Discitis and Osteomyelitis in a 63-Year-old Male
A 63-year-old man initially presented to an outside hospital with cholecystitis and thereafter developed a small pulmonary embolism. He was treated with an intravenous cholangiogram (IVC) filter which was then complicated by a wound infection at the femoral vein insertion site. The right groin area underwent surgical debridement and dressing changes begun. During the course of this treatment, he developed severe low back pain, recurrent fevers, and intermittent bacteremia.
Imaging studies revealed osteomyelitis of L3 and L4. Biopsy showed methicillin-resistant staph areus (MRSA). Despite intravenous antibiotics (Vancomycin), his back pain worsened and he continued to have intermittent fevers. He was transferred to our facility for further treatment.
On physical examination, he is awake and interactive. Temperature is 100.4 degrees and his heart rate is 95 with stable blood pressures.
He has weakness of the bilateral hip flexors at 4/5 bilaterally. His rectal tone is intact with no bowel or bladder incontinence. Gait is limited by back pain. His groin site is open with clean granulation tissue.
White blood count is 8.9, hematocrit 29.5, erythrocyte sedimentation rate 89, and C-reactive protein 18.1. Urinalysis is negative.
Chest radiographs are clear with no evidence of pneumonia.
Figure 1A. Radiograph of the lumbar spine shows disc space narrowing with endplate irregularity and sclerosis.
Figure 1B. Contrast-enhanced MR imaging shows involvement of the L2-L3 disc space along with increased signal of the L3 and L4 vertebral bodies.
Figure 1C. Axial image shows enhancement of the anterior soft tissues.
L3-L4 discitis with L3 and L4 osteomyelitis.
Suggest TreatmentIndicate how you would treat this patient by completing the following brief survey. Your response will be added to our survey results below.
Surgical treatment was pursued for the following indications: 1) continued infection despite IV antibiotics (fevers, elevated sedimentation rate and C-reactive protein); 2) debilitating back pain; and, 3) development of hip flexor weakness.
A one-stage, minimally invasive direct lateral transpsoas approach was utilized to perform corpectomies of L3 and L4 with insertion of an expandable cage. With this technique, the scarred abdominal vessels are avoided and any psoas abscess is debrided.
Posterior L3-L4 laminotomies were used to debride presumed epidural abscesses behind the L3 and L4 vertebral bodies (Figures 2A-C). Figures 2 below, show intraoperative fluoroscopic images with navigation of insertion of the initial dilator. The initial dilators are passed through the psoas muscle (Figure 2A) and docked at the center of the L3-L4 disc space using AP and lateral (Figure 2B) fluoroscopic images. Expandable tubular retractors (Quadrant DLIF System, Medtronic, Memphis, TN) are deployed through a 3.5cm outlined by the dotted oval (Figure 2C). As the skin opening is smaller than the corpectomy defect, the cage must be inserted in the collapsed position and expanded in situ.
Figure 2A. Initial dilators are passed through the psoas muscle.
Figure 2B. Initial dilators are docked at the center of the L3-L4 disc space.
Figure 2C. Expandable tubular retractors (Quadrant DLIF System, Medtronic, Memphis, TN) are deployed through a 3.5cm outlined by the dotted oval.
The patient tolerated the procedure well without complication. Blood loss was 400cc for the corpectomies and strut fusion. Minimal blood loss was encountered for the posterior procedure (less than 50cc).
Procedure: The minimally invasive direct lateral approach utilized a retroperitoneal, transpsoas approach to the lateral lumbar spine. The patient was positioned so that the bend in the table laid between the iliac crest and the greater trochanter. Once positioned, the patient was secured to the operating table at the level of the greater trochanter and the shoulders. The operative table was flexed to separate the iliac crest and the rib cage and to clear the iliac wing away from the L4-L5 level.
A sterile C-arm was then brought under the table, where it was locked at 180-degrees. The surgical table was rotated to obtain a perfect AP image of the spine, thereby placing the patient in a true lateral position relative to the floor. By maintaining the trajectory of the instrumentation in a vertically straight position, the need for repetitive lateral C-arm imaging can be avoided.
The entry point to the center of L3-L4 disc space was identified with fluoroscopy and a 3.5cm skin incision was made. From the lateral incision, blunt dissection was performed through the soft tissue until the retroperitoneal space was reached. Care was taken not to perforate the peritoneum. Finger dissection was then used to bluntly sweep the peritoneum anteriorly. In some cases, a small posterior incision may be used to clear the retroperitoneal space in an effort to ensure the safe placement of the dilators on the psoas muscle. As the blunt finger dissection deepens, the surgeon will appreciate the psoas muscle, which is characteristically spongy and smooth to the touch. The transverse processes should also be palpated to confirm the exposure is in the anterior compartment.
The first dilator was placed through the lateral incision, guided toward the psoas muscle with the index finger. The blunt dilator initially rests anteriorly on one-third of the spongy psoas muscle. Lateral fluoroscopy aids the surgeon to position the dilator directly over the disc space. A gentle back and forth "twirling" motion was used to spread the fibers of the psoas muscle as the first dilator was advanced to the disc space. With the initial dilator positioned directly over the mid-point of the disc, a guide wire was inserted into the disc.
The C-arm was then brought under the table for an AP image and the subsequent dilators were advanced with a similar back and forth twirling motion with great care not to injure the lumbar plexus that lies in the posterior third of the psoas muscle.
As the dilators were placed, an evoked electromyographic monitoring system allowed detection of nerve roots within the surgical path. The nerve monitoring system emitted an electrical signal that searches for the stimulus threshold of the lumbar nerves. When the nerve is triggered, it was detected via EMG monitoring on the distal extremities. It is preferable to use this nerve detection system during insertion of the retractor, particularly at L4-L5.
Once the disc was reached, a tubular retractor was placed over the last dilator. The dilators were removed leaving the guide wire in place. The guide wire allowed for subtle adjustments of the retractor based on the fluoroscopic position. For example, if the dilators were slightly too posterior as seen on fluoroscopy, the surgeon can reference from the guide wire and move the retractor slightly more anterior. In a similar fashion, the vertical alignment of the retractor should be referenced from the guide wire. Once a thorough evaluation of the retractor's position was completed by visual inspection and fluoroscopy, the retractor was solidly fixed by the retractor arm and the guide wire was removed.
Visualization through the tubular retractor was enhanced by a fiber optic light source. Further, blunt dissection with bayoneted instrumentation was important to delineate the vertebral anatomy and expose the disc.
Long curettes, paddle shavers, and dilators were utilized for the discectomy. A Cobb elevator can be used to release the contralateral annulus. The release of the contralateral annulus is necessary if disc space restoration is desired. Extreme care should be used to pass the elevator in a perfect lateral trajectory in order to avoid inadvertent injury to the great vessels anteriorly.
The retractor was then carefully expanded proximally and distally until the L2-L3 cephalad and L4-L5 caudally was reached. The retractor blades were fanned open to allow visualization of the L2-L3 and L4-L5 disc space. During retractor deployment, the segmental vessels were encountered over the bodies of L3 and L4. The neuro-sucker, along with endoscopic peanut sponges, was used to open the psoas muscle vertically along the course of its fibers. The trough at the midpoint of the vertebral body was identified radiographically and a slow, back-and-forth sweeping technique was used to coagulate the segmental vessels. In cases of infection, these vessels are often thrombosed. A subtotal discectomy was performed at L2-L3 and L4-L5 using long curettes and rongeurs. Osteotomies were used to break apart the vertebral bodies of L3 and L4. Long pituitary rongeurs and curettes were used to perform the corpectomies. The endplates L2 and L5 were prepared for fusion.
Strut fusion was accomplished using an expandable cage, as the incision was smaller than the corpectomy defect. The cage was inserted into the defect in the shortened position and expanded in situ (Figure 2C, above).
After wound closure, the patient was placed into a prone position. Posterior instrumentation was inserted percutaneously. The patient was extubed and transferred to the ward in stable condition.
Post-operatively, back pain improved, fevers abated, and the C-reactive trended down. The patient was discharged home on post-operative day 5. At 3 month follow-up, his C-reactive protein level was less than 0.8. He was ambulating independently and only taking over-the-counter analgesics for pain.
Figures 3 (below) are the patient's standing radiographs obtained at 3 months post-surgery. There is maintenance of coronal alignment on the AP radiograph (Figure 3A). The lateral radiograph (Figure 3B) shows mild settling of the cage into the superior endplate of L5, but this is stable compared to 6 weeks post-operative radiographs.
Figures 3A, 3B
James S. Harrop, MD
Jefferson Medical College, Department of Neurosurgery
The author presents a novel technique, "Minimally Invasive Direct Lateral Corpectomy for the Treatment of L3-L4 Osteomyelitis."
History: The initial treatment strategy consisted of identifying the organism causing the infection. In this case, it was achieved through a biopsy (MRSA), typically done with CT guidance. In some cases, the patient may present with positive blood cultures and this then can identify the offending agent, thus avoiding an initial biopsy. However, if the patient does not respond to treatment, then a biopsy should be considered.
It is important to also note the patient's initial laboratory work. The white blood cell count is typically not helpful, but an elevated erythrocyte sedimentation rate and more specifically, C-reactive protein are suggestive of infection. This infection was persistent despite appropriate antibiotics for presumably a 6-week course. If the course of antibiotic was less than 6-weeks, it is important to review pre-treatment C-reactive protein levels. In the setting of persistently declining laboratory values, one may consider continuing medical therapy.
Examination: The patient's vital signs are typically normal in that he does not have fevers, tachycardia, or hypotension. If these signs are present, one should be concerned for a septicemia, which is unusual to be caused by a spine infection. Therefore, an additional source of the bacteremia should be sought in that discitis is typically a secondary infection. In this case, the patient was noted to have 4/5 hip flexor weakness. This is most likely secondary to the inflammation of the psoas muscles rather than due to direct neural compression.
Imaging: Lateral plain radiographs show collapse of the L3-L4 disc space with sclerosis of the endplates. Although plain radiographs tend only to present findings late in the course of the disease, I find them helpful to assess the patient's overall alignment, any deformity, or instability.
The MRI is the radiographic modality of choice and illustrates a L3-L4 discitis with enhancement along the lateral borders and no psoas collection. In addition, there appears to be some hyperintense signal on the T2 sagittal MRI at L2-L3. It is important to recognize that the presence of a psoas collection is almost uniformly associated with a discitis. I also usually obtain a CT scan to assess the degree of bone destruction.
Interpretation of MRI after initiating treatment can be difficult in that the MRI typically lags behind the patient's clinical picture. Therefore, treatment decisions should be made based on the patient's clinical signs and symptoms, and serology (C-reactive protein trend) in addition to imaging studies such as MRI.
It should be specifically noted that in these elderly patients one must evaluate renal function prior to gadolinium administration due to the possibility of nephrogenic systemic fibrosis (NSF).
Discussion: This is an elderly patient with a complicated medical history and persistent infection despite intravenous medication presented for treatment. Due to the concurrent illnesses, cholecystitis and pulmonary embolism, the author sought to treat through less invasive strategies. The overall treatment goals of osteomyelitis remain unchanged and are addressed through this strategy. The goals consisting of: (a) Decompression of the neural elements; (b) debridement of the infection; (c) correction of any spinal deformity; and (d) stabilization/arthrodesis of the involved segment(s).
The author used a minimally invasive technique to achieve these goals. It should be noted that the approach can be difficult due to altered anatomy in the setting of infection. The spine typically forms large lateral osteophytes if the infection has been persistent for some period of time, which can alter the local anatomy and landmarks. Also, the endplates become sclerotic which can make decompression and debridement more difficult. Therefore, this case should only be approached after the surgeon has gained significant exposure and is comfortable with a minimally invasive technique.
I would have approached the decompression through an anterior exposure and 2-level discectomy and debridement. L2-L3 and L3-L4 tend to have the aorta anteriorly providing a lateral approach with concurrent debridement of psoas abscess if present. A pre-operative CT scan will outline the anatomy.
As pointed out above, the endplates tend to be sclerotic and can hold a graft without difficulty. In some cases, osteomyelitis destroys the endplates typically affecting the inferior body more than superior. This is important in that it can destroy the pedicle such that a posterior instrumented fusion may require an additional level inferiorly. Also, I would utilize a structural iliac crest autograft. This provides the greatest chance for arthrodesis and offers the least risk of rejection or recurrent infection.
In the setting of an active infection, I do not rely on anterior instrumentation but will place posterior instrumentation.(1) The author choose to proceed with the placement of an expandable cage, which is supported in the literature in the setting of an active infection.(2, 3) I have concern for potential recurrence in such elderly and immunocompromised hosts. Therefore, again I prefer anterior reconstruction with autograft with posterior supplemented instrumented fusion.
The author should be commended for achieving the surgical goals through this minimally invasive surgical strategy. The author concludes the case with radiographs and serology on the patient confirming resolution of the infection and maintaining the patient's alignment. It is important to continue and follow these patients and C-reactive protein is one of the most helpful tests, as the author noted.
1. Faraj AA, Webb JK: Spinal instrumentation for primary pyogenic infection report of 31 patients. Acta Orthop Belg 66:242-247, 2000.
2. Kuklo TR, Potter BK, Bell RS, Moquin RR, Rosner MK: Single-stage treatment of pyogenic spinal infection with titanium mesh cages. J Spinal Disord Tech 19:376-382, 2006.
3. Rauzzino MJ, Shaffrey CI, Nockels RP, Wiggins GC, Rock J, Wagner J: Anterior lumbar fusion with titanium threaded and mesh interbody cages. Neurosurg Focus 7:e7, 1999.
Harel Deutsch, MD
Co-director, Rush Spine and Back Center
Rush University Medical Center
Overall, I believe the direct, or extreme lateral approach is an excellent approach to access the lumbar disc spaces at L2-L3, L3-L4, and L4-L5. The approach could be applied to cases with discitis.
I have several reservations concerning the technique described. For a single disc level, a transpsoas approach is appropriate. In opening a retractor to expose three disc spaces, significant psoas damage will occur. The lumbar plexus travels through the psoas muscle. Nerve root monitoring can help avoid damaging the plexus, but with such a large psoas muscle opening, damaging the lumbar plexus is unavoidable.
A retroperitoneal approach requires a larger incision, but works well in the setting of the case described. I do not see a significant benefit to the patient by employing a transpsoas approach to perform a 2-level vertebrectomy. A transpsoas debridement of the discitis level, in conjunction with posterior instrumentation, would be an effective strategy.