Introduction

The most frequent complication in posterior fossa surgery is postoperative cerebrospinal fluid (CSF) leakage presented as pseudomeningocele, incisional CSF leak, CSF rhinorrhea, subsequent subcutaneous abscess, and meningitis [9, 14,15,16, 18, 20, 23]. Some of those patients may require additional surgical interventions; therefore, watertight dural reconstruction is a crucial step in preventing CSF leaks and relevant infections [8, 17]. Microvascular decompression (MVD) can be performed through a small craniotomy. It is not always feasible to suture the dural edges in a watertight fashion in such a small craniotomy due to shrinkage of the dura mater [16, 22]. A fascial flap harvested from the suboccipital muscles is widely used as a patch graft for watertight duraplasty in MVD. Collagen matrix dural grafting has been available for more than a decade,but dural repair with single layer grafting in posterior fossa surgery has not been popularized due to cost increase and relatively high CSF leak rate [9, 12,13,14,15]. We developed a double grafting method of inlay and onlay collagen sheets for dural closure in MVD and evaluated the potential alternative use instead of the conventional method of harvesting an autologous fascial flap.

Material and method

Patient cohort

The objectives are 120 patients who underwent MVD in our institutes from June 2018 to June 2020 with more than a 6-month follow-up period (Fig. 1). Two different methods of dural reconstruction were compared. Collagen matrix grafts (DuraGen®, Integra Lifesciences, Plainsboro, NJ, USA) were used in 60 patients (the collagen matrix group), and autologous fascial flaps were grafted in the remaining 60 patients (the fascia group). Twenty-four patients with combination use of a fascial flap and a collagen matrix graft were excluded in this study.

Fig. 1
figure 1

Participant flow diagram. MVD microvascular decompression

Full size image

Operative technique

Patients were operated through the retrosigmoid approach in the lateral position. For the collagen matrix closure, we use “inlay and onlay” double grafting technique (Fig. 2). A larger size of a collagen sheet (approximately 1.5 times larger in diameter than the craniotomy size) was inserted intradurally (inlay placement) to have sufficient contact with the inner surface of the dura mater. The inlay graft was fixed to the dural edge with a running 4–0 braided nylon suture and fibrin glue to prevent migration. Then, the onlay collagen graft was placed extradurally to cover all areas of the exposed outer dura mater. For the fascia closure, a sufficient size of a thin fascial flap was harvested from the surface of the suboccipital muscles before making the muscle split and the craniotomy. The harvested fascia was kept in a moist gauze until use for the dural closure. The fascial flap was sutured in a watertight fashion with 4–0 braided nylon sutures. In both methods, no CSF leakage was confirmed with the Valsalva maneuver after completion of the dural reconstruction. In case a CSF leak was found, several sutures were added to ensure watertight closure. An artificial bone cement, calcium phosphate paste (BIOPEX-R®, HOYA Technosurgical, Tokyo, Japan), was then applied for cranioplasty in most patients except 8 patients who underwent re-exploration. No drainage tube was placed in all patients. Postoperative care, including a period of bed rest, the method of wound dressing, and the selection of antibiotics, was as per our standard protocols, which did not differ between the two groups.

Fig. 2
figure 2

Double grafting technique using inlay and onlay collagen sheets. a The retrosigmoid approach on the left in microvascular decompression for trigeminal neuralgia. A dural defect near the transverse-sigmoid junction is shown. The craniectomy size is approximately 3 cm in diameter. The transverse and sigmoid sinuses are illustrated in blue. The mastoid air cells are opened during a craniotomy and sealed with bone wax. b A larger collagen matrix graft (4.5 cm in diameter) than the craniectomy size is prepared for inlay use. The center of the inlay graft is marked with pyoctanin (purple). c The inlay graft is placed intradurally in the middle of the dural defect with reference to the center marker (purple). The yellow dashed circle illustrates the larger size of the inlay graft inserted intradurally. The inlay graft is lifted with forceps (black arrowhead) to attach the inner surface of the dura and fibrin glue is applied for fixation (white arrowhead). d The intradural space is filled with a sufficient amount of artificial cerebrospinal fluid (CSF) through a small gap (black arrowhead). e The inlay graft is bulging with water pressure and attached to the inner surface of the dura mater. f The dural edge is sutured to the inlay graft to ensure fixation (black arrowheads). This picture shows bulging of the inlay graft during Valsalva maneuver over 30 mmHg. CSF leakage is hardly seen at this step. g An onlay graft of the same size with the craniectomy is placed onto the outer dural surface for secure watertight closure. h Dural reconstruction is completed and reconfirmed watertight with repeated Valsalva maneuvers

Full size image

Assessment of outcomes

Computed tomography (CT) was taken on postoperative day one in all patients. The maximum diameter of craniotomy was measured with axial images. Fluid density in the mastoid air cells is inspected and classified into four degrees (0: no fluid density, 1: fluid density limits posterior to the mastoid antrum, 2: fluid density entering the mastoid antrum, 3: fluid density reaching the tympanic cavity) (Fig. 3). Observation of the operative wound was monitored during the inpatient hospital stay and at the follow-up examination in our clinic. Patient information, surgical and radiological findings, and the incidence of wound complications, such as pseudomeningocele, incisional CSF leak, CSF rhinorrhea, subcutaneous abscess, and meningitis, were assessed between the collagen matrix group and the fascia group. In describing demographic characteristics, p values across dura closure categories were obtained by linear regression for continuous variables and by a χ2 test or the Mantel–Haenszel test for categorical variables. Each statistical test was set to be significant at p < 0.05 (2-sided p value). SAS software (version 9.4; SAS Institute, Inc., Cary, North Carolina) was used for all statistical analyses.

Fig. 3
figure 3

Postoperative radiological findings. a An axial slice of computed tomography on postoperative day 1 shows water density in the mastoid air cells over the sigmoid sinus (grade 1, white arrowhead). The water density may include saline used during the craniotomy. An artificial bone is used for cranioplasty. The yellow arrows indicate the size of the craniotomy. b Magnetic resonance imaging on postoperative day 11 demonstrates no increase of water in the mastoid air cells (white arrowhead), indicating no cerebrospinal fluid leakage occurs. The inlay collagen graft (black arrowheads) is securely attached to the inner surface of the dura mater with larger coverage size than the craniotomy (yellow arrows). The onlay graft is difficult to be recognized as it is compressed between the artificial bone and the dural surface

Full size image

Results

Patient characteristics and radiological findings

Patient characteristics, radiological findings, CSF leakage, and infectious complications of each group are summarized in Table 1. The patient characteristics were not significantly different in both groups. The mean diameter of craniotomy was 21.2 mm in total. Mastoid air cells were opened in more than half of the cases of each group (57% in the fascia group, 53% in the collagen matrix group). Fluid density observed in the mastoid air cells postoperative day one showed mostly grade 0 to grade 1. Artificial bone cement was used in most cases (93%). Statistical analyses showed no significant difference in radiological findings of the two groups.

Table 1 Summary of patients, radiological findings, surgical outcomes, and follow-up
Full size table

Operative results and follow-up outcomes

The follow-up period was not significantly different between the two groups (median 12 months). All complications were noticed within 3 months postoperatively, and no occurrence was noted thereafter. Overall CSF leaks were observed in 5 patients (4.2%) in total (5.0% in the fascia group, 3.3% in the collagen matrix group). Most of them (4 patients) were presented as pseudomeningoceles. No CSF rhinorrhea was found in this series. All CSF leaks were treated conservatively with repeated punctures or continuous lumbar drainage except one patient who required a repair surgery in the fascia group. There was no significant difference in CSF leaks between the two groups in the statistical analyses.

Infectious complications were found in 4 patients (3.4%) in total (5.1% in the fascia group, 1.7% in the collagen matrix group). Subcutaneous abscesses were found in 2 patients (3.3%) only in the fascia group. These two patients required debridement with the removal of the bone cement. No subcutaneous abscess was observed in the collagen matrix group. Aseptic meningitis was observed in one patient in each group. The collagen matrix group showed non-inferiority to the fascia group in infectious complications.

Discussion

Posterior fossa surgery is associated with a higher risk of CSF leak than supratentorial surgery. The incidence is reported between 0.3 and 33.3% in the literature (Table2) [2,3,4,5,6, 9,10,11,12, 15, 16, 18,19,20,21]. A CSF leak remains a potentially devastating complication of a retrosigmoid craniotomy that should be avoided. Primary dural closure is frequently not feasible due to shrinkage of the dura mater by desiccation and bipolar coagulation [5, 10]. An autologous graft, such as a fascial flap harvested on-site, is a practical solution in achieving watertight dural reconstruction to prevent CSF leak in posterior fossa surgery [3, 4, 11, 16, 20]. This method is well established in MVD, and it is therefore widely used at many institutions including our institute [3, 4, 11, 23]. Conventional fascial grafting at our institute demonstrated 5.0% of CSF leaks, which is a relatively low rate of complication when compared with previous reports; therefore, it is difficult to demonstrate significant differences in the statistical analyses due to a small number of complications (3.3%) in the collagen matrix double grafting even though the collagen matrix double grafting improved the rate of CSF leakage in this study. This simple method can be an effective alternative to conventional fascial grafting.

Table 2 Summary of CSF leak rate in posterior fossa surgery
Full size table

The disadvantage of collagen matrix grafting is its additional cost ($700–800 for an MVD craniotomy size) than conventional on-site autologous fascial grafting. Although the application may be limited due to the cost, it is valuable to have an alternative surgical technique. Harvesting a sufficient size of the fascia may have the potential risk of wound infection by creating a dead space, postoperative headache by damaging the lesser occipital nerve, and local atrophy of the subcutaneous and muscle layer [1, 7]. Double grafting is a useful technique for those with subcutaneous or muscle layers that are thin, such as re-do cases, elderly patients, patients who need to avoid infection risks such as diabetic patients, and those who care about cosmetic appearance. In this study, the double grafting collagen matrix demonstrated a decreased rate of wound infection. The fascial defect over the suboccipital muscles may cause donor site morbidity against infection and paucity of local tissue in repeat surgery or debridement [9]. We speculate that a double collagen matrix grafting is benefited from keeping the suboccipital fasciomuscular layers intact without harvesting the fascial layer.

All previous reports of the use of a collagen matrix graft were single layer use as a patch for a dural defect [9, 12, 14, 15], but there is no report of double grafting in the literature. Kshettry et al. reported the results of dural reconstruction by a single use of onlay graft in 84 cases of posterior fossa surgery [9]. They found relatively higher risks of complications: CSF leak in 11.9%, wound infection in 4.8%, aseptic meningitis in 7.1%, and pseudomeningocele in 10.7%. The present study regarding a small craniotomy for MVD surgery demonstrates that double layer grafting can provide a favorable CSF leak and infection rates as compared with previous studies of a single layer collagen matrix graft. The inlay graft may play a pivotal role in preventing CSF leaks because the inlay graft can have a larger area of contact with the dura mater rather than the onlay graft (Fig. 3). Narotam et al. reported histological evaluation of implanted collagen grafts in the long-term perspective [13]. There was no inflammatory reaction or cell infiltration found in the underlying brain section, suggesting the inlay collagen matrix may not be harmful to brain tissue. It is necessary to accumulate cases and do further research to ensure its safety in the long-term.

Limitations of the present study include those that are inherent to studies of retrospective design with the small number of patients and complications. Another limitation is the relatively short follow-up period of the patients. Although the complication rates in the current series were acceptable as compared with those in previous studies, a longer period of observation and accumulation of the number of cases are needed to confirm our conclusions.

Conclusion

A combination use of inlay and onlay collagen matrix grafts is comparable to a sole use of a fascial flap in terms of CSF leaks and infections. Although the application is restricted due to the cost, our double grafting method can be a safe and effective alternative to a fascial flap for dural reconstruction in some limited conditions.