Use of Beta-2-Transferrin to Diagnose CSF Leakage Following Spinal Surgery

INTRODUCTION

The spine surgeon is often confronted with a draining operative wound and/or deep-lying fluid collection. The most frequently ascribed diagnoses are seroma, infection and cerebrospinal fluid (CSF) leakage. The diagnosis is determined by review of clinical data, the appearance of the wound, and laboratory tests such as white blood cell count, erythrocyte sedimentation rate, C-reactive protein and microbiology specimens. MRI can show fluid collection but does not necessarily differentiate between diagnoses. Often, a percutaneous needle aspiration is necessary for fluid analysis.

Beta-2-transferrin is a protein found only in CSF and perilymph. Meurman¹ first described its use in the detection of CSF leakage in 1979. Since that time, beta-2-transferrin has been used extensively by otolaryngologists in the diagnosis of CSF rhinorrhoea and skull-base cerebrospinal fluid fistulas. With sensitivity of 94% - 100%, and specificity of 98% - 100%⁵, this assay has become the gold standard in detection of CSF leakage. While the beta-2-tranferrin assay has been briefly mentioned in an orthopaedic review article¹⁶ as a useful test in the postoperative diagnosis of CSF leakage, its broad potential in helping spine surgeons deal with postoperative complications has not been adequately emphasized. The following case demonstrates the utility of the beta-2-transferrin assay in identifying CSF within a postoperative fluid collection.

CASE REPORT

A forty-seven-year-old female schoolteacher presented with a two-week history of diffuse bilateral leg numbness, greatest on the anterior aspect of both thighs. She complained of progressive bilateral leg weakness and loss of balance, which limited her gait capacity to distances less than one hundred meters. These symptoms were significantly relieved in rest. She had no bowel or bladder incontinence, and no constitutional symptoms. There was no history of previous trauma or back pain. The patient was otherwise feeling well and had no other known medical problems.

On physical exam, the patient's spine was well aligned both on the coronal and sagittal planes. The range of motion of her back was within normal limits, and her gait was also normal. On neurological examination, she had normal motor strength in all four extremities. She had diffuse hypoesthesias in an L2 to S1 distribution bilaterally, most profound in the L2-3 distribution. Hyperactive deep tendon reflexes were noted in both lower extremities. She had two to three twitches of clonus, and absent Babinski reflex. Rectal tone examination was normal.

Plain radiographs of the thoracic spine demonstrated a bony spur projecting posteriorly into the spinal canal, at the T10-T11 interspace. A CT scan revealed a calcified T10-T11 disk space and a calcified mass, which occupied at least 50% of the AP diameter of the spinal canal. The posterior longitudinal ligament (PLL) was ossified at the level of the T10-T11 disk space, and both proximally and distally, extending posterior to most of the T11 vertebral body (Figure 1). The MRI revealed marked posterior displacement of the cord at this level with signal changes on T2-weighed images within the cord (Figure 2). The differential diagnoses were calcified thoracic disk herniation versus ossification of the posterior longitudinal ligament (OPLL). The patient's symptoms were compatible with spinal intermittent claudication, as described by Jellinger.^9,10 This represents a vascular compromise of the spinal cord¹⁰, and as described in the OPLL literature, once clinical manifestations begin, the course may be rapidly progressive.¹¹ For this reason, anterior canal decompression and spinal fusion were indicated.

Figure 1.

A: CT-scan axial cut and B: sagittal reconstruction at T10-T11. Bony spur and ossification of the PLL occupying a significant amount of spinal canal diameter.

Figure 1A.

Figure 1B.

Figure 2.

T2-weighed MRI sagittal image reveals severe spinal cord compression and posterior displacement of the cord.

The patient underwent a left tenth rib thoracotomy, T10-11 corpectomy, and spinal cord decompression. Decompression was very challenging, due to significant adhesion of the dural sac to the bony prominence. Once adequate decompression had been obtained, the spine was reconstructed with allograft fibula, autograft rib, and a low profile titanium plate (Figure 3).

Figure 3.

Post-operative plain films show T9 - T12 anterior spinal fusion with structural graft and low-profile titanium implant.

Intraoperatively, the offending bony prominence was intimately pushed against the dura and, in some areas, adhered. Removal of the mass provoked some CSF leak, although no discrete tear in the dura was identified. Once the cord had fallen anterior, to its normal position, the leak appeared to resolve spontaneously. Upon closure, the parietal pleura could not be completely closed due to the prominent hardware. The wound was closed without difficulty around two chest tubes.

On postoperative day four, the chest tubes were pulled. On postoperative day five, the patient developed orthostatically induced vertigo, headache, and nystagmus. The chest x-ray revealed only a small residual left-sided pleural effusion. With the clinical suspicion of symptomatic CSF leakage, the patient was placed on seven days of flat bed rest, and subsequently, slowly mobilized without any further orthostatic problems. On postoperative day fourteen, she was discharged to home without any dizziness, headaches, or respiratory difficulty. Her preoperative neurological symptoms had dramatically improved. A follow-up CT-scan revealed significant decompression of the spinal canal (Figure 4).

Figure 4.

Post-operative CT-scan shows significant decompression of the spinal canal.

One week after discharge, the patient presented to clinic with progressive dyspnea and orthopnea. She had a mild, non-productive cough, but she had no fevers, chills, or night sweats. On exam, she was afebrile with normal vital signs. Her respiratory rate was sixteen, but she was visibly uncomfortable at rest. With mild exertion, she became tachypneac and clearly dyspneic. Pulmonary exam revealed significantly diminished breath sounds in the entire left lung field and dullness to percussion. The surgical scar was healing well without signs of infection. Her neurological exam was normal, except for two beats of clonus, predominantly on the right. The remainder of physical exam was normal. Complete blood count and blood chemistries were within normal limits. Chest X-ray demonstrated a large left pleural effusion with mediastinal shift to the right (Figure 5).

Figure 5.

Post-operative x-ray on post-operative day #20, reveals massive fluid collection in the left chest field with mediastinal shift to the right.

The differential diagnosis of this fluid collection included subarachnoid pleural fistula, pleural fluid accumulation, hemothorax, chylothorax, and empyema. A thoracentesis was performed removing one liter of straw colored fluid. The patient's respiratory symptoms improved immediately. Analysis of the fluid revealed a pleural to serum creatinine ratio less than one, suggesting a transudate. The patient's postoperative symptomatic CSF leak strongly suggested that the transudate was a result of a subarachnoid pleural fistula. A beta-2-transferrin test of the fluid was performed and proved comfirmatory.

The patient remained asymptomatic, but her chest x-ray still demonstrated a massive effusion one week later. An additional 1.6 liters of fluid were removed. The patient was followed with regular chest x-rays over ensuing weeks, and the effusion slowly resolved without further intervention. She returned to work two months after surgery. Six months after surgery, she remained symptom free and returned to all regular activities. The neurological evaluation was completely normal, and her chest x-ray revealed completely clear chest fields.

DISCUSSION

The presence of a subarachnoid pleura fistula (SPF) upon the patient's presentation to clinic on her first post-operative visit is not surprising given that she had a known intraoperative CSF leak as well as concordant symptoms in the immediate post-operative course. When confronting a massive hydrothorax, the knowledge gained with the beta-2-transferrin assay was very useful. The ability to rule out empyema, chylothorax, and pleural fluid accumulation allowed us to focus our potential treatment modalities upon the CSF leakage. Heller¹² reported on two cases, which required trans-diaphragmatic pedicled greater omental flaps. Ido¹³ reported on three cases, which were repaired by either surgical closure with substitute dura mater and fibrin adhesive sealant or cyanoacrylate adhesive, or percutaneous intrapleural administration of OK-432, which has been shown to decrease pleural effusions in carcinoma patients through a local inflammatory response. The difficulty posed by our patient was the pressure gradient between the intrathecal and intrapleural space, which facilitates CSF leakage.¹² We were preparing to attempt a direct repair when spontaneous resolution occurred. A possible explanation for the spontaneous resolution would be migration of the spinal cord towards the dural defect, which tamponaded the flow of CSF. This may have allowed the defect to heal. A follow-up MRI taken at six months post-op showed the spinal cord to be completely centered in the spinal canal and surrounded by CSF, as well as absence of pseudomeningocele.

While this case illustrates a dramatic CSF leak into a large cavity, the pleural space, the beta-2-transferrin assay will likely be of more common use in posterior spinal surgery complications, such as deep lying fluid collections or wound drainage. In lumbar decompression surgeries, Wang¹⁴ reported a dural tear incidence of 14%, and in a review of 412 primary open discectomies, Stolke¹⁵ reported a 5.3% incidence of tears. This rate was over three times greater for revision procedures. In posterior spinal fusions with pedicle screw placement, West¹⁶ has described 5.6% incidence of dural tears.

The vast majority of these tears are identified intraoperatively and repaired. However, an unknown number of tears will persist, creating fluid collections, wound healing problems, and sometimes may even form subarachnoid-cutaneous fistulas. Persistent CSF leakage can lead to meningitis, epidural abscess, and pseudomeningocele.

When faced with treating a postoperative wound complication that involves persistent fluid drainage or fluid accumulation in the subcutaneous or submuscular planes, the beta-2-tranferrin assay and its ability to confirm that the fluid is CSF helps the surgeon plan the appropriate intervention. If the assay is negative, the surgeon can treat the wound with appropriate incision, debridement, and closure. However, if the assay is positive, the site of leakage should be sought out and repaired in an appropriate manner.

In summary, the beta-2-transferrin assay is a highly sensitive and specific test for the presence of CSF in body fluids. This test has been employed successfully by otolaryngologists in the diagnosis of skull-base CSF leaks. To our knowledge, the usefulness of the beta-2-transferrin assay has only been mentioned once in the spine surgery literature.¹⁶ Orthopedic spine surgeons are encouraged to take greater advantage of this method in appropriate clinical scenarios.