other leads.¹⁴ ST elevation, predominantly in the precordial leads is probably the most frequent ECG presentation in patients with complete occlusion of the LMCA, followed by right bundle branch block.^{5, 6,7 and 17}

The current recommendations by the AHA/ACCF/HRS for "resting ECGs that reveal ST-segment depression greater than 0.1mV in 8 or more body surface leads coupled with ST-segment elevation in aVR and/or V1 but are otherwise unremarkable” are that the automated interpretation should suggest "ischemia due to multivessel or left main coronary artery obstruction”.¹ Although in the text it is written that "It has been reported that in patients with angina at rest, ST-segment depression in 8 or more body surface ECG leads, combined with ST elevation in aVR and V₁, is associated with a 75% predictive accuracy of 3-vessel or left main stenosis”,¹ the recommendations do not mentioned symptoms during ECG recording. Obviously, currently we cannot add symptoms or risk factors to the computerized automated interpretation. Thus, the current recommendations suggest that the ECG should be interpreted independent of the clinical presentation.

Nikus et al. on behalf of an International Society of Holter and Non-Invasive Electrocardiology (ISHNE) working group recommended that the AHA/ACCF/HRS recommendation is supported by prior studies, "but needed minor modifications: 1.) The presence of left ventricular hypertrophy, left bundle branch block with QRS > 130 msec and/or tachycardia (rate > 100/min) invalidates these criteria. 2.) The number of required leads with ST depression should be reduced from 8 to 7. 3.) The threshold for ST depression in the V leads should be adjusted for age and gender. 4.) The threshold for ST elevation in aVR should be specified as 0.05 mV”.⁸ Nikus et al recommended that urgent CA should be considered if the patient is hemodynamically unstable or ischemic symptoms persist, but there are not enough data to recommend urgent CA based on the ECG findings alone.^5 and 8

The current study questions the accuracy of interpreting the ECG pattern of diffuse ST segment depression with ST elevation in aVR as a sign of diffuse three vessel disease or LMCA/LMEQ disease, using the above mentioned recommendations. Here we collected consecutive patients with ECGs showing diffuse ST depression and ST elevation in lead aVR. LMCA or LMEQ disease was found in only 23% of the patients that underwent CA. Thus, the positive predictive value of the ECG pattern for LMCA/LMEQ was significantly lower in our cohort than that reported in previous studies.¹³ In contrast to the previous studies, most of our patients were not diagnosed with ACS. Interestingly, the prevalence of LMCA/LMEQ was not higher among patients with presentation compatible with ACS as compared to all the 57 patients that underwent CA. It should be remembered that in about 30% of the patients with acute myocardial infarction, presentation can be atypical, especially in the elderly, diabetic patients and women.^2 and 18 Shortness of breath, hypotension, epigastric pain and arrhythmia, all can be manifestation of ACS are common in patients with cardiomyopathies, hypertensive heart disease and significant valvular disease that cause diffuse ST depression with ST elevation in aVR. Moreover, the prevalence of LMCA/LMEQ was not significantly higher among patients with dynamic ECG changes as compared to those with chronic pattern of diffuse ST depression and ST elevation in aVR. It should be remembered that the amount of ST depression in patients with repolarization abnormalities (LVH, cardiomyopathies, intraventricular conduction delay, etc) is dependent on the heart rate and the degree of QRS prolongation and therefore, may show dynamic changes. In the present study, presence of symptoms suggestive of ACS or dynamic ECG changes did not improve the association between the ECG pattern and the presence of LMCA, LMEQ or three vessel disease.

An important component of the recommendations is that the ECG should be "otherwise unremarkable”.¹There is no specification what is actually considered "unremarkable”. It is clear that repolarization changes secondary to LVH or left bundle branch block may cause diffuse ST depression with ST elevation in lead aVR. We did not include patients with left bundle branch block or QRS duration of > 130 msec in our cohort. Interestingly, after exclusion of the 14 patients with ECG signs of LVH, the prevalence of LMCA, LMEQ and/or three vessel disease did not significantly changed compared to the whole cohort. Even when we included only patients without ECG confounders, the percentage of patients with LMCA, LMEQ and/or three-vessel disease did not increase. Obviously, diffuse coronary artery disease may result in pathological Q waves, QRS widening and QRS axis changes. Moreover, hypertension, a major risk factor for coronary artery disease, may results in LVH. As these changes are considered "abnormal”, excluding high-risk patients with such abnormalities decreases the clinical applicability of the ECG sign by excluding a large number of high-risk patients.

Nikus et al emphasized the importance of the T wave polarity in leads V4-V5 in predicting LMCA/LMEQ disease. Patients with negative T waves had higher prevalence of LMCA/LMEQ and poorer outcome than patients with positive T waves.^13 and 19 Hence, the position manuscript by Nikus et al suggests that "sudden occlusion of the LMCA may present as widespread ST depression and inverted T waves with ST elevation in lead aVR.”⁵ We used strict criteria for the definition of T wave polarity. Investigators are using different definitions for the polarity of the T waves.²⁰ For example, Nikus et al considered the T wave to be positive or negative if it was 1 mm or more above or below the isoelectric line, measured more than 120 ms after the J point.¹³ As ST depression may affect the initial part of the T wave and during ischemia QRS may prolonged, we defined negative T wave based on the terminal part of the T wave. Interestingly, most of our patients had positive T waves per this definition. Moreover, two out of the four patients with negative T waves did not have significant coronary artery disease. Indeed, diffuse horizontal, and especially upsloping ST depression with prominent tall positive T waves may signify different coronary anatomy (regional subendocardial ischemia)²¹; however, small positive terminal T waves may carry the same significance as negative T waves. Further studies are needed to clarify this issue, especially to see if the definition used by Nikus et al.¹³ has better prediction.

Although the prevalence of LMCA, LMEQ or diffuse three-vessel disease was relatively high in our patients, it is clear that not all patients with such ECG pattern had LMCA/LMEQ disease. When combined with the right clinical scenario, this ECG pattern is associated with high prevalence of LMCA/LMEQ disease. It is plausible that in patients presenting with typical symptoms, this ECG pattern is more predictive for LMCA/LMEQ, especially if their baseline ECG is completely normal.^5 and 8 Gorgels et al have shown that during chest pain the number of leads with ST depression and the amount of ST deviation in the ECG correlate with the number of diseased coronary arteries.⁹ The authors mentioned that the absolute shift of ST deviation from the baseline ECG, obtained when the patient did not have symptoms, is important as ST abnormalities could be present at the baseline ECG even if the patient is free of ischemia.⁹ In particular, they suggested that ST depression in leads I, II and V4–V6 and ST elevation in lead aVR are predictive of LMCA or three vessel disease.⁹ However, if the abovementioned recommendations are adopted, urgent activation of the catheterization laboratory for patients with atypical presentation (shortness of breath, epigastric pain, hypotension, etc.) solely based on ECGs showing diffuse ST depression with ST elevation in lead aVR in order not to misdiagnose acute myocardial infarction due to LMCA/LMEQ/three-vessel disease, will probably lead to high rate of false activation. Similar high rates of false activation of the catheterization laboratory have been shown for patients with atypical symptoms presenting with presumably new left bundle branch block.^{22, 23 and 24}

We cannot exclude the possibility that small vessel disease or supply/demand imbalance leads to diffuse subendocardial ischemia in some of our patients. It has previously been reported that other medical conditions, such as hypothermia²⁵ and various neurological disorders²⁶ may present with such transient ECG pattern. Thus, the term "circumferential subendocardial ischemia” as suggested by Samuel Sclarovsky or "circumferential subendocardial stress” may be better than "ischemia due to multivessel or left main coronary artery obstruction”.¹

Study limitations

This is a retrospective review of patients based on ECGs that were identified during routine reading at the St. Luke's Episcopal Hospital ECG laboratory. Just over half of our patients presented with chest pain and/or shortness of breath. All ECGs met the pattern of diffuse ST depression with elevation in aVR, which is the pattern described by the current AHA/ACCF/HRS guidelines. We did not limit ourselves to patients presenting with clinical scenario that may be compatible with ACS, as we tested the accuracy of the ECG pattern for predicting LMCA or LMEQ-induced ischemia. Larger prospective studies are needed to confirm our results and assess the predictive value of this particular ECG pattern for LMCA/LMEQ disease in patients with and without typical clinical presentations. An alternative approach to compare the incidence of such ECG pattern among patients with ACS with or without LMCA/LMEQ/3VD has been reported (27). However, it is obvious that not all patients with such coronary anatomy have diffuse ischemia due to LMCA or LMEQ lesions. They may have culprit lesions leading to ACS in more distal locations. In many patients with NSTE-ACS, ischemia is intermittent and may not be detected by sporadic ECG recordings. In others, such pattern may be related to etiologies other than acute diffuse subendocardial ischemia.

Conclusion

The Texas Heart Institute at St Luke Episcopal Hospital is a tertiary center for heart disease, treating numerous patients with complex heart diseases, cardiomyopathies and heart failure. In our patient population, the pattern of diffuse ST depression in > 7 leads with ST elevation in aVR was associated with acute coronary syndromes in only in 28% of the patients.

Although coronary artery disease is prevalent among the patients that were selected to undergo CA, only 23% of the patients had LMCA/LMEQ disease. We cannot exclude that this pattern represents "circumferential subendocardial ischemia” due to "small vessel disease”, vasospasm or altered supply/demand ratio. Yet, the AHA/ACCF/HRS recommendation for interpreting this ECG pattern as representing "ischemia due to multivessel or left main coronary artery obstruction,” implying that these patients should be referred for urgent coronary angiography based on the ECG pattern alone may not be supported by our findings and the term "circumferential subendocardial ischemia” is probably more accurate.