Asystole During Endovascular Embolization of a Dural Arterio-Venous Fistula in the Brain

Abstract

Endovascular embolization has become a convenient and reliable technique in the treatment of cerebrovascular malformations in the neurological and neuroradiological field (1–8). Despite the option to reach almost every region of the brain without craniotomy, numerous complications have been described (3,5–10). We report a case of asystole during endovascular embolization of a dural arteriovenous fistula (d-AVF) in the brain. Case Report A 40-yr-old, 100-kg athletic man, with a d-AVF and otherwise unremarkable medical history, was scheduled for sessions of endovascular embolization. In the first session, the patient underwent endovascular embolization under general anesthesia resulting in a 50% embolization of the d-AVF. The procedure produced no complications. Ten days later, the patient underwent a second session of embolization under general anesthesia. Anesthesia was induced IV with 3 mg of midazolam, 200 μg of fentanyl, and 300 mg of propofol. Vercuronium (10 mg) was administered to facilitate orotracheal intubation. Anesthesia was maintained with 400 mg/h propofol IV. The lungs were ventilated by a controlled-pressure ventilation mode with a gas mixture of 70% nitrous oxide and 30% oxygen. Lactated Ringer's solution was administered IV at a rate of 700 mL/h. Approximately 100 min after the induction of anesthesia, the ascending pharyngeal artery feeder to the d-AVF was catheterized. Using standard microcatheter systems, routine preembolization was performed under scanner control by injecting 3 mL of 33% glucose at room temperature (approximately 20°C) to rinse out the vessel. This is done to remove the threat of continued polymerization with any ionic solution such as contrast material or blood vessel intima. Immediately after this procedure (i.e., before the histoacryl embolization), the patient experienced a complete sinus arrest for 40 s. No pulse could be detected by radial artery monitoring, pulse oxymetry, or palpation of the carotid artery. After the administration of 0.5 mg of atropine, sinus rhythm occurred with approximately the same frequency as before. Because asystole had not been attributed to the injection of 33% glucose, a further attempt at embolization was started. Asystole recurred, the procedure was interrupted, and the patient was admitted to a neurosurgical intensive care unit. Blood gas analysis and laboratory analysis of electrolytes were within the normal range. The patient's further stay was uneventful, and after a few days, he was discharged from the hospital. A follow-up examination 1 wk later revealed no persistent adverse sequelae. Discussion General reactions, such as radiocontrast reactions, pulmonary embolism of particles (5–8), and neurological catastrophes, such as aneurysm perforation (1–10), unintentional parent artery occlusion (3), obliterations of physiologic arteries, compression to the vasa nervorum of cranial nerves (11), or embolization via “dangerous” transcranial arterious anastomoses (12), are recognized and often reported complications of endovascular embolization. In this report, a patient who underwent endovascular embolization for the second time manifested asystole twice after injection of 33% glucose. Although reflex vagal dysrhythmias are known to occur during eye surgery (13), maxillofacial surgery (14), tracheal suction, and bronchoscopy (15), a review of the literature showed that cardiac arrest in conjunction with endovascular procedures has not been reported. In our case, the vagus nerve was obviously activated by the injection of 33% glucose infusion into the ascending pharyngeal artery via an artery feeder. Several possible neuroanatomic explanations may account for this event. First, the external carotid artery branch (i.e., the ascending pharyngeal artery) injected with 33% glucose supplies the vasa nervorum of cranial nerves IX–XII via the neuromeningeal branch (16). Selective transcatheter injections into this vascular territory may have affected the vagus nerve and may, in turn, have triggered asystole. This is the highly likely mechanism of the observed changes. Second, several “dangerous” transcranial anastomoses from the external carotid artery system to both the internal carotid artery territory and the vertebral artery have been described. Among these, the hypoglossal artery and “odontoid arch system” connect the ascending pharyngeal artery with the vertebral artery territory (12) and, thus, connect with branches supplying the nuclei of cranial nerves IX–X. However, it should be assumed that the volume of the blood through the vertebrobasilar system would quickly dilute the injected material. Although both mechanisms are unusual, the reproducibility of asystole suggests a causal relationship. The arrhythmia was terminated approximately 20 seconds after the first injection of atropine. Because we did not observe a subsequent tachycardia and because the atropine was administered by a long infusion line as a result of the remote position of the anesthesiologist, it is possible that the atropine did not reach the central circulation. The physiological mechanism of stimulation remains speculative. In our case, asystole was temporarily related to the injection of 33% glucose. Because glucose was continuously infused, oxygen supply was interrupted for approximately five seconds. Further, the infusion was injected at room temperature (approximately 20°C), which is about 14 degrees lower than body core temperature; in addition, the osmolarity of 33% glucose is 1828 mosmol/L, 6 times more than blood osmolarity, the pH value is between 3.5 and 6.5. We conclude that the vagus nerve was activated by temperature or hypoxic or osmotic stimulation. Asystole was rapidly detected only because the procedure was performed under general anesthesia, and therefore, sufficient monitoring and adequate therapy by an anesthesiologist was ensured. Even though endovascular embolization of arterio-venous malformations can result in stroke and death, the documented procedure is characterized as “minimally invasive neurosurgery.” In any case, it cannot be assumed that the anesthetic risk is minimized (5–8). It is the role of the anesthesiologist to provide adequate monitoring and sufficient and opportune therapy during such procedures, despite the minimally invasive characteristic. This article underlines the necessity of effective communication between neuroradiologist and anesthesiologist and the importance of incorporating anesthesia care in the routine management of endovascular neurosurgery, even if these procedures are performed under local anesthesia.