Secondary Logo

Journal Logo

A Case Report of Onyx Pulmonary Arterial Embolism Contributing to Hypoxemia During Awake Craniotomy for Arteriovenous Malformation Resection

Tolly, Brian T. MD*; Kosky, Jenna L. MD*; Koht, Antoun MD*†‡; Hemmer, Laura B. MD*‡

doi: 10.1213/XAA.0000000000000436
Case Reports: Case Report

A healthy 26-year-old man with cerebral arteriovenous malformation underwent staged endovascular embolization with Onyx followed by awake craniotomy for resection. The perioperative course was complicated by tachycardia and severe intraoperative hypoxemia requiring significant oxygen supplementation. Postoperative chest computed tomography (CT) revealed hyperattenuating Onyx embolization material within the pulmonary vasculature, and an electrocardiogram indicated possible right heart strain, supporting clinically significant embolism. With awake arteriovenous malformation resection following adjunctive Onyx embolization becoming increasingly employed for lesions involving the eloquent cortex, anesthesiologists need to be aware of pulmonary migration of Onyx material as a potential contributor to significant perioperative hypoxemia.

From the Departments of *Anesthesiology, Neurology, and Neurological Surgery, Northwestern University Feinberg School of Medicine, Chicago, Ill.

Accepted for publication August 31, 2016.

Funding: None.

The authors declare no conflict of interest.

Address correspondence to Laura B. Hemmer, MD, Department of Anesthesiology and Neurological Surgery, Northwestern University Feinberg School of Medicine, NMH/Feinberg Room 5–704, 251 E Huron St, Chicago, IL 60611. Address e-mail to

Craniotomy using local anesthesia and monitored sedation (awake) is common practice for neurosurgical tumor management, and it is increasingly used for lesions within or near eloquent cortex. Accumulating evidence, although lacking randomized controlled trials, suggests that awake brain tumor resection compared with general anesthesia is associated with improved outcome.1 Awake craniotomy for arteriovenous malformation (AVM) resection has been regarded as daunting because of diffuse vascular lesion borders involving eloquent cortex.2 However, 2 recent case series demonstrate functional outcome benefit in selected patients using awake cortical language and motor mapping intraoperatively during AVM resection.2,3 In recent years, endovascular management of AVMs has also gained recognition, either alone or as a preoperative adjunct to radiation or surgical resection for blood loss minimization.4 Ethylene-vinyl alcohol copolymer (Onyx, EV3, Irvine, CA) dissolved in dimethyl sulfoxide (DMSO) is an embolization material increasingly used for AVM management given its safety and target-vessel precipitation that facilitates regular angiographic assessment.5 We present a patient who underwent staged endovascular embolization of a right frontal AVM with Onyx followed by awake craniotomy and who developed significant late intraoperative hypoxemia.

Back to Top | Article Outline


The patient involved in the present case reviewed the text and figures and gave written permission for the authors to publish the report.

Back to Top | Article Outline


A 26-year-old, 71-kg man with new seizures and intrinsic left hand weakness presented for elective right supratentorial AVM (Spetzler-Martin Grade 2) embolization and resection. The patient had no other past medical history; home medication regimen included 500 mg of levetiracetam twice daily. On the evening before surgical resection, he underwent uneventful interventional embolization under general anesthesia using Onyx 34. After the patient demonstrated the ability to follow commands and recovery from neuromuscular blockade, the endotracheal tube was removed in the procedure suite. Upon arrival to the postanesthesia care unit, his vital signs were: heart rate (HR), 105 beats per minute (bpm); blood pressure (BP), 105/58 mm Hg; respiratory rate, 18 breaths per minute; and oxygen saturation (Spo2), 100% on 2 L O2 via nasal cannula. The patient’s heart rate remained elevated overnight between 90 and 120 bpm but with otherwise stable vital signs. The next morning, during a preoperative magnetic resonance imaging scan, he experienced an additional heart rate increase to 160 bpm. He denied dyspnea or chest pain but expressed anxiety. Electrocardiogram (ECG) demonstrated sinus tachycardia without the evidence of ischemia. Serum electrolytes, including potassium, magnesium, and calcium, were within normal range. In addition, serum troponin I was undetectable. With otherwise reassuring vital signs, the collaborative decision was made to proceed with surgery.

Upon arrival to the operating room, vital signs were: HR, 120 bpm; BP, 125/55 mm Hg; and Spo2 99% on room air. Intraoperative monitoring included intra-arterial blood pressure and urine output via Foley catheter, in addition to standard American Society of Anesthesiologists monitors. End-tidal CO2 was monitored via sampling nasal cannula while delivering 2 L of supplemental oxygen. Local anesthesia was provided with 6 bilateral scalp nerve blocks (supratrochlear, supraorbital, zygomaticotemporal, auriculotemporal, lesser occipital, and greater occipital) performed before pin fixation; total anesthetic doses equaled 50 mg of 1% tetracaine and 250 mg of 1% lidocaine with 1:200,000 epinephrine. The patient was placed in the supine position with the operating room table back elevated to 20 degrees. Sedation was maintained with remifentanil (0.05 to 0.12 mcg·kg−1·min−1) and propofol (25 to 55 mcg·kg−1·min−1) infusions titrated to maintain a respiratory rate between 8 and 12 breaths per minute. Surgical exposure was aided by 100 g of intravenous (IV) mannitol administered over 15 minutes. Esmolol infusion (100 to 250 mcg·kg−1·min−1) or phenylephrine infusion (15 to 60 mcg/min) was administered as needed for intraoperative tachycardia or hypotension, respectively.

At 7 hours operative time, Spo2 readings were persistently 93% to 95%. The sedation was minimized, and we provided 10 L of supplemental O2 via an anesthesia machine circuit blow-by. At approximately 10 hours operative time (coinciding with near-complete AVM resection), there was an acute progression of arterial desaturation, with Spo2 of 85% noted. At this time, 10 L of O2 was provided via standard facemask. The patient remained alert, and hypoxemia was responsive to continual encouragement to take deep breaths. Respiratory rate reached a maximum of 28 to 30 breaths per minute. Lung auscultation revealed faint inspiratory crackles bilaterally. Arterial blood gas (ABG) analysis obtained when the patient was transiently on room air showed severe arterial hypoxemia: pH 7.43, Paco2 36 mm Hg, Pao2 40 mm Hg, %Hbo2 78%, and serum bicarbonate 24 mEq/L. Intraoperative fluid balance over 12 hours was 6500 mL of crystalloid and 2000 mL of 5% albumin administered, 800 mL of estimated blood loss, and 7000 mL of urine output. IV furosemide (5 mg) and aerosolized albuterol (4 puffs) were administered empirically with no immediate improvement.

With persistent encouragement in deep breathing, oxygen saturations were maintained at around 90%. Although preparations were in place, an artificial airway was not immediately required. Notably, low-dose remifentanil infusion was necessarily continued until the end of the case in light of progressive pain as scalp blocks lost effectiveness.



Postoperatively, the patient was transferred to neurosurgical intensive care unit for observation and further management. Upon arrival, vital signs were: oral body temperature, 37.3°C; HR, 130 bpm; BP, 130/60 mm Hg; respiratory rate, 20 breaths per minute; and Spo2, 88% on 10 L via standard facemask. Off of sedation, the patient was noted to have a nonproductive cough. Overnight, he required humidified venturi mask and high-flow nasal oxygen to maintain adequate oxygenation. ABG at this time was pH, 7.53; Paco2, 30 mm Hg; Pao2, 78 mm Hg; and serum bicarbonate, 25 mEq/L. ECG demonstrated sinus tachycardia with T-wave inversions in the inferior leads. Because of the persistent tachycardia, a CT pulmonary angiogram was obtained that revealed Onyx pulmonary arterial embolism (Figure). During the next 24 hours, the patient remained afebrile and normotensive, with HR ranging between 90 and 120 bpm. At 36 hours after surgery, the patient was weaned to room air and discharged home on postoperative day 4. At a 3-month postoperative clinic visit, the patient was recovering well and did not express any respiratory complaints.

Back to Top | Article Outline


Transient intraoperative hypoxemia is a frequent challenge encountered during awake craniotomies with a broad differential diagnosis.1 Prolonged sedation increases the risk for atelectasis and hypoventilation without the possibility of performing recruitment maneuvers, such as in an intubated patient. Pulmonary aspiration, venous air embolism or thromboembolism, neurogenic pulmonary edema, intravascular volume overload, inadequate oxygen delivery, and bronchospasm are also possible etiologies. Routinely used in neurosurgery, high-dose mannitol administration carries a risk for pulmonary edema in patients with poor systolic reserve. Our case illustrates the importance of including Onyx migration as a potential contributor to hypoxemia during surgical resection of AVM after embolization.

To our knowledge, this is the first case of Onyx pulmonary artery embolism during awake or asleep craniotomy contributing to significant hypoxemia. However, other respiratory complications in the setting of Onyx use for cerebral AVM and aneurysm embolization have been described previously in the perioperative period. Asouhidou et al6 suggest that mild hypoxemia surrounding Onyx use may be attributable to the copolymer solvent DMSO, which is partially excreted via the lungs, causing a transient alveolar oxygen dilution. Two case reports of acute respiratory distress syndrome after Onyx embolization postulate that DMSO can also incite end-organ endothelial necrosis and inflammation, leading to severe pulmonary exudates.7,8 In the present case, five 1.5-mL vials of Onyx 34 copolymer and associated DMSO were used during the embolization procedure, which is below the recommended maximal dose of 200 mg/kg.5,9

Vascular migration of Onyx material from the site of intracranial embolization has also been reported. Pukenas et al10 described a case of asymptomatic Onyx pulmonary embolism identified on a routine chest CT performed 20 days after dural AVM embolization. A cerebral arteriovenous fistula in a neonate was treated with endovascular Onyx, and Onyx was later found in the right pulmonary artery.11 Both cases of embolized Onyx were incidentally identified without apparent respiratory sequelae. Crusio et al12 reported superior vena cava syndrome causing upper airway obstruction 2 weeks after dural AVM embolization. They hypothesized that an in situ central venous stent served as a nidus for migrating Onyx material.

The profound arterial hypoxemia illustrated in the present case was likely multifactorial in etiology, with substantial contribution from the Onyx pulmonary embolism. The first sign of significant embolic phenomena was likely the persistent sinus tachycardia after Onyx embolization. In addition, the patient developed evidence of right heart strain on postoperative ECG. We hypothesize that Onyx embolization may have occurred as early as the interventional procedure, with perhaps additional occurrences of emboli during the latter third of surgery, when larger AVM nidus vessels containing Onyx were manipulated and excised. Compared with cyanoacrylate embolization glues, Onyx interacts relatively less with the vascular endothelium, substantiating the concern for intraoperative dislodgement.5

In addition to repeated pulmonary embolization of Onyx perioperatively, progressive atelectasis and pulmonary edema also likely contributed to the profound arterial hypoxemia seen in this case; indeed, the CT scan showed evidence of mild consolidation in the dependent lung fields (Figure). A component of atelectasis is highly probable because of the 12-hour duration of monitored sedation in a recumbent patient without positive pressure ventilation, although this alone unlikely caused the nadir Pao2 of 40 mm Hg that acutely developed toward the end of the case in an otherwise healthy nonobese patient taking repeated deep inspirations.

Intravascular volume overload and bolus mannitol administration leading to pulmonary edema also warrant mention in the neurosurgical patient. After administration of 1.4 g/kg of mannitol, our patient briskly diuresed 1700 mL of urine in the first 2 hours of the case, followed by approximately 500 mL/h thereafter. High volumes of crystalloid solutions (3200 mL of 0.9% sodium chloride solution and 3300 mL of Plasma-Lyte) were administered to match urine output in a 1:1 ratio. As a result of the high-volume diuresis and 0.9% saline administration, the patient’s serum sodium concentration gradually increased from 133 mEq/L (obtained after intravenous mannitol administration reflecting dilution) to 144 mEq/L over the 12-hour intraoperative course. Despite this increase in serum sodium concentration, 7 hours after surgery, the patient’s serum electrolyte profile did not demonstrate a hyperchloremic metabolic acidosis (sodium, 139 mEq/L; chloride, 104 mEq/L; bicarbonate, 27 mEq/L).

We targeted intravascular euvolemia in this case to mitigate the elevated risk of rapid surgical hemorrhage from the AVM. Large volumes of the aforementioned crystalloid solutions were administered to match urine output in a 1:1 ratio and to wean the patient’s phenylephrine infusion requirement that was initially suggestive of intravascular volume depletion. Serial ABG analyses further guided intravascular volume therapy, with an initial base deficit of −4.2 mEq/L, decreasing to −0.3 mEq/L before operating room departure. The hemoglobin concentration also progressively decreased from 11.3 to 8.8 g/dL by the end of surgery. This decrease in hemoglobin concentration, attributable in part to an estimated blood loss of 800 mL, is also consistent with some degree of hemodilution and intravascular volume overload that may have led to latent pulmonary edema. However, given our patient’s normal cardiac function, high urine output, and prolonged fasting period, we do not believe that pulmonary edema caused by intravascular overload was solely responsible for the severe perioperative hypoxemia.

With increasing use of Onyx as an adjunctive AVM therapy, anesthesiologists need to be aware of perioperative pulmonary migration as a potential significant contributor to hypoxemia. Similarly, with increasing frequency of awake craniotomy for tumor resection, previous tumor embolization with Onyx may also be encountered for a variety of lesions.13 Perhaps embolization occurs more frequently than appreciated, with its physiological effects masked by general anesthesia and mechanical ventilation. However, as awake surgical techniques are increasingly employed, this phenomenon merits consideration in the differential diagnosis of perioperative hypoxemia in AVM resection patients.2

Back to Top | Article Outline


Name: Brian T. Tolly, MD.

Contribution: This author was the anesthesiology resident for the case, helped with literature review for case preparation, and manuscript preparation.

Name: Jenna L. Kosky, MD.

Contribution: This author was the anesthesiology attending for case, and helped with manuscript preparation.

Name: Antoun Koht, MD.

Contribution: This author was the anesthesiology attending for case, and helped with manuscript preparation.

Name: Laura B. Hemmer, MD.

Contribution: This author was the assisting anesthesiologist for case, helped with the literature review for case preparation, and manuscript preparation.

This manuscript was handled by: Hans-Joachim Priebe, MD, FRCA, FCAI.

Back to Top | Article Outline


1. Meng L, Berger MS, Gelb AW. The potential benefits of awake craniotomy for brain tumor resection: an anesthesiologist’s perspective. J Neurosurg Anesthesiol. 2015;27:310317.
2. Gabarrós A, Young WL, McDermott MW, Lawton MT. Language and motor mapping during resection of brain arteriovenous malformations: indications, feasibility, and utility. Neurosurgery. 2011;68:744752.
3. Gamble AJ, Schaffer SG, Nardi DJ, Chalif DJ, Katz J, Dehdashti AR. Awake craniotomy in arteriovenous malformation surgery: the usefulness of cortical and subcortical mapping of language function in selected patients. World Neurosurg. 2015;84:13941401.
4. Bruno CA Jr., Meyers PM. Endovascular management of arteriovenous malformations of the brain. Interv Neurol. 2013;1:109123.
5. Saeed Kilani M, Izaaryene J, Cohen F, et al. Ethylene vinyl alcohol copolymer (Onyx®) in peripheral interventional radiology: indications, advantages, and limitations. Diagn Interv Imaging. 2015;96:319326.
6. Asouhidou I, Katsaridis V, Meng L, et al. Desaturation during Onyx embolization. Br J Anaesth. 2010;105:385386.
7. Murugesan C, Saravanan S, Rajkumar J, Prasad J, Banakal S, Muralidhar K. Severe pulmonary oedema following therapeutic embolization with Onyx for cerebral arteriovenous malformation. Neuroradiology. 2008;50:439442.
8. Tawil I, Carlson AP, Taylor CL. Acute respiratory distress syndrome after Onyx embolization of arteriovenous malformation. Crit Care Res Pract. 2011;2011:918185
9. Thiex R, Williams A, Smith E, Scott RM, Orbach DB. The use of Onyx for embolization of central nervous system arteriovenous lesions in pediatric patients. Am J Neuroradiol. 2010;31:112120.
10. Pukenas BA, Satti SR, Bailey R, Weigele JB, Hurst RW, Stiefel MF. Onyx pulmonary artery embolization after treatment of a low-flow dural arteriovenous fistula: case report. Neurosurgery. 2011;68:E1497E1500.
11. Herman TE, Siegel MJ, Vachharajani A, Masand P, Cross D. Cerebral arteriovenous fistula to pulmonary artery Onyx embolization. J Perinatol. 2007;27:238240.
12. Crusio R, Ramachandran K, Ramachandran K, Kupfer Y, Tessler S. Superior vena cava syndrome caused by embolisation of liquid Onyx. BMJ Case Reports. 201110.1136/bcr.11.2010.3502
13. Ashour R, Aziz-Sultan A. Preoperative tumor embolization. Neurosurg Clin N Am. 2014;25:607617.
Copyright © 2016 International Anesthesia Research Society