Acute stroke is the abrupt onset, within seconds to hours, of neurologic deficits resulting from occlusion or rupture of arteries or veins that supply the central nervous system (CNS). By convention, this is a clinical definition and the radiologic–pathologic correlation is infarction. Acute strokes are classified as either hemorrhagic (15%–20%) or ischemic (80%–85%).¹ Of the hemorrhagic strokes, intracerebral hemorrhage is three times more common than subarachnoid hemorrhage. Additionally, transient ischemic attacks are temporary episodes of focal neurologic deficits to the brain or retina followed by complete recovery. Also, by convention, the definition of transient ischemia attack (TIA) encompasses full recovery without imaging evidence of infarction, within 24 hours. Most TIAs last for 5 to 20 minutes and events with persistent deficits for several hours are often associated with infarction. As such, a new definition for TIA suggested that the time window for TIA be reduced to less than 6 hours.² Awareness of these definitions is important in perioperative cerebrovascular disease as the goal of all acute stroke therapies is to recognize stroke symptoms, as well as differentiate strokes, and TIA from other acute, focal, neurologic, perioperative deficits such as focal seizures or complicated migraine, and ultimately, through rapid and aggressive interventions, convert all putative strokes into TIAs.

“Time is brain” as regards the treatment of acute ischemic stroke.³ Jeffrey Saver quantified the ongoing damage in an acute stroke and calculated that each minute during an acute stroke, 1.9 million neurons and 14 billion synapses are lost. Furthermore, for those patients with a large vessel (i.e., carotid, middle cerebral [MCA], or basilar artery occlusion) acute ischemic stroke, 120 million neurons, 830 billion synapses, and 714 km (447 miles) of myelinated fibers are lost each hour. Thus, rapid diagnostic evaluation and intervention, whenever possible, is critical regardless of whether the patient presents from home or has an in-hospital stroke. The challenge, however, for the patient in the perioperative period is recognition of an event. Whereas acute stroke in hospitalized patients potentially offers a greater likelihood of treatment, as there is no delay in arrival to the hospital, determining the onset of stroke symptoms in the perioperative period may be difficult, as the timing of onset may be hard to define because of patient sedation, pharmacologic paralysis, or a delirious state.⁴ Additionally, while time of arrival to hospital is not an issue, patients who are stable and transferred out of the postoperative care unit or intensive care unit may not be seen as frequently by nursing staff and thus the time window for intervention may be lost. Finally, surgery is typically an absolute exclusion criterion for intravenous (IV) thrombolysis and alternate acute interventions may not always be possible.¹

The incidence of stroke in the perioperative period is low. In one series, clinic strokes occurred in 3.6% of cases within 9 days following surgery though the frequency of acute infarct, as diagnosed by CT scan or MRI, was only 2.5%.⁵ In this series, 74% of stroke patients were diagnosed on the day of surgery, and 91% within the first 3 postoperative days. The authors reported that the stroke distribution was predominantly in the anterior circulation with 78% of acute infarcts located in the MCA distribution and 7% in the anterior cerebral artery distribution. However, there were more posterior circulation infarcts than otherwise would be predicted with 28% in the posterior cerebral artery distribution, 28% specifically in the cerebellum, reflecting with 36% of strokes occurring in multiple vascular territories probably as a result of cardiac embolism. Differences in cerebral distribution between infarct types were statistically significant with embolic strokes occurring in the right hemisphere in 34% of patients, left in 20%, vertebrobasilar in 19%, multiple in 26%, while watershed strokes occurred in the right hemisphere in 19%, left in 28%, vertebrobasilar in 0%, and multiple distributions 53% (p < 0.0001). This reflects the relative preservation of blood flow to the brainstem in the context of hypotension.

The risk of stroke after noncardiac procedures is low with the risk being significantly higher in cardiac and vascular (especially aortic and cervico-cerebral) procedures.^6,7 The most recent data indicate that strokes in surgical patients have the highest incidence in those undergoing cardiovascular-related procedures (1%–9%) followed by head and neck surgery (5%), vascular surgery (1%–3%), and lastly general surgical procedures (0.1%–0.7%).^6,7,8,9 Very important differences exist with respect to the type of cardiovascular procedure. Combined cardiac/valvular/aortic procedures have the highest associated risk as opposed to coronary artery bypass grafting (CABG) alone. Salazar et al.⁵ found a 3.6% incidence of clinical stroke in a prospective study involving a population of 5971 patients undergoing various isolated or combined cardiac and valvular and/or carotid procedures, performed between 1992 and 1997. Combined CABG and carotid endarterectomy (CEA) procedures were associated with the highest incidence (17.3%), followed by CABG/valve (6.7%), aortic (4.2%), CABG (3.2%, p < 0.0001), and valvular (2.8%) procedures. Dacey et al. reported a 1.61% incidence of perioperative stroke in 33 062 patients undergoing CABG alone between 1992 and 2001.¹⁰ Valvular surgery alone has a risk of between 5% and 9%.^7,9,11 CEA alone has a postoperative risk of 6% in symptomatic patients and 2% in asymptomatic patients.¹² However, review of studies encompassing simultaneous CEA-CABG totaling 1923 subject with a combination either symptomatic or symptomatic carotid artery stenosis and stable or unstable coronary disease (CAD) noted that an average perioperative complication rate of 3.0% stroke (range 0%–9%), 2.2% myocardial infarction (MI) (range 0%–6%), and 4.7% death (range 2.6%–8.9%).¹² Three studies reported long-term survival and the 5- to 6-year survival among 492 subjects ranged between 73% and 91%. In the only study with more than 50 subjects where CEA preceded the CABG, 257 patients with stable coronary artery disease were studied and the perioperative stroke rate was 1.9%, for MI 4.7%, and for death 1.6% though the data was observational and retrospective. Thus, the death rate for combined CEA-CABG is higher but the overall complication rates did not appear significantly increased in this instance compared to CEA alone.^9,10,11,12 Furthermore, while the risk of stroke is increased in both general surgical patients and those undergoing cardiovascular procedures in the context of greater than 50% cervical carotid artery stenosis, there was no association with increasing stenosis and stroke risk.⁶ Because of the risk of stroke from general surgical procedures related to carotid stenosis is estimated at 3.6% and the risk of stroke from surgery on an asymptomatic carotid artery stenosis is variably estimated at 2% to 3%, prophylactic CEA or carotid stenting cannot be supported at this time.⁶

Perioperative ischemic strokes are generally classified by pathophysiologic mechanism into cardio-embolic, atheroembolic, atherothrombotic, hypoperfusion (watershed), and hypercoaguable/other strokes. Most of the perioperative strokes are cardio- or atheroembolic and hemorrhagic stroke are much less common. Atherothrombotic (In situ thrombus) strokes, while common in de novo cerebra ischemia, are much less common in the perioperative period.^{7,8,9,10,11,12,13} Salazar et al.⁵ reported that out of 151 patients with acute infarction diagnosed on brain imaging, 71% had a pure embolic stroke, 12% had a watershed (border) zone infarct, with a mixed pattern (embolic + watershed) being present in 12%. By contrast, a Johns Hopkins series of 98 cardiac surgery patients identified with an acute clinical stroke found that 48% of patients had bilateral watershed infarcts on MRI.¹⁴ Fat or air emboli are additional rare causes of stroke as are hyperextension or flexion injuries to the posterior circulation.

About half of all strokes occur within the first day following surgery and are typically because of manipulation of vessels or debris from cardiac surgery, whereas a large proportion of delayed stroke is typically related to atrial fibrillation.^13,15,17 Hypotension is typically not a cause of stroke in the perioperative period even in patients with severe large vessel cervicocerebral atherosclerosis (i.e., extracranial carotid stenosis or vertebrobasilar stenosis). While patients with carotid artery stenosis are at higher risk for perioperative stroke, especially in the context of cardiac surgery, most of the strokes are not directly referable to the stenotic artery suggesting that the carotid disease is simply a surrogate marker for athero-embolism from the heart or aorta during manipulation of the heart or great vessels.¹⁷ In one series, postoperative strokes in watershed areas as a result of hypoperfusion occurred in less than 10% of cases following coronary artery bypass graft surgery (CABG).¹⁶

Several authors have enumerated a number of distinct preoperative, intraoperative, and postoperative risk factors for stroke.^6,7,8,9 Risk factors for perioperative stroke include age, metabolic syndrome risk factors of hypertension, diabetes, hyperlipidemia or obesity, renal failure, aortic or cervico-cerebral atherosclerosis, and a history of arrhythmias. Intraoperative factors include type of surgery and duration of surgery, duration of cardiac bypass cardiac and aortic cross-clamping (for cardiac surgery patients), intraoperative hypotension, intraoperative hypertension, and transient arrhythmias. There is some debate about the comparative benefit of regional versus general anesthesia in decreasing stroke risk.^{7,8,9,10,11,12,13,14,15,16,17,18} In general, volatile anesthetic agents, barbiturates, and propofol reduce ischemic neurologic injury and may have a neuroprotectant effect and the main issue related to general anesthesia is the induced hypotension that may occur with these agents.¹⁹ Postoperative risk factors include hyperglycemia, dehydration or blood loss, heart failure or myocardial ischemia or low cardiac ejection fraction, or cardiac arrhythmias.^6,7,8,20

Dacey et al.¹⁰ found that in a population of 35 733 patients undergoing isolated CABG, the characteristics of the patients who suffered stroke were age (mean value 70.7 years versus 65.4, p < 0.001), ejection fraction (49.7% versus 52.9%, p < 0.001), diabetes (42.4% versus 30.5%, p < 0.001), renal failure or creatinine >2 mg/dL (8.4% versus 3.2%, p < 0.001), and priority of surgery. A preoperative risk prediction model was developed by Charlesworth et al.²¹ based on the same population from New England of 33 062 consecutive patients undergoing isolated CABG. This risk prediction, model developed by the Northern New England Cardiovascular Disease Study Group to predict the risk of stroke after CABG, notes the relative weigh of the variables of age, elective versus urgent surgery, female sex, ejection fraction less than 40%, diabetes mellitus, creatinine greater than two and history of prior stroke or TIA, or prior history of peripheral vascular procedures.