Management of Perioperative Complications During AVM Treatment

Hemorrhage

Perhaps the most obvious complication during resection of an AVM is hemorrhage. AVMs are high-flow, low-resistance vascular lesions that are capable of profound bulk flow of blood. They often have multiple large arterial feeders. It is now common practice to perform staged INR embolization prior to iAVM resection for the purpose of decreasing AVM size, reducing flow, allowing staged acclimation of surrounding vascular beds, and embolizing deep feeding vessels that pose hemostatic challenges due to poor visibility and accessibility. Despite this practice, sudden blood loss may still occur, and it may be profound and prolonged. Standard preparation should include large-bore intravenous access for rapid transfusion and immediate blood product availability in the OR prior to AVM manipulation. If hemostasis is not rapidly achieved, initiation of a rapid transfusion protocol and mobilization of additional resources should occur for such purposes as running blood from the blood bank to the OR, cross-checking blood units, and running equipment such as a Belmont rapid infuser or cell saver, if available. Care must be taken during rapid blood loss resuscitation not to neglect maintaining adequate platelet and clotting factor levels, while avoiding profound hypothermia. Blood pressure reduction with rapid-onset short-acting agents such as clevidipine or nicardipine should be offered to assist with visualization and hemostasis, with the caveat that controlled blood pressure reduction is very difficult in a patient who is hypovolemic. Blood pressure reduction should always be limited in duration and degree to the minimum necessary to gain surgical control. Hemorrhage can also occur after resection, both in the immediate postoperative period and in a delayed fashion in the intensive care unit (ICU).

Malignant cerebral edema

Malignant cerebral edema is not necessarily specific to iAVM management; however, it has a strong association with iAVM treatment, and the underlying pathophysiology may be unique. Historically, malignant edema occurred in the OR with such frequency that treatments were initiated preemptively in anticipation of edema development, including barbiturate loading to burst suppression and moderate hypothermia. Anticipation was based primarily on factors such as large AVM size, prolonged microsurgery (> 10 hours), and the need for multiple intraoperative angiograms (contrast agents were ionized and profoundly hyperosmolar with significant chemotoxic effects). Although there were many advances in operative management during this period, intraoperative use of barbiturates and hypothermia seemed to help. The primary rationale for these therapies was to suppress cerebral metabolism and reduce cerebral blood flow (CBF) such that the acute cerebral hemodynamic changes ensuing from microsurgical resection might be attenuated. The delayed and gradual emergence from anesthesia in the ICU due to the prolonged half-life of barbiturates may also have had a favorable effect. However, with increased use of staged INR embolization and other management advances, the frequency of severe sudden intraoperative swelling has decreased such that the empiric use of these therapies is not supported.

As discussed in Chapter 25 , there are two leading theories to explain the etiology of acute malignant edema and hemorrhage that may manifest anytime in the operative and early postoperative period: (1) normal perfusion pressure breakthrough (NPPB) and (2) occlusive hyperemia. The first proposed mechanism involves either perfusion pressure in excess of the local autoregulatory limits or small perinidal areas of impaired cerebral autoregulation resulting in hyperemia. Several studies have demonstrated that autoregulation and CO ₂ reactivity appear to be preserved following iAVM resection, but with a left shift of the autoregulatory curve ( Fig. 26.1 ) due to adaptation to the low pressure shared with the arterial feeder ( Fig. 26.2 ). Interestingly, increased postoperative CBF appears to be a global phenomenon. Despite globally increased cerebral blood flow, hemorrhage and edema tend to occur ipsilateral to the iAVM resection. Some have suggested that absent astrocytic foot processes, such as seen in capillaries in animal models near the iAVM, may be responsible for local edema and hemorrhage. The second mechanism, occlusive hyperemia, is presumably caused by stagnation within veins and arteries. Arterial and venous remodeling due to prolonged mechanical stress predispose to stagnation once flow is acutely reduced following iAVM resection. Both supportive and contradictory evidence can be found for each of the proposed mechanisms. Residual AVM may also contribute to postoperative hemorrhage; however, the use of intraoperative angiography and/or fluorescein angiography has greatly diminished this as a contributing factor.

Although sudden malignant edema is quite rare in the modern era, a management plan for its occurrence in the OR should be considered. Of utmost importance is clear two-way communication between the surgeon and the anesthesiologist to determine the cause and correct management of this complication. Any blood pressure deviation from the predetermined target should be communicated and corrected promptly. Simultaneously, the anesthesiologist should ensure adequate anesthetic depth, normal airway pressures, and the absence of patient coughing or straining against the endotracheal tube. In the absence of obvious anesthetic-related factors, the possibility of a surgical cause, such as sacrifice of an arterialized vein, should be considered. In this situation, the surgical team must work quickly to remove the remaining arterial feeders to the AVM given the newly impaired venous drainage and imminent risk of rupture. A brief discussion with the surgical team regarding other interventions, such as additional blood pressure reduction, should occur. Acute blood pressure reduction can be achieved with rapid- and short-acting cardiovascular agents or anesthetic agents that have the benefit of simultaneously reducing blood pressure and cerebral metabolic rate (CMRO ₂ ). In the OR, propofol is the anesthetic agent most readily available for this purpose. Although barbiturates may also be considered, there will likely be a greater treatment delay, increased hemodynamic instability, and emergence delay due to prolonged drug half-life. Additional hyperosmolar therapy, hyperventilation, and hypothermia may be considered depending on severity and response to early management interventions.

As alluded to earlier, the two leading theories on the pathophysiology of malignant edema and hemorrhage imply potentially opposing physiology and medical treatment requirements. Stagnation and ischemia may potentially benefit from higher perfusion pressure, while hyperemia requires blood pressure reduction. Therefore standard iAVM management at our institution and many other centers involves tight blood pressure control within a narrow normotensive range that is based upon preoperative blood pressure, comorbidities, and intraoperative findings. In the event of postoperative hemorrhage or edema, determination of the underlying etiology will help fine-tune medical treatment. For instance, if imaging demonstrates reduced perfusion due to arterial stagnation, more liberal blood pressure limits may be beneficial. If hemorrhage or edema occurs secondary to venous thrombosis, anticoagulation treatment may be beneficial despite its risk. Finally, if hyperemia appears to be the primary etiology, more restrictive blood pressure control is likely indicated. Jugular venous oxygen saturation monitoring, if available, may be a useful adjunct, particularly in cases in which preoperative embolization is not feasible. NPPB would be associated with an elevated saturation, whereas venous stagnation would more likely be associated with a normal or reduced saturation.