Management of Post-Craniotomy Persistent CSF Subgaleal Collection in Skull Base Procedures: Local Experience

Background: Post-craniotomy CSF collection is a problem that may cause severe complications as meningitis, wound disruption, prolonged hospitaliza-tion, and additional surgeries. Objective: To evaluate our cases with resistant post-operative subgaleal CSF collection, trying to identify causes and optimal management. Methods: Retrospective review of elective skull base cases during the period of January 2104 to January 2019 identification of cases with post-operative CSF subgaleal collection, which either managed conservatively or needed a second surgery. Results: 219 patients, 30 of them suffered subgaleal CSF collection, 22 patients improved with non-operative measures, eight patients needed second surgery with pericranial graft augmentation, and obliteration of subgaleal space resulted in resolution of CSF leak with no mor-bidities. Conclusion: Meticulous tensionless dural closure, obliteration of subgaleal space, tethering of dural grafts to bone edges are useful techniques in preventing post-operative CSF leak.

We retrospectively revised our skull base cases in the period from January 2014 to January 2019 with post-operative subgaleal CSF collection that mandates a second intervention to close the fistula to find out the etiology behind persistent leakage.

Patients & Methods
This is a retrospective case study conducted at our skull base unit, Institutional Review Board approval was obtained to perform a retrospective analysis of patients who underwent elective skull base surgeries between May 2014 and June 2019. Patients' medical record were investigated and analyzed for persistent CSF subcutaneous collection that mandate second surgical intervention. Among 219 skull base procedures done during the study period, 30 cases suffered from post-operative CSF collection, 22 of them responded to conservative measures, and 8 of them required revision dural closure. Operative notes of the primary and second surgeries were revised to identify causes of persistent subgaleal leakage/collection. Our basic dural closure was based on primary suture using polyglactin 4 -0, if insufficient and dural edges are identified, interposition dural graft is sutured to dural edges. In cases with absent dural edges due to resection or deep location, an on-lay collagen-based dural substitute is used augmented with autologous fibrin sealant.
In cases with post-operative collection, firstly, we apply compressive dressing for 3 -7 days plus Acetazolamide 250 -500 mg TID, after assessment with contrast-enhanced CT scan of the head to exclude infections. If persistent, we apply lumber drain for 3 -5 days, if failed to reduce the collection, exploration and routine post-operative prophylaxis-is usually for three days.
In five cases, water-tight dural closure was attempted using polyglactin 4 -0; four of them were primary closure and one closed using pericranial patch sutured to dural defect. In two cases, the deficient basal dural edge was closed by applying glycogen based synthetic dural graft. The eighth case was a redo surgery for atypical meningioma at which dura was removed in the previous surgeries; on-lay synthetic glycogen based graft was applied. In the three non-water tight closed cases, fibrin sealant was sprayed to support closure.
We found dural suture tears to be the cause of CSF leak in six cases. Poor graft coaptation was found in three cases, especially at the lower basal border. In revision surgeries, in cases with available dural edges all around, a pericranial patch graft is applied and sutured over the primary closure with overlying sealant. In cases of absent one or more dural edges, on overlying pericranial patch is sutured to the dura and the bone edges if the dural border is missing through holes in the bone with sealant augmentation. In one case, the dura was missing, in which a fascia lata was tethered to the bone all around and sealed with fibrin sealant.
The subgaleal space was obliterated by tethering skin to the underlying bone using the technique described by Kato et al. [8], with a subgaleal drain under mild negative pressure.

Results
Out of 219 skull base procedures performed during the period of the study (from May 2014 to June 2019); thirty cases (13.7%) suffered post-operative subgaleal CSF collections. 22 cases out of them responded to non-surgical intervention.
We found eight cases not responding to conservative treatment requiring revision surgery. The patients required revision surgery were two females and six males with age ranged from 19 to 70 years. The primary pathology was meningiomas in 7 cases (one of which was a recurrent case) and the 8 th case was an invasive pituitary adenoma (Table 1).
Dural tears caused by sutures found in 5 cases, and a closure defect (any length of free dural edge that could not be anchored to opposing dura and only graft was not enough to seal it) found in 2 cases, and a mix of the above etiology found in one case.
Three lesions were operated via fronto-orbital approach, two via fronto-temporal approach, one through frontoparietal craniotomy, one through suboccipital craniotomy, and one via retrosigmoid approach.
We did not encounter revision surgery-related complications. All patients were discharged 2 -3 days after the second surgery except one patient who was kept for rehabilitation care because of left hemiparesis.

Discussion
Dura mater is a natural barrier, protecting the intracranial contents. Failure to sion-0.5 of them was initial approach [12]. In light of our developing health system with inadequate patient service after hospital discharge and poor communication with primary care providers, such problems are dealt with before discharging the patient. We start with a non-invasive approach and upgrade to invasive maneuvers up to revision surgery if the problem persists.
Non-surgical management of post-operative subcutaneous CSF collections includes padded compressive bandage, which is released 3 -4 times per day for 30 minutes to prevent flap ischemia and pressure ulcers, diuretics mainly Acetazolamide [13], decreasing CSF hydrostatic pressure through a lumbar puncture or lumbar drainage. Since introduced by Vourc'h, spinal CSF drainage has been used successfully in treating both spinal and cranial CSF fistulae [14]. The reported success rate for cranial fistulae is 85% -88% [15]. We found this was successful in 22 out of our 30 patients, which go along with the general trend of the non-invasive approach [12]. Considering 73% of our patient who had a good response to non-surgical lines of treatment, one can argues for more prolonged conservative measures. However, as mentioned, we are trying to deliver all possible treatment before discharging patients because of poor follow up capabilities and to shorten hospital stay at the same time.
Fibrin glue, a tissue glue and hemostatic composed of fibrinogen and thrombin, is proved to be effective in reducing post-operative subcutaneous fluid collection through dural sutures [9] [16]. Although we tend to augment synesthetic graft closure with fibrin sealant, we found displacements in some points off the underlying bone edges. This might be due to interruption of the healing process of the graft as proposed by Zachary et al., who found higher CSF leakage if bovine pericardium is combined with an overlying gelatin sponge [7].
The ideal dural graft should be available in terms of size, affordability, rapid tissue intake without rejection reactions. Pericranium has been used for decades with good results for years. It has advantages over the costly synthetic grafts: usually harvested from the same incision, inexpensive, free or pedicles, ease of application, and rapid uptake by tissues. Fandino et al., in an in vitro mechanical testing of different dural grafts and sealants to repair a porcine dural defect, found the best combination is pericranium plus Tisseal [17]. We tend to use pericranial grafts whenever possible because of its superior tissue compatibility, rapid take-up by surrounding tissues, the feasibility of on-lay application or suturing it, and lower cost. We harvest the graft usually through the same incision after due undermining of neighboring flaps.
In the current study of eight cases necessitating redo surgery for persistent subgaleal CSF collection, we did pericranial graft augmentation directly sutured to underlying dura or underlying dura and adjacent bone in cases of absent dur-  [18]. Before revision surgeries, we excluded post-operative infections or hydrocephalus. The dural closure bed is explored to identify the leaking points.
The bed is cleared from blood clots and hemostatic materials.

Conclusions
Although primary water-tight dural closure is debated in some studies, it is the trend to do it whenever possible to guard against serious preventable complications like pseudo-meningocele formation, wound dehiscence, meningitis, prolonged hospital stay, and additional surgeries, especially in skull base approaches. Water-tight closure can be achieved through tensionless closure by applying dural release dissection and graft closure in case of shrinkage. Rounded needles of the smallest possible suture materials should be used.
In the cases of absent dural edges, tethering the graft to bone edge prevents graft migration and helps to prevent the leakage.
Obliteration of subgaleal space, by tethering skin to the underlying bone, releases the pressure of the skin closure and decreases the potential of developing subgaleal collection.

Limitation of the Study
There are some limitations to current study; mainly, the small sample size (number of treated patients). This might encourage to plan for future large multi institutional cohort to collect more data for more in-depth analysis. Other limiting factor is the heterogeneity of the different variables associated with CSF leakage/collection as regard the pathology, age, location of the lesion, and surgical approach; further research plan is considered to analyze each factor separately.

Financial Support and Sponsorship
Nil.

Conflicts of Interest
There are no conflicts of interest.