Iatrogenic Spinal Subdural Hematoma due to Apixaban: A Case Report and Review of the Literature

In the last decade, the clinical relevance for developing safer oral anticoagulants prompted the development of new classes of drugs that have shown a lower risk of life-threatening bleeding events as compared to standard warfarin. Nontraumatic spinal subdural hematoma is an uncommon urgent complication that can be associated with the use of these agents. An unusual case of spinal subdural hematoma related to apixaban treatment for nonrheumatic atrial fibrillation is reported here.


Introduction
Over more than 50 years, warfarin has been the only oral anticoagulant. Warfarin is a vitamin K antagonist with welldocumented therapeutic e cacy, despite potential limitations including drug interactions, delayed e ect onset of action, narrow therapeutic window, regular measurement of the international normalized ration (INR), and a number of associated hemorrhagic adverse events. For this reason, there has been a great interest in the last decade for developing novel oral anticoagulants to reduce bleeding risks.
e new generation of nonvitamin K antagonist oral anticoagulants (NOACs) such as dabigatran, apixaban, and rivaroxaban has more predictable anticoagulant responses and has been shown to be e ective in the prevention and treatment of venous thromboembolism and in the prevention of stroke and systemic embolism in patients with nonvalvular atrial brillation [1].
Intracranial hemorrhage is one of the most severe complications associated with NOACs as well as posttraumatic spinal hematoma. By contrast, development of a nontraumatic spinal hematoma as a consequence of anticoagulant treatment is an exceptional complication, the natural history of which has been poorly described. Only a few cases of nontraumatic spinal subdural hematoma related to treatment with NOACs have been reported in the literature [2][3][4]. We report what appears to be the rst case of nontraumatic spinal subdural hematoma in a woman with atrial brillation during treatment with apixaban. In all previous cases, rivaroxaban was the drug involved in the hemorrhagic adverse e ect. e case described here adds further evidence to the development of this life-threatening neurosurgical entity in the context of treatment with the new oral anticoagulants.

Case Presentation
A previously healthy 75-year-old woman with paraparesis was referred to our hospital for diagnostic studies based on tentative diagnoses of low spinal vascular malformation or secondary spinal bleeding. She had a history of hypertension treated with enalapril 5 mg/daily and atrial brillation treated with NOACs. Ten days before admission to the hospital, she presented to the emergency room of a regional hospital with abdominal pain. On physical examination, a spontaneous hematoma 3 × 4 cm in diameter in the left iliac muscle was found. e patient was taking dabigatran at the dose of 150 mg twice a day. Change of dabigatran to apixaban (2.5 mg twice a day) was recommended because of the high cardioembolic risk associated with underlying atrial brillation. e dose of apixaban was reduced because of bleeding risk. No other risk factors such as hemostatic or spinal disorders were present. Results of laboratory tests including coagulation tests and renal function tests were unrevealing. Ten days after discharge from the emergency room, she developed multiple cutaneous hematomas in the upper extremities and a sudden onset of predominantly distal bilateral crural motor de cit. She was admitted to the hospital where a predominantly left paraparesis was conrmed. A spinal magnetic resonance imaging (MRI) revealed a hemorrhagic focus suggestive of spinal vascular malformation, which was not con rmed at operation.
On admission to the Department of Neurology of our hospital, neurological examination showed paraparesis, predominantly distal and of the left side, with discrete hyperre exia, hypoesthesia of the low dorsal deep sensitivity, indi erent left plantar re ex, and normal response on the right side. MRI of the spine revealed a large subdural hemorrhage with discontinuous cervical-dorsal-lumbarsacral involvement with secondary spinal cord compression but without signs of associated vascular malformation (Figures 1 and 2). e patient was treated surgically, and D1 to D3 vertebral laminectomy was performed. e subdural hematoma was evacuated (36 hours after hospital admission), and the diagnosis of subacute spinal hematoma was con rmed. Postoperatively, she was treated with increasing doses of dexamethasone 12 mg/24 h and standard analgesic medication. e clinical course was satisfactory, but a control MRI scan performed one week after surgery showed persistence of subdural hemorrhage and spinal compression at D4 to D7 segments with resolution of hematoma at D1-D3 (Figure 3). e patient was reoperated and underwent D4 and D7 partial laminectomy and D5 to D7  complete vertebral laminectomy. Postoperatively, paraparesis improved slowly, and the patient was discharged 3 weeks after the last surgical procedure with assisted standing and initiation of walking with aids. Six months after the hospital discharge, the patient's clinical condition had largely improved. She was able to walk alone with the aid of a stick, with mild weakness graded 4/5. A control MRI scan showed D1-D3 and D5-D7 laminectomies without hemorrhagic signs in the spinal cord.

Discussion
Spinal hematoma is an uncommon clinical entity that can cause sudden and sometimes irreversible neurological impairment if diagnosis and treatment are not promptly established. Apparently, the rst clinical case of spinal hematoma was reported by Jackson in 1869 under the term of "case of spinal apoplexy" [5], and the rst case of spinal hematoma, which was successfully evacuated by a surgical operation, was reported in 1911 [6]. Depending on its location, the bleeding can be classi ed as epidural, intradural (subdural), subarachnoid, or intramedullary. Epidural hematomas are the most frequent (75% of cases), with subdural hematomas being extremely rare. e exact underlying mechanism leading to spontaneous nontraumatic subdural hematoma remains controversial. It is postulated that sudden increases of pressure in the thoracic and/or abdominal cavities can raise the pressure inside the subdural vessels, with rupture of these vessels and subsequent formation of the hematoma [6]. Haines et al. [7] studied the dura-arachnoid junction with electron microscopy. e dura is composed of elongated, attened broblasts and copious amounts of extracellular collagen, whereas the dural border layer (dura-arachnoid junction) is characterized by attened broblasts, no extracellular collagen, extracellular spaces, and few cell junctions. Under normal conditions, there is no evidence of a naturally occurring space at the dura-arachnoid junction, but a space may appear at this point subsequent to pathological/traumatic processes that result in tissue damage with a cleavage opening of the structurally weakest plane in the dural border cell layer. erefore, subdural hematomas should be considered intradural or hematomas of the dural inner cell border.
Spinal subdural hematomas, as in the case reported here, are usually secondary to vascular lesions, underlying tumors, or result from invasive procedures, such as lumbar puncture and spinal anaesthesia (or even acupuncture) [8], associated with blood dyscrasias and coagulation defects [9]. In a review of 106 cases of nontraumatic acute subdural spinal hematoma, there was a defect in the hemostatic mechanism in 57 patients (54%) and a history of spinal puncture in 50 patients (47%) [10]. Primary or spontaneous nontraumatic spinal hematomas are very uncommon. A few cases of spinal hematoma directly related to anticoagulant treatment have been reported, most of them presenting as epidural hematomas [11]. Only three cases of spontaneous spinal subdural hematoma associated with the use of rivaroxaban have been reported in the literature [2][3][4], but to our knowledge, no case due to apixaban has been described. Salient features of these cases including the present case are summarized in Table 1.
Acute and sudden onset of symptoms, with dorsolumbar pain followed by symptoms of spinal cord compression, is characteristic clinical manifestations [11], although multiple neurological de cits have been reported, such as sensory and motor de cits, sphincter dysfunction, and Brown-Séquard syndrome. Our patient did not have dorsal pain, and her clinical picture was mostly a mixed motor-sensory de cit.
MRI is the technique of choice for diagnosis and preferable to computed tomography (CT) [12]. MRI has a high sensitivity for de ning the type of bleeding and assessing the craniocaudal extension of the hematoma. Changes in the intensity of MRI signal after development of spinal hematoma are similar to those of cerebral hematomas, although it seems that metabolism of hemoglobin in the spinal canal is faster than that in the intracranial space. us, T1-isointense foci and T2-hyperintense lesions due to the presence of oxyhemoglobin can be seen during the rst 24 hours, followed by hypointense T1 and T2 signals on the second and third days and subsequent greater signal intensity in both sequences due to the appearance of methemoglobin. Cervicodorsal and dorsolumbar levels are the most frequently a ected regions, with predominantly cervical and dorsal involvement in young patients, whereas dorsolumbar involvement is more common in adults. Our patient showed extensive involvement especially in the D1 to D7 region. e management of nontraumatic spontaneous spinal subdural hematoma has not been de nitively established, probably because of the rarity of this condition. e management usually includes emergent surgical evacuation and discontinuation of NOACs. Conservative management may be considered in cases with mild neurological decits, early spontaneous recovery, or high surgical risk [4]. Prognosis is poor in the presence of associated subarachnoid hemorrhage with spinal arachnoiditis as a potential complication [13].

Conclusion
Nontraumatic spontaneous spinal subdural hematoma is a potential complication of the use of apixaban, and this appears to be the rst case reported in the literature. is entity should be distinguished from spinal subarachnoid hemorrhage, spinal epidural hemorrhage, or intraspinal hemorrhage [14]. Physicians should be aware of the increasing incidence of severe bleeding complications associated with the increasing use of NOACs. Serious spinal hematomas secondary to rivaroxaban and apixaban therapy are a devastating complication for the patient and the family. Recovery of the neurological de cit depends on prompt diagnosis with MRI scans and early surgical evacuation of the hematoma.

Conflicts of Interest
e authors declare that they have no con icts of interest regarding the publication of this paper.

Authors' Contributions
Alba Colell participated in background research, literature review, care of the patient, and writing of this manuscript. Adrià Arboix assisted in literature review, diagnosis of the patient, and writing of the manuscript. Francesco Caiazzo and Elisenda Grivé participated in the care of the patient and review of the manuscript for intellectual content.