Spinal subarachnoid haematoma after neuraxial anaesthesia in a patient with polycythaemia vera

Summary

Polycythaemia vera is a common haematological proliferative disorder. It is characterised by uncontrolled red cell production with ensuing peri‐operative vaso‐occlusive and haemorrhagic complications. Spinal haematoma after neuraxial anaesthesia is rare; most cases are associated with technical difficulties or bleeding disorders. Current consensus opinion suggests that neuraxial anaesthesia in patients with polycythaemia vera is safe due to a lower risk of thrombotic events and hypoxaemia compared with general anaesthesia. We report a case of a spinal subarachnoid haematoma after uneventful neuraxial anaesthesia in a patient with optimised polycythaemia vera. Despite an emergent laminectomy, the patient developed permanent motor deficits. This report highlights that although neuraxial anaesthesia is recommended by many authors, patients with polycythaemia vera can paradoxically have an increased haemorrhagic risk from platelet dysfunction and acquired von Willebrand disease. Clinicians proceeding with surgery under neuraxial anaesthesia should appreciate these risks even in patients with normal or apparently elevated thrombotic states. This case also demonstrates that traditional coagulation tests may need to be complemented by pre‐operative platelet function tests and screening for von Willebrand disease. Finally, the importance of the patient participation in the choice of the anaesthesia technique cannot be understated, with specific attention paid to this frequently unrecognised risk.

Introduction

Polycythaemia vera is a myeloproliferative disease of hematopoietic stem cells, characterised by an overproduction of all myeloid lineage cells secondary to a mutation in the JAK‐2 gene. Although polycythaemia vera is one of the more common myeloproliferative disorders, it is still rare, with incidence estimated at 1.9 in 100,000 [1, 2, 3, 4]. The resultant erythrocytosis leads to blood hyperviscosity, which in turn impairs microcirculation. This explains the most common symptoms: headache; dizziness; visual disturbance; angina; and intermittent claudication [1, 2, 3, 4].

During the peri‐operative period, patients are at increased risk of thrombo‐embolic complications and paradoxical haemorrhagic events [1, 2, 3, 4]. This dichotomy presents an anaesthetic challenge, particularly for neuraxial anaesthesia. Nevertheless, there are many reports of successful regional anaesthesia in these patients, and many authors consider this the technique of choice, as regional anaesthesia has been reported to reduce thrombotic risk relative to general anaesthesia [5, 6, 7]. Additionally, haemorrhagic risk has been reported to be negligible when polycythaemia vera is controlled [3, 4].

We present a case of a spinal subarachnoid haematoma after uneventful neuraxial anaesthesia in a patient with optimised polycythaemia vera. Furthermore, we reflect on the peri‐operative approach to, and the implications of, anaesthetising a patient with this condition.

Report

An 83‐year‐old man was scheduled to undergo an elective total knee arthroplasty. His past medical history included polycythaemia vera, (managed with aspirin and regular venesection, and hypertension, controlled with ramipril and felodipine. No history of thrombotic or haemorrhagic complications was recorded. Pre‐operative blood tests showed a haemoglobin concentration of 160 g.l⁻¹; haematocrit of 52% platelet count of 336 × 10⁹.l⁻¹; and coagulation parameters within the normal reference range. Aspirin was stopped 5 days before surgery and a precautionary venesection was performed. Another blood test performed on day of surgery showed reductions in haemoglobin concentration (130 g.l⁻¹); haematocrit (40%); and platelet count (233 × 10⁹.l⁻¹).

With the patient in the sitting position, a spinal anaesthetic was performed in what was thought to be the L2–3 intervertebral space with a 27‐G Whitacre needle. The technique was successful on the first attempt with free flow of clear cerebrospinal fluid. The injectate contained 12.5 mg of hyperbaric bupivacaine and 2.5 μg of sufentanil. This achieved an adequate block for surgery, extending to the T10 dermatome.

The intra‐operative period was uneventful with an estimated 200 ml of blood loss. The block had completely resolved after 2 h with no motor or sensory deficits at discharge from the recovery room.

At 12 h postoperatively, the patient developed flaccid paraplegia of the lower limbs and diminished sensation, corresponding to a neurological level of T12. Magnetic resonance imaging showed a subarachnoid haematoma extending from T11 to L1 (Fig. 1). Emergency laminectomy was performed and, immediately after, there was a recovery of sensory deficits, although lower limb weakness remained. The patient was discharged home with a significant motor deficit (muscular force 2/5), which did not improve despite physiotherapy. One year after surgery, the patient remains in a wheelchair, with no further improvement and consequent severe limitation to his autonomy and activities of daily living.

Figure 1.

Magnetic resonance imaging demonstrating the T11 to L1 subarachnoid haematoma (white arrows).

Discussion

Due to an increased risk of thrombotic and haemorrhagic complications, anaesthetic management of patients with polycythaemia vera may be challenging (Table 1). In the first instance, the hyperviscosity due to erythrocyte proliferation predisposes patients to vaso‐occlusive and thrombotic complications. The peri‐operative thrombotic risk is increased 100‐fold, particularly when the haematocrit is greater than 55%. Moreover, hyperviscosity also affects the pulmonary capillaries, leading to a sluggish circulation that impairs oxygenation and predisposes to hypoxaemia. To minimise this, haematocrit should be maintained at less than 45% [2, 3, 4, 5].

Table 1.

Anaesthetic considerations in patients with polycythaemia vera.

Pre‐operative
Obtain history of thrombotic or haemorrhagic complications
Treatment
Aspirin – consider suspension only if high surgical haemorrhagic risk
Myelosuppressive agents (hydroxyurea)
Regular venesection
Ensure haematocrit < 45% and platelet count < 400×109.l−1
Avoid dehydration and administer intravenous fluids
Avoid prolonged fasting
Intra‐operative
Promote adequate hydration Avoid hypoxia General anaesthesia versus regional anaesthesia
General anaesthesia	Regional anaesthesia
↑Risk of hypoxia ↑Risk of thrombosis	↓↓Thrombotic risk Risk of potentially catastrophic complication due to haemorrhagic risk (neuraxial anaesthesia)
Postoperative
Observe for thrombotic or haemorrhagic complications

Paradoxically, polycythaemia vera is associated with an increased risk of haemorrhagic complications. The mechanism is poorly understood and may be multifactorial. It has been hypothesised that, despite thrombocytosis, there are functional and qualitative platelet abnormalities which increase haemorrhagic risk [3, 4]. In vitro studies of affected platelets have shown reduced aggregation in response to adrenaline, adenosine diphosphate and collagen [4].

Additionally, acquired von Willebrand disease has been associated with polycythaemia vera and may be the principal cause of the haemorrhagic diathesis seen in patients with polycythaemia vera and high platelet counts [4]. Von Willebrand factor is a plasma protein released from platelet and endothelial storage in response to vascular injury. It plays a crucial role in primary haemostasis as it promotes platelet adhesion to sites of endothelial injury. Its active form is a multimeric protein that undergoes progressive proteolysis by a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13 (ADAMTS13) enzyme to a smaller and less active forms [4]. In polycythaemia vera, thrombocytosis reduces the concentration of large von Willebrand factor multimers in the circulation. The mechanism for this phenomenon is not fully understood, but it is thought that affected platelet binding of von Willebrand factor promotes a conformational change which facilitates von Willebrand factor proteolysis by ADAMTS13 [4]. Although important, haemorrhagic complications are much less frequent, especially in controlled disease. It is generally accepted that for platelet counts under 400 × 10⁹.l⁻¹, haemorrhagic risk is negligible [2, 4, 5].

Pre‐operative optimisation is crucial to mitigate risk. Myelosuppressive agents, such as hydroxyurea, or venesection may be necessary to achieve the recommended haematologic cut‐offs before surgery. In order to reduce blood viscosity, adequate hydration must be promoted, prolonged fasting periods avoided, and, if necessary, intravenous fluids administered. Many of these patients will be on antiplatelet therapy, and peri‐operative management should be individualised to patient risk [1, 2, 3, 4].

Neuraxial anaesthesia is contra‐indicated in patients with bleeding disorders, and therefore the safety of neuraxial techniques in patients with polycythaemia vera has been questioned. However, as early as 1950, authors have stated that regional anaesthesia techniques should be used whenever possible due to the problems with hypoxaemia under general anaesthesia in this patient population [1, 2]. Moreover, it is well established that regional anaesthetic techniques reduce thrombotic risk in normal patients [1]. This may be an advantage of regional anaesthesia in polycythaemia vera, where patients have a high risk of thrombosis [1], and there have been several reports on the safe use of neuraxial anaesthesia in patients with controlled polycythaemia vera [1, 2, 3, 4, 5, 6].

Spinal haematomas are rare, but well‐recognised complications of neuraxial anaesthesia [7]. Incidence is estimated as 1 in 18,000 for epidural anaesthesia and 1 in 775,000 for spinal anaesthesia [7]. Epidural haematomas represent the most frequent presentation of this pathology (74% of cases) and subarachnoid haematomas are extremely rare (less than 10% of all spinal haematomas) [8]. The vast majority of haematomas after spinal anaesthesia are associated with coagulation disorders or technical difficulties during insertion [7, 8]. Needle tip and size also influence the incidence of spinal haematoma, with smaller needles carrying a lower risk [7, 8].

Although the patient in this report had no history of thrombotic events, he nevertheless had a high peri‐operative basal risk of thrombotic complications [9]. In addition, this risk was theoretically further increased by polycythaemia vera, making neuraxial anaesthesia an attractive option. The decision to proceed with spinal anaesthesia was based on a normal coagulation profile and no history of bleeding diathesis. The patient was optimised before surgery with venesection, and the immediately pre‐operative platelet count was below the recommended safety threshold. Aspirin was stopped to further minimise platelet dysfunction. A small spinal needle was used and an uneventful first attempt spinal was performed. However, despite the implementation of recommended thrombotic and haemorrhagic risk reduction strategies, the patient developed a spinal subarachnoid haematoma.

Peripheral nerve blocks could have been a safe alternative to neuraxial or general anaesthesia in this case. Postoperative analgesia and functional outcomes seem to be similar when neuraxial techniques or peripheral nerve blocks are used [10]. Moreover, our patient had no medical conditions that predisposed him to pulmonary complications and his pre‐operative haematocrit was 40%, which is below the value recommended to minimise microcirculation sludge. Therefore, the risk of hypoxaemia and thrombotic complications was theoretically minimised.

This case highlights that although neuraxial anaesthesia is recommended, the increased haemorrhagic risk from platelet disfunction associated with polycythaemia vera, even in apparent controlled disease should not be overlooked. In our opinion, platelet functional tests, like aggregometry, and von Willebrand disease screening may be of value to complement traditional coagulation tests in patients with polycythaemia vera before surgery. More studies are needed to demonstrate the efficacy of screening tests for pre‐operative haemorrhagic diathesis in patients with polycythaemia vera. Moreover, this case is also a good example of the importance of informed consent for anaesthesia, tailored to the individual or subjective patient risk profile.