Over the past 2 decades, along with developments in endoscopic equipment and anatomy exploration, neurosurgery has accomplished great success in operative treatment for skull base tumors. Lesions located in the cerebral ventral region only used to be touched through a complicated transcranial approach, which might induce extra damage and prolonged hospitalization. Now, by using an endoscope with an expanded endoscopic endonasal approach (EEA), neurosurgeons can not only easily explore the ventral cerebrum region but are also capable of resecting tumors with minimal interference to other adjacent structures. As a coin holds two sides, the disadvantage of the expanded EEA is as obvious as its advantage, which is the induction of cerebrospinal fluid (CSF)-related complications. Due to the EEA’s special trajectory and the location of the skull base lesion, isolation between the nasal cavity and the intracranial space is a vital procedure for decreasing the morbidity of CSF complications after resection. In the past, the rate of postoperative CSF complications was >20% until the introduction of the Hadad-Bassagasteguy flap (HBF) for expanded EEA (1). Currently, it has dramatically decreased to only <5% (2–6), which is comparable with that of the traditional transcranial approach (7). However, along with the ubiquitous utilization of expanded EEA worldwide, for some selected but not rare patients, CSF complications are still a huge challenge. This challenge might be life-threatening for certain patients who have to undergo a secondary EEA surgery. In prior surgery, neurosurgeons initially harvested an HBF to prevent anticipated CSF leakages. This HBF occupies the ipsilateral nasal septal artery (NSA) as its blood pedicle. Meanwhile, to fully visualize vital anatomical landmarks located at the posterior wall of the sphenoid sinus, sphenoidectomy and nasal septectomy are usually performed to gain binaural and bimanual access. Such a maneuver always interrupts the contralateral NSA, which sacrifices the capability of employing the contralateral septal mucosa as a rescued HBF for secondary surgery (8). Given the essential role of the vascularized flap for the reconstruction of the skull base, to obtain a vascularized flap again, Kassam et al. employed a pedicled inferior turbinate flap (IPTF) to restore the boundaries between the nasal cavity and subarachnoid space (9). Carnevale et al. introduced a method with a pedicled middle turbinate flap (MPTF) to cover the skull base defect (10). In addition, a temporoparietal facia flap was employed to restore dural defects after expanded EEA by Gardner (11). Surely, these kinds of methods are good substitutes for reconstruction when an HBF is absent in repeated EEA surgery; however, when considering this consensus that improving patients’ postoperative life quality and minimizing surgery-related injury as much as possible, they all hold some drawbacks. After harvesting the MTPF and ITPF, severe crusting and nasal pain frequently arise postoperatively due to a more denuded bone surface. Furthermore, the sizes of the MTPF and ITPF are smaller than that of the HBF. Due to their irregular shape of inferior and middle turbinates, MTPF and ITPF are more fragile and can be easier to rupture when elevating from the bone surface. Regarding the temporoparietal facia flap, an extra hemicoronal incision is needed, and a series of destructive maneuvers are performed to transfer it into the nasal cavity. Thus, regarding secondary EEA, how to properly confront CSF complications and maintain patients’ quality of life is a dilemma. Upon this delicate issue, recycling a prior HBF is hardly an unreasonable option. When elevating a prior HBF in secondary surgery, however, it is always inevitably ruptured due to a tight adhesion between the flap and the bone or dura, which leads to a shortened flap that is not capable of fully covering the dura defect. Thus, a small autologous or artificial “patch” is needed to compensate for its shortened size. Here, we are the first to introduce our clinical experience of comprising a prior HBF with free autologous nasal septal mucosa to rebuild the skull base and share our results about the independent risk factors for CSF complications in secondary EEA surgery.

After scrutinizing the related medical files, 69 patients who underwent secondary EEA surgery were enrolled in this study. Approximately 37 male and 32 female patients were included in this cohort, whose mean age was 50.8710.90 years. Based on the “patch” option, the patients were assigned into 2 groups: prior HBF incorporated with autologous mucosa (n=41) and prior HBF incorporated with artificial dura (n=28). The lesions involved in this study included meningioma (n=5), craniopharyngioma (n=5), chordoma (n=6) and pituitary adenoma (n=53). There was no significant difference between these 2 groups regarding age (49.509.54 years, 51.8011.77 years, p=0.392), sex (p=0.628), BMI (p=0.323), diabetes mellitus (p=0.933), intraoperative communication of the 3^rd ventricle (p=0.315), pathological diagnosis (p>0.38), or the rate of total resection (p=0.698). There were no postoperative mortalities (death in the immediate postoperative course or within 1 month after operations). All patients were clinically followed until death or loss of contact. The mean follow-up was 30.615.8 months. Detailed information about the patients is listed in Table 1 .

As shown in Table 1 , in the artificial patch group, there was 1 patient with a pseudomeningocele (3.6%), 3 patients with a CSF leak (10.7%) and 6 patients with meningitis (21.4%). In the autologous patch group, there were 2 patients with a pseudomeningocele (4.9%), 1 patient with a CSF leak (2.4%) and 1 patient with meningitis (2.4%). A significant difference was observed between these 2 groups regarding postoperative CSF-related complications (p=0.008). In particular, the employment of autologous patches dramatically reduced the incidence of postoperative meningitis (p=0.031).

As shown in Figure 2 and Table 2 , according to the SNOT-22 test, there was no significant difference between the 2 groups preoperatively (p=0.698). The 1-month and 2-month postoperative SNOT-22 test scores were obviously increased compared with the preoperative scores (p<0.001). However, 3 months after the secondary EEA surgery, the score of the SNOT-22 test gradually returned to the preoperative level (p=0.056, p=0.08). Meanwhile, no significant difference was observed between the 2 groups regarding the SNOT-22 test at 1 month, 2 months and 3 months postoperatively (p=0.409, p=0.131, p=0.318).

To further evaluate the effectiveness of the autologous patch and detect related independent risk factors for postoperative CSF complications, multiple variables, such as age, sex, BMI, pathology type, diabetes mellitus, gross total resection ratio (GTR), duraplasty material type and intraoperative perforation of the 3^rd ventricle, were included in our study according to the primary literature and clinical experience (14–16). Univariate and multivariate logistic regression were used to analyze multiple variables. Variables with p<0.05 detected by univariate logistic regression were included into the multivariate logistic regression. In our series, as shown in Table 3 , artificial patches (p=0.013), diabetes mellitus (p=0.011), craniopharyngioma (p=0.011), and chordoma (p=0.043) were confirmed to be independent risk factors through the univariate and multivariate logistic regression. Age (p=0.716), sex (p=0.768), BMI (p=0.621), GTR ratio (p=0.414) and intraoperative perforation of the 3^rd ventricle (p=0.414) were not independent risk factors for secondary EEA surgery.

To further assess the influence of the independent risk factors above on each of the CSF complications (pseudomeningocele, CSF leak, meningitis), multinomial logistic regression was employed to detect the association among them. As shown in Table 4 , diabetes mellitus and the utilization of artificial patches were more likely to induce CSF leak (p=0.008, p=0.05) and meningitis (p=0.02, p=0.02) than pseudomeningocele (p=0.066, p=0.466). The patients suffering from craniopharyngioma were more susceptible to meningitis (p=0.038) than CSF leak (p=0.112) and pseudomeningocele (p=0.196). Chordoma indiscriminately increased the risk for all 3 kinds of postoperative CSF complications (p=0.1, p=0.157, p=0.254).

In recent years, an increasing number of neurosurgeons have chosen EEA as their priority when confronting brain ventral tumors. Regarding related postoperative CSF complications, specific techniques and clinical data have been elucidated. However, to the best of our knowledge, few studies have considered the needs of patients who might experience a secondary EEA operation. These patients are minority but are not rare. During a secondary EEA surgery, a more extensive bone window was sometimes performed to gain a better view. Meanwhile, due to the lack of an accessible vascularized graft, a water-tight closure might not be accomplished when compared to that of the initial surgery. Thus, exploring an accessible, effective and safe substitute for HBF in secondary EEA surgery is a crucial issue. In the present study, we aimed to introduce a method that could not only confine the damage as much as possible but also hold considerable effectiveness in secondary EEA surgery. In our view, recycling the prior HBF is a reasonable option. According to our results, a recycled HBF incorporated with autologous free mucosa was effective in preventing CSF complications for secondary EEA surgery, which was comparable to that of the initial EEA surgery that rebuilt the skull base with a whole HBF (14–17). Moreover, we consecutively recorded the SNOT-22 scores before and 1 month, 2 months, and 3 months after the operation for the 2 groups. Although the SNOT-22 score was distinctly increased 1 month and 2 months postoperatively compared to those pre-operation, no significant difference was observed between pre-operation and 3 months postoperation for each group, and no significant difference was confirmed between the 2 groups at the 1-month, 2-month and 3-month postoperative SNOT-22 tests. Thus, our solution did not aggravate patients’ sinonasal quality of life. Considering the irreplaceable role of a vascularized flap in restoring the skull base, many scholars have presented multiple alternative resolutions regarding the lack of a whole HBF in secondary EEA surgery. These alternative vascularized flaps mainly originated from middle/inferior turbinate or pericranial tissues. In an anatomic dissection study, a pedicle middle turbinate flap was capable of covering a defect caused by EEA surgery (18). In clinical studies, some interesting results have been reported from a series of small cohort studies. Julián et al. proposed the utilization of a middle turbinate flap as a replacement when an HBF was not available (19). None of the patients suffered from postoperative CSF leakage in their cohort. Another study that enrolled more patients introduced by Carnevale et al. also evaluated the satisfactory effect of pedicled middle turbinate flaps in repeated EEA surgery (10). Furthermore, a pedicled inferior turbinate could also be employed as a reliable backup for HBF (9, 20). According to these studies, the pedicled middle/inferior turbinate flap is trustworthy when a whole HBF is absent. However, harvesting a pedicled middle/inferior turbinate flap will inevitably contribute to a more denuded nasal bone surface, which always leads to more severe nasal morbidity and poor sinonasal quality of life (18, 21). However, this issue has not attracted the full attention of neurosurgeons (13), and few neurosurgeons have specifically assessed alterations in nasal quality of life before and after EEA surgery. In addition, limited by the irregular shape and size of the turbinate, middle/inferior turbinate flaps are easily ruptured during the harvest process and might only fit a standard defect (22). Thus, when confronted with an expanded defect in secondary EEA surgery, pedicled middle/inferior turbinate flaps might not be sufficient. Apart from these pedicled nasal flaps, other scholars introduced a series of pedicled pericranial flaps, such as split-frontal pericranial flaps (23, 24) and temporoparietal fascia flaps (11), as substitutes for HBF in secondary EEA surgery. Although a watertight closure could be obtained by these pedicled pericranial flaps, they all need an extra incision to harvest the flap and a small bone window to rotate the flap into the nasal cavity, which inevitably causes extra injury for patients. In sum, for secondary EEA surgery, our strategy was not only a reliable solution but also prevented to bring extra damage to patients’ nasal quality of life.

In the present study, diabetes mellitus, artificial dura, craniopharyngioma and chordoma were detected to be independent risk factors for CSF complications in secondary EEA surgery.

According to Kiril E et al., diabetes mellitus patients are more vulnerable to certain infectious diseases, and the incidence of meningitis in diabetes patients is 2-fold higher than that in non-diabetes mellitus patients (25). Helene et al. declared that cell-mediated immunity might be severely disturbed in diabetics, especially when the normal function of polymorphonuclear leukocytes was obviously interfered with (26). Delamaire et al. also noted that neutrophil chemotaxis was weakened compared with controls after stimulation in both type I and type II diabetics (27). According to Sharipov et al. and Yu Jin et al., diabetes mellitus is one of the risk factors for nosocomial meningitis after endoscopic transsphenoidal surgery (28, 29). In the present study, for secondary EEA surgery, diabetes mellitus was an independent risk factor for CSF complications detected by the univariate and multivariate logistic regression and was more likely to induce meningitis and CSF leakage than pseudomeningocele according to multinomial logistic regression, which was consistent with previous studies.

Amanda et al. stated that clival chordoma was highly associated with postoperative CSF leakage in pediatric patients (30). According to a systematic review by Ethan et al., dural defects are always viably reconstructed by HBF, except those located in the clival region (17). Calvin et al. noted that dural defects located at the central skull base (sella and clival region) have an increased risk of CSF leakage compared to those of the anterior skull base (14). This phenomenon could be associated with multifactorial reasons. They proposed that an anterior skull base dura defect was always located at a horizontal plane, and the potential gap around the defect could be eliminated by counter pressure brought by the frontal lobe due to gravity. In contrast, regarding the clival and sella regions, a dural defect is always located at a vertical plane where there is a lack of tight counter pressure between the defect and the reconstructed material in fear of inducing more injuries to vital neural structures such as the optic nerve and brain stem. Shannon et al. also declared that lesions located in the posterior fossa are associated with a high rate of CSF leakage (16). In our cohort, craniopharyngioma and chordoma, which are both mainly located at the central skull base, were detected to be independent risk factors for postoperative CSF complications in secondary EEA surgery. Meanwhile, through multinomial logistic regression, the risk of each CSF complication was indiscriminately increased in patients suffering from chordoma. Moreover, based on our results, craniopharyngioma was more likely to induce meningitis, which might contribute to the prolonged stretching of brain tissue due to a lack of a clear interface between normal structures and lesions during secondary EEA surgery. Such injuries might subsequently induce astrocytic and microglial infiltration and related inflammatory responses (31, 32), which develop and persist during the perioperative period. Furthermore, a cystic craniopharyngioma could lead to postoperative chemical meningitis attributed to a leakage of cyst contents (33, 34).

According to previous literature, infection in situ, immune response, delayed healing, hemorrhage, bacterial and viral colonization are obviously associated with artificial duraplasty materials (35, 36). Other scholars also stated that artificial dura could trigger related chronic inflammation and the formation of granulation tissue due to its xenogeneic nature (37). Malliti et al. declared that dural defects reconstructed with synthetic dura are more susceptible to wound infection and CSF leakage. They also suggested that the utilization of synthetic dural grafts should be reserved except when autologous materials are unavailable (38). Consistently, the employment of artificial material was an independent risk factor for postoperative CSF complications, as confirmed by the univariate and multivariate logistic regression in the present study. Moreover, as detected by multinomial logistic regression, duraplasty with artificial material is more likely to increase the incidence of infection and CSF leakage than pseudomeningocele in secondary EEA surgery.

The major limitation of our study is its retrospective nature. Some patients’ first endonasal surgeries were not performed in our tertiary hospital. We could only speculate its first possible approaches and reconstruction strategies according to historical CT or MRI images. Thus, an accurate depiction about the approaches and reconstruction strategies performed in the first surgery is somewhat far-fetched, which might bring certain bias for our research. Meanwhile, patients who had previously experienced radiotherapy after initial surgery were excluded from our cohort, which might result in a certain population bias in our study. Thus, further prospective randomized controlled trials are needed to validate our results.

Recycling an initial HBF along with a small piece of free autologous mucosa to reconstruct the skull base in secondary EEA surgery is a reasonable and effective method, especially for patients simultaneously suffering from diabetes mellitus and central skull base tumors.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Ethics Committee of the Seconds Affiliated Hospital of Air force Medical University of China. The patients/participants provided their written informed consent to participate in this study.

Material preparation, data collection and analysis were performed by RW, GZ, and ZZ. The first draft of the manuscript was written by RW and all authors commented on previous versions of the manuscript. All authors contributed to the article and approved the submitted version.