Effect of pituitary adenoma consistency on surgical outcomes in patients undergoing endonasal endoscopic transsphenoidal surgery

Most pituitary adenomas (PAs) are considered to have a soft tumor consistency. However, there is a non-negligible percentage (5–13%) of tumors presenting or exhibiting a fibrous consistency that would entail a more difficult and complicated surgical excision with higher surgical morbidity and mortality rates. To analyze the clinical consequences of PA tumor consistency on the surgical outcomes in patients undergoing endonasal endoscopic transsphenoidal (EET) pituitary surgery. An ambispective study of patients with PAs operated on through an EET approach in two Spanish tertiary hospitals over the last 12 years. A total of 226 consecutive interventions were carried out in the Neurosurgery Departments of the Hospital Universitario Ramón y Cajal (HURC) and the Hospital Universitario Puerta del Sur by the same neurosurgeon. PAs were grouped into soft (n = 150) and fibrous (n = 76). All patients underwent hormonal and magnetic resonance imaging (MRI) studies before and after surgery. In addition, neurosurgical complications were recorded in each patient. Fibrous adenomas were independently associated with lower resection rates compared to soft adenomas (fibrous gross total resection [GTR] rate 48.7% vs. 76.3%, p < 0.001), even in those adenomas without invasion of the cavernous sinus (Knosp grades 0, I, and II). There were more intraoperative cerebrospinal fluid (CSF) leaks in patients with fibrous PAs. Moreover, fibrous PAs showed higher rates of postoperative hypopituitarism, permanent diabetes insipidus (DI) and postoperative treatments (hormonal treatment and radiotherapy). The excision of a fibrous PA required a longer surgical time (22.5 min more than soft PAs, p = 0.014), regardless of other factors. The consistency of the PAs significantly conditions both the results of surgery (lower resections rates), complications (higher incidence of postoperative hypopituitarism, permanent DI), and the prognosis (higher incidence of postoperative treatments) of the patient undergoing EET.


Introduction
Pituitary adenomas (PAs) are one of the most frequently diagnosed central nervous system (CNS) tumors [1]. Surgery is a well-established first-line treatment for most symptomatic subtypes of PAs (aside from prolactinomas). In the last decade, the classical surgical approach (transsphenoidal microsurgery or transcranial craniotomy) has been displaced by the endonasal endoscopic transsphenoidal (EET) approach in most pituitary tumor centers.
The consistency of PAs has not been analyzed in-depth in recent literature [2][3][4]. Most PAs are considered to be soft in consistency, facilitating complete excision. However, there is a non-negligible percentage (5-13%) of tumors with a hard/fibrous consistency that, in theory, are more difficult to remove completely [5] and may be accompanied by greater morbidity and operative mortality [6,7]. To date, most studies have focused on trying to determine tumor consistency preoperatively by means of imaging studies with disparate and contradictory results [2,6,7]. However, almost no reports examined the effects of a fibrous consistency in surgical outcomes in PA surgery. Our objective was to analyze the effects of tumor consistency on surgical outcomes in a homogeneous series of patients with PAs operated on using the EET approach by the same neurosurgeon.

Material and methods
Patients An ambispective study in 226 consecutive surgical interventions (Table 1) at the Neurosurgery Departments of the Hospital Universitario Ramón y Cajal (HURC) and at the Hospital Universitario Puerta del Sur by the same neurosurgeon (VRB) was carried out.
All the clinical data were obtained from a database created for this purpose in 2012, including all patients who have been operated on since 2008 retrospectively and recruited prospectively from then until 2019. Patients included underwent surgery for a PA using an EET approach. All patients were followed for at least 6 months after surgery. PA consistency was reported during surgery by the same neurosurgeon (VRB). Patients with a diagnosis different from PA (Rathke cleft cyst, craniopharyngioma, and meningioma) and those with pituitary apoplexy were excluded.
A baseline hormonal study was performed in each patient with determinations of anterior pituitary hormones (corticotropin (ACTH), thyrotropin (TSH), prolactin (PRL), growth hormone (GH), and luteinizing and follicle stimulating hormones (LH and FSH)), as well as their peripheral hormones (free thyroxine (FT4), insulin-like growth factor I (IGF-1), cortisol, and testosterone in males and estradiol in females) before and after surgery.
Hormonal assessment included preoperative measurement of 8 a.m. serum cortisol, ACTH, TSH, FT4, PRL, FSH, LH, GH, IGF-1, and total and free testosterone in males. Preoperative morning serum cortisol <5 µg/dL or a peak serum cortisol <18 µg/dL (21 µg/dL in women under oral hormonal contraceptives) after 250 µg of cosyntropin administered intravenously was taken as evidence of ACTH  [10]. FSH/LH deficit was diagnosis based on hypothalamic amenorrhea in pre-menopausal women and low/ normal FSH and LH levels in postmenopausal women, and in the presence of testosterone levels below the reference ranges with normal or low gonadotropins in men. TSH deficit diagnosis was established when FT4 was below reference ranges and TSH in normal or low range. The diagnosis of GH deficit was performed in the presence of an IGF-1 level of -2 standard deviations for age and sex in the presence of three or more hormonal deficits. GH stimulation testing was not routinely performed. Panhypopituitarism was defined as complete anterior pituitary failure, with or without posterior pituitary involvement. Hyperprolactinemia was defined when prolactin levels were higher than 30 µg/L in women and 25 µg/L in men in the absence of dopaminergic therapy and other causes of hyperprolactinemia. No cases of diabetes insipidus were reported in the preoperative stage. The diagnosis of acromegaly, Cushing disease, TSH-secreting PA and prolactinomas was established following the current guidelines [11][12][13][14]. Serum cortisol, LH, FSH, prolactin, free and total testosterone, FT4, and TSH were measured by chemiluminescent immunoassay microparticle (ARCHITECT®), intraassay coefficient of variation (CV) was <10%. ACTH by chemiluminescence (LIAISON®), intraassay CV was <20%. GH and IGF-1 were measured by chemiluminescence assays. Before May 2013 they were measured with IMMULITE 2000®; between May 2013 and October 2018 with Isys (IDS Vitro®) and after then, by Liaison XL (DIASORIN®). The intraassay CV was <10% with all these methods. The assays were calibrated according to the WHO international standard for GH and IGF-1 with code 98/574 and 02/254, respectively.
In addition to hormonal assessment by the aforementioned tests, clinical signs and symptoms of hypopituitarism were evaluated by means of a standardized questionnaire and physical examination.
Patient follow-up consisted of daily clinical visits while in-hospital and then at 1-month, 3-months, 6-months, and every 12 months. However, this schedule could have been modified due to clinical criteria.
The patients were evaluated by an expert ophthalmologist before and after the intervention, performing formal visual field testing in most cases and studies with optical coherence tomography (OCT) since 2012. An otolaryngologist evaluated all patients prior to surgery [15].
The study was reviewed and approved by the local ethics committee of the HURC (approval date: October 4, 2019, Code: Acta 372).
Gross total resection (GTR) was defined as when the surgeon's intraoperative vision and postoperative imaging document no residual tumor. The resection was considered subtotal (STR), otherwise.

Surgical technique
All surgeries were performed by the senior author (VRB) through an EET approach and four-hand technique, with anterior sphenoidotomy and partial posterior septectomy. Nasosphenoidal stage, opening, and closure were done collaborating with an otolaryngologist.

Tumor consistency
The senior author described PA consistency in every case among the two groups, along with some of the most relevant studies describing PA consistency [2,6,7,16]. We divided the consistency into "soft" (n = 150), which includes easily aspirated tumors with conventional 8-10 french suction tip, and "fibrous" (n = 76), which implies that the tumor had to be fragmented previously with ring curettes and occasionally needing extracapsular dissection for its complete removal. Within this group, an additional "very hard" subgroup was reported (n = 21), requiring sharp instruments to debulk them and extracapsular dissection for their complete removal.
Following the novel proposed grading system for pituitary adenoma consistency by Rutkowski et al. [7], we group grades 1 and 3 ("freely suckable tumor; minimal curettage required", "requires some curettage or mechanical debulking") as soft adenomas and grades 4 through 5 as fibrous adenomas ("not suckable, curettage or mechanical debulking required"). Our "very hard" consistency group would be included within grades 4 and 5 ("extracapsular technique typically required", "sharp of on bloc removal required").

Complications
We have carefully reported surgical complications [17]. We grouped them into severe (life-threatening or nonreversible) (postoperative intracranial bleeds, meningitis, visual deterioration, and death) and minor (reversible or medically treatable) (presence of permanent diabetes insipidus [DI], loss of the anterior pituitary hormonal axis, postoperative cerebrospinal fluid leakage, and medical complications). We analyzed the complications after overcoming the learning curve (first 40 cases).

Statistical analysis
The statistical analysis was performed using Stata 16 (StataCorp. 2019). In the descriptive analysis, categorical variables were expressed as percentages (absolute values of variables), and quantitative variables were expressed as means (standard deviations [SDs]). The normality assumption was studied with the Shapiro-Wilk test. Levene's test was used to test for homogeneity of variances.
Student's t-test and multiple linear regression models were performed accordingly to compare differences in continuous parameters between two subgroups. The chisquared test was applied to compare categorical variables between independent samples. A logistic regression model was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs). In addition, multivariate analysis with logistic regression was performed, and adjusted models were chosen following best-adjustment criteria and those changing measures of association <10%. The significance level was set at p < 0.05.

Extent of resection and consistency
In the univariate analysis, fibrous adenomas showed a higher incidence of incomplete resections (OR 3.79, 95% CI 2.07-6.92, p < 0.001), with a GTR percentage among fibrous PAs of 45.8% vs. 76.2% among soft PAs (Table 2). To assess the effect of the fibrous consistency of the adenoma, we chose a logistic regression model that included the Knosp grade and its interaction with the consistency and the size of the adenoma (mm) (craniocaudal), as well as having received prior treatment. According to this multivariate-adjusted model, fibrous adenomas presented an OR of 8.22 (95% CI 1.93-35.12, p = 0.004) of having an incomplete resection. An additional subgroup analysis was done, grouping Knosp grades according to cavernous sinus (CS) invasion. Analyzing the subgroup of Knosp grades 0, I, and II (without CS invasion), an adjusted OR of 7.13 was obtained for incomplete resection in those fibrous adenomas (95% CI 1.53-33.25, p = 0. 012) ( Table 2; Fig. 1).
The effect of adenoma consistency was also analyzed in those adenomas classified as very hard (n = 20), resulting in an OR of 11.4 (95% CI 2.06-63.49, p = 0.005) for  incomplete resection, adjusting for the previous model ( Table 2).
Regarding the biochemical cure rate of functioning adenomas, there was a particular trend towards a higher number of no remissions in fibrous adenomas, with an OR of 2.68 (95% CI 0.97-7.46, p = 0.058). However, this difference was not maintained in the adjusted model of the analysis according to the Knosp grades (Table 2). In the subgroup of Knosp grades with no CS invasion, the seven fibrous adenomas (100%) had a biochemical cure; hence, no adjusted OR is presented in the corresponding table.
A subgroup of adenomas of intermediate size (macroadenomas excluding giant adenomas) greater than 10 mm and less than 40 mm in the craniocaudal axis was analyzed separately. The differences between the number of subtotal resections (total and grouping by CS invasion according to Knosp grade) and biochemical remissions were similar to those described above. An adjusted OR of 9.01 was obtained for incomplete resection in those fibrous adenomas (95% CI 1.65-49.27, p = 0. 011), and an adjusted OR of 11.4 was obtained for incomplete resection in those fibrous adenomas without CS invasion (95% CI 1.76-74.41, p = 0. 011) ( Table 2; Fig. 2).

Intraoperative variables and consistency
Several intraoperative variables and complications were evaluated after getting past the learning curve (after carrying out the first 40 cases).
There were 186 recorded operative times. Fibrous adenomas involved longer times (
Adjusting for adenoma size, there was no statistically significant association between a fibrous consistency and greater number of total and severe complications. No relationship was found in the crude or multivariate analysis regarding minor complications (including pneumonia, epistaxis, and syndrome of inappropriate antidiuretic hormone secretion).
There appeared to be a higher number of cases of postsurgical panhypopituitarism in patients with fibrous adenomas. The OR obtained in the univariate analysis was 15.12 (95% CI 1.51-108.11, p = 0.019) and 10.36 (95% CI 1.18-91.12, p = 0.035) in the adjusted model (Table 4; Fig. 3).
There was no relationship between consistency and transient DI. The excision of fibrous adenomas seems to imply a greater number of cases of permanent DI (Table 4; Fig. 3).  We analyzed the subgroup of adenomas without prior treatment and with preserved pituitary function (52 cases, 78.9% soft and 21.1% fibrous). Pituitary dysfunction occurred in 4.9% (2/41) of soft consistency adenomas and in 9.09% (1/11) of fibrous consistency adenomas. Fibrous PAs showed a higher incidence of transient DI, with an OR of 6.8 (95% CI 1.26-36.47, p = 0.026) obtained in the bestadjusted model. There were no cases of permanent DI in this subgroup of patients.
In this surgical series, a total of three cases of death were reported. Only one of them took place after getting through the learning curve (1/184, 0.54%). All three patients had giant fibrous adenomas, two of them with significant comorbidities. The other patient suffered from complete blindness with an aggressive recurrent pituitary tumor requiring multiple operations and reirradiation. Palliative surgery was considered, but further added complications made the patient's family reject aggressive treatments (CSF leak, fungal meningitis and various surgical interventions).

Clinical outcomes, posterior treatments and recurrence rates
First, we considered the crude analysis of postoperative changes after getting through the learning curve. We found a lower percentage of visual clinical improvement in fibrous adenomas (all those patients with altered visual fields at diagnosis were analyzed), 88.0% (22/25) vs. 100% (30) of improvements in soft adenomas, although this association was not statistically significant (p = 0.088). When patients without prior treatment were analyzed, the percentages maintained the trend, improving 85.7% (18/21) of patients with fibrous adenomas compared to 100% (28) of patients with soft adenomas, almost reaching a significant association (p = 0.07). Three cases of visual deterioration were found (3/62, 4.8%) in patients with fibrous adenomas compared to 0% in patients with soft adenomas; these cases of deterioration had an adenoma of intermediate size, between 20 mm and 35 mm in the different axes.
There were no significant differences in the improvement or worsening of headache symptoms, neither in the crude nor subgroup analysis. All patients with fibrous adenomas (100%, n = 10) had improved headache symptoms compared to 90% improvement in those with soft adenomas (18/20).
In the univariate analysis, patients with fibrous adenomas received a more significant number of treatments after surgery (OR 2.73, 95% CI 1.34-5.57, p = 0.006). The subgroup of those adenomas that had apparent complete resection was evaluated with the same regression model, with an association that lost statistical significance. We observed a higher frequency of medical treatments and radiotherapy after surgery in those with fibrous adenomas. In the multivariate analysis, adjusting for degree of resection, Knosp grade, and previous treatments, statistical significance was not reached, except for postoperative radiotherapy (OR 2.76, 95% CI 1.08-7.01, p = 0.033) ( Table 5). In the very hard PA group, 7/17 (41.2%) received treatment after surgery (Fisher's exact test, p = 0.04).
We found a small sample of 12 cases with recurrences. Of which, only five cases had not received any previous treatment. There was no evidence of association with tumor consistency, despite analyzing subgroups according to the degree of resection.

Discussion
There are several factors, such as tumor size or CS invasion (assessed through Knosp grade), that may alter the outcomes of EET surgery for PA [18][19][20]. Among them, some authors have also considered the possible influence of the consistency of the PA [2,4,6,21]. Nevertheless, there is no universally accepted classification for the consistency of adenomas [7,22], although most authors differentiate easily aspirated (soft) tumors from those that require prior fragmentation [2,4,23]. We have considered the group of easily aspirated (soft) vs. the rest (fibrous), as it is more easily identifiable and represents a clear difference in the excision technique (use of a conventional aspirator vs. the use of curettes or sharp instruments). However, we additionally differentiated a subgroup of adenomas with a very hard consistency, which required previous extracapsular dissection, fragmentation with a scalpel or the use of an ultrasonic aspirator [22]. According to this, our series contains 33.6% fibrous (non-suckable) adenomas, being comparable to that presented by certain authors (15-35%) [2,6,7,24]. It should be noted that some authors reported a lower prevalence of fibrous tumors (5-13.5%) [3,25], which would correspond with our obtained figure of 9.3% (21/226) (considering the subgroup of very hard PAs). These differences might be explained due to the variability in tumor consistency definition. As proposed by Rutkowski et al. [7], probably, some tumors included in grade 3 should be considered fibrous due to the difficulties in differentiating between the amount of curettage needed to distinguish soft from fibrous adenomas. Further validation is needed to determine if soft adenomas, which consistency might not influence surgical outcomes, are above or below that figure and if a higher amount of fibrous adenomas should be described due to the consequences that manipulation and debulking might have on surgical outcomes. Surgery on fibrous adenomas seems to have a lower number of complete resections. Different authors have reported similar findings [24,26]; in particular, Cappelletti et al. [2] reporting a GTR of 41.7% in fibrous adenomas compared with 76.2% in soft adenomas, our series shows similar results (GTR of 45.8% for fibrous adenomas vs. 76.2% of soft adenomas). In the subgroup of 21 adenomas of very hard consistency, the GTR was even lower, 25.0% compared to 70.3% in the rest of the series; this downward trend of GTR was also reported by Rutkowski et al. [7]. On the other side, Yamada et al. [27] reported no relationship between tumor consistency and degree of resection.
When analyzing the subgroup of macroadenomas (excluding microadenomas and giant PAs) and adjusting for Knosp grade, tumor size, and previous treatment, a lower number of complete resections was also observed in fibrous adenomas vs. soft adenomas (51.9% [27/52] vs. 76.6% [72/ 94], p = 0.011). This analysis was extended by stratification according to CS invasion, yielding an OR of 11.4 (95% CI 1.76-74.41, p = 0.011) for incomplete resections in those fibrous adenomas with the previously described conditions and Knosp grades between 0 and II. This subgroup analysis aimed to show the influence of consistency in those tumors where other possible confounding factors were eliminated, such as invasion of the CS [7,21,26], giant adenomas [28] or adenomas not visible on imaging tests [29]. Stratified analysis in PA with very hard consistency could not be reported due to very small number of cases.
In our series, the percentage of biochemical remissions for functioning fibrous adenomas tended to be lower than that found in soft adenomas, although the relationship was not statistically significant (Table 2). Only five functioning PA were found in the very hard group, no analyses were further conducted. No studies have been found reporting the association of consistency and biochemical remissions [30]. The biochemical cure rate is closely related to the degree of tumor resection [31]; therefore, fibrous functioning PAs might be associated with lower biochemical cure rates [32], especially in higher Knosp grades (invading CS) as reported above in the subgroup analysis (Table 2). We assessed the complications [33] of the 226 interventions, differentiating major and minor complications. The main surgeon learning curve factor, within the first three years of EET surgical experience, was evaluated, because it is considered that complications or intraoperative variables in the first 40 cases could be overestimated by this factor [34]. A multivariate analysis was performed adjusting for Knosp grade, previous treatments, and tumor size once the learning curve was through [35]. Generally, more total complications occurred in fibrous adenoma surgeries, although this relationship lost statistical significance in the multivariate analysis (OR 1.42, 95% CI 0.74-2.75, p = 0.294). It is not surprising that such a disparate group was not directly related to the consistency of the PA.
If only major complications were considered, interventions on fibrous adenomas seem to have a more significant number of complications according to the univariate model (Table 4), despite having passed the learning curve. This may be mainly due to the fact that surgical removal in these tumors is more complex, requiring greater glandular manipulation [6] and, in many cases, an extracapsular dissection [36]. However, statistical significance disappeared in the multivariate model. Considering PA with very hard consistency, only one major complication was encountered. No further analysis was performed. There was no evidence of a relationship between minor complications and tumor consistency [6].
The fibrous consistency implies more aggressive excision and dissection techniques and could lead to the appearance of a greater number of complications, as described by other authors [6,7,32]. However, tumor size (craniocaudal axis) indeed seems to be the most determinant factor according to our adjusted model [37,38] in the analysis of pooled complications. An important association was seen in a greater number of intraoperative CSF leaks in surgeries for fibrous adenomas (Table 4). Adjusting for tumor size and previous treatments, this difference persisted (OR 5.23, 95% CI 2.52-10.87, p = 0.001). We did not find this relationship reported in the literature review, even though some authors have defined different factors (surgical, anatomical, etc.) predicting the opening of the sellar diaphragm [39]. This is justified by the difficulty in handling these lesions and the need to perform an extensive extracapsular dissection. However, even once the learning curve has been completed, this does not translate into a greater number of postoperative CSF leaks (where a surgical closure technique is also involved), as mentioned by other authors [7].
In the present study, there seem to be more cases of postsurgical panhypopituitarism in patients with fibrous PAs (Table 4). Although our sample size was small (eight cases), some authors have reported similar trends [6,7]. We did not find a statistical association between transient DI and adenoma consistency. Fibrous adenomas were associated with higher rates of persistent DI ( Table 4). The apparent higher frequency of persistent DI in fibrous adenomas could be due to greater manipulation by more aggressive dissection techniques, a higher incidence of diaphragm openings, or the interaction between the two. Few authors have specifically analyzed the influence of adenoma consistency and postoperative hormonal complications [6,7,40]. Some authors have reported a relationship between the size of the adenoma and the incidence of postoperative hormonal deficits [41,42] although not taking into account PA consistency or the learning curve.
The mean surgical time in our series is comparable to others [43]. Fibrous PA surgery lasted longer, with a mean of about 42 min, despite completion of the learning curve [44]. This association maintained its statistical significance in the multivariate analysis (Table 3), lengthening the surgical time 22.5 min (p = 0.014). It is noteworthy that, despite adjusting for various factors and having overcome the learning curve, the fibrous consistency of the adenoma continued to be a factor to take into account in the surgical act. Very hard PA also lasted longer, with a mean surgical time of 28.9 min higher compared to non-very hard PA (p = 0.04). This longer duration of surgery could have implications for surgical results and increased anesthetic risk; a recent meta-analysis showed a 14% more probability of complications for every 30 min of additional prolongation [45].
It cannot be concluded that the hospitalization time varies directly, depending on the consistency. A trend towards a longer hospitalization time can be identified in those patients with fibrous adenomas, despite the apparent decrease in the time of hospitalization concerning the learning curve [44]. This trend could be influenced by the more significant number of postoperative complications associated with fibrous tumors (Table 4), which sometimes require more observation or in-hospital care. In the multiple linear regression analysis, the presence of complications had a significant influence (p = 0.001), being an independent predictor, lengthening the hospitalization time 3.7 days. Asemota et al. reported the incremental costs and the increase in hospitalization time in those patients who presented with complications after surgery although without describing tumor characteristics [17].
Regarding the prognosis and postoperative changes [46], we did not find significant differences in the improvement/ worsening of visual symptoms due to the small sample size. Despite this, three cases of visual impairment in ours study had a fibrous consistency and intermediate size, between 20 mm and 35 mm in the different axes. Different authors [47,48] consider the size of the adenoma was a very determinant factor of the possibility of both worsening and improvement of visual symptoms. Therefore, it cannot be ruled out that the consistency of the adenoma was not related to a worsening of visual symptoms in these patients. Thomé and Zevgaridis [49] propose a theory that vascular compromise of the optic pathway could be related to impaired vision in certain patients. In addition to what they comment on, this vascular compromise could be due to excessive manipulation with the fibrous consistency of the adenoma.
We found no studies analyzing the relationship of consistency with clinical prognosis.
According to the univariate analysis, patients with fibrous adenomas received more postoperative treatments (OR 2.76, p = 0.004). In the univariate analysis of the subgroup of fibrous adenomas with apparent complete resection, the OR was 3.26 (95% CI 0.88-12.11, p = 0.078). Patients with very hard PA received more subsequent treatments (P = 0.04), although most of those cases were partially resected. These results are likely related to fibrous adenomas and their tumor size, although the bestadjusted multivariate model to measure effect only included Knosp grade and prior treatment.
No significant differences were obtained in terms of the number of recurrences or elapsed time until the recurrence occurred. Although the sample size of recurrences was small (n = 14), it cannot be ruled out that the fibrous consistency was associated with a higher recurrence rate. No studies have been found in this regard, with most only analyzing the degree of invasion and tumor size [50].
This study represents the results of a nonrandomized design, implying certain limitations. All patients included in the study were operated on with the same surgical technique and prospectively classified by the same neurosurgeon. There is no universally accepted classification of PA consistency, so there is an inherent difficulty when comparing data between different reports. Adopting a standard for classification/definition of adenoma consistency will make it easier to compare surgical series. Moreover, several confounding factors mentioned above could bias our results. The limited number of patients is another limiting factor, specially within the very hard consistency group. More cases will allow statistical analyses with greater power and narrower CIs. Ideally, a multicenter prospective study would yield more scientific evidence in this regard.

Conclusion
In conclusion, fibrous PAs are independently associated with lower resection rates, even in those without CS invasion (Knosp grades 0, I and II). There is more intraoperative CSF leakage in fibrous PA surgeries, and there appears to be an association between this consistency and more cases of postoperative hypopituitarism, permanent DI and the need for subsequent treatments. Removal of a fibrous PA requires a longer surgical time, regardless of other factors. Therefore, the consistency of the PAs could significantly condition the surgery and the patient's prognosis. Identifying these challenging fibrous adenomas preoperatively could encourage centralizing these high complexity tumors in reference centers.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Ethical approval The studies involving human participants were reviewed and approved by the Ethical Committee of the Ramoń y Cajal Hospital.
Informed consent The patients/participants provided their written informed consent to participate in this study.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.