Sinusitis Independently Predicts the Presence of Headache Following Endoscopic Transsphenoidal Resection for Pituitary Adenomas

Purpose Headache is common among patients with pituitary adenomas undergone endoscopic endonasal surgery (EES), but was seldomly concerned before. The present study aims to investigate the incidence and prole of risk factors of headache after EES. Methods A meta-analysis was performed to evaluate the occurrence proportions of postoperative headache in patients with pituitary adenomas. Then, a cohort of 101 patients undergone EES were enrolled for analyzing risk factors of headache. The Headache Impact Test (HIT-6) was used to score the headache preoperatively, 1 month and 3 months postoperatively. Results A total of 18 studies and 4442 participants were included for meta-analysis. The pooled occurrence proportion of postoperative headache was 29% (95% condential interval: 20-38%). For the 101 patients enrolled in the present study, 26 (25.74%) of them had a HIT-6 scores of > 55 preoperatively, but decreased to 22 (21.78%) at 1 month, and 6 (5.94%) at 3 months, postoperatively. Multivariate analysis showed that pituitary apoplexy (OR=3.591, 95%CI 1.219-10.575, p=0.020) and Hardy's grade C-D (OR=21.06, 95%CI 2.25-197.02, p=0.008) were independently risk factors for preoperative headache. In contrast, postoperative sinusitis (OR=3.88, 95%CI 1.16-13.03, P=0.028) and Hardy's grade C-D (OR=10.53, 95%CI 1.02-109.19, P=0.049) independently predicted the presence of postoperative headache at 1 month. At 3 months postoperatively, the proportion of sinusitis tended to be higher in the headache group than the one in non-headache group (100% vs. 30.0%, p=0.070). Conclusion management


Introduction
Pituitary adenomas (PAs) are among the most common benign primary tumors of the central nervous system, accounting for 10% to 15% of all intracranial tumors [1][2][3]. At present, transsphenoidal surgery is the primary treatment for most of PAs, except prolactinomas [4]. Endoscopic endonasal surgery (EES) has shown advantages in removing pituitary adenomas, and become the mainstream choice for the most of PA patients [5][6][7]. In the long time, complications with endoscopic transsphenoidal surgery include epistaxis, sinusitis, hyposmia and hypopituitarism [8,9]. Chronic headache is also very frequent postoperatively, and affects the quality of life of patients, but was seldomly concerned before. The incidence of headaches in patients with pituitary adenomas has been reported to range between 33% and 72% [10][11][12], and only half of them were relieved postoperatively [13] . The underlining reasons still unclear.
Previous studies tried to investigate the potential risk factors associated with chronic headache post transsphenoidal surgery, and larger tumor volume, cavernous sinus invasion, optic chiasm compression and functioning PAs were shown to be independent risk factors [12,14]. However, some studies reported the negative results, where cavernous sinus invasion, suprasellar extension, and optic chiasm compression were not associated with the incidence of postoperative headache [15,16]. Several factors may account for the inconsistence of previous studies. Firstly, both groups of patients treated by transsphenoidally microsurgical approach and EES were enrolled for analysis, which might partly account for the variation of conclusions from different studies [17,18]. Secondly, the evaluation scale for headache varied among the studies [12,[15][16][17][18]. Last and the most importantly, the evaluation timing was also different among the studies, and few studies have evaluated the headache dynamically.
To address these limitations, a cohort of 101 PAs patients treated only through EES were enrolled to elaborate the risk factors of chronic headache. Moreover, to better characterized the pro le of risk factors of headache, we performed a dynamic evaluation of headache at preoperative, 1 month and 3 months postoperative. Interestingly, the pattern of risk factors of pre-and post-operative headache was different. The inclusion criteria are as follows Participants were pituitary adenomas, older than 18, had surgery intervention; Studies with detailed information on the number of headaches after surgery. The exclusion criteria are as follows: Participants were pregnant, had craniectomy. Study selection and data collection process were performed by two researchers separately. We extracted occurrence proportions as outcome indicators and pooled them through R version 4.0.5 (Shake and Throw).

Population for the retrospective study
Our study retrospectively collected clinical data of 101 patients who underwent endoscopic transsphenoidal resection of pituitary adenomas between July 2018 and December 2020. This study was reviewed and approved by the Research Ethics Committee. All the patients have known and agree to be involved in the present study.
Patients underwent pituitary adenoma removal through EES, with pathological con rmation and complete follow-up data, were included. Patients underwent craniotomy resection for pituitary adenomas, or those with incomplete follow-up data were excluded. EES resection of PAs was performed by an experienced neurosurgeon. In general, posterior nasal septectomy was performed to allow binasal access. The face of the sphenoid and intersphenoidal septations was ground away. The tumor was resected after the dura was incised. Complex intraoperative cerebrospinal uid leaks were repaired in a layered fashion, which composed of a combination of Gelfoam, fat, facia, bone, and a vascularized naso-septal ap. A catheter balloon was placed in the posterior nostril for reconstruction.
All patients underwent preoperative CT plain scan and contrast-enhanced MRI of the head. Clinical data, including age, gender, pituitary apoplexy, diameter of tumor, clival osseous destruction, sinusitis, extent of tumor resection and postoperative intracranial hemorrhage, were collected. The diagnosis of sinusitis was based on head MRI. The invasiveness of PAs was graded with Knosp's [19] and Hardy's [20] grade system. The Headache Impact Test (HIT-6) was used to evaluate the extent of headache preoperatively, 1 month and 3 months postoperatively [21,22]. Patients were with a HIT-6 scores of >55 was considered to have headache [23,24].
Nose care during the follow-up of patients post-EES for PAs.
A follow-up MRI will be performed 1 month, and 3 months postoperatively to assess residual tumor and sinusitis. One month after the operation, all the patients will be suggested to start nose rinsing every day for 2 months. In addition, maxillary natural ostium and sphenoidal sinus will be cleaned endoscopically, 1 month postoperatively.
Statistical analyses SPSS 20.0 (IBM Corp.) was used for data analysis. The quantitative data were expressed as the mean ± SD. Independent sample t test and Pearson chi-square test or Fisher's exact test were used to analyze the statistically signi cant difference between the two groups. Risk factors of headache were analyzed by univariate and multivariate analysis. A p value less than 0.05 was considered signi cantly.

Results
Meta-analysis for the occurrence proportion of headache postoperatively.
A total of 619, 735 and 47 studies were retrieved through PubMed, EMBASE and Cochrane library, respectively. As shown in Fig. 1A, 18 studies with 4442 participants were nally included in metaanalysis. Information on the number of headaches was shown in Fig. 1B. Collectively, the pooled occurrence proportion were 29% (95% con dential interval: 20-38%) (Fig. 1C).
Dynamic analysis for the risk factors of pre-and post-operative headache.
Basic clinical and pathological characteristics of patients.
The baseline clinical characteristics of patients with pituitary adenomas are shown in Table 1. A total of 101 patients were enrolled for the present study, with a mean age of 45.71 ± 11.90 years. The mean maximum diameter of the tumor was 28.25 ± 11.04 mm. Thirty-nine patients (38.6%) were with tumor of Knosp's grade 3-4, and 6 patients had those in Hardy's Grade C-D. Gross total resection was achieved in 85 patients (84.2%). Twenty-four cases (23.8%) were functioning pituitary adenomas, and 15 cases of them were growth hormone secreting adenomas. There were 24 samples (23.8%) with positive staining of transcription factor TBX19. The average value of Ki-67 index was 2.56 ± 1.64%. Preoperatively, 32 patients were found to have slight sinusitis, based on head MRI. At the follow-up at 1 and 3-month postoperatively, 54 (53.5%) and 17 (15.9%) were con rmed to have sinusitis.
Preoperative headache was signi cantly relived 3-month after operation.
HIT-6 scores of pre-and post-operations are compared in Fig. 2B. On the whole, the HIT-6 scores of 3 months postoperatively was signi cantly lower than the preoperative one (41.12 ± 0.72 vs. 46.06 ± 1.06, P= 0.0002) and the 1-month postoperative one (41.12 ± 0.72 vs. 46.82 ± 0.90, P< 0.0001). However, there is no statistically signi cant difference between preoperative HIT-6 scores and the 1-month postoperative one. These data suggested that patients with headache preoperatively could be signi cantly relived 3month after operation.
Univariate and multivariate analysis of risk factors associated with headache post EES for PAs.
To compare the pattern of risk factors of pre-and post-operative headache, we rst made the analysis on the preoperative HIT-6 scores. Univariate and multivariate analysis showed that both pituitary apoplexy (OR=3.59, 95%CI 1.22-10.58, p=0.020) and Hardy's grade C-D (OR=21.06, 95%CI 2. .02, p=0.008) were independently risk factors (Table 3). However, there was no statistically signi cant difference in gender, presence of sinusitis, clival osseous destruction and Knosp's grade, as well as growth hormone secreting adenomas and TBX19 positive (Table 3).
Then, we analyzed the risk factors associated with the headache 1 month postoperatively. Intriguingly, the pattern of risk factors is different from the preoperative one. As displays in Table 4, Hardy's grade C-D (OR=10.53, 95%CI 1.02-109.19, P=0.049) remains an independent risk factor of postoperative headache.
However, postoperative sinusitis (OR=3.88, 95% CI 1.16-13.03, P=0.028) also independently predicts the presence of postoperative headache at 1 month, which strongly recommends the prevention of sinusitis after EES for PAs.
Correlation analysis of sinusitis and headache in the subgroup of patients with headache.
Considering the clinical signi cance of the causal correlation between sinusitis and headache, we further analyzed their potential correlation in the subgroup of patients with headache. A total of 35 patients suffered from headache in the current study, including 26 preoperative and 9 postoperative ones. The dynamic change of HIT-6 scores and relevant presence of sinusitis were present in Fig. 2C-E.
Then, we further tracked the status of patients 3 months after operation, which are displayed in Fig. 3. Among the 26 patients with headache preoperatively, 12 patients relieved and 14 patients remain with headache 1-month postoperatively. The proportion of sinusitis was signi cantly higher in the headache group than the one in non-headache group (85.7% vs. 41.7%, p=0.038). At 3 months, 10 of the 14 patients relieved and 4 patients remain with headache. The proportion of sinusitis was tended to be higher in the headache group than the one in non-headache group (100% vs. 30.0%, p=0.070).
In the 9 patients who did not have headache preoperatively, 8 patients developed headache 1 month postoperatively, and 6 (85.7%) of them had sinusitis. The patient developed headache at 3 months also was also with sinusitis con rmed by MR. Collectively, these data strongly suggested that postoperative sinusitis partly accounts for the presence and relief of chronic headache following EES for PAs.

Discussion
Chronic headache is a common complication following PAs resection transsphenoidally. However, its incidence and risk factors remain inconsistent. By systematically reviewing previous literature, we declared a pooled occurrence proportion of headache was 29%, following transsphenoidal surgery for PAs. By evaluating the headache at multiple points, we found a different pattern of risk factors of headache between pre-and post-operation, which may help to explain the inconsistent conclusions among previous studies. Furthermore, postoperative sinusitis, but not the preoperative sinusitis, was rst shown to be in-dependently risk factors for the occurrence of headache, which strongly suggested that measures are taken to prevent postoperative sinusitis, and thus decrease the occurrence of postoperative headache.
In order to evaluate whether hypothalamus repression will induce headache, patients were divided into Hardy's grade C-D and Hardy's grade A, B and E groups. As shown in table 1, the ratio of patients with headache was signi cantly higher in the Hardy's grade C-D group (22.7% vs. 1.3%, p=0.002). In multivariate analysis, Hardy's grade C-D was also independently predicted the presence of headache, in both preoperative and postoperative analysis. Hypothalamic regulation of has been shown to be involved with migraine and cluster headache, and indeed a structure in close proximity to the hypothalamus has been identi ed to play a crucial role in attack generation [25]. Therefore, for the PAs patients with headache induced by mass effect on hypothalamus, headache could be treated postoperatively. And the function of hypothalamus won't be recovered until 1-3 months after operation.
Headache associated with pituitary adenomas may be induce through displacement of intracranial painsensitive structures located in the blood vessels, cranial nerves and dura mate, in ammation involving pain-sensitive structures or meningeal irritation, or hormonal dysregulation. [12,15] The cavernous sinus contains pain-producing structures, such as the internal carotid artery and trigeminal nerve and ganglion, invasion of which might be expected to cause pain. Different studies have different opinions on pain caused by cavernous sinus invasion [12,15,16]. The evaluation timing for headache, and scale used may partly accounting for the inconsistence. In our study, we evaluate the risk factors at multiple points, including preoperative, 1 month and 3 months postoperatively. Knosp's grade was not shown to be signi cantly associated with headache at all points. Much more deliberated studies are warranted to declare their correlations.
Postoperative sinusitis is more common due to the opening of the sinuses in the nasal cavity and the need to open the sphenoid sinus for endoscopic resection of pituitary adenomas [26]. Sinusitis can cause headaches [27]. Given that the nasal mucosa is extremely sensitive, the feeling of pain is more obvious after surgery via endoscopic transsphenoidal surgery. In our study, preoperative sinusitis was determined by craniocerebral MRI in 31.7% of patients, and the proportion of postoperative sinusitis increased to 53.5%. The proportion of postoperative sinusitis in headache group was higher than that in non-headache group. Multivariate logistic regression analysis showed that the OR of postoperative headache in patients with postoperative sinusitis was 3.883. When analyzing data of headache and sinusitis, we found that the proportion of sinusitis in patients with postoperative headache at the same period is always higher than that without headache. Patients with postoperative sinusitis are more likely to develop postoperative headaches.
In some studies, patients with functioning pituitary adenomas are more likely to develop postoperative headaches [15], especially nociceptive tumors like growth hormone secreting tumors and prolactinomas. They can potentially demonstrate improvement with endocrinological treatments [28,29]. In our study, functioning pituitary adenomas and growth hormone adenomas were not associated with postoperative headache in patients with pituitary adenomas. We suspect that some patients with functioning pituitary adenomas have achieved biochemical remission after surgery and their hormone levels have decreased to normal or close to normal and therefore do not present with headache. We need to further verify the relationship between postoperative hormone levels and headache.
Our study also has some limitations. Firstly, the sample size of this study is still insu cient and the follow-up time is limited. it is necessary to further expand the sample size and extend the follow-up time to investigate the risk factors of long-term headache. Secondly, we failed to evaluate the status of pituitary function completely, which is an important issue for the patients' quality of life. Finally, the causation relationship between sinusitis and postoperative headache remains unknown until prospective studies performed.

Page 8/18
Headache is a very common complication following EES for pituitary adenomas. Postoperative sinusitis and Hardy's grade are independent risk factors for postoperative headache. Prophylactic treatments may help to prevent postoperative sinusitis, and thus reduce the incidence of headache. Prospective and randomized clinical studies including more patients are suggested to con rm our ndings.

Competing interests
The authors declare that they have no competing interests.

Availability of data and material
The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.
Authors' contributions JG, XC, SY and SC participated in data collection and follow-up of patients. YZ performed meta-analysis. XJ, DL, KS, and YM performed surgery. JG, XC, ZZ and XJ participated in writing and proofreading the manuscript. ZZ and XJ designed the study, were involved in the whole process and were guarantors of the integrity of the entire study.

Ethics approval
This study was reviewed and approved by the Research Ethics Committee of Sun Yat-sen University Cancer Center.

Consent to participate
All the patients have known and agree to be involved in the present study.

Consent for publication
Written informed consent for publication was obtained from all participants. Tables Table 1 The     Flow chart of headache and sinusitis preoperatively, 1 or 3 months postoperatively (Numbers in brackets indicate the number and proportion of patients with sinusitis at the same period).