Subarachnoid and subdural haematoma after attempted spinal anaesthesia for caesarean section

Summary

We report a case of a previously healthy patient who developed a vertebral canal haematoma in the subarachnoid and subdural spaces after a spinal puncture for elective caesarean section. Vertebral canal haematomas are rare. There are different mechanisms for haematoma formation, but coagulation disturbances and trauma, most often due to needle punctures, are the most important. Vertebral canal haematoma may warrant emergent surgical decompression. In this case report we discuss vertebral canal haematomas, including possible mechanisms, clinical diagnosis, imaging modalities, methods for management and advice for patients. We consider the possible association between a vertebral canal haematoma and non‐steroidal anti‐inflammatory drugs, and draw attention to an existing black box warning for ketorolac. In this case, we explain why a conservative approach was chosen with a good outcome. We also report the effect of this complication on the patient experience.

Introduction

Vertebral canal haematoma (VCH) is a rare but potentially catastrophic complication following neuraxial puncture. The Third National Audit Project of the Royal College of Anaesthetists (NAP 3) reported six VCHs amongst 707,425 central neuraxial blocks (CNB) [1]. A meta‐analysis of obstetric patients reports an incidence of one VCH per 168,000 epidurals [2]. Most VCHs occur in the epidural space, representing 74.2% of all cases. The remaining 25.8% are subarachnoid or subdural haematomas [3]. Vertebral canal haematomas at any site can result in neurological injury.

Vertebral canal haematomas may occur spontaneously or may be associated with anticoagulants or antiplatelet drugs; congenital or acquired coagulopathies; difficult neuraxial punctures; and anatomical abnormalities affecting the spinal cord or column. The Third National Audit Project mentions both tissue disruption (‘trauma’) and coagulation impairment frequently associated with causation [1]. The clinical presentation of VCH includes back or buttock pain, with or without radicular symptoms, motor weakness, sensory alterations and/or urinary or faecal incontinence.

The preferred imaging modality to confirm the diagnosis is magnetic resonance imaging (MRI), as computed tomography may provide false negatives. Additionally, spinal MRI is more accurate in determining the location, extension and age of the haematoma; assessing for spinal cord compression; and identifying vascular malformations [4].

The neurological sequelae are due to mechanical compression or inflammation of the neural tissue or the meningeal membranes, and treatment for this condition ranges from conservative management to emergency surgical decompression within less than 12 h [1]. Conservative treatments have been reported with satisfactory results in patients with back pain and leg weakness without paralysis [5].

Case report

A 32‐year‐old healthy, para 1 pregnant woman was admitted for elective caesarean section at 36‐weeks gestation owing to breech fetal presentation. Her only medical history was anaemia, requiring iron supplementation, during pregnancy. The patient had no personal or family history suggestive of coagulopathy. She had received an epidural for labour analgesia six years previously without complication. There was no history of other neuraxial technique or of previous surgery. Physical examination was normal. Relevant preoperative laboratory testing, as per our institutional protocol, revealed a normal platelet count, normal prothrombin time and normal activated partial thromboplastin time. The anaesthetic technique chosen was a subarachnoid block.

After informed consent, a subarachnoid puncture with a 27 G Quincke needle was attempted at two different intervertebral levels. The initial attempt was at the presumptive intervertebral space L3–L4 and a dural puncture was achieved. However, the cerebrospinal fluid (CSF) was mixed with blood that did not clear. The needle was withdrawn and another attempt made at the intervertebral space below (presumed L4–L5). Again, the CSF was mixed with blood that did not clear. Consequently, the anaesthetic plan changed to general anaesthesia.

General anaesthesia was induced with propofol, with rocuronium for muscle relaxation. Maintenance was with sevoflurane (1%) in nitrous oxide and oxygen (50/50%). A 5 IU oxytocin bolus was given following delivery, followed by an infusion at 10 IU.h^‐1. Intraoperatively, fentanyl 200 μg; paracetamol 1 g; tramadol 150 mg; ketorolac 30 mg; and ondansetron 4 mg were administered. Muscle relaxation was reversed with sugammadex 200 mg prior to tracheal extubation. The surgery was uncomplicated, and the estimated blood loss was less than 500 ml. No abnormalities were recorded during the immediate postoperative period.

In the ward, the patient recovered normally. She did not receive pharmacological thromboprophylaxis as per our institutional risk assessment and guidelines. Postoperative pain was treated with regular paracetamol and ketorolac, plus tramadol as required. She was discharged home at the end of the second postoperative day without any complaints.

The next day (third postoperative day), she began to experience progressive lower back pain, lower limb paraesthesia and holocranial headaches. However, she did not seek medical attention until the sixth postoperative day, when she was hospitalised for investigation.

A complete neurological examination revealed an antalgic gait. No other anomaly was detected. She had normal leg strength; normal cranial nerve examination; no nuchal rigidity; negative Babinski, Brudzinski and Kernig signs; normal deep tendon reflexes; and normal anal sphincter tone.

The patient underwent computed tomography imaging of the spine that suggested a possible VCH. Spinal MRI confirmed “a subarachnoid/subdural hematoma, between T11 and L3, with a filiform extension into the lumbosacral region, causing deviation of the spinal cord” (Fig. 1). It was decided not to manage this surgically because there were no motor deficits, and the clinical manifestations were stable. The patient was kept under observation and bed rest. Her brief reflections on her experience are presented in Table 1.

Figure 1.

T2‐weighted dorso‐lumbar sagittal MRI of the spinal cord showing a haematoma between T11 and L3 (arrows).

Table 1.

Patient perspective.

“The experience was terrifying, and it has caused me great concern the prospect to have to be submitted to spinal surgery and the possibility that I could have become permanently disabled. It caused me great disturbance not to be able to care for my baby in his first days as I had expected. I never thought I would live such a nightmare. I am afraid this might happen again in a future anaesthesia, because I can´t understand why this has happened to me, considering I was assured everything was performed correctly.”

The symptoms resolved completely within 14 days. Before discharge, further spinal MRI showed partial reabsorption of the dorso‐lumbar haematoma. Eight months later, a third MRI revealed complete absorption.

Discussion

In our case, the haematoma in the subarachnoid and subdural space may have been caused by the neuraxial punctures. However, this rarely occurs in patients without anticoagulant therapy. The exact mechanisms by which subarachnoid haematomas develop after spinal anaesthesia are still unclear [6]. Several attempts at spinal anaesthesia have been implicated as causing rupture of spinal cord vessels. Haemorrhage may result from injury to small root vessels that enter the subarachnoid space with each segmental nerve root. A factor to consider is the needle diameter [7]. Thinner needles may result in a lower incidence of post dural puncture headache (PDPH); however, these needles may also deviate more easily from the intended path or cause delayed CSF flow through the needle leading to unnecessary attempts and consequently more neuraxial trauma. In our case, a 27 G Quincke needle was used in two different intervertebral spaces. However, there was no report of great difficulties during the procedure and the reason to abandon the regional anaesthesia was that blood was found in two consecutive spinal taps.

Considering other mechanisms for VCH, the MRI did not reveal any vascular malformations, but it is conceivable that small malformations could pass unrecognised. Also, the patient was not aware of a family history of coagulopathy and she had no identifiable risk factors for acquired coagulopathy, such as pre‐eclampsia; eclampsia; hepatic failure; thrombotic thrombocytopenic purpura; renal failure; or a haematological disease. However, subclinical coagulopathy is a possibility. For example, the symptoms of von Willebrand disease may start at any age and can vary in severity.

An intriguing aspect demonstrated by the MRI is the anterior location of the haematoma in relation to the spinal cord, and its higher level of distribution (T11‐L3) relative to the puncture site (Fig. 1). Movement with gravity could explain the cephalad dislocation of the haematoma in a recumbent patient; however, this explanation is inconsistent with its anterior location. The patient was not in the recumbent position for prolonged times during the first days after the neuraxial punctures. There may be two possible explanations. The first is the possibility that the patient suffered a coincidental spontaneous haematoma in the subdural space during those days, unrelated to the spinal punctures. The second and more plausible explanation is that a small arterial vessel in the subdural space could have been damaged by the spinal needle during the block placement, and this would explain the cephalad and anterior spread, caused by the leakage of blood under higher pressure.

Another element worthy of reflection is that the patient received ketorolac intra‐ and postoperatively for analgesia. The Association of Anaesthetists of Great Britain & Ireland, Obstetric Anaesthetists' Association and Regional Anaesthesia UK state that “no additional precautions” are required for patients taking non‐steroidal anti‐inflammatory drugs (NSAIDs) around the time of regional anaesthesia [8]. However, in 2005, the American Food and Drug Administration (FDA) issued a black box warning that states “TORADOL (ketorolac tromethamine) inhibits platelet function and is, therefore, contraindicated in patients with suspected or confirmed cerebrovascular bleeding, patients with haemorrhagic diathesis, incomplete haemostasis, and those at high risk of bleeding” [9]. The recent Joint European Society of Anaesthesiology and Intensive Care/European Society of Regional Anaesthesia guidelines for regional anaesthesia in patients on antithrombotic drugs does not fully address NSAIDs; however, they note the time interval between a spinal tap and the next antithrombotic drug should possibly be increased [10]. Considering these statements, it may be appropriate to delay the administration of ketorolac after a haematic spinal puncture.

Conservative medical management was appropriate in our case, given that the neurological changes were limited and stable and considering the risks of spinal surgery. A spinal haematoma is frightening to every anaesthetist, but we should remember that the most effective treatment for PDPH is a medically induced epidural haematoma. It is also important to know that there are reports of subarachnoid haematomas associated with mild complaints, which were successfully treated conservatively [6]. The presence of VCH can be found on postoperative imaging in 33–100% of patients after spinal surgery, but only rarely will patients experience any neurologic deficit [4]. However, even a small haematoma must be monitored closely and urgently decompressed if needed, because irreversible neurologic deficits may occur rapidly. During the admission, the patient's neurological function was monitored via regular clinical examination including testing sensory dermatomes and motor function. In case of worsening symptoms, she would have undergone an urgent MRI and evaluation by a spinal surgeon. This highlights the importance of ensuring MRI and spinal surgery availability in the same facility when caring for these patients.

The coagulation studies (platelet count, prothrombin time and activated partial thromboplastin) remained within normal values during the second hospital stay. After discharge, the patient was referred to a haematology consultation for a thorough investigation. Unfortunately, the patient missed the medical appointments, despite repeated invitations to attend. The cause of VCH formation therefore remains unknown, and this raises the question of what should be done in case of any future surgery where neuraxial anaesthesia may be contemplated. Considering the reason for the haematoma formation remains unknown, we feel that neuraxial anaesthesia and the use of drugs with antiplatelet effects should be avoided in the peri‐operative setting. However, labour analgesia and anaesthesia are a more challenging situation. In this case, informed consent and acknowledgement of risks is important. Intravenous alternatives such as patient controlled use of short‐acting and fast‐onset analgesics like remifentanil may be a safer approach, with careful consideration for neuraxial procedures for delivery anaesthesia.

It is noteworthy that the symptoms appeared gradually a few days after the spinal puncture. This highlights the importance of informing patients of the symptoms of vertebral canal haematomas and how to seek appropriate help.

In summary, a high level of suspicion of VCH should be maintained after a neuraxial puncture whenever new neurological symptoms and signs present. Vertebral canal haematoma after a neuraxial block may have a late presentation and patients should be informed of this possibility. A subarachnoid/subdural haematoma may be conservatively managed if there are no motor deficits. Finally, this report highlights a possible association between the use of ketorolac and the development or amplification of a vertebral canal haematoma. Delaying or withholding the administration of NSAIDs after a haematic puncture or a difficult block requires further investigation.