Cortical spreading ischemia in the absence of proximal vasospasm after aneurysmal subarachnoid hemorrhage: evidence for a dual mechanism of delayed cerebral ischemia

Abstract

There are longstanding inconsistencies in the evidence thought to link vasospasm in the major branches of the Circle of Willis with delayed cerebral ischemia and poor outcome from aneurysmal subarachnoid hemorrhage (aSAH). The demonstrations, first in the laboratory, and more recently in patients with aSAH, of cortical spreading ischemia based on an abnormal response of the cerebral microcirculation to spreading depolarization offer an additional possible mechanism for delayed ischemia. That such events can occur in the substantial absence of proximal vasospasm is compatible with this concept, but the preliminary evidence needs support from more extensive studies.

The nature and cause of delayed ischemic neurologic deterioration or delayed cerebral ischemia (DCI) that occurs 3 to 14 days after subarachnoid hemorrhage (SAH) from rupture of an intracranial aneurysm (aSAH) has been much studied and debated, most specifically since the angiographic features of spasm of the major branches of the Circle of Willis were first described by Ecker and Riemenschneider (1951). The natural logic would be for DCI to develop as a consequence of the cerebral blood flow reduction from proximal vasospasm. This is consistent with the clinical experience of generations of neurosurgeons (Crowley et al, 2011) and more recently of neurointerventionalists (Kimball et al, 2011). However, inconsistencies in this ‘a priori' concept emerged, for example, from the findings (1) that nimodipine improves outcome without influencing proximal angiographic vasospasm (Dorhout Mees et al, 2007), (2) that the predictive value of angiographic vasospasm for DCI is <50% in some series (Dankbaar et al, 2009)—arguably worse than from tossing a coin, and (3) from meta-analyses of SAH clinical trials suggesting that pharmacological prevention of angiographic vasospasm is not associated with improved clinical outcome (Etminan et al, 2011); yet, cerebral infarction is highly correlated with poor outcome (Fergusen and Macdonald, 2007; Vergouwen et al, 2011). However, other series show a high correlation between angiographic vasospasm and DCI (Crowley et al, 2011). This paradox of prevention of vasospasm without improvement in outcome in combination yet a high correlation of vasospasm with poor outcome led to the hypothesis that DCI contributes to poor outcome by a ‘double hit' process including some mixture of large artery vasospasm and constriction in the cerebral microcirculation (that would not be evident on angiography—save perhaps as a delay in dye transit).

The new story starts with the hypothesis that cerebral ischemia can be induced by the products of hemolysis in the subarachnoid space (Macdonald and Weir, 1991) examined in a group of papers from the laboratory of Dreier and colleagues in Berlin. They showed, in summary, that the normal vasodilator response to induced cortical spreading depression was replaced by a highly abnormal vasoconstrictor response, ‘cortical spreading ischemia', in a rat preparation in which the cortex was superfused with artificial cerebrospinal fluid with a composition designed to replicate approximately that of human postSAH cerebrospinal fluid, and that this was associated with cortical necrosis similar to that seen in patients dying from aSAH (Dreier et al, 2000). This provided the first experimental evidence for a contribution to cortical ischemia from distal vasoconstriction in the cortical microcirculation—potentially additive with the effects of proximal vasoconstriction—angiographic vasospasm—the basis for a ‘double hit' concept of DCI.

Woitzik et al (2012) now report—in a promising initial clinical aSAH study based on the use of continuous electrocorticographic monitoring to detect spontaneous spreading depolarizations in 13 patients—that if proximal vasospasm is minimized by placement of nicardipine pellets around the middle cerebral artery at open aneurysm surgery, then spontaneous depolarization events nevertheless occurred over the cortical surface in 10 of 13 patients (77%), accompanied by an abnormal (biphasic or monophasic decrease) tissue pO₂ response to 82% of depolarizations, where both electrocorticographic and tissue pO₂ were recorded. The number of depolarizations per day of recording time was significantly (P<0.01) higher in those patients with delayed cerebral ischemia than in those without, described correctly by the authors as a possible association.

Where does this leave us? The new body of work does not remove proximal vasospasm as a cause of DCI: it merely offers a persuasive explanation for some longstanding inconsistencies in that model. This offers some support for the concept that proximal large artery vasospasm and distal cortical events might be required in varying combination to induce DCI. Woitzik et al (2012) provide the basis for a more extensive study examining the incidence, behavior and correlates (e.g., hemodynamic, tissue oxygen, and tissue glucose (Feuerstein et al, 2010) responses) of depolarization events, and, most critically, their relationship with outcome and with established risk factors for adverse outcome. They contribute to the study of SAH by raising fundamental questions: whether depolarizations in aSAH (‘Killer waves of depolarization' Iadecola, 2009) are an independent risk factor for DCI, and how they interact with proximal vasospasm. The new results should stimulate development of methods to detect depolarizations less invasively than with subdural electrocorticographic strips placed at open craniotomy—thus perhaps allowing study of the large numbers of patients now treated by endovascular procedures and, at a later stage, initial treatment studies.

The authors declare no conflict of interest.