Reducing Risk for Perioperative Stroke

INTRODUCTION

Although strokes can and do occur at any age, the incidence of stroke doubles with every decade after the age of 45 years, with three-fourths of all strokes in the United States occurring in people over the age of 65 years. People over 65 represented 13.1% of the population in the year 2010 but are expected to grow to be 20.6% of the population by 2030. Furthermore, the population of those undergoing surgery is aging at a faster pace than the general population is. ^, Perioperative stroke is a potentially devastating complication of surgery that is associated with substantial morbidity and a 5- to 10-fold greater likelihood of in-hospital mortality.

A perioperative stroke can occur intraoperatively or in the postoperative period and is commonly defined as a brain infarction of ischemic or hemorrhagic etiology with focal or global neurologic deficits that persists beyond 24 hours occurring within 30 days of the initial surgical procedure. Further classification of perioperative stroke (most commonly applied in the cardiac surgery population) consists of three subtypes based on the timing of the clinical presentation: intraoperative, early postoperative, and late postoperative. These subtypes have distinct risk factors and outcomes related to the mechanism of injury. The first is intraoperative stroke, which is diagnosed upon anesthesia emergence and is primarily caused by thromboembolism or hypoperfusion. The next is early postoperative stroke, which occurs within the first 7 days after the initial surgery and is generally caused by postoperative arrhythmias or hemodynamic factors. The third is late postoperative stroke, which occurs after 7 days but before 30 days after surgery and is commonly related to the general cerebrovascular risk profile of patients. ^,

The incidence of clinically recognized perioperative stroke varies widely with the type and timing of surgical procedure and the method of stroke detection. Comprehensive reviews on this topic typically illustrate the representative incidences based on surgical procedure. ^, Surgical categories commonly include cardiac surgery (1.0%–8.7%), carotid endarterectomy (1.8%–4.8%), ^, neurosurgery (0.5%–3.0%), ^, and noncardiac/noncarotid/nonneurologic procedures (0.1%–0.8%). ^, Certain surgeries in the latter category have a higher risk for perioperative stroke than others, such as thoracic, transplant, and major vascular surgery. Furthermore, the incidence of clinically unrecognized stroke, termed “covert stroke,” after noncardiac, noncarotid, and nonneurologic surgery may be as high as 7% in patients 65 years and older. Underreporting of stroke and the high rate of covert strokes are thought to be related to a potential masking of deficits because of comorbid factors and insufficient neurologic evaluations. For example, the highest clinical incidence of stroke after surgical aortic valve replacement was reported to be 17% in the Determining Neurologic Outcomes from Valve Operations Study by Messé and colleagues, in which a neurologist assessed the study patients preoperatively and performed serial postoperative evaluations. Within the same cohort, the incidence of stroke reported to the Society of Thoracic Surgery was 6.6%, and the overall rate of perioperative stroke after aortic valve replacement published by The Society of Thoracic Surgeons National Database Annual Reports was 1.3% to 1.5% over a similar time period. Of note, 54% of the patients without clinical stroke in Messé’s study had evidence of silent cerebral emboli found with diffusion-weighted magnetic resonance imaging (MRI). This variability in perioperative stroke incidence likely reflects the underlying surgical anatomy, the risk for vascular compromise and injury, the patient’s overall preoperative health status, and the methods employed to detect stroke. As such, there are likely no simple solutions to prevent this complex perioperative complication. The problem has been approached by different specialties with a variety of preventive measures, including the development of predictive models, preprocedural optimum medical therapy, intraoperative neuromonitoring, novel approaches to the surgical procedure, and multidisciplinary postprocedural care. Despite innovations and strategies, the incidence of perioperative stroke has remained a concern.

PATHOPHYSIOLOGY

Perioperative stroke after nonneurologic surgery is predominantly ischemic, rather than hemorrhagic, and the proposed mechanisms include thrombotic, embolic, lacunar, hematologic (hypercoagulable state), and hypoperfusion processes. ^{, , ,} Intraoperative stroke during cardiac surgery is because of macroembolism from atherosclerotic aorta or cardiac chambers during manipulation in 70% to 80% of cases, and watershed stroke attributable to hypoperfusion and cases of mixed etiology make up the remaining 20% to 30%. Early postoperative stroke after cardiac surgery is typically the result of emboli related to postoperative arrhythmias, specifically new-onset or preexisting atrial fibrillation (AF) or hypoperfusion insults related to hemodynamic factors such as low cardiac output syndrome and bleeding. ^, Late postoperative stroke is most commonly thromboembolic in nature and related to the general atherosclerotic risk profile of patients, including intracranial atherosclerotic disease, hypercoagulability, and AF.

The location of infarct and the distribution pattern often provide clues to the embolic origin of thromboembolic stroke, which are helpful in the perioperative period. Emboli originating from atherosclerotic plaque of the carotid bifurcation affect the anterior cerebral circulation, whereas emboli from plaque in the subclavian or vertebral arteries affect the posterior cerebral circulation. Intracardiac thrombi or a ruptured atheroma in the aortic arch may result in thromboembolic stroke in multiple vascular territories. Typically, the anterior circulation is involved in almost three-quarters of all instances of thromboembolic stroke among the general population, with occlusion of the middle cerebral artery or one of its branches accounting for approximately 90% of cases. Posterior circulation infarcts occur far less often. Nevertheless, thromboembolic stroke occurs twice as often in the posterior circulation in cardiac surgery patients compared with ischemic infarcts that occur in the general population. This is consistent with embolized atheromas of the distal ascending aorta, which are commonly manipulated (cannulation, cross-clamping, proximal aortic anastomoses) during coronary artery bypass grafting (CABG). Additionally, intraoperative stroke has been reported to occur more commonly in the right hemisphere, which can be attributed to the high-velocity jet emerging from the aortic cannula. ^,

Stroke after noncardiac, noncarotid, nonneurologic surgery more often presents after a variable recovery time, in either the early or late postoperative period, and is less often evident on emergence from anesthesia, which suggests that intraoperative proceedings may be contributory rather than causal for stroke in these surgical patients. The mechanisms for perioperative stroke in this population aside from AF-related embolic events, or fat or air embolism, may include surgery-mediated inflammatory responses and hypercoagulability. Although intraoperative hypotension is often assumed to be a major cause of stroke, data to support this mechanism are lacking, largely because hypotension during surgery is ill-defined and commonplace. ^, Although a prolonged period of time with a critically reduced blood pressure will certainly result in cerebral hypoperfusion and stroke, the temporal relationship between hypotension and infarction is difficult to establish perioperatively and blood pressure thresholds for cerebral ischemia likely vary by individual. Furthermore, hypotension and underresuscitation in the postoperative period may go unrecognized because most patients are less intensively monitored than during surgery.

Stroke after neurologic surgery is most commonly ischemic in origin. Unlike patients undergoing nonneurologic procedures where stroke is predominantly arterial in nature, neurosurgical patients may suffer either arterial or venous cerebral infarctions. Arterial ischemia may result from traumatic laceration or from intentional sacrifice of an artery for either hemostasis, aneurysmal ligation, or surgical access. Venous infarcts may similarly result from traumatic laceration (i.e., when a major venous sinus is disrupted by the craniotomy or when a bleeding vein is coagulated to provide hemostasis). Venous occlusion may also occur with compression from intraoperative or postoperative cerebral edema, which compromises cerebral perfusion and prevents venous outflow. The increased venous pressure further reduces effective drainage of affected brain tissue. This leads to increased cerebral blood volume and an even further reduction in cerebral perfusion pressure with subsequent oxygen deprivation and eventual infarction. Resection of tumors located near cerebral venous sinuses, especially parasagittal, convexity, parafalcine, or tentorial locations, increase the risk for venous injuries. Cerebral venous sinus thromboses frequently lead to hemorrhagic infarctions, which are driven by venous congestion and subsequent rupture of venules and capillaries. Cerebral venous infarction should be considered in cases of perioperative neurosurgical stroke (with or without hemorrhage) that do not correspond to a typical arterial territory. Clinically significant intracranial hemorrhages complicate 0.5% to 6.9% of craniotomies, and both the hematoma location and etiology are dependent on the preoperative pathology. ^, Although most postoperative bleeds usually occur within the first 24 to 48 hours, the first 6 hours has been identified as a critical period within which an acute postoperative hematoma may become clinically evident. The pathophysiology of postoperative intracranial hemorrhages varies and may be related to the underlying surgical anatomy, reperfusion injury, perioperative hypertension, cerebrospinal fluid loss, or hyperosmolar therapy leading to a parenchymal shift, or coagulopathy. ^,

OPTIONS/THERAPIES

The implication from the previous discussion of the pathophysiology of perioperative stroke is clear. An appreciable reduction in the incidence of stroke requires both universal and selective improvements by each surgical and anesthesiology subspecialty. Available techniques and methods to reduce perioperative stroke include measures to be taken early in the preoperative setting, such as predictive modeling and identification of modifiable risk factors, as well as medication and surgical optimization. Intraoperative measures to reduce the risk for perioperative stroke are increasing and involve preoperative and intraoperative strategic decision making, sophisticated detection techniques and surgical expertise, and multimodality neuromonitoring. Lastly, identification of perioperative stroke requires collaboration among an interdisciplinary and multiprofessional perioperative team to initiate early therapy, such as embolectomy.

EVIDENCE