Thoracolumbar Junction Osteomyelitis with Kyphotic Deformity
The patient is a 54-year-old woman with a prior history of IV drug abuse. She was hospitalized in 2007 with spinal osteomyelitis at T11-T12. This was treated with IV antibiotics. She was admitted multiple times in the last several months for chronic refractory osteomyelitis (MRSA on biopsy) at T11-L1 despite prolonged courses of antiobiotics. A kyphotic deformity was now obvious as well (Figure 1).
Globally weak and disoriented secondary to recent meningitis. No focal motor or sensory deficits. No evidence of myelopathy.
Several courses of prolonged IV antibiotics therapy for chronic MRSA osteomyelitis of the thoracolumbar spine.
Figure 1: Pre-op sagittal CT demonstrating kyphosis at the thoracolumbar junction (T12-L1 epicenter of infection and T11-T12 nearly fused). Note Cobb angle of 32.5°.
The patient was diagnosed with chronic refractory osteomyelitis of the thoracolumbar junction with kyphotic deformity.
Suggest TreatmentIndicate how you would treat this patient by completing the following brief survey. Your response will be added to our survey results below.
The patient underwent a multilevel mini-open lateral corpectomy followed by image-guided posterior fixation.
After fluoroscopic-guided localization in the lateral decubitus position, a 7 cm oblique incision was made along the left mid-axillary line. The overlying rib was resected and the dissection proceeded retropleuraly.
As the lateral vertebral column was encountered, adhesions and inflamed tissue precluded a total retropleural dissection.
Segments of the lateral attachments of the diaphragm to the spine (medial and lateral arcuate ligaments) were detached at the T12 and L1 levels and a retractor system was inserted (Figure 2). This was secured to the table top mounted arm. Vertical smooth blade retractors were used rostrally and caudally and an "egg-beater" retractor was used against the visceral pleura.
The corpectomies (at T11, T12, and L1) proceeded in the standard fashion (ie, ligation of segmentals followed by discectomies and vertebral body bone removal). The canal was decompressed along the entire segment.
After the corpectomies were performed, the kyphosis did not reduce. At this point, we opted to open up the wound in the midline posteriorly and using subperiosteal dissection, dorsal elements from T10 to L1 were exposed. Facetectomies were performed at T10-11, T11-T12, and T12-L1.
MEP's and SSEP's were closely monitored and suggested no deviations from baseline. With this performed, the kyphosis was reducible.
An expandable cage was fitted in a reversed lordotic fashion. Morsalized rib cage was used for grafting. A transverse plate was utilized and secured across the segment with 4 vertebral body screws. The chest cavity was irrigated and closed in standard fashion. Valsalva maneuvers during closing obviated the need for a chest tube.
The patient was taken back to the OR the following day for image-guided T9-L3 pedicle screw fixation to augment the long anterior construct (Figure 4). Post-operative CT demonstrated improved correction of kyphosis and proper hardware placement. Final uprights were also adequate (Figures 5A and 5B).
Figure 2: Intra-operative fluoroscopic image demonstrating placement of the retractor system and initial deployment of cage.
Figure 3A: Final AP intra-operative fluroscopoic image demonstrating cage and plate placement.
Figure 3B: Final lateral intra-operative fluoroscopic image demonstrating cage and plate placement.
Figure 4: Post-operative CT demonstrating kyphosis reduction (32.5° to 18.3°) and cage placement
Figure 5A: Immediate post-operative upright lateral x-ray
Figure 5B: Immediate post-operative upright AP x-ray
Figure 6: Photograph demonstrating the lateral and posterior skin incisions from Stage 1.
Post-operatively, the patient had no new neurologic deficits but continued to experience back pain in the immediate post-operative period.
At 3 months follow-up, she was noted to have an asymptomatic, mild compression fracture at T7 that was treated conservatively.
At 6 months follow-up, she is ambulating with minimal back pain and with radiographic evidence of fusion on flexion-extension x-rays. She is off antibiotics.
First, I would like to commend the surgeons on tackling a complex reconstructive case and not resorting to half measures or underestimating the need for surgical eradication of her infection.
I would strongly agree that non-operative measures would be doomed to failure in this particular clinical scenario. Aggressive pre-operative nutritional support can be critical, particularly with her likely not eating secondary to her current meningitis and possible long-term poor access to health care of adequate protein.
I believe that any plan that involved only anterior column or only posterior column reconstruction would risk early failure.
Perhaps more important is the recognition of deformity correction as a significant goal of the operation, and when faced with suboptimal correction, it appears the team was admirably able to think on their feet and perform an extensive posterior release while in the lateral position. While not ideal to be faced with this situation, it is far better than simply accepting a corpectomy and instrumentation without increasing the anterior column height.
Often times reviews of this sort devolve into critiques based solely on surgeon preference. While I will try to avoid that, the following criticisms are significantly based on my preference treating a number of similar cases through an all posterior vertebral column resection approach.
This has a number of advantages, including: obviating the need for a difficult anterior approach which was unable to stay retropleural and necessitated division of the diaphragm, and in most cases requires a chest tube simultaneous control of the anterior and posterior columns, obviating the need for an unplanned posterior approach in the lateral position and providing total deformity control a single stage procedure on a medically and likely nutritionally compromised patient.
The advantage of (2), complete control of the entire spine through a single approach, is hard to stress enough. I have been able to routinely bring kyphotic deformity of 40° to 80° into neutral alignment with this approach.
Whatever the approach, the concepts remain the same: Eradication of infected bone and disc, lengthening of the anterior column, resection, shortening, or at least significant release of a partially fused posterior column, and excellent immediate stability in a sagittally balanced position. Of course, solid bony union is also a prerequisite for any long term success.
Judging the above case on this criteria, I think they have accomplished all but proving sagittal balance, as no long cassette films are provided.
With 18°of kyphosis over the TL junction (T10 to L2), which should be close to neutral alignment, and the near certainty of at least a small amount of settling over the first 3 to 6 months, this patient may be at risk of global imbalance at final follow-up.
Also of significant concern in these cases, regardless of approach, is achieving solid union across a very large defect. In my experience, the best union is not through the center of the expandable cage, but all around the cage and posteriorly, where there does not appear to be robust bone on the post op CT.
Again, long cassette films at minimum 1 year follow-up, and perhaps a CT, would provide that answer.
It is easy to critique from afar. There is, of course, no strong data on the best approach to these quite unique situations. On balance, I commend the team for treating a challenging but no longer rare clinical problem in a difficult patient population.