The American Journal of Forensic Medicine and Pathology 1124: 149-153, 1990. e199) Raven Press, Lid. New York International Registry of Unusual Cases Barotrauma and Air Embolism in Hyperbaric Oxygen Therapy Helmut K. Wolf, M.p., Richard E. Moon, M.p., Phillip R. Mitchell, M.p., and Peter C. Burger, M.D. A 69-year-old woman underwent hyperbaric oxygen ther- apy because of a nonhealing ulcer of her foot. During decompression, she developed a left-sided hemiplegia and confusion. Recompression resulted in transient neu- rologic improvement, but she eventually became coma- tose. Ventricular dysrhythmias developed and she died without regaining consciousness 17 h after onset of symp- toms. An autopsy revealed diffuse interstitial pulmonary fibrosis with severe paracicatricial emphysema, chronic interstitial inflammation, and chronic bronchitis with abundant intrabronchial mucus. There was extensive multifocal ischemic injury of the cerebral cortex. The hippocampi. basal ganglia. and cerebellum were spared. Scattered acute myofiber necrosis was present in the heart. Clinical presentation and autopsy findings strongly support the diagnosis of air embolism and illustrate a potential risk of hyperbaric oxygen therapy in patients with preexisting pulmonary disease. Key Words: Pulmonary barotrauma—Hyperbaric oxygen therapy—Air embolism—Ischemic cortical brain injury. From the Department of Pathology (H.K.W.. P.C.B.), the Department of Anesthesiology, Division of Hyperbaric Medicine (R.E.M., P.R.M.), and the Department of Medicine (R.E.M.). Duke University Medical Center, Durham, North Carolina. Address correspondence and reprint requests to Helmut K. Wolf, M.D., Box 3712, Department of Pathology, Duke Univer- sity Medical Center. Durham, NC 27710, U.S.A, Barotrauma of the lung is defined as rupture of pulmonary parenchyma secondary to increased in- trapulmonary air pressure. This is a serious and frequently lethal hazard to divers that is often seen in the presence of obstructive airway disease when a ball valve effect prevents the egress of expanding air during ascent. Acute and chronic bronchitis, emphysema, bronchiectasis, asthma, pulmonary fi- brosis, foreign bodies, tumors, granulomas, and congenital cysts are reported causes of such air trapping that may lead to barotrauma in divers (1.2). The rupture of pulmonary parenchyma in baro- trauma may cause pneumothorax if the visceral pleura is involved, but barotrauma may also result in passage of air from alveoli into pulmonary capillaries and veins. After passing through the left side of the heart and embolizing via systemic arteries, the air bubbles may block the micro- circulation and cause ischemic injury. The brain and the heart are the organs most sensitive to this occlusive effect of air emboli, and functional dis- turbances may occur immediately following embo- lization. Barotrauma with dissection of intraalveolar air into blood vessels due to rupture of pulmonary parenchyma must be distinguished from decom- pression sickness. In decompression sickness, gas bubbles form in blood and other tissues due to the release of dissolved nitrogen or other inert gases in rapid decompression. Decompression sickness does not occur during hyperbaric oxygenation be- cause of the lack of dissolved nitrogen or other inert gases, since the patient breathes pure oxygen. Although pulmonary barotrauma with subsc- 150 HK. WOLF ETAL. quent air embolism is most frequently observed in divers, a recent case studied here illustrates that severe preexisting pulmonary disease may also lead to barotrauma with arterial air embolism during decompression in hyperbaric oxygen treatment. CASE REPORT A 69-year-old white woman was admitted to a hospital with a nonhealing ulcer of the left foot and severe peripheral vascular disease. She was also known to have steroid-dependent chronic obstruc- tive lung disease. asthma, episodic hemoptysis, bronchiectasis, and coronary artery disease. De- bridement and autologous skin grafting of the ulcer were performed. Hyperbaric oxygen therapy in a monoplace hyperbaric chamber was administered twice daily. The patient tolerated the first 19 treat- ments well. However, following her 20th oxygen exposure [30 min at 60 ft (18 m) seawater pressure], she became less responsive during decompression [1 ft (0.3 m)/min]. Evaluation after removal from the hyperbaric chamber revealed left hemiplegia and left facial weakness. The patient was mildly con- fused although she was cooperative and answered questions. Radiologic examination did not reveal a pneumothorax. Within 30-45 min from onset of symptoms, the patient was returned to the hyper- baric chamber and repressurized to 60 ft pressure, which resulted in improvement of her left hemiple- gia with ability to raise the left arm and the left leg. Decompression to 30 ft (9 m) again resulted in confusion and inability to raise the left arm and leg. Repressurization to 60 ft revealed initial improve- ment, but the patient complained of temporal head- aches and became increasingly confused. The hyperbaric therapy was then terminated and de- compression was performed, Evaluation after de- compression revealed a dense left hemiplegia and decreased responsiveness. Her heart rate was 95- 100/min with a blood pressure of 150/100 mm Hg. Ventricular tachycardia was noted and lidocaine therapy was initiated. The patient was transferred to this hospital via helicopter for treatment in a multiplace hyperbaric chamber. Upon arrival, she was unresponsive, but hemodynamically stable. Recompression to 60 ft with 100% oxygen did not improve her neuro- logic status. During decompression, she developed multiple episodes of ventricular tachycardia and progressive hypotension requiring defibrillation, antiarrhythmic drugs, and pressor agents. After removal from the hyperbaric chamber, the patient was transferred to the medical intensive care Am J Forensic Med Pathol, Vol. 11, No. 2, 1990 unit where she again developed ventricular arrhyth- mias that were refractory to therapy. The patient was pronounced dead ~ 45 min after the hemo- dynamic deterioration occurred and ~ 17 h after onset of neurologic symptoms. An autopsy was performed. POSTMORTEM FINDINGS At autopsy, great care was taken to identify intravascular air bubbles. Postmortem fundoscopy was performed in order to detect possible bubbles in retinal vessels. however no intravascular air bubbles or foamy blood indicative of air embolism were identified in the vessels of the brain, heart, or other organs. Examination of the lungs revealed diffuse interstitial fibrosis with a prominent honey- comb pattern involving all lobes to approximately the same degree and chronic interstitial inflamma- tion, There was severe paracicatricial emphysema with large bullae measuring up to 4 cm in greatest dimensions throughout all lobes (Fig. la). Addition- ally, there was marked chronic bronchitis. Many of the bronchi and bronchioli were occluded by tena- cious plugs of mucus (Fig. Ib). No increase in eosinophils was present and no Charcot-Leyden crystals, or Curschmann’s spirals was present. There was no pulmonary hemorrhage. The brain was macroscopically unremarkable ex- cept for a small focus of atherosclerosis with 75% narrowing of the left middle cerebral artery just proximal to the trifurcation. Microscopic examina- tion revealed extensive acute ischemic injury of the cerebral cortex involving all lobes. The injury consisted of numerous small, well-circumscribed round, oval, or irregular foci showing vacuolization of the neuropil and neuronal nuclear pyknosis. The lesions were surrounded by unremarkable brain tissue and were distributed in a peculiar geographic pattern, frequently in relation to small blood ves- sels, many of which showed prominent perivascular spaces (Fig. 2). The injury primarily involved the gray matter with only focal extension into the subcortical white matter. The changes were some- what more severe in the right motor cortex com- pared with the left. The right and left hippocampi, brainstem, and cerebellum were unremarkable. The lesions were anemic and there were no hemor- rhages. The spinal cord was unremarkable. The coronary arteries revealed multifocal severe atherosclerosis with ~99% narrowing of the right coronary artery and ~80% narrowing of the left anterior descending coronary artery. The left ven- tricle showed patchy old and organizing myocardial BAROTRAUMA AND AIR EMBOLISM 151 fibrosis. No confluent infarcts were present. Addi- tionally, there were widely scattered foci of acute myofiber necrosis with some myofibers showing contraction bands. There was no acute inflamma- tory reaction. Other findings included moderate systemic ath- erosclerosis, severe atheroscleros ulcer of the left foot, left nephrolithiasis, status post right nephrectomy, severe generalized osteoporo- sis, and multiple recent rib fractures (status post resuscitation). DISCUSSION Barotrauma and air embolism have received con- siderable attention in diving medicine, but, to our knowledge, no such case has yet been documented as a result of hyperbaric oxygen therapy. although “* (AA FIG. 1. Pulmonary pathology underlying air trapping and barotrauma in this case. (a) Hilar section of the left lung after formalin fixation. Honeycomb pattern with severe paracica- triceal emphysema and large bullae in both lobes. (b) His- tologic section of the right lower lobe. Abundance of mucus in small airways. Marked interstitial fibrosis and chronic inter- Stitial inflammation. Hematoxylin—-eosin, x 40. the absolute pressure changes are the same in divers and hyperbaric oxygenation. There are sev- eral possible reasons for this discrepancy. In the first place, a slower rate of decompression is used after hyperbaric oxygen therapy than that used by ascending divers. This allows more time for gas exchange and helps to prevent barotrauma. Sec- ondly, in divers, panic situations may induce laryn- gospasm that prevents exhalation during ascent and results in barotrauma (1). Laryngospasm is less likely to occur in patients in the hyperbaric cham- ber. Thirdly, during expiration, small airways close earlier if the body is immersed in water compared with nonimmersed individuals (3). Therefore, baro- trauma due to expansion of entrapped gas in the terminal respiratory unit is more likely to occur in divers who are immersed in water compared with patients being treated in the hyperbaric chamber. Am J Forensic Med Pathol, Vol. 11, No, 2, 1990 152 H. K. WOLF ET AL. FIG. 2. Superior right precentral gyrus. Well-circum- scribed perivascular focus of cortical necrosis with nuclear pyknosis of neurons, vacuolization of the neuropil, and prominent perivascular spaces. Luxol—fast blue/hematox- ylin-eosin, x 40. Finally. hyperbaric therapy is administered under the supervision of personnel trained to identify those patients at risk for air embolism, and to intervene carly with potentially life-saving recom- pression therapy. In patients who died several hours after the embolic event for several hours, the postmortem diagnosis of air embolism depends on the clinical presentation and morphologic changes. Intravascu- lar air bubbles will have dissolved during this period of time (4), especially if the patient was subject to repeated recompression therapy as in this case. Acute onset of hemiplegia and altered mental status during or shortly after decompression from a hyper- baric environment as seen in our case is a typical clinical presentation of pulmonary barotrauma with air embolism to the brain (5). Initial improvement under recompression therapy with subsequent worsening or recurrence of symptoms and the de- velopment of coma as seen in this case are also frequently observed (5.6). Reports of neuropathologic findings in patients Am J Forensic Med Pathol, Vol. 11, No, 2, 1990 with air embolism are limited in number because many patients die before characteristic histologic changes develop, but the results of numerous ex- perimental studies in animals are available. The brain lesions in our case closely resemble the changes reported in the literature. In combination with the clinical presentation, they strongly support the diagnosis of air embolism in the present case, although there are no histologic findings that are absolutely specific for air embolism (7-9). Multiple small, well-circumscribed foci of corti- cal necrosis, as were seen in the present case, have been described in human brains (10-13) and in experimental studies in animals (14-17) and seem to be fairly typical for air embolism. The foci of necrosis are round, oval, or irregular and are often related to small blood vessels as was noted also in the present case. Other areas of the same brain may show a laminar pattern of necrosis (10,11). Simi- larly, one may see diffuse necrosis of the cerebral cortex with varying degrees of neuronal changes in different areas. Again, the areas with the most advanced ncuronal damage may be associated with small blood vessels, resulting in an irregular geo- graphic pattern (7). Prominent perivascular spaces may be present microscopically (7,18-20). As in our case, the hippocampal formation that is usually involved in generalized hypoxia or hypotension may be entirely spared in air embolism (10,12,13). This may reflect a tendency of emboli to involve the middle cerebral arteries rather than the posterior cerebral arteries that supply the hippocampi. Dif- ferences in the involvement of different vascular distributions in air embolism have also been ex- plained by the position of the patient during the embolic event (13). The observation of more severe changes in the right side of the brain compared with the left side as scen in our case may be due to the anatomy of the aortic arch. As the brachiocephalic artery on the right is the first large branch of the aorta, air bubbles may preferably embolize to this vessel (13). As in our case, there may not be any hemorrhage (13.14.17). In other cases, occasional small foci of hemorrhage (10,16,21) or numerous perivascular hemorrhages (7,$2,17.22) are seen. Studies in dogs have shown that rapid injection of air into the carotid artery results in numerous perivascular hemorrhages whereas the same vol- ume of air injected slowly results in cortical necro- ses but not in hemorrhages (17). These studies suggest that the presence or absence and the degree of hemorrhage depend on the time course of the embolic events and perhaps the size of the air bubbles. BAROTRAUMA AND AIR EMBOLISM 153 In animals, multifocal necrosis has been demon- strated in the myocardium of the left ventricle if the time of survival after air embolism was of sufficient length (21). In our case, the temporal correlation between neurologic symptoms and ventricular dys- rhythmias suggests that the air embolized not only to the cerebral vessels, but also to the coronary arteries, which were already markedly narrowed by atherosclerotic disease. This assumption is sup- ported by the presence of multiple foci’ of acute myofiber necrosis at autopsy. It can not be ruled out, however, that these changes were caused by the terminal resuscitation efforts with high-dose pressor support, chest compressions, and defibril- lation. REFERENCES 1. Novomesky F, Ehm OF. Tédliche Luftembolie beim Tauchen. Eine morphologische Untersuchung. Z Rechtsmed 1985:95: 105-11. . Smith FR. Air embolism as a cause of death in scuba diving in the pacific northwest. Dis Chest 1967.5 3. Prefaut Ch, Dubois F, Roussos Ch, tral-Marques R, Macklem PT, Ruff F. Influence of immersion to the neck in water on airway closure and distribution of perfusion in man. Respir Physiol 1979;37:313-23. 4. Walcher K. Uber die gerichtlich-medizinische Beurteilung der Luftembolie im kleinen und grossen Kreislauf mit beson- derer Beriicksichtigung der cerebralen Luftembolie. Disch Z Gesamte Gerichtl Med 1925;5:561-73. 5. Ah-See AK. Review of arterial air embolism in submarine escape. In: Smith G. ed. Proceedings of the sixth interna- tional congress on hyperbaric medicine 1977, Aberdeen: Aberdeen University, 1977:349-51. 6. Pearson RR, Goad RF. Delayed cerebral edema complicat- ing cerebral arterial gas embolism: case histories. Undersea Biomed Res 1982;9:283-96. 7. Weissenrieder M. Histologische Verinderungen bei cere- ie) . Hendry WT. Autopsy appearances in braler Luft- und Fettembolie [Inaugural dissertation}. Munich: University of Munich, 1934. . Kohn K. Zum Nachweis der arteriellen Luftembolie des Gehirns. Dixch Z Gesamte Gerichtl Med 1983:42:301-7. . Brierley JB. Neuropathological findings in patients dying after open-heart surgery. Thorax 1963.18:29 1-304, . Neubiirger K. Uber cerebrale Fett- und Luftembolie. Z Gesamte Neurol Psychiatr 1925.95;279-318. . Lhermitte J. Barrelet. Embolic gazeuse cérbrate d'origine périphérique: étude anatomique. Rev Neurol (Paris) 1934: 62:85 1-7, . Ameuille P, Lhermitte J. Kudelski. Embolie gazeuse con- sécutive a une insufflation pleural mollissement cérébral massif. Rey Neurol (Paris) 1935;63:867-76, . Brion S, Psimaras A, Gallissot MC. Neuropathologie de Vembolie gazeuse humaine au cours da la chirurgie car- diaque. In: Artel G. Naquet R, eds. L’embolie gazeuse du systéme carotidien. Paris: Doin, 1974: 194-8, . Spielmeyer W. Uber die anatomischen Folgen der Luftem- bolie ins Gehirn. Verh Dtsch Kongr Inn Med 1913:30:359- 62. . Nishimoto K, Wolman A, Spatz M, Klatzo I. Pathophysio- logic correlations in the blood-brain barrier damage due to air embolism. Adv Neurol 1978,20:237-44. » Schoenmackers J, Die Folgen intrakardialer Injektion Kleiner Luftmengen im Kaninchenversuch. Virchows Arch JA] 1950:318:48-60. . Fries CC. Levowitz B. Adler S. Cook AW. Karlson KE, Dennis C. Experimental cerebral gas embolism. Ann Surg 1957:145:461-70, . Roessle R. Uber die ersten Veriinderungen des menschli- chen Gehirns nach arterieller Luftembolie. Virchows Arch [A] 1948:315:461-80. . Hekmatpanah J. Cerebral microvascular alterations in arte- 3. iving accidents. In: Smith G. ed. Proceedings of the sixth international congress on hyperbaric medicine. Aberdeen: Aberdeen University, 1977:364-70. rial air embolism. Adv Neurol 1978 . Adelbahr G, Staak M. Morphologischer Beitrag zur Frage der Verbrauchscougulopathie bei Luftembolie. Virchows Arch JA] 1969:346:224-38. . Jansen W. Zur Pathogenese und forensischen Bewertung von Hirnblutungen nach cerebraler Luftembolie. Diseh Z Gesanute Gerichtl Med 1967:61:62-80. Am J Forensic Med Pathol, Vol. HL. No. 2, 1990