96

Neuroanesthesia and Intensive Care Cerebral arterial gas embolism following diagnos- tic bronchoscopy: delayed treatment with hyper- baric oxygen [Embolie gazeuse de l’artère cérébrale suivant une bronchoscopie diagnostique: traitement différé avec l’oxygène hyperbare]

Chris G. Wherrett MDFRCPC,* Reza J. Mehran MDFRCSC,† Marc-Andre Beaulieu MDFRCPC‡

Purpose: To describe a clinical scenario consistent with the diag- Objectif:Décrire un scénario clinique caractéristique du diagnostic nosis of cerebral arterial gas embolism (CAGE) acquired during an d’embolie gazeuse de l’artère cérébrale (EGAC) survenue pendant outpatient bronchoscopy. Our discussion explores the mechanisms une bronchoscopie réalisée en clinique externe. Notre discussion and diagnosis of CAGE and the role of hyperbaric oxygen therapy. explore les mécanismes et le diagnostic d’EGAC et le rôle de Clinical features: A diagnostic bronchoscopy was performed on a l’oxygénothérapie hyperbare. 70-yr-old man who had had a lobectomy for bronchogenic carci- Éléments cliniques:Une bronchoscopie diagnostique a été réalisée noma three months earlier. During the direct insufflation of oxygen chez un homme de 70 ans qui avait subi, trois mois auparavant, une into the right middle lobe bronchus, the patient became unrespon- lobectomie pour un carcinome bronchique. Pendant l’insufflation sive and developed subcutaneous emphysema. Immediately, an directe d’oxygène dans la bronche du lobe droit moyen, le patient est endotracheal tube and bilateral chest tubes were placed with resul- devenu comateux et un emphysème sous-cutané s’est développé. On tant improvement in his oxygen saturation. However, he remained a immédiatement placé un tube endotrachéal et des drains tho- unresponsive with extensor and flexor responses to pain. Later, in raciques bilatéraux, ce qui a amélioré la saturation en oxygène. the intensive care unit, he exhibited seizure activity requiring anti- Cependant, le patient est demeuré comateux et réagissait à la convulsant therapy. Sedation was utilized only briefly to facilitate douleur par des mouvements de décérébration et de décortication. controlled ventilation. Investigations revealed a negative computer- Par la suite, à l’unité des soins intensifs, on a enregistré une activité ized tomography (CT) scan of the head, a normal cerebral spinal épileptique qui a nécessité une thérapie anticonvulsive. La sédation, de fluid examination, a CT chest that showed evidence of barotrauma, courte durée seulement, a visé à faciliter la ventilation contrôlée. Les and an abnormal electroencephalogram. Fifty-two hours after the examens ont révélé une tomodensitométrie négative du crâne, un event, he was treated for presumed CAGE with hyperbaric oxygen examen normal du liquide céphalo-rachidien, un barotrauma pul- using a modified United States Navy Table 6. Twelve hours later he monaire mis en évidence par une tomographie, un électroencéphalo- had regained consciousness and was extubated. He underwent gramme anormal. Cinquante-deux heures après l’incident, il a été two more hyperbaric treatments and was discharged from hospi- traité pour une EGAC présumée avec de l’oxygène hyperbare en uti- tal one week after the event, fully recovered. lisant une Table 6 modifiée de la Marine américaine. Douze heures Conclusion: A patient with presumed CAGE made a complete plus tard, il a repris conscience et a été extubé. Il a reçu deux autres recovery following treatment with hyperbaric oxygen therapy even traitements hyperbares et, une semaine après l’événement, il a quit- though it was initiated after a significant time delay. té l’hôpital complètement rétabli. Conclusion:Un patient, victime d’une EGAC présumée s’est com- plètement rétabli à la suite d’une oxygénothérapie hyperbare même si ce traitement a été différé de manière significative.

From the Departments of Anesthesiology,* Surgery,† and Medicine,‡ The Ottawa Hospital Hyperbaric Unit, Ottawa, Ontario, Canada. Address correspondence to: Dr. Chris Wherrett, Department of Anesthesiology, The Ottawa Hospital, General Campus, 501 Smyth Road, Ottawa, Ontario, K1H 8L6, Canada. Phone: 613-737-8187; Fax: 613-737-8189; E-mail: [email protected] No financial support was used for this study. Accepted for publication March 9, 2001. Revision accepted June 4, 2001.

CAN J ANESTH 2002 / 49: 1 / pp 96–99 Wherrett et al.: ARTERIALGASEMBOLISMANDHYPERBARICOXYGEN 97

EREBRALarterial gas embolism (CAGE) alized epileptiform discharges (PLEDS). A computer- is a rare but potentially fatal event. The ized tomography (CT) scan of the head was normal. exact incidence is unknown but neurolog- Serum creatine kinase (CK) was elevated at 281 U·L–1. Cic sequelae are thought to occur in one Two days after the bronchoscopy the patient third of patients who experience CAGE after cardiac remained unconscious. An examination of the cere- surgery.1 In addition to resuscitative and supportive brospinal fluid was done and was normal. CT chest care, hyperbaric oxygen therapy (HBO2) is a thera- showed a residual loculated right pneumothorax, peutic option for CAGE. We report a case of severe pneumomediastinum, and subcutaneous emphysema. neurological impairment associated with barotrauma While the possibility of CAGE had previously been during a diagnostic bronchoscopy. Hyperbaric oxygen entertained, at this time there was a clear mechanism was initiated two days following the event. for CAGE to have occurred and no apparent alterna- tive explanation for the patient’s condition. Therefore, Case report hyperbaric oxygen therapy was requested and com- A 70-yr-old man underwent a diagnostic bron- menced approximately 52 hr after the respiratory choscopy three months following a right upper lobec- arrest. The chamber is a dualplace, capable of com- tomy for bronchogenic cancer. He had a history of pression to 66 feet of sea water (fsw). Attendants hypertension, smoking, and chronic obstructive pul- accompany critically ill patients and ivmedications monary disease. In an attempt to expand atelectasis of may be infused from outside the chamber. The patient the right middle lobe, oxygen from the hospital was treated with a U.S. Navy Table 62with extensions pipeline 12 L·min–1at 50psi (345 kPa) was insufflat- at both 60 fsw and 30 fsw (284 and 193 kPa respec- ed into the right middle lobe bronchial orifice. During tively). Total treatment time was seven hours and 25 the procedure, the patient developed severe respirato- min. During the course of the treatment the dose of ry distress and acute oxygen desaturation. He became propofol was increased on several occasions to control unresponsive and subcutaneous emphysema was patient agitation. There were no adverse effects noted noted. The bronchoscopist promptly intubated the in the patient or attendants. patient’s trachea and inserted bilateral chest tubes. Gas Ninety minutes after completion of hyperbaric was found in the left pleural cavity. Oxygen saturation therapy, the patient could obey commands. Ten hours immediately improved. He remained hemodynamical- later he was alert and his trachea was extubated. He ly stable throughout the event and was transferred to was oriented to place but not time. He was treated the intensive care unit (ICU). with hyperbaric oxygen two more times at 2.5 atmos- Seven hours later, he remained unconscious and pheres absolute (253 kPa). One week later he was dis- withdrew to pain only. Until then he had received fen- charged home. Two weeks later his complete tanyl 50 µg, morphine 4 mg, and midazolam 4 mg. A neurologic recovery was verified. propofol infusion of 10–15 µg·kg·min–1was started at this time to facilitate mechanical ventilation and mini- Discussion mize airway pressures. At this time CAGE was not CAGE most often occurs in two situations, SCUBA considered in the differential diagnosis. (self contained underwater breathing apparatus) div- The next day there were flexor and extensor ing or medical procedures. Gas embolization requires responses to pain in the extremities. Brainstem reflex- the presence of two factors: a portal of entry into a es were intact. The propofol infusion (10 µg·kg·min–1) blood vessel and a pressure gradient to drive gas into was temporarily discontinued. Forty-five minutes the blood vessel. The gas can be entrained from the later, he had a transient episode of ventricular tachy- atmosphere by negative pressure within a vein (for cardia followed by both generalized and right-sided example the jugular vein in the sitting position). Gas seizures. This occurred during routine endotracheal can also be forced into a vessel by positive pressure. tube care involving manual ventilation of the lungs. Gas may enter the arterial circulation directly or from Treatment with thiopentone, midazolam, and dilantin the venous circulation via a right-to-left cardiac shunt was initiated. The patient was immediately turned on or intrapulmonary shunt. Passage of bubbles through his right side and the head of the bed was lowered, the pulmonary capillaries to the left heart is well rec- and he was given 100% oxygen. The seizure activity ognized. Embolization of even small amounts of gas and the ventricular tachycardia resolved. Fundoscopic to the brain can produce neurologic damage thereby examination was normal and there were no localizing producing the syndrome of CAGE. signs. An electroencephalogram (EEG) showed a dif- In SCUBA divers, aside from trauma or other fuse disturbance of cerebral activity and periodic later- forms of decompression sickness, there are usually no 98 CANADIANJOURNALOFANESTHESIA factors to confound the diagnosis of CAGE. ment in rabbits has shown a protective effect from pro- Symptoms typically develop within 20 min of surfac- phylactic heparin.13 ing. As well, there is often accompanying evidence of There are several proposed beneficial effects of pulmonary barotrauma (pneumothorax, pneumome- HBO2 in CAGE. Firstly, increased ambient pressure diastinum, subcutaneous emphysema, and interstitial directly compresses the gas bubbles. Bubble volume emphysema). Iatrogenic gas emboli, however, are and therefore diameter is reduced in accordance with more elusive. They tend to occur during diagnostic Boyle’s Law (P1*V1=P2*V2). Secondly, a resulting and therapeutic procedures involving vascular cannu- increase in the bubble surface area to volume ratio lation or intracavitary gas insufflation. Patients experi- allows more rapid resorption of nitrogen from within encing iatrogenic emboli are more likely to have the bubble. Finally, volume reduction is hastened by the concurrent medical problems such as cerebrovascular diffusion of nitrogen down a steeper bubble-plasma disease, cardiovascular disease, and metabolic disor- partial pressure gradient. All three mechanisms promote ders. These factors, as well as pharmacological seda- restoration of distal blood flow. A reduction of bubble- tion, can all produce neurological impairment and endothelial interface results in less “foreign body” therefore expand the differential diagnosis. exposed to the vascular endothelium. This decreases The presenting symptoms and signs of CAGE can vascular endothelial injury and therefore reduces the be generalized or focal. Cardiopulmonary arrest, “no reflow” phenomenon, which is believed to be a altered level of consciousness, seizures, hemiparesis, cause of delayed symptom onset or failure of HBO2 hemianopsia, paresthesiae, pupillary asymmetry, dizzi- treatment.5,6Decreased reperfusion injury is thought to ness, nausea, and headache are all reported.1,3–5 result from reduced leukocyte aggregation.5A high dis- Unfortunately, these features are non-specific and can solved oxygen content allows better oxygen delivery to be attributed to other etiologies. Retinal gas bubbles neural tissue compromised by marginal blood flow (the on retinoscopy and Liebermeister’s sign (a blue or “ischemic penumbra”).1,5As well, cerebral edema can 6 pale tongue secondary to lingual artery emboli) are be reduced by HBO2. pathognomonic of arterial gas embolism but are Hyperbaric oxygen is recommended for CAGE12 extremely rare signs.6 although there are no randomized controlled trials A CT scan of the head is usually not helpful. Dexter reported in humans. Studies in cats show significant et al.7 demonstrated that only macroscopic bubbles somatosensory evoked potential recovery following

(1.3 mm radius) would be identifiable and only if the induced CAGE when HBO2 is given. This compares to CT cuts, usually at 1-cm intervals, coincidentally a poor recovery in animals kept at atmospheric pressure intersect the appropriate level. Therefore, absence of breathing air.11,14 In humans, two early studies compar- air on CT scan does not exclude the diagnosis. ing mortality rates suggest a beneficial effect. Magnetic resonance imaging, single photon emission Retrospective data published in 1964 showed a decrease computed tomography scanning, or Xenon enhanced in mortality from 93% with no treatment to 33% with CT are only helpful in delineating regional blood conventional aggressive treatment (left lateral decubitus flow.8The finding of gas in an arterial blood sample is position, vasopressors, and oxygen by positive pressure).8 specific but infrequent.6Elevated CK levels have been A later study showed a mortality rate of only 7% in 30 noted in divers with CAGE. Elevation of serum patients treated when hyperbaric oxygen was utilized.8 gamma-glutamyl transpeptidase, aspartate amino- In our patient, there were two possible mechanisms transferase, and lactate dehydrogenase can also occur of CAGE. Firstly, oxygen administered under pressure in CAGE or decompression sickness.4 during the bronchoscopy caused significant barotrauma Initial management of CAGE includes immediate manifested by pneumothoraces, subcutaneous emphy- respiratory and hemodynamic support. Trendelenberg sema, and pneumomediastinum. The oxygen could positioning may prevent movement of arterial gas then have passed into the pulmonary venous circula- cephalad. Although it is frequently recommended, tion, resulting in arterial embolization. As well, the jet there is no evidence that left lateral decubitus position- of oxygen could have entrained air present in the alve- ing improves outcome with venous gas embolism.9 oli. Secondly, the barotrauma may have created a per- Tight control of serum glucose is advocated as this may sistent open pathway for gas to travel from alveoli into limit the extent of neuronal damage in areas of the circulation. Thus the patient may have experienced ischemia.8Lidocaine has been shown in animal models ongoing gas embolization during mechanical ventila- to improve recovery of somatosensory evoked poten- tion after the initial event. Nitrogen gas is more likely tials following CAGE.10,11There is no proven benefit of than oxygen to persist in the circulation because it is not steroids and they are not recommended.12An experi- metabolized or bound by hemoglobin. Wherrett et al.: ARTERIALGASEMBOLISMANDHYPERBARICOXYGEN 99

The diagnosis of CAGE is primarily based on clini- Company, 1993. Chapter 8. cal features. As described, current diagnostic tests for 3 Dunbar EM, Fox R, Watson B, Akrill P. Successful late CAGE are likely not sensitive enough, particularly treatment of venous air embolism with hyperbaric oxy- when performed hours after the event. However, in gen. Postgrad Med J 1990; 66: 469–70. our patient, the following clinical features were con- 4 Pao BS, Hayden SR. Cerebral gas embolism resulting sistent with this diagnosis: impairment of conscious- from inhalation of pressurized helium. Ann Emerg ness immediately following massive barotrauma, Med 1996; 28: 363–6. seizures, an EEG with PLEDS which are indicative of 5 Pereira P.A fatal case of cerebral artery gas embolism acute cerebrovascular dysfunction,1,3–5 an elevated following fine needle biopsy of the lung. Med J Aust CK, and the complete recovery of the patient. There 1993; 159: 755–7. were no metabolic, infectious, or structural causes 6 Jain KK. Cerebral air embolism. In: Jain KK (Ed.). identified that could explain the clinical findings. The Textbook of Hyperbaric Medicine, 2nded. Kirkland patient did not have a cardiac arrest nor was there any WA: Hogrefe & Huber Publishers, 1996: 137–45. evidence of a post anoxic-ischemic encephalopathy. 7 Dexter F, Hindman BJ. Recommendations for hyper- The propofol sedation is a possible confounding fac- baric oxygen therapy of cerebral air embolism based on tor. However, it was not started until seven hours after a mathematical model of bubble absorption. Anesth the event and was maintained at a relatively low dose Analg 1997; 84: 1203–7. that does not account for the patient’s degree of neu- 8 Layon AJ.Hyperbaric oxygen treatment for cerebral air rologic impairment. As well, the EEG did not show embolism – Where are the data? (Editorial). Mayo Clin the characteristics of propofol sedation.15 Proc 1991; 66: 641–6. There are several factors that support our impres- 9 Alvaran SB, Toung JK, Graff TE, Benson DW.Venous air embolism: comparative merits of external cardiac sion that HBO2was therapeutic in this case. During the course of treatment the dose of propofol had to be massage, intracardiac aspiration, and left lateral decubi- increased. This suggests that the patient’s level of con- tus position. Anesth Analg 1978; 57: 166–70. sciousness improved during treatment, a finding 10 Dutka AJ, Mink R, McDermott J, Clark JB, Hallenbeck reported by others.1,3The patient’s neurologic status JM. Effect of lidocaine on somatosensory evoked remained dramatically improved after treatment with response and cerebral blood flow after canine cerebral no other explanation. There had been no signs of air embolism. Stroke 1992; 23: 1515–21. improvement over the preceding two days. 11 McDermott JJ, Dutka AJ, Evans DE, Flynn ET. There are several other reports of successful out- Treatment of experimental cerebral air embolism with comes after late hyperbaric treatment up to 30 hr after lidocaine and hyperbaric oxygen. Undersea Biomed CAGE.2,3,16An effect of late treatment is supported by Res 1990; 17: 525–34. evidence that bubbles may persist in the cerebral cir- 12 Muth CM, Shank ES. Gas embolism.New Engl J Med culation for 40 hr or longer.8 2000; 342: 476–82. In summary, we present a previously unreported 13 Ryu KH, Hindman BJ, Reasoner DK, Dexter F. mechanism of pulmonary barotrauma followed by Heparin reduces neurological impairment after cerebral impaired consciousness likely due to CAGE. There arterial air embolism in the rabbit. Stroke 1996; 27: 303–10. was a significant delay in commencing HBO2yet the 14 McDermott JJ, Dutka AJ, Koller WA, Flynn ET.Effects patient made a complete recovery. HBO2is the only available treatment that can reverse the underlying of an increased PO2during recompression therapy for pathophysiology of CAGE. We suggest that urgent the treatment of experimental cerebral arterial gas embolism. Undersea Biomed Res 1992; 19: 403–13. HBO2 therapy be considered in any patient when CAGE is suspected, even after a significant time inter- 15 Tomoda K, Shingu K, Osawa M, Murakawa M, Mori K. val of one to two days. Comparison of CNS effects of propofol and thiopen- tone in cats. Br J Anaesth 1993; 71: 383–7. References 16 Mader JT, Hulet WH.Delayed hyperbaric treatment of 1 Armon C, Deschamps C, Adkinson C, Fealey RD, cerebral air embolism: report of a case. Arch Neurol Orszulak TA.Hyperbaric treatment of cerebral air 1979; 36: 504–5. embolism sustained during an open-heart surgical pro- cedure. Mayo Clin Proc 1991; 66: 565–71. 2 U.S. Navy Diving Manual Volume 1 (Air Diving) - Published by the Direction of Commander, Naval Sea Systems Command. Flagstaff, AZ: Best Publishing