P22. Recurrent CSF leak following repair of incidental durotomy in lumbar spinal stenosis surgery

Abstract

BACKGROUND CONTEXT Incidental durotomies (ID) are a known occurrence in lumbar spine surgery. In most cases, an ID can be effectively managed with direct repair that is often supplemented by fibrin glue or a synthetic or biologic patch. Generally, there is minimal adverse effect to the patient other than a longer than expected postoperative hospitalization accounting for a period of recumbence following the dural repair. However, on occasion, a recurrent CSF leak occurs postoperatively which can cause troubling symptoms and may require additional surgery to repair. PURPOSE To identify the frequency of recurrent CSF leaks in patients who underwent surgery for lumbar spinal stenosis that was complicated by a dural tear and subsequently to both define risk factors for the occurrence of a recurrent leak and to assess the effect upon the ultimate clinical outcome. STUDY DESIGN/SETTING A retrospective analysis of a prospectively-treated cohort of patients. PATIENT SAMPLE A total of 1,480 consecutive patients who underwent elective, primary surgery for lumbar spinal stenosis over a 5-year period by one of four surgeons at a single academic, tertiary care medical center. Clinical follow-up was a minimum of 2 years (range: 2-8 years). OUTCOME MEASURES VAS Score (1-10) for back and leg pain, Oswestry Disability Index, Patient Satisfaction Assessment, and Retrospective Willingness to Repeat Surgery. METHODS A dural tear (DT) occurred in 103 of 1,481 patients (7%). Fourteen of 103 patients (14%) developed a symptomatic, postoperative CSF leak requiring additional surgery. Overall, a recurrent CSF leak occurred in 14 of the 1,481 patients (0.9%) who underwent a primary lumbar decompression for spinal stenosis who were evaluated in the study. All dural tears were identified at the time of the index surgery. Dural repairs were directly repaired with either a 4-0 silk or 6-0 polypropylene suture in a water tight fashion and assessed with a Valsalva maneuver prior to closure. Repairs were augmented with fibrin glue in 70% of cases, a collagen patch in 32%, and both in 20%. Patients were kept on bedrest following surgery for 1-3 days. A subfascial drain was employed in 82% of patients. Choice of suture, augmentation of repair, use of drain and duration of bedrest were based on surgeon discretion. Multi-variate regression analysis was employed to identify significant risk factors for a postoperative CSF leak. RESULTS Of the 14 patients with a recurrent CSF leak, symptoms arose during the postoperative hospitalization in 11 patients (79%) and after discharge in 3 (21%). Each patient had a palpable subcutaneous mass and 4 patients (29%) had a percutaneous CSF leak. Each patient was symptomatic with postural headache and 10(71%) experienced photophobia. Significant risk factors for development of a CSF leak included age > 70, body mass index > 30 (obese), history of diabetes and chronic use of corticosteroids. Notable factors not significant included duration of surgery, the additional diagnosis of degenerative spondylolisthesis or degenerative scoliosis, presence of a fusion, numbers of levels decompressed or fused, use of BMP or particular bone graft choice, nor preoperative use of pharmacologic anticoagulation. The likelihood of a recurrent leak was not significantly affected by a particular type of surgical dural repair, the suture material employed, nor the use of a synthetic patch or fibrin glue. As well, the use of a postoperative, subfascial drain nor the duration of recumbancy after the initial surgery did not significantly affect the development of a recurrent CSF leak. At 3-month follow-up, two patient reported outcomes (ODI, patient satisfaction) were slightly less in patients who had a recurrent CSF leak, compared with both durotomy patients with no recurrent leak and those with no durotomy, although this did not reach statistical significance. At 6 month, 12 month, and final follow-up, there were no statistical differences in patient-reported outcomes between the recurrent CSF leak, durotomy without recurrent leak, and no durotomy cohorts. Length of hospitalization was an average of 4.2 days longer in patients with a recurrent CSF leak compared with patients with no durotomy and 2.2 days longer than patients with a durotomy but no recurrent leak. Cost of hospitalization was $65,000 greater in patients with a recurrent leak compared to those with no durotomy and $35,000 greater than those with a durotomy but no recurrent leak. CONCLUSIONS Incidental durotomy is a relatively uncommon occurrence in primary surgery to treat lumbar spinal stenosis. Recurrent CSF leak following repair of a durotomy is also uncommon, with an incidence of 14% of patients with a durotomy in this large single center study. Prompt recognition of a recurrent CSF leak and subsequent surgical repair effectively treats this problem and adverse sequelae did not occur and ultimate clinical outcomes were not altered compared to those without a durotomy or a recurrent CSF leak. Hospital length of stay and cost were greater in those with this uncommon complication. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs.