Introduction

Vertebral compression fractures (VCF) can result from minimal to no trauma in osteoporotic bone or oncologic lesions. Though cement augmentation procedures for VCF, vertebroplasty and kyphoplasty, are thought to be safe [1,2,3,4] and to improve outcome [5], complications have been described including local extravasation into paraspinal soft tissues, perivertebral veins, disc space, and spinal canal [6]. A review of neurological deficits due to cement extravasation showed that though cement extravasation into the spinal canal is common, clinical sequelae are rare, identifying 21 cases of neurologic compromise [7] because of direct compression, thermal injury, or nerve reaction to cement especially when the cement is intradural [8, 9]. Other reported causes of paraparesis following cement augmentation procedures include arterial cement embolization causing spinal artery syndrome [10, 11], epidural hematoma [12], and spinal artery syndrome [13], latter of which described a transient paraplegia following kyphoplasty without cement extravasation and without report of MR confirmation of spinal artery syndrome. The purpose of this report is to present a case of progressive paraparesis following kyphoplasty without arterial cement embolization or epidural hematoma but with non-compressive extravasated cement that improved after surgical treatment and with post op MRI confirmation of spinal artery syndrome.

Case presentation

A 77F with multiple myeloma and remote history of deep venous thrombosis on apixaban presented with T12 compression fracture and persistent pain despite conservative treatment. She underwent T12 kyphoplasty with two SpineJack implants (Stryker, Kalamazoo, MI, USA) [14, 15] inserted into the T12 vertebral body under fluoroscopy using transpedicular approaches. During polymethylmethacrylate (PMMA) injection, posterior cement extravasation was noted, and the procedure was stopped (Fig. 1a, b). There were no pulmonary or hypotensive episodes during the procedure. In the recovery room, the patient was found to have left worse than right lower extremity weakness and altered sensation in bilateral lower extremities. She was started on steroids and then underwent urgent computed tomography (CT) of the thoracic spine that showed that on the right side, the approach was likely medial to the right pedicle (Fig. 1d) with cement leakage into the canal and without evidence of arterial embolization (Fig. 1e). The SpineJack implants were positioned well within the vertebral body, surrounded by cement. The patient was emergently transferred to a tertiary care center where her initial physical exam was notable for 0–1/5 strength in all left lower extremity muscle groups and 4/5 strength on the right side, ASIA D according to ASIA/ISCoS International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI). Patient reported that the sensation in the left leg was 10% and the right leg was 50% normal. She had baseline lower extremity neuropathy which made sensory exams including pain, temperature, and vibratory sense difficult. Left lower extremity (LLE) also showed long tract signs of hyperreflexia and sustained clonus, but not in the right LE. The rectal tone was diminished, but she had not had episodes of urinary or fecal incontinence. Urgent MRI of the thoracic spine showed that the extravasated cement was extradural and just adjacent to the spinal cord and no obvious signs of epidural hematoma [12] (Fig. 2a, b). After multidisciplinary discussions among intensivists, neurologists, spine surgeons, and radiologists, we considered iatrogenic injury during kyphoplasty, intradural cement, spinal cord infarct, and spinal artery syndrome, though diffusion-weighted imaging (DWI) that could have identified the syndrome was not obtained. Medical management included steroids and elevated mean arterial pressure (MAP) goals ≥85. We had decided against anticoagulation initially because epidural hematoma was a possibility and she had an increased risk for bleeding because of medial positioning of the cannula during the kyphoplasty and because she was on anticoagulation pre-op though she did report she did not take the medication for 3 days. We also discussed additional imaging such as MRI with DWI or angiography. Serial examinations showed progression of her deficit to 0–1/5 strength in the right tibialis anterior and extensor hallucis longus (ASIA C). Given the progression of her symptoms, we recommended and went ahead with surgical treatment 9 h after the completion of kyphoplasty.

Fig. 1: Cement leakage in the spinal canal.
figure 1

Kyphoplasty (a) was done with SpineJacks (Stryker, Kalamazoo, MI, USA) using the transpedicular approach and cement extravasation into the canal was noted (b) during cement injection and the injection was stopped. CT scan showed right-sided intracanal cement extravasation (c, d, e) and on the right side, the approach appears to be medial to the pedicle (d).

Full size image
Fig. 2: Pre- and post-op imaging.
figure 2

MRI of T spine was done that confirmed that the extravasated cement was just next to the spinal cord (a, b). Because the patient had progression of right leg weakness, she underwent decompressive surgery and intracanal cement was excised (c) that corresponded to the intracanal cement seen in the axial images (b). Post op MRI showed excision of the intracanal cement (d) and a close-up of the same image (e) showed T2 hyperintensities consistent with spinal artery syndrome (white arrows).

Full size image

Operation

After T11 and T12 laminectomies, the dorsal surface of the dura was carefully inspected. We found no evidence of clots that would be consistent with epidural hematoma and no evidence of dural injury on the dorsal surface that would be seen if the dura were violated during kyphoplasty. Using a right T12 transpedicular approach, we confirmed that the dura did not appear to be violated and visualized the extravasated cement with gentle retraction of the dura. The dura easily fell away from the cement. The stalk of the cement was cut from the vertebral body using a curved osteotome and the cement in the canal was removed (Fig. 2c). The size of removed cement corresponded to that seen in the CT/MRI and we concluded that the cement was extradural and decided not to proceed with a durotomy. The remainder of the procedure, T11 to L2 instrumented fusion, was uncomplicated.

Post-operative course and recovery

Post-operatively, she was admitted to the intensive care unit to maintain MAP over 85 mm Hg for spinal cord perfusion. We obtained an MRI of the thoracolumbar spine on post op day 1 that showed no evidence of residual protruding or intradural cement (Fig. 2d) but did show T2 hyperintensities of the spinal cord at the level of cement extravasation consistent with sequelae of spinal artery syndrome (Fig. 2d, e) [16]. Immediately post op, in the recovery room, she started to have improvement of R LE motor function and by the next morning, R LE motor function improved to 4/5 as well as improved R LE sensation to her baseline. Her LLE strength remained 1/5 with diminished sensation. On post-operative day 30, she had regained global 4/5 LLE and 5/5 RLE strength with improved bilateral LE sensation and had returned to ambulating. The final follow-up was at 6 months by telephone at which time she stated that her back pain was better and that her walking was back to baseline. Her health failed from her multiple myeloma, and she passed 8 months after the surgery.

Discussion

We presented a case of progressive paraparesis due to spinal artery syndrome following kyphoplasty in the setting of a non-compressive extradural cement extravasation with near-complete neurological recovery following surgical decompression. We searched the Pubmed and Pubmed Central for {vertebroplasty or kyphoplasty} and {spinal artery syndrome}and found three cases, two of which were due to cement embolization during vertebroplasty [10, 11] and one with transient paraplegia after kyphoplasty without cement extravasation with clinical diagnosis of spinal artery syndrome without imaging confirmation who was given thrombolytic treatment with resolution of symptoms within 6 h [13].

In our present case, the multidisciplinary team initially treated the patient with steroids and elevated MAP goals and decided against anticoagulation given the concern for epidural hematoma and our patient’s increased risk for bleeding. There is general agreement that maintaining MAP greater than 85 mmHg for 7 days has been associated with improved outcomes in spinal cord injuries [17]. Though IV heparin would be indicated in cases with thromboembolism, other treatments such as steroids remain controversial [18].

As her right leg weakness continued to worsen despite medical management, we felt justified in our decision to proceed with surgery and her almost immediate improvement of symptoms post op suggested that thrombosis was unlikely to be the cause of her symptoms and supported our approach. Post op MRI showed hyperintensities in the spinal cord at the level of the extravasated cement consistent with spinal artery syndrome. To the authors’ best knowledge, this is the first case of MRI-confirmed spinal artery syndrome following cement augmentation procedure without cement embolization.

What would be the mechanism of spinal artery syndrome in our present case? The acute motor deficit and decreased lower extremity sensation would be consistent with anterior spinal artery syndrome. The preop MRI showed that the cement was just adjacent to the spinal cord on the right side, where the hyperintensities were seen in the post op MRI. Though there has not been a reported case of spinal artery syndrome in patients with cement extravasation, in three cases with spinal artery syndrome where the thoracic disc herniation just touched the spinal cord, two had anterior [19, 20] and one had posterior [21] spinal artery syndrome and the authors’ proposed mechanism was compression of the radicular arteries that arise from the segmental arteries. We posit a similar mechanism in our case.

Limitations of this report include the lack of angiographic evidence of the disruption of the blood supply associated with the extravasated cement. Upon presentation to the tertiary care center, the multidisciplinary team discussed repeat MRI but the progression of her symptoms resulted in surgical treatment before MRI could be obtained. Even if pre-op MRI with DWI was obtained that showed spinal artery syndrome [16], because the localization of the blood supply disruption would be limited with DWI, one could not completely rule out the possibility that the spinal cord injury occurred at the time of the kyphoplasty and that cement extravasation and MRI findings were incidental findings or that the hyperintensities seen were due to the surgical procedure. However, if that were the case, one would not expect to see a quick recovery of the neurological deficits after the surgery that was observed. Authors believe that the simplest explanation is that the cement extravasation caused spinal artery syndrome and that its removal resulted in prompt improvement of the neurological deficits.

We also do not know for certain whether the patient’s neurological deficit would not have improved given more time. For example, in the three cases spinal artery syndrome with thoracic disc herniations that just touched the spinal cord, patients were successfully treated non-operatively [19,20,21]. If our patient’s symptoms were stable, we might have continued to observe the patient initially. However, in the case of spinal artery syndrome following kyphoplasty without cement extravasation, the patient recovered after 6 h [13]. In the review of neurological deficits after cement extravasation [7], two patients with motor deficits who did not get surgery had no improvement of their deficits and a thoracic disc herniation causing spinal artery syndrome was treated successfully with surgery [22]. Therefore, given the progression of her deficits, we felt justified proceeding with the surgery 9 hours after kyphoplasty.

Another limitation is that we did not provide an explanation for why the patient’s symptoms were worse on the lower extremity contralateral to the extravasated cement. Though this had not been previously reported with cement extravasation, there is a body of literature on patients with disc herniations who present with contralateral pain and symptoms. Recent reviews noted that lumbar disc herniation with contralateral symptoms was rare and suggested that the pathophysiology of contralateral symptoms was multifactorial [23, 24]. In cervical disc herniation, the role of lateral spinothalamic tract was discussed as a possible source of pain [25] and the tract could have played a role in this patient’s symptoms. In the present case, left-sided symptoms improved after the removal of the cement on the right side.

Discussions on how to avoid complications during cement augmentation procedures are beyond the scope of this report but are found in the literature including indications and contraindications [26] and techniques [27].

Neurological deficits following vertebral cement augmentation procedures are rare but when they occur especially without severe cord compression, spinal artery syndrome should be considered, and the patient should be treated with elevated MAP goals and IV heparin if thromboembolism is likely. If the symptoms do not improve, then surgical management should be considered even if the extravasated cement is non-compressive and is contralateral to the more symptomatic extremity.