Complications of cerebral air embolism associated with pleural lavage for empyema

Abstract

Standard initial treatment for acute empyema involves antibiotic administration and chest tube drainage; however, pleural lavage with saline is another treatment that mitigates the need for surgical drainage. Although this treatment is recognised as non-invasive and safe, the complications of cerebral air embolism are less recognised.

In this case, a man in his late 40s was diagnosed with acute empyema and treated with chest tube (28 Fr) drainage and antibiotics. Empyema remained on follow-up chest imaging; thus intrapleural fibrinolytic therapy (urokinase 120 000 units/day for a total of 3 days) and pleural lavage (0.9% saline 1000 mL/day daily) were administered. During the 10th pleural lavage, the patient suddenly became unconscious. Head imaging revealed a cerebral air embolism. Consequently, he received urgent hyperbaric oxygen therapy and improved without any neurological sequelae.

Clinicians should be aware of the complications of sudden cerebral air embolism due to pleural lavage for empyema.

Background

The standard treatment for acute empyema is appropriate antibiotics and chest tube drainage. Unfortunately, about one-third of cases fail treatment and require additional intervention.¹ Intrapleural fibrinolytic therapy (urokinase, streptokinase and tissue plasminogen activator (TPA)), combination therapy of fibrinolytic and mucolytic agents (TPA–deoxyribonuclease (DNAse)) and pleural lavage with saline are effective additional treatments that circumvent the need for surgical drainage.^2–5 These non-surgical alternatives are often used because they are relatively safe and have few adverse events. However, they are very rarely associated with dangerous cerebral air embolism.

Case presentation

A man in his late 40s visited the hospital for fever, chest pain and dyspnoea. The fever and dyspnoea had previously gradually worsened for 1 week. He had no history of hospitalisation, living in nursing homes, sick contacts, travelling abroad, using hot springs or public baths, contacting animals, smoking or drinking. His medical history included oesophageal achalasia and schizophrenia, and his drug history included lansoprazole (30 mg/day), risperidone (2 mg/day) and olanzapine (10 mg/day). He had occasional heartburn in the supine position. The physical findings at the time of visit were alert, systolic blood pressure of 70 mm Hg, heart rate of 136 beats/min, respiratory rate of 30 breaths/min, body temperature of 38.9 °C and oxygen saturation of 94 % (on 8 L oxygen via a non-rebreather mask). The breath sounds in the left chest were diminished. There were no other notable details from the physical exam.

Investigations

Laboratory investigations revealed the following: PaO₂ of arterial oxygen of 55.8 mm Hg (breathing 8 L oxygen via non-rebreather mask), lactic acid of 1.1 mmol/L, white blood cell count of 22.2 cells/nL, platelet count of 550 cells/nL, creatinine of 2.45 mg/dL, total bilirubin of 0.5 mg/dL and C reactive protein level of 49.29 mg/dL. No other remarkable blood test findings were noted. Rapid screening test findings were negative for COVID-19, influenza A antigen/B, urinary pneumococcal capsular and urinary Legionella antigens. Radiographic chest examination revealed massive pleural effusion in the left lung field and right shift of the mediastinum (figure 1). Pleural effusion investigations revealed the following: appearance of yellow opacity, cell number of 6184 (neutrophils 92 %), pH of 6.67, total protein of 5.7 g/dL, lactate dehydrogenase of 581 IU/L and glucose of 1 mg/dL. The patient was diagnosed with empyema-induced sepsis (sequential organ failure assessment score of 8 points) and severe respiratory failure. As an initial treatment for empyema, antibiotics (ampicillin/sulbactam (ABPC/SBT) 9 g/day) were administered and chest tube (28 French) drainage was performed. Chest tube insertion was performed in the empyema space near the left diaphragm after ultrasonographically confirming the availability of sufficient space in the chest cavity (figure 1). After the chest tube was inserted, pus was drained, but no blood or air was aspirated. Infusion resuscitation was performed for circulatory failure due to sepsis. Subsequently, circulatory failure, respiratory failure and organ damage gradually improved. The Streptococcus anginosus group (sensitive to ABPC/SBT) was isolated from the sputum and pleural effusion culture tests at the time of admission, and ABPC/SBT (9.0 g/day) was continued. On day 9, a follow-up chest CT scan showed partial empyema. Therefore, additional treatment was necessary, and treatment options such as video-assisted thoracic surgery, open thoracostomy, additional chest tube insertion, fibrinolytic therapy and saline pleural lavage were considered. The patient chose non-surgical, non-invasive treatment; thus, he was treated with fibrinolytic therapy (120 000 units of urokinase for 3 days) and pleural lavage (500 mL of 0.9% saline two times per day). Daily pleural lavage was administered intrathoracically with a slow drip in a semisitting position. On day 21, during the 10th pleural lavage with saline, he developed sudden disturbance of consciousness, right conjugate deviation of the eyes and paralysis of both lower limbs.

Figure 1.

Chest X-ray examination on the day of admission. Radiographic chest examination revealed massive pleural effusion in the left lung field and a right shift of the mediastinum. A chest tube (28 Fr) was immediately inserted (arrow) after the diagnosis of empyema.

Differential diagnosis

The patient’s impaired consciousness and neurological abnormalities were not presymptomatic and developed suddenly. Therefore, cerebral infarction, cerebral haemorrhage, subarachnoid haemorrhage, epilepsy and head injury were differential diagnoses. Hypoxic encephalopathy, metabolic encephalopathy and sepsis were also considered. The patient’s breathing, circulation and temperature were stable, and blood tests, electrocardiograms and echocardiography were normal. A CT examination revealed a small amount of air in the chest cavity (figure 2), and head CT and MRI revealed a cerebral air embolism and ruled out cerebral infarction, cerebral haemorrhage, subarachnoid haemorrhage and head trauma (figure 3). Transthoracic colour Doppler echocardiography revealed no apparent atrial and ventricular septal defects with left-to-right or right-to-left shunts. However, since transoesophageal echocardiography of the Valsalva manoeuvre was not performed, atrial septal defect or a patent foramen ovale cannot be excluded. Nonetheless, venous catheter-related cerebral air embolism was ruled out because there was no air intrusion from the indwelling peripheral venous catheter (diameter 0.8 mm×length 25 mm) in the patient’s upper limbs. Therefore, the patient was diagnosed with a cerebral air embolism associated with pleural lavage.

Figure 2.

Chest CT examination on day 21. Chest CT examination after pleural lavage with saline revealed a small amount of air in the chest cavity (arrow).

Figure 3.

Head CT examination on day 21. Head CT after pleural lavage with saline revealed scattered air-isodense spots in the right and left cerebral hemispheres (arrows).

Treatment

Initially, the patient was placed in a supine position, and airway, respiratory and circulatory stability was confirmed. High-flow oxygen administration (FiO₂ 100%) was then initiated. Next, he was urgently treated with hyperbaric oxygen therapy (HBO) for 60 min at 2.8 atmospheres absolute (ATA) (2.8 ATA is equivalent to the pressure at 18 m under sea water). Recompression therapy using US Navy Treatment table 6 (US Navy Recompression Therapy Protocol; 75 min at 2.8 ATA, followed by 150 min at 1.9 ATA) was then applied. In addition, edaravone (60 mg/day) was administered as a brain-protective drug and fosphenytoin sodium hydrate as an epilepsy-preventive drug (initial dose: 22.5 mg/kg/day, maintenance dose: 7.5 mg/kg/day) for 3 days.

Outcome and follow-up

After HBO, the patient’s impaired consciousness and neurological abnormalities improved. After that, he continued rehabilitation and intravenous antibiotic treatment for a total of 6 weeks and was then discharged from the hospital. One year after discharge, the patient remains healthy with no neurological sequelae.

Discussion

This case report highlights two clinically useful points. First, severe cerebral air embolism is rarely associated with pleural lavage with saline. Second, HBO is effective for cerebral air embolism.

The initial treatment for acute empyema is administration of appropriate antibiotics and drainage with a chest tube, but for residual empyema that is resistant to initial treatment, there is also intrathoracic injection of fibrinolytic drug±DNAse and/or pleural lavage with saline available as a non-surgical additional treatment. These additional treatments have shown no difference in mortality compared with that of the placebo group, but may help remove the need for surgical intervention.^{1–3 5} Bleeding is a serious side effect of fibrinolytic drugs, but pleural lavage with saline has been reported to be safe and without serious side effects.^{5 6} Severe cerebral air embolism is extremely rare during pleural lavage with saline.

Air embolism can occur during iatrogenic complications such as those associated with endovascular catheterisation, surgical procedures, radiological procedures and positive pressure ventilation of the ventilator. Other causes can include trauma and decompression sickness. The air and vascular passages and the pressure gradient of air and blood flow make it possible to develop an air embolism,⁷ which are classified into venous air embolisms that affect the pulmonary circulatory system, arterial air embolisms that affect the arterial circulatory system and arterial air embolisms that cause ischaemic organ damage. The most vulnerable organs are the brain and heart, wherein arterial ischaemia can be fatal.⁸

For this case study, we investigated existing cases of cerebral air embolism due to non-surgical additional treatment of empyema in the literature. There are seven English-language documents (eight patients) reported in PubMed, and this case is the eighth (ninth patient, table 1).^9–15

Table 1.

Literature reports of empyema or parapneumonic pleural effusion with cerebral air embolism due to non-surgical additional treatment

Reference number	Age, sex	Thoracic procedures	Drain size Patient position Time until onset	Site of cerebral air embolism	Treatment of air embolism	Neurological outcome
9	40, Nr	Pleural lavage Saline 1000 mL/day Fibrinolytic therapy Streptokinase 250 000 IU/day	20 Fr. Nr. 3rd time.	No detection	HBO (2.5 ATA, 90 min)	Improvement
9	56, Nr	Pleural lavage Saline Fibrinolytic therapy Streptokinase	Nr. Nr. 8th time.	No detection	HBO (2.5 ATA, 90 min)	Neurological sequelae
10	50, M	Pleural lavages Saline Fibrinolytic therapy Streptokinase	Nr. Nr. 3rd time.	Right cerebral hemisphere	HBO (2.8 ATA, 135 min+2.3 ATA, 70 min)	Improvement
11	78, M	Pleural lavage	Nr. Sitting position. Nr.	Right cerebral hemisphere	Only supportive care	Neurological sequelae
12	77, M	Pleural lavage Saline 500 mL/day	Nr. Nr. 1st time.	Right cerebral hemisphere	Therapeutic hypothermia (33°C, 24 hours)	Improvement
13	79, F	Fibrinolytic therapy Streptokinase 250 000 IU/day	Nr. Nr. 1st time.	Right cerebral hemisphere	HBO (280 kPa, 180 min+190 kPa, 240 min)	Neurological sequelae
14	53, M	Pleural lavage Saline 1000 mL/day Fibrinolytic therapy Urokinase 60 000 IU/day	24 Fr. Sitting position. 10th time.	Right and left cerebral hemisphere	Only supportive care	Improvement
15	69, F	Pleural lavage Saline 500 mL/day	14 Fr. Nr. 7th time.	Left cerebral hemisphere	HBO (1.9 ATA, 50 min)	Improvement
This case	48, M	Pleural lavage Saline 1000 mL/day Fibrinolytic therapy Urokinase 120 000 IU/day	28 Fr. Sitting position. 10th time.	Right and left cerebral hemispheres	HBO (2.8 ATA, 60 min+table 6)	Improvement

Table 6, US Navy Recompression Therapy Protocol 75 min at 2.8 ATA followed by 150 min at 1.9 ATA.

ATA, atmospheres absolute; F, female; Fr, French; HBO, hyperbaric oxygenation; M, male; Nr, not reported.

Non-surgical treatment in the literature varies along with timing of the onset of the cerebral air embolism and drain size. Symptoms of the cerebral air embolism range from impaired consciousness, convulsions, conjugate deviation of the eyes, visual field disorders, aphasia, movement disorders and sensory disorders, depending on the location and severity of air embolism. Similarly, the site of air embolism in the head CT and MRI varies. In some cases, air embolisms were not detected by head imaging, which suggests an effect of the amount of air and the timing of imaging. Cerebral air embolism is more likely to occur in the right cerebral hemisphere because the brachiocephalic artery is the first branch of the aorta.¹⁶ Head CT findings showed air-isodense spots, and head MRI findings showed a high signal on diffusion-weighted images and low signal on T2 star-weighted images.^{11 12 14} Treatment was HBO in most cases, and the neurological prognosis was improved in most cases.

Evidence of the mechanism of development of cerebral air embolism during pleural lavage in this case is limited. Nevertheless, based on previous literature, we present the following mechanisms for air-vascular passage and the pressure gradients required for development of air embolism.^9–15 In this case, pneumonia with empyema may have caused a minor lung injury. Thus, it is possible that there were small fistulas in the peripheral pulmonary veins and bronchi that could not be confirmed by chest CT examination. Alternatively, fibrinolytic therapy may have dissolved the fibrin membrane of the visceral pleura, exposing the peripheral pulmonary veins within the pleura. The subsequent pleural lavage and coughing may have triggered an increase in intrapleural pressure and air may have flowed into the peripheral pulmonary veins. Accidental minor lung injuries due to chest tubes cannot be completely ruled out. However, there was no bleeding or air aspiration in the chest tube, and there were no complications prior to the 10th pleural lavage. The tip of the chest tube was close to the lungs but did not penetrate the tissue (online supplemental figures 1; 2). Inflammation due to pneumonia and empyema, fibrinolytic therapy and frequent pleural lavage with saline may have weakened the tissue and accidentally promoted air invasion into the peripheral pulmonary veins.

Although cerebral air embolism is rare, it requires prompt recognition and early intervention because of its life-threatening and/or neurological sequelae. The initial treatment is to maintain the airway and respiratory and circulatory stability, eliminate the cause and administer high-flow oxygen. Simultaneously, the patient is placed in the supine position to avoid further embolism. Next, early HBO therapy (preferably within the first 4–6 hours) is considered for reducing the size and reabsorption of air embolism.^{7 8 17} However, the hospitals that can offer HBO are limited, and it is necessary to weigh the risks and benefits of transfer time. This case was diagnosed with cerebral air embolism, and HBO could be performed immediately at the hospital, achieving rapid improvement of neurological symptoms and complete cure without sequelae.

Fibrinolytic therapy and pleural lavage with saline are well known as effective and safe treatments as additional treatments for empyema that can avoid surgical treatment. However, very rarely, it is associated with dangerous and serious cerebral air embolism. Although not a routine treatment for air embolism, early HBO is an option for treatment of neurologically impaired cerebral air emboli.

Learning points.

• Pleural lavage and fibrinolytic therapy are additional treatments for empyema that may avoid surgical treatment.

• Pleural lavage and fibrinolytic therapy are rarely associated with dangerous cerebral air embolism.

• Rapid hyperbaric oxygen therapy is a useful option for the treatment of neuropathic cerebral air embolism.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).