Overall, we found that ischemic-appearing ECG changes on arrival were observed in 41% of patients with ICH and that such changes independently predict myocardial injury. Therefore, clinicians faced with such changes should not disregard these findings as nonspecific in the setting of ICH but, rather, consider such patients as having true myocardial injury.
The first account in the Western literature of an association between acute stroke and ECG changes appeared in 1947 [30
]. Since then, many authors have observed ECG changes in patients with acute central nervous system events [12
]. Abnormalities noted include prolonged QT interval, ischemic-appearing changes, U waves, tachycardia, and arrhythmias. Some of these studies have evaluated these measures in ICH specifically [14
]. However, not all abnormalities are likely to be clinically relevant in the acute setting, and the emergency physician is faced with the question of which findings to take seriously.
Another question that arises is how to diagnose myocardial injury in the setting of ICH. Troponin elevations have been found to occur relatively commonly in this disease and are associated with poor outcome [48
]. However, this association may well reflect that there has been true myocardial injury. Both troponins I and T are highly specific for myocardial necrosis and currently considered the criterion standard for detection of myocardial injury [49
]. In addition, coronary angiography and cardiac stress testing may not be feasible in the acute setting for many patients with ICH.
Recent studies in patients with subarachnoid hemorrhage have shown cardiac troponin elevations in 17% to 28% of patients [50
]. In patients with stroke of varied origin, troponin elevations occurred in roughly 17% and were associated with an adverse prognosis over time [53
]. Overall, then, myocardial injury appears to be relatively common after cerebrovascular emergencies in general.
The source of myocardial injury in patients with central nervous system emergencies is not fully understood. Although one possibility is coronary artery occlusion, many have suggested that myocardial injury after stroke can be attributed to abnormally high levels of plasma catecholamines [54
]. The assertion that cardiac alterations are mediated by catecholamines is supported by the fact that disrupting the sympathetic chain at the cervical level prevents arrhythmias, whereas a vagotomy does not [56
]. These changes can be inhibited by catecholamine blocking agents [58
]. Furthermore, myocardial catecholamine concentrations rise and fall rapidly after an intracranial catastrophe [60
Catecholamine-mediated myocardial injury is believed to be multifactorial: tachycardia, coronary spasm and vasoconstriction, toxic effects on cardiac myocytes, and an increased intracellular concentration of calcium [61
]. The histologically confirmed lesions in the myocardium are small foci of subendocardial hemorrhage and myocytolysis surrounding epicardiac nerves. Myocytolysis represents acute cardiac muscle fiber stress, a phenomenon different from the coagulation necrosis seen in coronary ischemia [62
]. These observations provide evidence that pathologic changes in the myocardium as well as ECG changes may result from cerebral injury and may be mediated by the autonomic nervous system.
In an exploratory, hypothesis-generating analysis stratifying ECG findings by anatomical distribution, it appeared that changes in I and aVL as well as V5 and V6 were more specific for myocardial injury. However, it is possible that this finding was an artifact of multiple hypothesis testing; future studies are necessary to confirm whether specific anatomical ECG changes can be used to predict troponin elevations in this population.
It has traditionally been thought that diffuse ECG changes are more consistent with cerebrovascular changes [63
]. However, our study suggested no association between ischemic-appearing ECG changes among multiple anatomical distributions and myocardial injury. Thus, from an operational standpoint, we were unable to demonstrate that the clinician in the emergency setting can safely ascribe ischemic-appearing changes in multiple distributions to nonspecific neurocardiogenic influences.