HOWEVER rare, complications likely to occur in patients with pulmonary cysts may be life threatening. Any modification of ambient pressure is likely to induce pulmonary bulla to change in size or, if not possible, to rupture or leak air. We describe a case of severe cerebral air embolism occurring during an aircraft flight in a patient with a pulmonary cyst. A successful treatment with use of hyperbaric therapy was administered at the 48th hour.
A 43-yr-old man who smoked and had asthma experienced a sudden loss of consciousness followed by dyspnea and agitation 20 min after take-off of a commercial flight in a Boeing 737. On examination by a medical doctor on the plane, the man was pale and sweating, his blood pressure was 120/80 mmHg, and his heart rate was regular at 80 beats/min. A short-term episode of asthma was first considered, but physical examination did not reveal any wheezing, and the neurologic symptoms could not be explained by an exacerbation of asthma. Therefore, the flight was diverted toward an airport for emergency care of the patient.
On admission to the intensive care unit, the patient had a Glasgow Coma Score of 7, and was intubated and mechanically ventilated. Further neurologic examination revealed symmetric pupils that were reactive to light; left hemiparesis; and upper motor neuron dysfunction with increased muscle tone, brisk deep tendon reflexes, and Babinski sign. Tendon reflexes were symmetric at the lower limbs. Painful stimuli produced right decorticate posturing. No seizure activity was observed at any time, and the physical examination results were otherwise normal except for a core temperature of 38.5°C. No physician could recollect the presence of signs of air embolism in the skin, such as livedo reticularis, at that time.
A radiograph of the chest and a computed tomography (CT) scan showed a 12 × 8 × 10-cm right lower lobe bulla, with a thin hyperdense wall but no parenchymal infiltration (fig. 1). A cerebral CT image and an electrocardiogram were normal. Electroencephalography showed no focal signs.
Initial (i.e. , 8 h after the event) laboratory data showed signs of myocardial injury: The troponin Ic concentration was 9.3 μg/l (normal < 0.4 μg/l). Other laboratory data were within normal ranges.
The patient’s fever, neurologic status, and myocardial injury raised several diagnostic hypotheses: herpes meningoencephalitis, venous thrombosis of the cerebral sinus, or coma after heart attack. The results of all of the examinations performed to confirm these diagnoses, including lumbar puncture, herpes polymerase chain reaction, and a new cerebral CT scan with contrast injection, were negative.
Finally, because of the presence of the lower pulmonary lobe bulla, a diagnosis of air embolism involving the brain and the heart was made. The most important criterion to settle the diagnosis of arterial gas embolism was the patient’s medical history.1The temporal relation between the neurologic symptoms and the pressure variation seemed compatible with a sudden entry of air into the pulmonary vessels. A normal CT scan did not rule out the occurrence of arterial gas embolism. Therefore, a recompression treatment started 48 h after the initial accident in a hyperbaric chamber according to a D50 HeO2table. This involved immediate compression to 50 m seawater or 6 ATA and lasted 8 h while the patient was ventilated with pure oxygen or a mix of oxygen and helium adjusted to keep the partial pressure of oxygen less than the toxic level of 3.0 ATA (table 1). Before the first dive, a catheter was inserted into the bulla under radiographic guidance to avoid the risk of rupture during decompression. At the end of this first compression, the patient was awakening but still showing clouding of consciousness. However, the motor response elicited by pain was now purposeful. The left hemiparesis had improved, and the muscle tone was normal. One more D50 HeO2compression was performed 12 h after the first one, and the patient fully recovered from his neurologic symptoms. Afterward, a thoracotomy with resection of the right lower lobe was performed before a final recompression to 15 m seawater (2.5 ATA) during 2 h with pure oxygen. There were no postoperative complications, and the patient was discharged 10 days later. A month later, magnetic resonance imaging (MRI) was performed, and no abnormalities were found.
Although pneumothorax is one of the possible complications in patients with giant intrathoracic bulla,2air embolism is rare.3However, the mechanisms are similar. Decreasing the ambient pressure leads to an increase in the volume of the bulla or, if not possible, to an increased pressure gradient between the bulla and the pleural space or the surrounding vessels. Such a variation can be observed during an aircraft flight when ambient pressure varies from sea level to an altitude of 2438 m corresponding to the usual cabin altitude pressure of a B737.4This pressure is equal to approximately 0.75 atm. It is reached within 20 min and may be modified during the flight because of traffic or weather.5Whatever the amplitude of variation of pressure, this induces an increase in volume of gas in a noncommunicating space. Therefore, any parenchymal bulla is a classic contraindication of activities leading to sudden variations of ambient pressure6and is a compelling reason to administer surgical treatment. The current patient was aware of the presence of a parenchymal bulla, but no special recommendations regarding air flight had been made.
Pulmonary cysts produce conditions that tend to cause arterial gas embolism in case of pressure variation inside the cyst. Air from the bulla may enter the pulmonary veins, and then bubbles of gas are directly distributed into all organs. Small emboli in muscles or viscera are clinically unnoticed, whereas cerebral or coronary localizations may bring about severe clinical manifestations. This arterial gas embolism is different from the venous gas embolism that may occur during some surgical procedures in which gas enters the systemic venous system. In venous gas embolism, the main clinical manifestations are due to the migration of bubbles to the pulmonary circulation and the subsequent increase of right ventricle afterload. However, even in venous embolism, bubbles may be encountered on the arterial side as a result of a patent foramen ovale or by moving across the pulmonary vascular bed.1The consequences are twofold: First, local ischemia and edema due to the absence of blood supply may occur. Second, the interface between the gas bubbles and the endothelium engenders a local inflammatory response, possibly disrupting the blood–brain barrier, worsening edema and further impairing cerebral perfusion.7This physiopathology explains why the main initial clinical manifestations are the neurologic symptoms, whether associated with circulatory disorders or not.
There is no definite nonclinical criterion for air embolism. A CT scan may reveal small air bubbles, but because of the minute size of the bubbles, they may not be well visualized.8Theoretically, only a bubble with a radius exceeding 1 or 2 mm could be detected, provided that the bubble is situated on the tomographic cut.9Previous publications reporting the visualization of cerebral gas on CT scans describe massive arterial embolism occurring during the course of angiography8,10or pulmonary contusion11or venous gas embolism.12However, an initial cerebral CT scan can rule out a cerebral infarct or an intracerebral hemorrhage. Late CT scan changes are nonspecific and could be interpreted as ischemic infarction in case of massive gas embolism.13Other diagnostic methods, such as MRI and single-photon emission CT scanning, are markers of abnormal regional flow. However specific, the finding of gas in the cerebral circulation is uncommon.14MRI of the cerebrum has been described in decompression sickness as multiple punctate nonenhancing foci of prolonged relaxation times in subcortical white matter.15However, it was unclear whether these MRI findings reflected the acute decompression sickness injuries or the effects of cumulative diving episodes. One study reported a normal MRI scan obtained immediately after the onset of symptoms after gas embolism, and these findings were ascribed to the small dimensions of the injured areas.16In the same study, early single-photon emission tomography confirmed the diagnosis. A delayed MRI including diffusion-weighted imaging demonstrated abnormalities consistent with ischemic brain damage.
The primary point of interest in this case is the definitive role of recompression therapy. The mechanism of the beneficial effect of recompression therapy is threefold. First is the reduction of bubble volume according to the law of Boyle. Second is the increase of the surface of the bubble with respect to its volume and therefore an increase in the exchange gas surface. Finally, when pure oxygen or heliox is breathed, blood nitrogen gas tension is reduced to zero, creating a high gradient for the elimination of nitrogen from the bubble and thus the resolution of the bubble. This last mechanism requires that nitrogen be eliminated from the inspired gas mixture during recompression. This can be achieved with the D50 HeO2table or the modified (heliox) Comex 30 table (Comex Inc., Marseille, France). However, this mechanism requires the presence of some blood flow around the bubble. In case of arterial gas embolism, the bubbles fragment and stack longitudinally in small arteries. There is no perfusion around distal bubbles, explaining why they may persist in the cerebral circulation for a long time, 40 h or even longer.17This long duration may be enhanced by the platelet aggregation occurring at the gas–blood interface.18These changes likely explain why even late recompression therapy can be beneficial. This is also the reason why we used a long and deep table without nitrogen to increase the nitrogen gradient between bubbles and blood. This nitrogen gradient is 4.70 ATA with the recompression table we used, whereas it is 2.18 ATA with the United States Navy 6 and 3.12 ATA with the modified (heliox) Comex 30. Incidentally, the initial compression using the United States Navy 6A table is performed while the patient breathes air; therefore, the nitrogen gas concentration in the tissues at the end of recompression exceeds that observed at the beginning of hyperbaric therapy. These long and repeated treatment courses are likely to recruit a distal flow by progressively eliminating the stacked bubbles. In addition, a high dissolved oxygen content could increase oxygen delivery to the cerebral zones compromised by a marginal oxygen supply. Our observation proved that this therapy was of value even when performed with a delay beyond the initial embolism accident. Such beneficial effect of delayed treatment has been already reported in venous and arterial gas embolism.19–21In fact, we were unable to find any delay beyond which hyperbaric treatment provided no benefit. Some previous articles reported successful treatment of gas embolism symptoms with hyperbaric oxygen applied 13 days22and 24 days23after the initial injury. Even for patients in extremely poor conditions with coma, hyperbaric therapy in a chamber devoted to the treatment of critically ill patients should be considered.
The authors thank the nursing staff of the Hyperbaric Center of Hôpital E Herriot, Lyon, France, for incomparable expertise and support.