PMMA Extravasation Following Kyphoplasty
The patient is a 72-year-old female who underwent kyphoplasty to treat a T12 burst fracture two weeks ago at a different hospital. After the kyphoplasty procedure, her back pain increased and a new symptom of lower extremity weakness, right greater than left developed. She also lost bowel and bladder function. While postoperative imaging was performed, she was ultimately transferred to a rehabilitation center. While in rehab, her pain persisted, and she did not seem to be making any functional gains. Her daughters, therefore, requested she be transferred to our referral hospital.
She presents at our hospital with intractable back pain and dense R>L lower extremity weakness. Her proximal limb girdle muscles, including her iliopsoas, quads and adductors are 1/5 on the right and 2/5 on the left. Her distal extremity strength, including her hip abductors as well has her gastrosoleus and tibialis anterior are 3/5 on right and 4/5 on the left. Her lower extremity sensation is preserved, including the perineum. She exhibits weak, voluntary anal sphincter contraction. Her vascular exam is intact.
The patient’s medical history includes osteoporosis, chronic obstructive pulmonary disease, coronary artery disease, and steroid-dependent rheumatoid arthritis.
A lateral thoracic CT scan with contrast shows a T12 kyphoplasty with PMMA extravasation into the spinal canal (Figure 1).
Figure 1 (Below)
Axial images demonstrate the extent of the extravasation of PMMA; Figures 2-4 below.
Figure 2 (Below)
Figure 3 (Below)
Figure 4 (Below)
T12 osteoporotic burst fracture, status post T12 kyphoplasty with PMMA extravasation into the spinal canal, incomplete spinal cord injury, spinal stenosis at T11-T12, T12-L1.
Suggest TreatmentIndicate how you would treat this patient by completing the following brief survey. Your response will be added to our survey results below.
- Open reduction and internal fixation of the T12 burst fracture
- Pediculo-facetectomy; T12, right
- Thoracolumbar laminectomies: T10-T11, T11-T12, T12-L1
- Posterior thoracolumbar fusion: T9-L2
- Posterior segmental thoracolumbar instrumentation: T9-L2
- Vertebroplasties at T9, T10, and L2
Given the cranio-caudal extent of the PMMA extravasation, laminectomies to normal dura at the T10-T11 and T12-L1 interspaces was planned. Given the patient’s osteoporosis and the plan for resection of the right T12 pedicle and T11-T12 facet, an instrumented fusion from T9 to L2 was planned. We augmented the pedicle screw tracts at T9, T10 and L2 with PMMA. This was done by preparing the pedicle for screw placement using a Lenke probe. A ball tipped feeler was used to ensure cortical integrity. We placed pedicle markers and assessed their positions fluoroscopically. Next, the markers were removed, and we tapped the screw tracts. Bone void fillers were placed into the vertebral bodies. PMMA was placed under live image intensification guidance. Immediately thereafter, the screws were placed.
Initially, placement of a short screw into the left T12 pedicle was considered but ultimately the cement mantle prevented a screw much longer than 15 mm, so this was abandoned.
Laminotomies were performed at T10-T11 and T12-L1. Then an en bloc laminectomy was performed over the intervening laminae by creating bilateral troughs with a high speed burr. The troughs were completed with a micro Kerrison.
The right T11-T12 facet was removed in its entirety by extending the laminectomy through the lateral pars above and below the facet. The remaining, lateral portion of the right T12 pedicle was skeletonized and burred flush with the vertebral body. At this point, the PMMA was essentially free of the posterior elements and attached only anteriorly, just medial to the stump of the T12 pedicle.
Figure 5 (Below)
With the pedicle removed laterally, it was possible to break the PMMA away from the dura in a lateral and right posterior direction. There was some adhesion of the PMMA to the dura laterally, but no frank durotomy was encountered. The dura was thin and was ultimately augmented with DuraSeal.
Figure 6 (Below)
Figure 7 pictures the pieces and fragments of PMMA removed.
Figure 7 (Below)
Figure 8 (Below): T9 is top left, L2 bottom left, and T12 on right top and bottom.
We had discussed bone grafting options with the patient preoperatively. Given her medical issues and poor bone quality, autogenous graft harvest from the iliac crest was not advised. We were concerned that she would be at heightened risk for screw loosening and implant failure. We therefore employed local bone from the decompression to which 1 gm of vancomycin powder had been added. This bone graft was augmented with rh-BMP2 and allograft (MagniFuse).
Figure 9 (Below): Intraoperative imaging
Figure 10 (Below): Postoperative lateral and posterolateral x-rays
Although the patient's lower extremity MEP responses were poor (as predicted by the patient's weakness), they (and the SSEPs) were stable during surgery.
The patient was in the operating room for 105 minutes. Blood loss was 50 cc. There were no complications.
Given her medical issues, the patient spent her first postoperative night in the ICU, but she recovered from anesthesia well. She spent an additional 4 nights in the hospital followed by transfer back to the rehab center. Interestingly, her pain was under better control compared with preop by the second postoperative day. Her left leg weakness also improved fairly quickly. We placed her in a limited contact brace, which we plan to continue for 8-12 weeks.
At the time of discharge, her proximal limb girdle muscles, including her iliopsoas, quads and adductors are 2/5 on the right and 4/5 on the left. Her distal extremity strength, including her hip abductors as well has her gastrosoleus and tibialis anterior are 3/5 on right and 5-/5 on the left.
She was seen in the office just over one month from the procedure. In the interval, she had continued to do well symptomatically with marked improvement in pain and function. Her bowel function normalized, but, as of this writing, she continues to experience daily (though not complete) bladder incontinence.
Her current proximal limb girdle muscles, including her iliopsoas, quads and adductors are 3-/5 on the right and 4+/5 on the left. Her distal extremity strength, including her hip abductors as well has her gastrosoleus and tibialis anterior are 4+/5 on right and 5-/5 on the left. Her recovery has been challenged by a UTI and bouts of confusion, one of which led to a 2 day hospital readmission, at which time a postoperative thoracic spine CT was obtained.
The case presented should be considered a "never event". This unfortunate individual with multiple medical comorbidities underwent a vertebral augmentation procedure with extensive cement extravasation into the spinal canal. The intraoperative extravasation of cement is directly related to the initial practitioner not recognizing the leak on fluoroscopic images while the cement was being deposited, or simply not using fluoroscopy during the placement of the cement.
In addition, one can surmise that the postoperative symptoms were either ignored or not recognized, which resulted in a substantial delay of treatment. Dr. Truumees immediately recognized and investigated the issues, and initiated the most appropriate and least invasive method of addressing all the surgical variables, which included an elderly individual with severe osteoporosis and neurological compromise. Even with a delayed decompression and reconstruction he reports that this patient is making a recovery.
The only additional question in this case would be a consideration of prophylactic cement augmentation of the levels above and below the instrumentation and fusion. I would maintain that this individual is at high risk for proximal and distal junctional kyphotic failure, and as such prophylactic adjacent level cement augmentation may provide added benefit.
The lessons from this case include; 1) the importance of intraoperative fluoroscopy during any vertebral augmentation procedure, 2) the importance of proper history, examination and investigation if the expected outcome is not be realized after an intervention, and 3) even with a delayed decompression, there is still an opportunity for neurological recovery.