Cerebrospinal fluid fistula after endoscopic skull base surgery case series

Data from consecutive patients diagnosed with sellar region lesions with grade II-III Cerebrospinal fluid fistula after endoscopic skull base surgery during endoscopic endonasal surgery were collected from May 2015 to May 2020. Skull base reconstructions were performed with the cruciate embedding fascia-bone flap (CEFB) or the conventional Hadad-Bassagasteguy flap. Related clinical data were analyzed. The combined use of the CEFB and PNSF was applied to an additional supplemental case series of patients with grade III leak and multiple high-risk factors.

Results: There were 110 and 65 patients included in the CEFB and PNSF groups, respectively. The CEFB demonstrated similar effects on the incidence of postoperative CSF leak (2.7%), intracranial infection (4.5%), and lumbar drainage (LD) placement (5.5%) as PNSF (3.1%, 3.1%, and 6.2%), but with less epistaxis (CEFB: 0%, PNSF: 6.2%) and nasal discomforts (CEFB: 0%, PNSF: 7.7%). The LD duration (CEFB: 6.67 ± 2.16 days, PNSF: 10.50 ± 2.38 days), bed-stay time (CEFB: 5.74 ± 1.58 days, PNSF: 8.83 ± 3.78 days) and hospitalization time (CEFB: 10.49 ± 5.51 days, PNSF: 13.58 ± 5.50 days) were shortened in the CEFB group. The combined use of CEFB and PNSF resulted in 0 postoperative CSF leaks in the supplemental case series of 23 highly susceptible patients.

Conclusion: This study suggested that the new CEFB technique has the potential to prevent postoperative CSF leak in EES. The results indicated that it can be used effectively without PNSF in suitable cases or applied in addition to a PNSF with high compatibility when necessary. Its effectiveness should be further verified with a larger cohort and better design in the next step. Trial Registration Current Controlled Trials ChiCTR2100044764 (Chinese Clinical Trial Registry); date of registration: 27 March 2020. Retrospectively registered ¹⁾.

Intraoperative CSF leaks occurred in 45 (29%) of 154 patients. No risk factors for CSF leaks were identified using conventional statistical methods. The deep neural network-based prediction model classified 88% of patients in the test set correctly, with an area under the curve of 0.84. Sensitivity (83%) and specificity (89%) were high. The positive predictive value was 71%, negative predictive value was 94%, and F1 score was 0.77. High suprasellar Hardy grade, prior surgery, and older age contributed most to the predictions.

Staartjes et al., trained and internally validated a robust deep neural network-based prediction model that identifies patients at high risk for intraoperative CSF. Machine learning algorithms may predict outcomes and adverse events that were previously nearly unpredictable, thus enabling safer and improved patient care and better patient counseling ²⁾.

The aim of the study is to classify the risk of postoperative leak and to propose a decision-making protocol to be applied in the preoperative phase based on radiological data and on intraoperative findings to obtain the best closure.One hundred fifty-two patients were treated in our institution; these patients were divided into 2 groups because from January 2013 the closure technique was standardized adopting a preoperative decision-making protocol. The Postoperative CSF leak Risk Classification (PCRC) was estimated taking into account the size of the lesion, the extent of the osteodural defect, and the presence of intraoperative CSF leak (iCSF-L). The closure techniques were classified into 3 types according to PCRC estimation (A, B, and C).The incidence of the use of a nasoseptal flap is significantly increased in the second group 80.3% versus 19.8% of the first group and the difference was statistically significant P < 0.0001. The incidence of postoperative CSF leak (pCSF-L) in the first group was 9.3%. The incidence of postoperative pCSF-L in the second group was 1.5%. An analysis of the pCSF-L rate in the 2 groups showed a statistically significant difference P = 0.04.The type of closure programmed was effective in almost all patients, allowing to avoid the possibility of a CSF leak. Our protocol showed a significant total reduction in the incidence of CSF leak, but especially in that subgroup of patients where a leak is usually unexpected ³⁾.

Retrospective review of patients who underwent endoscopic endonasal surgery EEA for the resection of intradural pathology between January 1997 and June 2012 was performed. Basic demographic data were collected, along with patient body mass index (BMI), tumor pathology, reconstruction technique, lumbar drainage, and outcomes. RESULTS Of the 615 patients studied, 103 developed a postoperative CSF leak (16.7%). Sex and perioperative lumbar drainage did not affect CSF leakage rates. Posterior fossa tumors had the highest rate of CSF leakage (32.6%), followed by anterior skull base lesions (21.0%) and sellar/suprasellar lesions (9.9%) (p < 0.0001). There was a higher leakage rate for overweight and obese patients (BMI > 25 kg/m2) than for those with a healthy-weight BMI (18.7% vs 11.5%; p = 0.04). Patients in whom a pedicled vascularized flap was used for reconstruction had a lower leakage rate than those in whom a free graft was used (13.5% vs 27.8%; p = 0.0015). In patients with a BMI > 25 kg/m2, the use of a pedicled flap reduced the rate of CSF leakage from 29.5% to 15.0% (p = 0.001); in patients of normal weight, this reduction did not reach statistical significance (21.9% [pedicled flap] vs 9.2% [free graft]; p = 0.09). CONCLUSIONS Preoperative BMI > 25 kg/m2 and tumor location in the posterior fossa were associated with higher rates of postoperative CSF leak. Use of a pedicled vascularized flap may be associated with reduced risk of a CSF leak, particularly in overweight patients ⁴⁾.

Jeon et al. assessed the repair outcomes for high-flow CSF leaks based upon the tumor location, and analyzed the reasons for repair failure after endoscopic transnasal approach solely for skull base tumors.

From Feb. 2009 to Dec. 2014 they performed endoscopic endonasal surgery for a variety of skull base lesions in 788 patients at the institution. Among them, 95 patients with intradural skull base tumors underwent endoscopic transnasal surgery. They performed surgical repairs with a multilayered nonvascularized construct (38 patients) and a vascularized pedicled nasoseptal flap construct combined with a fascia graft (57 patients). Overall, 14 of 95 patients (14.7%) who underwent endoscopic transnasal surgery for skull base tumors developed postoperative CSF leaks. The major causes of repair failure included graft disruption by a lack of counter-pressure in the multilayered non-vascularized technique, and inadequate drilling of the sphenoid bone, displacement of the flap due to pressure from CSF or gravity, or disruption of flap integrity in the vascularized pedicled flap technique. Logistic regression analysis revealed that there was no significant association between repair failure and age, sex, type of reconstructive method used, and primary tumor type (p>0.05). Reconstruction after endoscopic endonasal surgery remained challenging, especially for non-pituitary skull base tumors requiring intra-arachnoidal dissection. Recent advances in reconstructive techniques require the accumulation of experiences with sufficient dexterity to achieve an acceptable morbidity rate ⁵⁾.

The intraoperative detection of CSF leaks during endonasal endoscopic skull base surgery is critical to preventing postoperative CSF leaks. Intrathecal fluorescein (ITF) has been used at varying doses to aid in the detection of intraoperative CSF leaks. However, the sensitivity and specificity of ITF at certain dosages is unknown. METHODS A prospective database of all endoscopic endonasal procedures was reviewed. All patients received 25 mg ITF diluted in 10 ml CSF and were pretreated with dexamethasone and Benadryl. Immediately after surgery, the operating surgeon prospectively noted if there was an intraoperative CSF leak and fluorescein was identified. The sensitivity, specificity, and positive and negative predictive power of ITF for detecting intraoperative CSF leak were calculated. Factors correlating with postoperative CSF leak were determined. RESULTS Of 419 patients, 35.8% of patients did not show a CSF leak. Fluorescein-tinted CSF (true positive) was noted in 59.7% of patients and 0 false positives were encountered. CSF without fluorescein staining (false negative) was noted in 4.5% of patients. The sensitivity and specificity of ITF were 92.9% and 100%, respectively. The negative and positive predictive values were 88.8% and 100%, respectively. Postoperative CSF leaks only occurred in true positives at a rate of 2.8%. CONCLUSIONS ITF is extremely specific and very sensitive for detecting intraoperative CSF leaks. Although false negatives can occur, these patients do not appear to be at risk for postoperative CSF leak. The use of ITF may help surgeons prevent postoperative CSF leaks by intraoperatively detecting and confirming a watertight repair ⁶⁾.