Convulsive syndrome as a manifestation of acute cerebral damage due to paradoxical air embolism in neurosurgical patients. Series of clinical cases and literature review

Lapteva K.N.; Anan’ev E.P.; Savin I.A.; Rasulova E.V.; Kozlova A.B.; Sazonova O.B.; Sokolova E.Iu.; Pitskhelauri D.I.; Pronin I.N.

doi:https://doi.org/10.17116/neiro20208402151

Paradoxical air embolism (PAE) is a rare potentially fatal complication followed by entering of air emboli from the right cardiac chambers and pulmonary artery to large circulation circle [1, 2]. In neurosurgical practice, PAE is common after procedures in patient’s sitting position [3]. According to the literature, the main causes of PAE are intracardiac bypass through patent foramen ovale and intra-pulmonary arteriovenous shunts. Typical target organs of air emboli are lungs, heart and brain [4]. Intraoperative PAE may be followed by postoperative cardiovascular and respiratory insufficiency, neurological aggravation, etc. [4]. Air embolism into cerebral artery may be associated with neurological symptoms including impaired consciousness, focal neurological deficit de novo and seizures [5].

Development of convulsive seizures after neurosurgical interventions complicated by PAE is described in single cases. However, the authors explain development of seizures by other concomitant factors rather air embolism [4, 6—8].

In this report, we analyzed 5 patients who underwent neurosurgical intervention complicated by PAE and early postoperative convulsive syndrome.

Material and methods

There were 5 patients operated at the Neurosurgery Center in 2014—2018. In all cases, neurosurgical intervention was complicated by PAE and subsequent early postoperative convulsive syndrome. These patients did not have preoperative convulsions.

Results

Intraoperative period. All patients underwent surgery in sitting position. Intraoperative transesophageal echocardiography was not performed in any case. Jugular vein compression test was used for intraoperative identification of the cause of embolism. In some cases, the following anesthetic measures were used in patients with intraoperative PAE to stabilize their condition: mechanical ventilation with FiO2 100%, external compression of cervical veins, forced infusion of 500—1000 ml within 20—60 minutes, norepinephrine infusion at a dose of 1 µg/kg/min.

Postoperative period. Postoperative intensive care included sedation, circulatory support, mechanical ventilation. Echocardiography, CT and MRI were carried out. Mean duration of ICU-stay in these patients was 20 days, duration of mechanical ventilation — 15 days. One patient died from purulent-septic complications (length of hospital-stay was 120 days). Autopsy was not performed in this case.

Convulsive syndrome and EEG-monitoring. The indications for video EEG-monitoring were impaired consciousness and/or convulsive syndrome [9—11]. Video EEG-monitoring was conducted on "Encephalan ABP 26/19" scanner (19 channels, "Medicom MTD", Taganrog, Russia). The electrodes were placed in accordance with international scheme of electrode placement "10—20%". Duration of EEG-monitoring was determined individually considering international recommendations [12].

Neuroimaging. All patients underwent MRI on the first or the second postoperative day (DWI, T2, T2-FLAIR, T1). Air bubbles in convexital CSF spaces and basal cisterns were visualized in all cases. Moreover, hyperintense T2, T2-FLAIR and DWI signal was determined in cortical and subcortical white matter. Measured diffusion coefficient (MDC) as a quantitative value determining spatial diffusion of water proton was 0.5—0.7*10-8 mm2/s in cortical gray matter (norm 0.8—1.0*10-8 mm2/s). Two patients (Table 3, 5) had similar foci of hyperintense MR signal in pulse sequences of DWI, T2, T2-FLAIR and subcortical nuclei. In all cases, there were no changes of MR signal in brain stem structures. MRI data are presented in Table 1. Comparison of neuroimaging and neurophysiological data in all cases showed that the focus of modified DWI signal coincided with the focus of ictal activity onset according to EEG data (Fig. 1).

Fig. 1. Patient 2. MRI in 2 days after surgery. Diffusion-weighted image. There are foci of hyperintense signal in hemispheric cortex. Focus of epileptiform activity (EEG-monitoring data) is marked on the scheme that corresponds to one of the foci of hyperintense signal (MRI data).

Case report 1

A 55-year-old patient K. has been ill since the age of 28 years when anaplastic ependymoma of the 4th ventricle was diagnosed. Surgical intervention was performed three times (1987, 2006, 2010), radiotherapy was administered too. The patient was hospitalized due to clinical aggravation with progression of general cerebral symptoms. Preoperative neurological status included general cerebral symptoms and dysfunction of the left cranial nerves V—VIII. Continued tumor growth was confirmed by MRI. Redo surgery was performed on July 30, 2014. Craniotomy was followed by midline suboccipital approach to posterior cranial fossa. Craniotomy was accompanied by injury of confluence of sinuses and single episode of venous air embolism with decrease of EtSO₂ up to 16%, SpO₂ up to 97%, arterial hypotension up to 90/60 mm Hg. Total resection of tumor was performed after stabilization of patient’s condition.

Impaired consciousness (GCS score 7), dysfunction of the left cranial nerves III—XII, tetraparesis (the most significant in the left hand up to 1 score) were observed in early postoperative period. CT on the 1st postoperative day did not reveal intracranial complications as a cause of impaired consciousness. MRI revealed hyperintense DWI signal in parietal, frontal and occipital cortex in both hemispheres. Apparently, this was a result of PAE (Table 1, Fig. 2).

Fig. 2. Patient 1. DWI scans on the 2nd postoperative day. Diffuse foci of hyperintense MR signal in parietal and frontal cortex due to restricted diffusion as a manifestation of ischemic brain damage.

EEG monitoring revealed continued epileptiform activity with a focus in the left frontal area. Moreover, there was a tendency to generalization throughout all brain regions without clinical seizures (epileptic unconvulsive status). Anticonvulsant therapy was prescribed under EEG monitoring (levetiracetam 500 mg i.v. bolus). Considering persistent epileptiform activity, infusion of valproic acid at a dose of 1.5 mg/kg/hour and levetiracetam 1.8 mg/kg/hour was started. Infusion of propofol at a dose of 4 mg/kg/hour was started 12 hours later due to the absence of the effect according to EEG-monitoring data. Regression of epileptiform activity was observed within 18 hours. Propofol infusion was carried out within 48 hours with gradual decrease of the dose until complete abolition.

Moreover, air embolism followed by acute lung damage required postoperative mechanical ventilation with PEEP±8—10 cm H2O and FiO2 35—45% for 7 days in order to maintain normal acid-base state. Then, mechanical ventilation was performed in support modes (SIMV+PS and PSV) for 24 days until weaning from ventilator.

The patient was transferred to the ward on the 39th day after surgery as soon as clinical stabilization and correction of infectious-septic complications were achieved. Symptoms of dysfunction of the left cranial nerves III—XII and tetraparesis up to 3 points were observed at discharge. Convulsive attacks were not observed. Patient's quality of life at discharge was assessed as Karnofsky score 30.

Case report 2

A 30-year-old patient K. is sick since 2012. The first symptoms were impaired hearing in the right ear and choking in swallowing solid food. The following symptoms were numbness of the right half of the face, tongue, diplopia in gaze to the right. Contrast-enhanced MRI revealed trigeminal neuroma on the right. Resection of neuroma of Gasser’s ganglion was performed on April 21, 2014 through retrosigmoid approach and osteoplastic craniotomy. There were 4 episodes of venous air embolism during soft tissue incision, craniotomy and tumor resection. Embolism was followed by decrease of EtCO2 up to 20 mm Hg, SpO₂ up to 97% and arterial hypotension up to 60/30 mm Hg. Clinical stabilization was followed by partial resection of tumor.

Early postoperative impairment of consciousness up to GCS score 8 and left-sided hemiparesis up to 3 points were observed. There were no CT data on intracranial complications explaining impaired consciousness on the first postoperative day. Hyperintense DWI signal in frontal and parietal cortex (mainly on the right), as well as diffuse small foci in the cerebellum were observed (apparently, as a result of embolism) (Table 1, Fig. 3a).

Fig. 3. Patient 2. a — DWI scan on the 2nd postoperative day. Diffuse foci of hyperintense MR signal in frontal and parietal cortex (more significant on the right) due to restricted diffusion as a manifestation of ischemic brain damage. b — EEG monitoring within the first postoperative day. Epileptic seizure is represented by “acute-slow wave complexes” with an accent in posterior parts of the hemispheres on the right and a tendency to generalization throughout the entire brain (15 μV / mm).

There were 3 epileptic seizures debuting from turning the head to the right and gaze upwards followed by fasciculation in the left hand and then in the right one. EEG monitoring was carried out according to indications. Sixteen electrographic epileptic seizures were recorded with initiation in the right occipital region in the form of spike-wave activity (Fig. 3b). Three out of sixteen seizures were accompanied by clinical convulsions. Other EEG seizures proceeded without clinical equivalent. Specialist for intensive care administered intravenous infusion of valproic acid 1.3 mg/kg/hour and levetiracetam 1.6 mg/kg/hour. Stabilization of vital functions was associated with restoration of clear consciousness, regression of hemiparesis and relief of epileptic syndrome.

Air embolism followed by acute lung damage required postoperative mechanical ventilation with PEEP±10—12 cm H2O and FiO2 35-45% for 3 days in order to maintain normal acid-base state. Then, mechanical ventilation was performed in support modes (SIMV + PS and PSV) for 5 days until weaning from ventilator. The patient was transferred to the ward in 14 days after surgery and discharged later for further rehabilitation. There was dysfunction of the right cranial nerves V—VI at discharge. Six months later, the patient underwent redo surgery for resection of residual tumor. Patient's quality of life at discharge was assessed as Karnofsky score 80.

Case report 3

A 32-year-old patient G. is ill for 4 months.

The first symptoms were numbness in the left half of the face and right limbs, diplopia. Neurological status upon admission included dysfunction of the left cranial nerve VI, bilateral dysfunction of cranial nerve V, right-sided hemiparesis up to 4 points. According to MRI data, left-sided cavernous angioma of pontine tegmentum with signs of hemorrhage was found. Surgical intervention was performed on January 17, 2017 through osteoplastic craniotomy and suboccipital approach. Resection of tumor was accompanied by venous air embolism, decrease of EtCO2 up to 22 mm Hg, SpO₂ up to 94% and arterial hypotension up to 70/30 mm Hg. Compression of the jugular veins did not reveal a clear source of air embolism. Clinical stabilization was followed by total resection of cavernous angioma.

Postoperative deterioration of neurological symptoms included impaired consciousness (GCS score 7), dysfunction of cranial nerves V, VI, VII on the left. There were no CT data on intracranial complications explaining impaired consciousness on the first postoperative day. MRI revealed hyperintense DWI signal in frontal-parietal-occipital cortex, cerebellar vermis and right hemisphere (apparently, as a result of air embolism) (Table 1, Fig. 4a).

Fig. 4. Patient 3. a — DWI scan on the 2nd postoperative day. Diffuse foci of hyperintense MR signal in occipital cortex due to restricted diffusion as a manifestation of ischemic brain damage. b — EEG monitoring within the first postoperative day. Continuous epileptiform activity (“acute-slow wave complex”) is recorded with an accent in posterior parts of hemispheres on the left with a tendency to generalization throughout the entire brain (7 μV/mm).

EEG monitoring was performed. There was continued epileptic activity (status epilepticus) represented by “acute-slow wave” complexes with an accent of amplitude in the left occipital-parietal region and spread to posterior parts of the right hemisphere (Fig. 4b).

The episodes of activity coincided with clinical phenomena: saccadic eyeball movements and gaze rotation to the right. Anticonvulsants were sequentially prescribed: levetiracetam 1 mg/kg/day, valproic acid 1.7 mg/kg/day. However, generalization of continued epileptic activity was noted. Prolonged infusion of propofol at a dose of 4 mg/kg/hour was prescribed for 48 hours under the control of video EEG monitoring. Thus, complete reduction of EEG epileptiform activity was achieved. A dose of propofol was gradually reduced until complete cancellation. Recovery of consciousness was noted.

Postoperative infusion of norepinephrine at a dose of 0.4 μg/kg/min was administered for 5 days to support systemic circulation. Then, circulatory characteristics achieved physiological values without vasopressor support. Supporting mechanical ventilation was carried out for 12 days. The patient was transferred to the ward in 14 days after surgery.

Neurological status at discharge included right-sided hemiparesis up to 2 points in the hand and 3 points in the leg, dysfunction of the left cranial nerves V, VI, VII. The patient was discharged for further rehabilitation. Patient's quality of life at discharge was assessed as Karnofsky score 60.

Case report 4

A 56-year-old patient S. has been ill for 20 years when headaches first appeared. MRI revealed a neoplasm of the left thalamus. There were no severe focal neurological symptoms at admission. Tumor resection was performed on November 16, 2017 (low-grade astrocytoma). Significant fusion of the bone with dura mater and the walls of venous sinuses was revealed after osteoplastic craniotomy and suboccipital approach. As a result, injury of confluence of the sinuses occurred. Air embolism was followed by decrease of EtCO2 up to 22 mm Hg, SpO₂ up to 94%, arterial hypotension up to 90/50 mm Hg.

Early postoperative deterioration of neurological symptoms included impaired consciousness (GCS score 10). There were no CT data on intracranial complications explaining impaired consciousness on the first postoperative day. MRI revealed hyperintense DWI signal in cerebral cortex, cerebellum and cerebellar vermis, white matter. Apparently, this was a result of air embolism) (Table 1, Fig. 5a—c).

Fig. 5. Patient 4. DWI scans in 2 days after surgery. Diffuse foci of hyperintense MR signal due to restricted diffusion as a result of ischemic brain damage: a) frontal and parietal cortex; b) occipital cortex; c) cortex of the right cerebellar hemisphere; d) EEG monitoring within the first day after surgery. Epileptic seizure is represented by “acute-slow wave complexes” with an accent in posterior parts of the hemispheres on the right with a tendency to generalization throughout posterior parts of the hemispheres (15 μV/mm).

A 16-hour EEG monitoring was carried out. Six epileptic seizures were recorded (Fig. 5d). Seizures manifested stereotypically as spike-wave activity in the right occipital-parietal region with spread to the left occipital area and gradual generalization throughout the entire brain. Ictal clinical manifestations occurred in 1.5—2 min after EEG onset of seizures (turning head to the left, tonic tension in the left hand, clonic oscillations in the left hand with generalization of convulsions).

Convulsive seizures were cured by intravenous injection of benzodiazepines 10 mg. Anticonvulsant therapy also included valproic acid 1.5 mg/kg/day and levetiracetam 1.8 mg/kg/day. Epileptic seizures did not recur. Mechanical ventilation was performed in support modes (SIMV + PS and PSV) for 24 hours until weaning from ventilator.

The patient was transferred to the ward in 7 days after surgery and then discharged in satisfactory condition without focal neurological deficit. Patient's quality of life at discharge was assessed as Karnofsky score 80.

Case report 5

A 65-year-old patient I. underwent nephrectomy for right kidney cancer 13 years ago. Headache and dizziness occurred 10 months ago. MRI revealed a neoplasm of the fourth ventricle. Preoperative neurological status included general cerebral symptoms, bilateral dysfunction of cranial nerves IX, XII.

Resection of the tumor of the 4th ventricle was performed on March 14, 2018. Osteoplastic craniotomy was followed by suboccipital approach to posterior cranial fossa. Craniotomy was accompanied by venous air embolism, severe hemodynamic disturbances (arterial hypotension up to 60/40 mm Hg), reduced saturation (90%) and EtCO2 (25 mm Hg). A long period of time was required to stabilize clinical state. Next, total resection of tumor (metastasis of renal cell adenocarcinoma) was performed.

Discontinuation of anesthetic drug infusion at the ICU was not followed by recovery of consciousness (GCS score 5). There were no CT data on intracranial complications explaining impaired consciousness on the first postoperative day. MRI revealed hyperintense DWI signal in the cerebral cortex (more significant on the right), foci in the white matter, subcortical nuclei, cerebellum. Apparently this was a result of embolism (Table 1, Fig. 6).

Fig. 6. Patient 5. DWIs in 2 days after surgery. Diffuse foci of hyperintense MR signal in frontal and parietal cortex on the left and right hemisphere cortex due to restricted diffusion as a manifestation of ischemic brain damage.

EEG monitoring confirmed continued epileptiform activity (“acute-slow wave complexes”) with a source in the left occipital region and tendency to generalization throughout the brain. Nonconvulsive status epilepticus was diagnosed considering the absence of obvious clinical manifestations. Anticonvulsant therapy (levetiracetam 1.6 mg/kg/day, valproic acid 1.4 mg/kg/day) for 2 hours was not effective in accordance with EEG monitoring data. Next, propofol was administered as continued infusion 3 mg/kg/hour for 48 hours. These measures were followed by EEG-confirmed regression of convulsive activity and propofol infusion was gradually canceled.

Moreover, norepinephrine infusion at a dose of 0.8 μg/kg/min has been used for 6 postoperative days for hemodynamic support. Then, characteristics of systemic hemodynamics corresponded to physiological norms without vasopressor support. Mechanical ventilation was performed for 24 days until weaning from the ventilator. The patient was transferred to the ward after 27 days with left-sided hemiplegia and bilateral dysfunction of cranial nerves III—XII. The patient died in 120 days after surgery from purulent-septic complications.

Discussion

We report a rare complication of venous air embolism in neurosurgery (convulsive syndrome). In our sample, injury of confluence of sinuses occurred in 3 cases as a result of various intraoperative circumstances. This complication was followed by embolism. The source of embolism was confirmed by jugular vein compression. There were no clear sources of venous air embolism in other cases.

The most common causes of air embolism in systemic circulation are patent foramen ovale and intrapulmonary shunts. However, according to the literature, the absence of interatrial septal defect does not exclude the possibility of intraoperative venous air embolism in patient’s sitting position [25, 26]. The second known mechanism of embolism is intrapulmonary shunts.

Pathophysiological mechanisms of organ damage associated with venous air embolism are not clear. Some authors emphasize mechanical damage to the endothelium by air bubbles, endotheliitis, arterial spasm, platelet activation, thrombosis, prolonged vessel occlusion followed by infarction [5, 13].

Brain damage as a target organ can result ischemic lesion with appropriate neurological symptoms including convulsive attacks. Seizures followed by various clinical and electrographic manifestations after air embolism in cardiac surgery (cardiopulmonary bypass), liver and lung transplantation are reported in the literature [14—16].

There are few reports devoted to convulsive syndrome after surgery for posterior cranial fossa diseases. Moreover, the authors did not establish a relationship between convulsive syndrome and patient’s position or development of intraoperative complications including venous air embolism [17—19].

Several reports are devoted to resection of tumors of posterior cranial fossa complicated by intraoperative venous air embolism and no convulsive syndrome in early postoperative period [19—22].

Other authors described convulsive syndrome after neurosurgical intervention in patient’s sitting position. However, these authors did not observe correlation between venous air embolism and convulsive syndrome [6, 17]. Sachidanand J Bharati. et al. analyzed postoperative complications in 63 patients who underwent surgery for malignancies in prone position (57%), sitting position (27%), supine position (13%), lateral position (3%). Postoperative seizures occurred in 4 (6.3%) patients. However, the authors associated convulsions with hyponatremia rather venous air embolism [23].

Shin-Tseng Lee et.al. retrospectively analyzed 726 patients who underwent posterior cranial fossa surgery. In 13 (1.8%) patients, convulsions occurred within 2 weeks after surgery, in 5 cases – within 24 hours after surgery. There were no data on intraoperative venous air embolism. Surgery in patient’s sitting position was performed in 7% of cases, prone position — 53%, lateral position — 40%. The authors associate postoperative seizures with metabolic acidosis and hyponatremia [18]. Other authors also report the influence of water-electrolyte and other homeostatic disorders on the development of convulsive syndrome in these patients [6].

There are researches with determined correlation of seizures with intraoperative episode of venous air embolism. Suri et. al. retrospectively analyzed occurrence of seizures in patients who underwent posterior cranial fossa surgery. This study included 511 patients. In 30 (5.9%) cases, seizures occurred within 14 days after surgery. Intraoperative episode of venous air embolism was observed in 10 out of 30 (33%) patients with seizures. The authors determined the following predictors of postoperative seizures: sitting position, age <20 years, histological structure of tumors (cochlear neuroma, medulloblastoma, astrocytoma) and intraoperative air embolism [6].

According to few studies, localization of focal brain lesion associated with air embolism is mosaic. The authors do not distinguish certain brain areas as high risk regions of ischemic damage [26, 27]. These features are true for patients undergoing venous air embolism in sitting [28] and supine position [27, 29]. In our series, we observed multiple foci of hyperintense MR-signal in hemispheric cortex with somewhat predominance of ischemic foci in the right hemisphere. Perhaps, this is associated with anatomy of cervical vessels, blood flow mechanics and sequence of air bubbles entering into innominate artery and left subclavian artery (ascending aorta → innominate artery, right subclavian artery → left subclavian artery). However, the focus of epileptiform activity was usually localized in occipital cortex despite the diffuse nature of cortical damage. There are single case reports describing EEG-patterns in patients with convulsive syndrome after venous air embolism. Thus, Menkin et. al. reported spike-wave activity and spike-slow wave complexes in the right occipital region [13]. These results are comparable with our data. Probably, patient’s sitting position with bent neck and inclined head is associated with embolism of relatively small air bubbles into posterior cerebral arteries, and occipital-parietal lobes are less affected by ischemia. This is manifested by "irritation" as a sign of epileptiform activity. Analysis of large sample size is required to confirm this hypothesis.

Thus, there is currently no consensus on the direct impact of venous air embolism during neurosurgery on the development of postoperative convulsive syndrome and pathophysiology of seizures. However, postoperative convulsions and status epilepticus are possible after intraoperative venous air embolism. Therefore, early EEG-monitoring is required in these patients if aggravation of general cerebral symptoms, impaired consciousness and/or clinical convulsions are observed [9, 10, 24].

Conclusion

Neurosurgery complicated by venous air embolism may be accompanied by early postoperative cardiovascular, respiratory, multiple organ failure and focal neurological symptoms not corresponding to the area of surgical intervention. Convulsive syndrome is also possible. In our sample, we observed a tendency to recurrent convulsive seizures and severe course of convulsions (status epilepticus). Major parameters of homeostasis remained normal (water-electrolyte state, serum glucose, acid-base state, etc.) and, accordingly, did not cause seizures. Therefore, venous air embolism followed by focal ischemic lesion was the most probable cause. The focus of epileptiform activity coincided with one of the foci of hyperintense MR signal qualified as restricted diffusion zone in all cases. In most cases, epileptiform activity was initiated by ischemic foci in occipital cortex.

Thus, intraoperative venous air embolism can cause convulsive syndrome. In some cases, seizures were resistant to standard anticonvulsant therapy and required propofol infusion for complete electroencephalographic regression of epileptiform activity. In some cases, there were only electrographic patterns of ictal activity combined only with consciousness impairment without clinical equivalent.

The authors declare no conflicts of interest.

Commentary

The report of Lapteva K.N. et al. is devoted to an urgent problem — intraoperative cerebral paradoxical air embolism. The authors raise three extremely important clinical problems in neurosurgery:

1. The problem of intraoperative cerebral paradoxical air embolism.

2. The problem of acute cerebral damage due to intraoperative cerebral paradoxical air embolism.

3. The problem of status epilepticus following acute cerebral damage in patients with intraoperative cerebral paradoxical air embolism.

Intraoperative air embolism of the pulmonary artery is required in a patient with abnormal bypass between systemic and pulmonary circulation for clinical manifestation of intraoperative cerebral paradoxical air embolism. The problem of intraoperative air embolism is classical for neurosurgery [1]. Neurosurgical and anesthetic approaches to prevention and correction of this event are well known. These measures enroll preferable use of intraoperative prone position of the patient, timely surgical closure of probable venous cranial sources of air embolism, short-term compression of the jugular veins, adequate correction of mechanical ventilation and infusion therapy [2]. K.N. Lapteva et al. also briefly consider these aspects in their manuscript. In my opinion, advanced attention is not necessary for this aspect, since these approaches have already become routine. However, the authors, unfortunately, paid undeservedly little attention to the problem of abnormal bypass between systemic and pulmonary circulation. At the same time, this abnormality is a cornerstone phenomenon in the development of intraoperative cerebral paradoxical air embolism and its complications (brain damage and seizures).

The authors reasonably emphasize that patent foramen ovale and pulmonary arteriovenous shunts are the most common causes of abnormal bypass between circulatory circles [3]. Postoperative transthoracic echocardiography without contrast bubbles did not reveal patent foramen ovale in any patient. Thus, the authors conclude that air embolism was caused by intrapulmonary shunts in all cases. One cannot agree with this interpretation of the cause of cerebral embolism. Firstly, patent foramen ovale is very common phenomenon and observed in about 25% of people while pulmonary arteriovenous shunts (i.e. arteriovenous fistula as a rule) is an extremely rare malformation. Secondly, transthoracic echocardiography without contrast bubbles is characterized by low sensitivity and specificity in the diagnosis of patent foramen ovale. This malformation may be present even in case of failed diagnosis after transthoracic echocardiography without contrast bubbles. These data determine indications for contrast-enhanced transthoracic echocardiography and, in the case of a negative result, transesophageal echocardiography without and with contrast bubbles. Sensitivity and specificity of the last method in the diagnosis of patent foramen ovale is almost 100%. Therefore, only this method is required to judge about the presence or absence of this variant of abnormal bypass between circulatory circles [4, 5].

Thus, patients undergoing neurosurgery in a sitting position should be allocated to the special group. Patent foramen ovale should be preoperatively excluded in these ones. Prone position should be preferred in patients with patent foramen ovale. Surgical occlusion of foramen ovale may be considered prior to neurosurgery if prone position is inadequate for neurosurgical procedure.

The authors focused on the morphological pattern of damage and MR features in patients with acute cerebral damage following intraoperative cerebral paradoxical air embolism. Undoubtedly, this discussion is partly interesting. However, morphological pattern of brain damage described by the authors in patients with intraoperative cerebral paradoxical air embolism could not be another, since we are talking about conventional cardioembolic type of ischemic stroke. Secondly, MRI semiotics of ischemic stroke is well known and cannot be the subject of serious scientific discussion in description of 5 clinical observations. Unfortunately, the authors ignored extremely important issues requiring discussion. This is the problem of intraoperative detection of cerebral air emboli and effective treatment of brain damage following intraoperative cerebral paradoxical air embolism.

Intraoperative detection of cerebral emboli can be successfully performed by prolonged transcranial Doppler monitoring [3]. This monitoring modality may be easily implemented in patients with high risk of intraoperative cerebral paradoxical air embolism, since the last one is typical for patients with lesion of posterior cranial fossa and pineal region who undergoing neurosurgery in a sitting position. Ultrasound sensors will be placed outside the area of neurosurgical manipulations in these cases. Timely detection of cerebral air emboli will be immediately followed by comprehensive neurosurgical and anesthetic measures aimed at stopping cerebral embolism and prevention of brain damage.

Thus, intraoperative transcranial ultrasound monitoring for emboli detection is probably indicated in all patients with patent foramen ovale regardless of the position on the operating table. Intraoperative sitting position of a patient with patent foramen ovale seems unreasonably risky if ultrasound transcranial detection of emboli is absent. Routine intraoperative ultrasound detection of emboli in all patients with or without patent foramen ovale undergoing surgery in sitting position is valuable to increase the patient's safety.

Brain damage following intraoperative cerebral paradoxical air embolism is a cardioembolic ischemic stroke due to occlusion of small cerebral artery. However, reperfusion therapy (thrombolytic therapy or mechanical thromboextraction) is impossible in patients with intraoperative cerebral paradoxical air embolism unlike conventional ischemic cardioembolic stroke. According to the literature, hyperbaric oxygenation is the only effective method in patients with intraoperative cerebral paradoxical air embolism followed by cerebral damage [6]. Earlier hyperbaric oxygenation is accompanied by better outcomes in these patients. Hyperbaric oxygenation is the most effective within 24–48 hours after surgery.

Thus, early MRI of the brain is indicated if acute postoperative cerebral damage following intraoperative cerebral paradoxical air embolism is suspected. MRI is required to decide whether hyperbaric oxygenation is appropriate.

The authors adequately described the problem of status epilepticus as a result of acute cerebral damage after intraoperative cerebral paradoxical air embolism. Unfortunately, clinical component of this urgent and interesting problem was almost completely ignored (definitions, treatment methods, assessment of treatment quality and discussion of the uniqueness of these patients). Indeed, the last one principally allows presentation of this clinical material as a small series of clinical observations.

The authors use the term “convulsive syndrome” to describe all 5 cases. As you know, convulsive syndrome should be classified into seizures, recurrent seizures and status epilepticus depending on duration of seizures and response to antiepileptic therapy. Status epilepticus is divided into simple epileptic status, refractory and super-refractory epileptic status [7]. Refractory epileptic status developed in all cases since all patients required dual anticonvulsant therapy with valproic acid and levetiracetam combined with intravenous anesthetics. According to modern guidelines, measurement of serum levels of anticonvulsant drugs is essential for adequate selection of the dose of these drugs. Unfortunately, the authors did not measure these concentrations. Moreover, they did not discuss this problem. At the same time, it is extremely important to emphasize once again that monitoring of serum anticonvulsant drugs together with prolonged EEG monitoring are essential in patients with refractory and super-refractory epileptic status. These monitoring modalities largely determine the outcomes of disease [8].

There is another interesting finding. Mean therapeutic doses of anticonvulsants usually recommended in patients with seizures, recurrent seizures and status epilepticus but not for refractory and super-refractory status, were ultimately effective [8]. This phenomenon requires an explanation and further research.

Thus, correction of status epilepticus as a result of intraoperative cerebral paradoxical air embolism followed by acute cerebral damage should be based on EEG and drug monitoring data. This monitoring is necessary for selection of adequate dose of anticonvulsant drugs.

Thus, the authors raised an interesting problem of complicated course of intraoperative cerebral paradoxical air embolism. This comment has two main objectives. The first theoretical goal is to emphasize the depth of the problem of intraoperative cerebral paradoxical air embolism in neurosurgical practice. The second practical purpose is to justify the need for guidelines on the prevention of intraoperative cerebral paradoxical air embolism, its early diagnosis and correction, as well as the treatment of its consequences and complications. Creation of these guidelines is a difficult task and requires multidisciplinary approach and cooperation of neurosurgeons, anesthesiologists, specialists for intensive care, neurologists, specialists for ultrasound diagnosis, endovascular surgeons, electrophysiologists and pharmacologists

K.A. Popugaev (Moscow, Russia)

References

1. Chandler WF, Dimcheff DG, Taren JA. Acute pulmonary edema following venous air embolism during a neurosurgical procedure. Case report. Journal of Neurosurgery. 1974;40:400-404.

2. Domaingue CM. Anaesthesia for neurosurgery in the sitting position: a practical approach. Anaesth Intensive Care. 2005;33:323-331.

3. Bedell EA, Berge KH, Losasso TJ. Paradoxic air embolism during venous air embolism: transesophageal echocardiographic evidence of transpulmonary air passage. Anesthesiology. 1994;80(4):947-950.

4. Le Moigne E, Timsit S, Ben Salem D, Didier R, Jobic Y, Paleiron N, Le Mao R, Joseph T, Hoffmann C, Dion A, Rousset J, Le Gal G, Lacut K, Leroyer C, Mottier D, Couturaud F. Patent Foramen Ovale and Ischemic Stroke in Patients With Pulmonary Embolism: A Prospective Cohort Study. Annals of Internal Medicine. 2019;170(4):756-763. https://doi.org/10.7326/M18-3485

5. Abdelghani M, El-Shedoudy SAO, Nassif M, Bouma BJ, de Winter RJ. Management of Patients with Patent Foramen Ovale and Cryptogenic Stroke: An Update. Cardiology. 2019;143(1):62-72. https://doi.org/10.7326/M18-3485

6. Tekle WG, Adkinson CD, Chaudhry SA, Jadhav V, Hassan AE, Rodriguez GJ, Qureshi AI. Factors associated with favorable response to hyperbaric oxygen therapy among patients presenting with iatrogenic cerebral arterial gas embolism. Neurocritical Care. 2013;18(2):228-233. https://doi.org/10.1007/s12028-012-9683-3

7. Sen A. Status epilepticus special edition. Seizure. 2020;75:129-130. https://doi.org/10.1016/j.seizure.2019.12.004

8. Der-Nigoghossian C, Rubinos C, Alkhachroum A, Claassen J. Status epilepticus — time is brain and treatment considerations. Current Opinion in Critical Care. 2019;25:638-646.

Commentary

The issue of venous air embolism and paradoxical air embolism as its secondary complication is still clinically relevant for modern neurosurgery. Therefore, the reviewed article immediately becomes interesting for the reader. The authors comprehensively describe five similar clinical cases occurred in the same hospital (Burdenko Neurosurgery Center) over a five-year period (from 2014 to 2018). All cases are characterized by three common phenomena: 1) intraoperative venous air embolism in a patient’s sitting position; 2) early postoperative brain damage localized at a distance from the surgical area; 3) clinical or electrographic seizures due to brain damage. In my opinion, these data are sufficient to report this series of rare clinical observations. Neurosurgeons, anesthesiologists and specialists for intensive care should be aware of the possibility of this complication and prepared for diagnosis and treatment. All other considerations of the authors are not so important. We can discuss the issue of venous air embolism in patients undergoing neurosurgery in a sitting position, the mechanisms of venous air embolism and associated cerebral damage for a long time. By the way, pulmonary bypass vessels between the pulmonary artery pool and bronchial arteries are normal collaterals rather some exclusive arteriovenous fistulae. Is it necessary in this report? I don’t think so. There are various available literature data on this problem.

I emphasize again that this manuscript is valuable due to demonstration of the complexity of venous air embolism and paradoxical air embolism in patients undergoing neurosurgery in a sitting position

A.Yu. Lubnin (Moscow, Russia)