Fatal postoperative intracranial hypotension‐associated venous congestion after elective laminectomy

Summary

Postoperative intracranial hypotension‐associated venous congestion is a rare complication that features radiologic characteristics suggestive of hypoxic ischaemic encephalopathy without systemic hypoxia or hypotension actually occurring. This makes the condition prone to being misidentified as a complication of anaesthesia. In this case, a patient undergoing emergency haematoma evacuation after a previous lumbar laminectomy lost cerebrospinal fluid rapidly and accidentally via a wound drain. She subsequently developed postoperative seizures and eventually died some days later having never regained consciousness. A magnetic resonance imaging scan acquired after the event showed signs initially thought to be indicative of hypoxic ischaemic encephalopathy but was later identified to be postoperative intracranial hypotension‐associated venous congestion due to temporary obstruction of the great cerebral vein, resulting in local rather than systemic hypoxia. Anaesthetists should be aware of this rare condition and be familiar with its pathophysiology and presentation.

Introduction

Postoperative intracranial hypotension‐associated venous congestion or pseudohypoxic brain swelling, is a rare and relatively unknown complication that occurs after uneventful brain or spine surgery where the dura mater has been opened intentionally or unintentionally [1, 2, 3, 4, 5, 6, 7]. Wound drainage systems and a rapid loss of cerebrospinal fluid (CSF) are usually involved. Its features include failure to awaken after surgery with or without convulsions, as well as radiological features that suggest hypoxic ischaemic encephalopathy even though no apparent surgical or anaesthetic cause can be identified. This can lead to misidentification as anaesthetic complication with potentially harmful consequences. We therefore would like to draw attention to this condition among anaesthetists.

Report

An 87‐year‐old woman with a history of chronic obstructive pulmonary disease, hypertension, chronic kidney disease and type 2 diabetes mellitus presented with lumbar spinal stenosis at L4–L5 and neurogenic claudication that was progressively worsening. It was refractory to conservative treatment including analgesics and physiotherapy. Because she was still independent in activities of daily living but severely impaired by her symptoms, she was scheduled for laminectomy. The procedure was performed on spinal levels L4 and L5. From a surgical and anaesthetic point of view, it proceeded without complication. Two days postoperatively the patient developed worsening pain around the incision. Soon thereafter she developed urinary retention and faecal incontinence, indicating cauda equina syndrome. An emergency lumbar magnetic resonance imaging (MRI) scan revealed haematoma with nerve compression, necessitating emergency decompressive surgery.

Anaesthesia was induced with propofol and sufentanil and maintained with sevoflurane. Succinylcholine was used to facilitate tracheal intubation. During the operation she maintained respiratory and cardiovascular stability. The patient's arterial oxygen saturation and end‐tidal CO₂ concentration remained in the normal range and there was no significant hypotension or hypertension. The haematoma was evacuated with the patient prone and the sources of bleeding were cauterised. The neurosurgeon left a drainage catheter in the cavity and closed the wound. The vacuum was subsequently activated, leading to immediate aspiration of at least 50 ml of clear liquid with a reddish tinge. The negative suction was deactivated and the wound re‐opened. A laceration of the dura mater, close to the drain tip was discovered and closed using a fibrinogen‐thrombin‐matrix whereupon the incision was resutured.

The patient was returned to the supine position and emergence from anaesthesia commenced. After return of spontaneous breathing, the patient's trachea was extubated. Shortly after tracheal extubation she developed generalised tonic clonic seizures, which were terminated with boluses of propofol (50 mg). During this period, she received oxygen 100% with a maximal desaturation to 83% for less than 1 min. She had no hypoglycaemia. Her airway was supported with an oropharyngeal airway and a computed tomography (CT) scan of the brain was performed. Besides slender ventricles, the CT revealed no significant structural abnormalities and no signs of cerebrovascular accident. She was admitted to the intensive care unit, where she experienced another convulsion. Her trachea was re‐intubated, and she was sedated with propofol and given levetiracetam, with phenytoin later added. The patient did not regain consciousness with sedation breaks, but instead developed convulsions requiring resedation after neurological assessment. An electroencephalogram (EEG) was conducted without sedation showing diffuse continuous slowing and no signs of epilepsy. All laboratory values remained within their reference ranges.

After 3 days, an MRI scan of the brain revealed symmetrical oedema of the deep grey matter, especially in the basal ganglia and thalami with some small punctate bleeds (Figure 1). Several parieto‐occipital cortical areas close to the ventricles showed diffuse perfusion restriction. Arterial and venous structures appeared open without thrombi or infarctions. There was no contrast enhancement indicative of encephalitis or meningitis. A hyperintense rim at the lateral and inferomedial lentiform nuclei with unknown clinical importance was observed (Figure 1). On postoperative day 8, sedation was discontinued without neurological improvement. The patient subsequently died after withdrawal of active treatment on postoperative day 10.

Figure 1.

Magnetic resonance imaging showing postoperative intracranial hypotension‐associated venous congestion, charactersied by (a) a T2‐weighted image showing bilateral oedema of basal ganglia and thalami (arrows); (b) a fluid attenuated inversion recovery image showing bilateral oedema and bilateral lentiform rim sign (marked unilaterally in green); and (c) susceptibility‐weighted image showing punctate bleeding in the basal ganglia.

The radiologist – who at the time of brain imaging did not know all the peri‐operative details – initially concluded that the clinical picture fitted most with hypoxic ischaemic encephalopathy. However, all continuously recorded electronic records including arterial blood pressure, heart rate, pulse oximetry and capnography showed there was no significant systemic hypoxia, hypercapnia or cardiovascular instability pre‐, intra‐, or postoperatively, or during the ICU admission that could explain how hypoxic ischaemic encephalopathy might have developed. The CT and MRI scans excluded haemorrhage, arterial occlusion, meningitis and encephalitis as possible causes of the comatose state but found ischaemic damage to the basal ganglia, creating a discrepancy with the aforementioned clinical records. The EEG had excluded subclinical status epilepticus from the differential diagnosis. The common laboratory values, including ammonia, electrolytes and glucose, did not show significant deviations or signs of infection that could help to explain the complication. The most important abnormality identified was the rapid and substantial loss of CSF during the second surgery. After an extensive review of the literature and re‐evaluation of the history and MRI scan with the involved physicians, the problem was determined to be postoperative intracranial hypotension‐associated venous congestion, formerly called pseudohypoxic brain swelling.

Discussion

Postoperative intracranial hypotension‐associated venous congestion is a rare complication after lumbar laminectomy that mimics typical hypoxic ischaemic encephalopathy [1, 2, 3, 4, 5, 6]. It usually occurs after loss of significant amounts of CSF following spinal or cerebral surgery with intended or unintended opening of the dura mater. Unexpected postoperative deterioration and failure to regain consciousness – sometimes accompanied by seizure activity – are typical and may result in death or leave the patient impaired permanently. After excluding the most common causes of postoperative altered conscious state, a MRI scan can show abnormal signal intensities in the basal ganglia, thalami and cerebellum. These findings initially appear typical for hypoxic ischaemic encephalopathy even though, as in our case, no systemic hypoxia or circulatory instability has been observed. Therefore, this condition may often result in a misdiagnosis.

One factor that has been identified and accepted as a cause in all known postoperative intracranial hypotension‐associated venous congestion cases is rapid CSF loss via drainage from cranial or lumbar wound drains [1, 2, 3, 4, 5, 6]. It is believed that rapid drainage of extensive amounts of CSF (the total CSF volume being approximately 150 ml with a daily production and turnover of approximately 450 ml [7]) leads to sudden intracranial hypotension with temporary brain sagging, leading to congestion and outflow impairment as the great cerebral vein close to the straight sinus and adjacent to the tentorium kinks, resulting in venous infarction [1, 2, 3, 4, 5, 6]. Intracranial hypotension is defined as CSF pressure <6 cm H₂O and usually caused by a dural leak [4]. The Monro‐Kellie hypothesis states that the inelastic cranial compartment has a fixed volume consisting of brain tissue, CSF and blood that remain in equilibrium, meaning that a decrease in one (CSF) results in a compensatory increase of the one or more of the others (venous blood) [8]. This means that after rapid loss of at least 50 ml of CSF (representing about one third of the total volume) an initial venous distension compensated for this ‘reduction’ in intracranial volume. Loss of the buoyant action of CSF allowed a temporary downward displacement of the brain, causing great cerebral vein compression [1, 2, 3, 4, 5, 6, 8, 9]. This explains the radiological findings indicating hypoxic damage without visible occlusion of vessels manifesting as oedema of basal ganglia and thalami, since this temporary occlusion is resolved by replenishment of the lost CSF. Here, the MRI was conducted 3 days postoperatively, giving sufficient time for CSF replacement. A recent publication by Sotoudeh et al. proposed the lentiform rim sign as MRI marker to distinguish postoperative intracranial hypotension‐associated venous congestion from hypoxic ischaemic encephalopathy [5]. A hyperintense rim at the lateral and inferomedial lentiform nuclei border in fluid attenuated inversion recovery sequences appears in several described PIHV cases and has not been seen in HIE or other pathologies. The lentiform ring sign seems to result from the pathophysiologic congestion of deep cerebral veins (especially the great cerebral vein) without impairment of the superficial veins draining the peripheral brain parenchyma. The difference in oedema between these venous regions appears to cause the sign to be expressed in postoperative intracranial hypotension‐associated venous congestion but not in hypoxic ischaemic encephalopathy [5]. We have demonstrated the lentiform rim sign in our case (Figure 1).

The occurrence of a dura mater laceration might have been facilitated in our case by the advanced age of the patient and decreased tissue elasticity [10]. The avoidance of wound drains with negative pressure suction has proven to be the most effective preventive measure for postoperative intracranial hypotension‐associated venous congestion [2, 3, 6]. Additionally, usage of softer drain tubes and paying extra attention to possible intra‐operative CSF leaks might reduce the risk of postoperative intracranial hypotension‐associated venous congestion. Unfortunately, even with these protective measures, unpredicted CSF leakage with fatal outcome can occur.

To conclude, postoperative intracranial hypotension‐associated venous congestion is a rare but potentially fatal complication following brain or spine surgery. The sudden loss of CSF results in obstruction of venous outflow from the brain, with the resultant local ischaemia producing radiological findings suggestive of hypoxic ischaemic encephalopathy without a history of systemic hypoxia or cardiovascular instability during surgery. Postoperative intracranial hypotension‐associated venous congestion is prone to being misidentified as an anaesthetic complication with failure to awaken and postoperative convulsions. It should be considered a differential diagnosis in appropriate circumstances after having excluded the most common causes of postoperative altered conscious state and seizures. The lentiform rim sign might be a potential MRI marker for postoperative intracranial hypotension‐associated venous congestion in the appropriate context.