Misdiagnosis in a case of non-compressive myelopathy due to a lumbar spinal intradural fistula supplied by the Artery of Adamkiewicz.

BACKGROUND: Spinal vascular malformations comprise a rare but significant group of spinal disorders where clinching the diagnosis early is absolutely necessary since the morbidity increases as the time goes by. These malformations present mainly with symptoms of myelopathy with a gradually worsening course and thus early diagnosis and intervention may revert the symptoms to some extent. Owing to ignorance, sometimes the diagnosis may be missed or delayed and this delay can make a significant difference in the final outcome. CASE DESCRIPTION: A 44-year-old male presented to us with an 8-month history of gradually worsening difficulty in walking and lower limb paraesthesias along with recent bladder complaints. Earlier, the imaging had revealed prolapsed lumbar disc and he had undergone L4-5 micro-discectomy few months back. As his symptoms worsened further, he developed paraparesis and then a more detailed analysis revealed a missed spinal arterio-venous fistula at L4-5 level causing congestive myelopathy. He was re-operated and the fistula was disconnected which led to an improvement months after surgery.
CONCLUSION: Thus, to differentiate between compressive and non-compressive myelopathy and detailed investigation of the latter to identify the actual cause remains imperative. Misdiagnosis leading to a wrong surgery caused further deterioration which could have been avoided by careful analysis of imaging. Open surgery remains the preferred treatment for the fistulas supplied by the artery of Adamkiewicz.

INTRODUCTION

Spinal vascular malformations are a rare group of disorders which require a very high index of suspicion for diagnosis especially in early stages where imaging is not too helpful usually. But, subtle findings are always present to clinch the diagnosis at the earliest. Awareness about these relatively rare vascular disorders is required for a good clinical outcome so that diagnostic angiography is ordered immediately in all suspicious cases and nothing remains undiagnosed. This becomes important as the degree of neurological recovery depends mainly on the pre-operative neurological status. We report a case of a 44-year-old male with a spinal intradural arterio-venous fistula (AVF) who underwent micro-discectomy due to an erroneous diagnosis and further deteriorated clinically till the actual cause was identified and treated. The history, examination, imaging and operative details are discussed along with relevant review of literature.

CASE DESCRIPTION

A 44-year-old male presented to us with an 8-month history of chronic backache and lower limb paraesthesias. Initially, only symptomatic treatment was given which did not help him and he developed bladder complaints in the form of urgency and poor flow. Magnetic resonance imaging (MRI) of spine was done and multiple disc bulges were identified in the lumbar region, the maximum being at L4-5 level. He was diagnosed to have a cauda equina syndrome and hence, L4-5 micro-discectomy was done. Clean intermittent self catheterization (CISC) was advised for bladder care and active physiotherapy was started. He continued to deteriorate and developed difficulty in walking with increasing stiffness and weakness in legs. He was admitted at our institute and a thorough examination was performed. He had an overall grade 3 spastic paraparesis with poor left foot dorsiflexion (grade 1). The reflexes were depressed and plantars were upgoing bilaterally. The anal tone and the peri-anal sensations were reduced. We reviewed the previous MRI and detected T2 hyperintense signal in the dorsolumbar cord which was missed at the first surgery [Figure 1]. The MRI was repeated and it showed progression of the cord signal to higher level and appearance of flow voids on T2-weighted images which enhanced post-gadolinium [Figure 2]. Suspecting a vascular pathology, a spinal angiogram was ordered which led to the diagnosis of spinal intradural AVF at L4-5 level which was supplied by the artery of Adamkiewicz (originating from right D9 segmental artery) [Figure 3]. The AVF was draining into an ′early filling′ peri-medullary vein leading to appearance of the dilated veins and flow voids on MRI. As the feeder was the artery of Adamkiewicz (D9 level) and the fistula was at was L4-5 level, open surgery was preferred over the endovascular therapy and the patient underwent re-exploration and L4-5 laminectomy. The tortuous and dilated veins were seen on the filum and the fistula was identified and disconnected [Figure 4]. Intraoperative angiography confirmed the obliteration of the AVF with non-visualization of the peri-medullary vein [Figure 5]. The post-operative course was uneventful and he started making gradual recovery. The CISC was continued and he was discharged with active physiotherapy. The patient continued to improve and started walking independently resuming his active lifestyle (Karnofsky performance scale >90). After 3 months, the post-operative MRI showed complete disappearance of the flow voids and cord signals [Figure 6].

Figure 1

Initial sagittal T2-weighted MRI showing hyperintense signal in dorsolumbar cord and lumbar disc bulges.

Figure 2

Repeat MRI showing ascended cord signals and appearance of flow voids which enhance after contrast.

Figure 3

Digital subtraction angiography showing spinal fistula at L4-5 level with cranially draining tortuous and dilated peri-medullary vein.

Figure 4

Intraoperative view of the dilated vein on the filum and disconnected fistula with clip in-situ.

Figure 5

Intraoperative angiography showing complete course of the artery of Adamkiewicz and non-visualization of the vein.

Figure 6

MRI after 3 months showing resolution of the cord signals and flow voids.

DISCUSSION

These peri-medullary intradural fistulae account for 20% of all spinal vascular lesions and are seen in younger patients. Spetzler has classified the spinal vascular pathologies in the most apt manner although various other classifications by Wyburn-Mason, Pia and Vogelsang, Rosenblum and Borden have been described in the past.[7] His classification includes neoplastic, aneurysmal and arterio-venous categories. The arterio-venous lesions have been further sub-divided according to extra and intradural sites. The ventral intradural AVF can be type A, B or C depending on the size. The dorsal ones are the most common dural arterio-venous fistulae (DAVF). Berenstein classified these malformations into genetic hereditary, non-genetic hereditary and single sporadic lesions.[2] Lasjaunias has separated spinal arterio-venous malformations (AVM) according to their anatomical location into DAVFs, extradural fistulas, paraspinal fistulas and spinal cord AVMs.[6] Earlier, Djindjian recognized that these spinal intradural fistulae are characterized by rapid arterio-venous transition without intervening capillary network or nidus like in an AVM.[3] Most commonly males are affected and they present with symptoms and signs of myelopathy with motor weakness, sensory complaints and bladder involvement depending on the site of fistula although sudden onset presentation with excruciating backache due to sub-arachnoid hemorrhage has also been described. According to Aminoff et al., the pathogenesis revolves around the venous hypertension transmitted directly from the artery involved causing cord engorgement and decreased arterio-venous pressure gradient with reduced tissue perfusion resulting in necrotizing myelopathy.[1] This venous stasis in the cord presents itself as T2 high-intensity signal and has a gradual ascending course. The veins become tortuous and dilated and appear as flow voids on T2-weighted MRI. Usually one should be very suspicious about these cord signals and must look out for flow-voids. The signals were missed and our patient underwent a surgery which was not required. The therapeutic aim is to disconnect the fistulous connection and thus reduce the cord ischemia due to venous hypertension. This can be either done by endovascular method or by surgical disconnection. Usually embolization is attempted first as the less-invasive alternative although high-recanalization rates have remained its chief drawback. In our case, an opinion from an experienced endovascular surgeon was sought but as the feeder was the artery of Adamkiewicz, also supplying the anterior spinal artery; it was difficult to access and had considerable risk of cord ischemia.[4] Surgery was planned and the fistula was disconnected. Only one such case has been reported in the past where the peri-medullary fistula at L2 level was supplied by artery of Adamkiewicz (right D9 level). Although, this fistula was successfully embolized, a new dural fistula was detected at L4 level draining in the same vein later.[5] Our case is the first case of surgical disconnection of lumbar AVF supplied by artery of Adamkiewicz after it was missed initially due to misdiagnosis. Surgery should be preferred over the endovascular treatment in all such cases as it provides direct access and results in definitive cure with minimal risk of cord ischemia. However, detailed study of selective angiography to confirm the level and identify the correct feeder is a must. This case report emphasizes the importance of careful pre-operative analysis of imaging and screening the whole spine in doubtful cases so that nothing is missed.

CONCLUSION

High index of suspicion is required to diagnose this potentially curable spinal vascular malformation as delay or wrong diagnosis can prove detrimental for the patient. These peri-medullary fistulas fed by the anterior spinal artery are best managed by surgical disconnection as this gives long-lasting cure with minimal risk. The intraoperative confirmation of the obliteration should always be done with angiography. A good outcome can be achieved if the surgery is performed at the right time by a vigilant neurosurgeon.