Identifying hormones and other perioperative risk factors for postoperative delirium after endoscope‐assisted transsphenoidal pituitary adenoma resection: A retrospective, matched cohort study

Abstract Objective As a complex and acute brain dysfunction, if postoperative delirium (POD) occurs in the postoperative period, it will lead to a prolonged length of stay in the critical care unit, with increased hospitalization costs and higher mortality. A few case reports inspired us to pay close attention to pituitary tumor‐associated delirium. We hypothesized that the changes in hormone levels after pituitary tumor resection might be associated with POD occurrence. Methods Retrospective analysis was performed on data from a single‐center cohort study conducted at Southwest Hospital between January 2018 and May 2022. A total of 360 patients with pituitary tumors who underwent endoscope‐assisted transsphenoidal pituitary tumor resection were divided into two groups at a 1:3 ratio, with 36 patients in the POD group and 108 patients in the non‐POD group matched by propensity score, age, sex, and tumor size. Basic characteristics, pituitary adenoma features, endocrine levels and other biochemical indicators, and Confusion Assessment Method for the Intensive Care Unit (CAM‐ICU) for postoperative delirium were documented for further analysis. Results Lower insulin‐like growth factor‐1 (IGF‐1, p = .024) and corticotropin‐releasing hormone (CRH, p = .005) levels were closely associated with postoperative delirium and with high levels of blood glucose (GLU, p = .023) after surgery. Subsequent analysis indicated that serum potassium (OR: 0.311, 95% CI 0.103–0.935), sodium (OR: 0.991, 95% CI 0.983–1.000), CRH (OR: 0.964, 95% CI 0.936–0.994), and GLU (OR: 1.654, 95% CI 1.137–2.406) levels in the perioperative period were independent risk factors for delirium. Conclusions Our study indicated that lower serum CRH, potassium, sodium, and GLU levels may be associated with the occurrence of POD after endoscopic‐assisted transsphenoidal surgery. These data provide preliminary evidence for the management of POD in pituitary adenoma patients after surgery. Further studies are needed to identify pharmacological and nonpharmacological multicomponent treatment strategies.


INTRODUCTION
It has been reported that pituitary adenoma is the second most prevalent primary tumor in the brain parenchyma. In accordance with the progress of neuroendoscopy technology, endoscope-assisted transsphenoidal surgery for pituitary tumors and other basicranial tumors has led to more radical excision with fewer complications. A recent meta-analysis showed that the most frequently reported postoperative complications included cerebrospinal fluid leakage (12.8%), diabetes insipidus (12.8%), postoperative hypopituitarism (9.2%), and secondary hematoma (8.5%) (Stefanidis et al., 2022). However, after resection of either functional pituitary adenomas or nonfunctional tumors, endocrine complications are still one of the important factors for patients in neurosurgical intensive care units (NICUs) during the early postoperative period (Keandoungchun et al., 2021). This hormonal status might lead to complex neurological complications, including postoperative delirium, in the NICU (Cruz-Flores, 2021), which has been reported to be tightly associated with worse cognitive outcomes, including short-term and long-term cognitive decline after surgery (Goldberg et al., 2020;Huang et al., 2021).
As a complex and acute brain dysfunction, postoperative delirium is characterized by acute and fluctuating changes in cognition and consciousness. Once delirium appears in the postoperative period, it will lead to a prolonged length of stay in the NICU, increased hospitalization costs and higher mortality; therefore, it is very important to recognize and explore the risk factors for postoperative delirium in patients. To date, only a few case reports have inspired us to pay close attention to pituitary tumor-associated delirium (Burne et al., 2021;Li et al., 2017;Weng et al., 2008). Even idiopathic pituitary insufficiency could present as prolonged delirium and a "terrible dream" (McAulay-Powell & Friedman, 2013). Based on these clues, we hypothesized that the changes in hormone levels after pituitary tumor resection might be associated with postoperative delirium occurrence.
We enrolled cohorts of patients who underwent endoscope-assisted transsphenoidal pituitary adenoma resection to analyze the potential perioperative risk factors for postoperative delirium in our NICU ward.

Delirium diagnosis
The diagnostic criteria for postoperative delirium according to The The patients were routinely evaluated twice a day at 9:00 am and 21:00 pm and additionally screened if the patients exhibited signs of altered mental status. After the patients were transferred out of the NICU, nurses performed the same assessment.

Cohort matching
The CAM-ICU delirium evaluation showed that 36 patients were positive and 312 patients were negative for delirium. To manage the common characteristics and variates of enrolled patients, we performed propensity score matching for age, sex, and tumor size using a 1:3 ratio. Then, the cohort was divided into two groups according to the presence or absence of postoperative delirium, with 36 patients in the postoperative delirium group (POD group) and 108 patients in the non-POD group. The flow diagram of cohort matching is also illustrated in Figure 1.

Data extraction
We retrieved the characteristics and clinical data of those matched cohorts via medical records, categorized as (1) basic characteristics,

Statistical analysis
The 1:3 propensity score match was performed by R version R 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria) on the enrolled subjects via the nearest-neighbor method without replacement, and the caliper value was set as 0.2 ( Figure S1).

Patient characteristics
The difference in tumor size between the delirium and nondelirium groups could not be ignored before propensity score matching (PSM), and the standard mean difference (SMD) was 0.249. However, there was an obvious decrease in the SMD value and no significant differences in age, female sex, or tumor size between the two matched groups (SMD = 0.076, 0.019, 0.063, Table 1 and Figure S1). In addition, after PSM, the comparisons of baseline data, including adenoma classification, hospital days, and operation times, were not significantly different (p = .604, 0.190, 0.729, Table 2).

Prognosis and characteristics of postoperative patients with delirium
The neural prognosis of postoperative pituitary adenoma patients was assessed by the modified Rankin scale (mRS) score at discharge, and all the enrolled subjects scored 0 or 1. Thus, we regarded patients with a mRS score = 1 as having a poor prognosis. A larger proportion of postoperative delirium patients had a poor prognosis at discharge than those without delirium (48.22% vs. 0.93%, χ 2 = 48.762, p < .0001, Figure 2A). In addition, CAM-ICU was used to determine the occurrence of delirium, while the ICDSC and RASS were used to evaluate the severity of delirium at five timepoints after surgery (i.e., within 6 h,  Logistic regression was used for univariate and multivariate analysis to establish a stepwise regression analysis model.
ICDSC scores between the two groups with or without poor prognosis exhibited no significant difference (median ± IQR, 6.00 ± 2.00 vs. 7.00 ± 3.00, Z = −0.632, p = .528, Figure 2B). Finally, we found a rapid increase in the percentage of delirium patients (CAM-ICU score = 1) and the proportion of patients with a high RASS score at 24 h after surgery during the hospital stay (χ 2 = 64.980, p < .0001, Figure 2C and D). Univariate and multivariate analyses confirmed that delirium was an independent risk factor for poor neural prognosis in postop-
23.03 ± 18.74, p = .005). Conversely, the GLU level in the blood of the delirium group was higher than that in the nondelirium group 72 h after surgery (7.26 ± 2.79 vs. 5.64 ± 1.23, p = .023). The two-group comparisons of other variables showed no significant difference (p > .05,

DISCUSSION
We investigated the potential perioperative risk factors for postoperative delirium in pituitary adenoma patients treated by endoscopeassisted transsphenoidal surgery. We found that lower IGF-1 and CRH levels were tightly associated with postdelirium and high levels of blood glucose (GLU) after surgery. Logistic regression analysis indicated that serum potassium and sodium levels and blood glucose levels in the perioperative period were independent risk factors for delirium.
Postoperative delirium is one of the most severe complications after surgery, with an acute confusion status characterized by inattention, cognitive dysfunction and an altered consciousness. It has been reported that delirium is an independent risk factor for the subsequent development of dementia (Richardson et al., 2021), and delirium in dementia patients can accelerate cognitive decline (Davis et al., 2017;Goldberg et al., 2020). Most patients in our cohort were approximately 50 years old and at high risk of developing cognitive impairment.
Despite pituitary adenomas, high-grade glioma was also reported to increase the risk of postoperative delirium, with insufficient compensation for injured brain regions involving cognition (Huang et al., 2022).
Therefore, predicting, identifying and intervening to minimize the risk of delirium might reduce or prevent long-term cognitive impairment (Fong & Inouye, 2022;Kong et al., 2022;Myles, 2020). In addition to the CAM-ICU and ICHSC we used in this study (Gusmao-Flores et al., 2012; von Hofen-Hohloch et al., 2020), regional cerebral oxygen saturation may also be helpful for identifying the risk of delirium in postsurgery patients (Mutch et al., 2020;Susano et al., 2021).
Perioperative hydrocortisone has been widely used for pituitary adenoma patients to avoid postoperative hypopituitarism due to compression of the pituitary gland by the tumor and intraoperative injury to the pituitary (Higham et al., 2016;Husebye et al., 2021). Therefore, perioperative glucocorticoid supplementation has long been a routine practice for patients with pituitary adenomas (Hattori et al., 2021;Molitch, 2017). A recent randomized clinical trial indicated that withholding hydrocortisone was safe and noninferior to the conventional hydrocortisone supplementation regimen regarding the incidence of new-onset adrenal insufficiency among patients with an intact hypothalamic-pituitary-adrenal axis undergoing pituitary adenomectomy (Guo et al., 2022). However, researchers did not take postoperative delirium into consideration in their study. Postoperative delirium is prevalent in patients after elective intracranial surgery and is associated with adverse outcomes and high cost .
However, the association between hormone levels and delirium is still controversial. Only a few case reports have inspired us to pay close attention to hormone-associated delirium (Burne et al., 2021;Li et al., 2017;Weng et al., 2008). In the present study, our data also indicate that lower IGF-1 and CRH levels are tightly associated with postoperative delirium. Consistent with previous studies, low levels of IGF-1 and high levels of GH are independently associated with the occurrence of delirium Li et al., 2018). Inversely, Chu et al. (2016) reported that serum IGF-1 level was nonspecific for predicting POD onset, and IGF-1 level alteration might be regarded as a disease biomarker rather than a risk marker. However, Chu et al. (2016) also noted that the results were influenced by factors such as the time of blood collection, basic medical disease, and sample size, which indicated that different study designs and data can lead to different conclusions. In this study, we used PSM to reduce the bias of age, gender and tumor size, attempting to provide more scientific evidence.
In parallel, higher preoperative cortisol levels in cerebrospinal fluid are not associated with postoperative delirium in elderly hip fracture patients (Witlox et al., 2020). Compared to patients with nonfunctioning tumors, GH-secreting tumor patients exhibit a higher incidence of sleep disturbance but not typical delirium symptoms (Kim et al., 2020).
With growing evidence to support the withholding of hydrocortisone management for postoperative pituitary adenoma patients, normalized preventions, and interventions for delirium are needed in further studies and clinical practice (Scicutella, 2020).
To date, no pharmacologic drugs have been approved to specifically treat postoperative delirium (Scicutella, 2020). Haloperidol, dexmedetomidine, statin, and ketamine are not recommended for the prevention of delirium in all critically ill adults in current guidelines, and the evidence is low quality (Devlin et al., 2018). Recently, dexmedetomidine seems beneficial for attenuating neuroinflammation and likely to reduce the incidence of delirium, but it is limited to application in ICU practice and surgical patients (Swarbrick & Partridge, 2022), especially those who are mechanically ventilated  more high-quality evidence regarding the prediction, identification, and management of postoperative delirium is urgently needed.
At present, 95% of pituitary adenoma patients are treated by endoscope-assisted transsphenoidal surgery, which introduces much less iatrogenic damage than transfrontal craniotomy or radiotherapy (Goudakos et al., 2011;Kanat et al., 2022;Yadav et al., 2012 to affirm whether endoscope-assisted transsphenoidal surgery contributes to the occurrence of POD. Regrettably, we did not enroll patients who had undergone other surgery approaches, which needs to be further studied in the future.

CONCLUSION
In summary, our study indicated that lower IGF-1 and CRH levels, as well as serum potassium and sodium levels and blood glucose levels, were potentially associated with postoperative delirium in pituitary adenoma patients treated by endoscope-assisted transsphenoidal surgery. These data might provide preliminary evidence for the management of postoperative delirium in pituitary adenoma patients after surgery, but further studies on pharmacological and nonpharmacological multicomponent interventions are needed.

CONFLICT OF INTEREST STATEMENT
The authors declare that they have no competing interests.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

HUMAN SUBJECTS
Informed consent to participate was waived by the Ethics Committee of Southwest Hospital of Army Medical University due to the observational nature of the study.