Introduction

Nonfunctioning pituitary adenoma (NFPA) is a benign tumor of the pituitary gland characterized by a lack of excessive secretion of pituitary hormones [6]. Endoscopic transsphenoidal pituitary surgery has been established as the treatment for pituitary adenoma, and is associated with a high gross total resection rate and low complication rate due to improved visualization [8, 15]. However, manipulation of the pituitary gland and neuroendocrine stress response can affect postoperative pituitary hormonal status in NFPA patients undergoing endoscopic transsphenoidal surgery [2, 38].

Postoperative hypocortisolism is a common complication of endoscopic transsphenoidal surgery in NFPA patients [3]. Hypocortisolism patients exhibit symptoms such as profound fatigue, weight loss, postural hypotension, and muscle and abdominal pain [24]. Regarding health-related quality of life, general health and vitality are reduced in hypocortisolism patients [13, 21]. Furthermore, hypocortisolism is a potentially serious clinical condition because hypocortisolism patients have an approximately 50% increased risk of adrenal crisis, which leads to hypotension, shock, and even death [31]. In addition, patients with hypocortisolism require steroid replacement, which is associated with various complications such as weight gain, glucose intolerance, and osteoporotic fracture [27]. Therefore, identifying risk factors for postoperative hypocortisolism after transsphenoidal surgery in NFPA patients is of clinical importance. In previous studies, 13–20% of NFPA patients with preoperative normal hypothalamic–pituitary–adrenal (HPA) axis newly developed hypocortisolism after surgery [15, 25]. Although the tumor size, preoperative urinary-free cortisol level, and morning serum cortisol level on postoperative day 2 were identified as predictors of postoperative hypocortisolism in patients with pituitary/peripituitary tumors in a few previous studies [20, 29], there is a lack of evidence regarding predictors of postoperative hypocortisolism in NFPA patients undergoing endoscopic transsphenoidal surgery, especially those with preoperative normal HPA axis.

Anesthetic technique affects perioperative neuroendocrine stress hormones, such as serum ACTH and cortisol, by modulating the neuroendocrine stress response and sympathetic tone. In previous studies, when the effects of inhalation and intravenous anesthesia on perioperative neuroendocrine stress hormones were compared in various surgical patients, the former was associated with higher intraoperative and early postoperative ACTH and cortisol levels [18, 22, 38]. Regarding the effects of anesthetics on intermediate-term postoperative neuroendocrine functional outcome, only one previous study conducted by Dr. Oh and co-workers showed no differences in 3-month postoperative neuroendocrine functional outcomes between inhalation and intravenous anesthetic techniques in NFPA patients. [25]. However, their study included inhomogeneous subjects with respect to preoperative HPA status. In other words, both patients with preoperative normal HPA axis and those with preoperative hypocortisolism were included in their study. In clinical practice, clinicians have an interest in newly developed postoperative hypocortisolism rather than persisting postoperative hypocortisolism in NFPA surgical patients. Therefore, the effects of inhalation and intravenous anesthetics on intermediate-term postoperative ACTH functional outcome are needed to be reinvestigated in NFPA patients with preoperative normal HPA axis.

In this retrospective study, we determined whether the sevoflurane-based inhalation anesthesia was predictive of 3-month postoperative hypocortisolism in patients undergoing endoscopic transsphenoidal surgery for NFPA with preoperative normal HPA axis.

Methods

Subjects

This retrospective single-center study was conducted on adult NFPA patients with preoperative normal HPA axis who underwent tumor removal surgery via the endoscopic transsphenoidal approach at Seoul National University Hospital, between March 2010 and December 2020. Sixty-three patients included in our previous study were again included, in which we investigated a difference in the neuroendocrine stress response between sevoflurane anesthesia and propofol anesthesia in patients undergoing endoscopic transsphenoidal tumor removal surgery between May 2017 and 2018 [38]. Patients with missing laboratory data on serum ACTH or cortisol were excluded from the study. Prior to data collection, this study was approved by the Institutional Review Board of Seoul National University Hospital (number: H-2110–126-1264, date: 27 Oct 2021). The requirement for written informed consent was waived because this was a retrospective study.

Data collection

Demographic data (age, sex, body mass index), American Society of Anesthesiologists physical status, comorbidities, previous treatment history (surgery, gamma-knife surgery, radiotherapy), laboratory data (preoperative and postoperative serum levels of pituitary hormones), oncologic data (tumor size, immunohistochemical staining, cavernous sinus invasion, optic chiasm compression), surgical data (duration and degree of resection), anesthetic data (duration of sevoflurane-based inhalation anesthesia and propofol-based intravenous anesthesia, total time of mean blood pressure [MBP] < 65 mmHg, total area under MBP < 65 mmHg, time-weighted average of MBP < 65 mmHg [total area under MBP < 65 mmHg/anesthetic time], and intraoperative arterial partial pressure of oxygen and carbon dioxide), postoperative complications (diabetes insipidus [DI], hyponatremia, infection, cerebrospinal fluid leakage, hemorrhage, and hydrocephalus), and postoperative steroid administration were respectively collected from electronic medical records. Pituitary hormones, such as ACTH and cortisol, were measured within 1 month preoperatively and at 3 months postoperatively.

Propensity score matching

Patients were divided into two groups based upon intraoperative anesthetic technique: inhalational and intravenous anesthesia group. Propensity scores were matched with a ratio of 1:1 to exclude the effects of confounders, which differed significantly between the sevoflurane-based inhalational and propofol-based intravenous anesthesia group, on 3-month postoperative hypocortisolism.

Anesthetic management

The main anesthetic agent was chosen at the discretion of attending anesthesiologists without special considerations. In all patients, target-controlled infusion of remifentanil was used during anesthetic maintenance. All patients were mechanically ventilated and managed to maintain arterial oxygen tension ≥ 100 mmHg and arterial carbon dioxide tension of 36–44 mmHg. Also, all patients were subjected to invasive blood pressure monitoring and fluids and vasoactive drugs were administrated to maintain mean arterial pressure within ± 20% of preoperative value. Intravenous fentanyl 100 µg was administrated for pain control at the end of surgery. Most patients were taken immediate postoperative brain CT and transferred to the neurointensive care unit without emergence and extubation. There was no difference in patient management in the intensive care unit or postoperative recovery unit between sevoflurane-based inhalation anesthesia group and propofol-based intravenous group.

Perioperative steroid management

Among the patients, 84% of the patients received 100 mg intravenous hydrocortisone in the morning and 50 mg intravenous hydrocortisone twice postoperatively on the day of surgery. On postoperative days 1 and 2, 50 and 25 mg of intravenous hydrocortisone, respectively, were administered twice daily. Postoperative morning cortisol levels were measured on postoperative days 2, 7, and 14 and postoperative 1 month, respectively. Oral hydrocortisone was prescribed to patients with morning cortisol < 8 µg/dL in each blood test. Postoperative hyponatremia was assessed on postoperative days 7 and 14, respectively. Oral hydrocortisone was given to patients with postoperative hyponatremia.

Definition and outcome

Hypocortisolism was defined as serum cortisol < 18 µg/dL in the insulin-induced hypoglycemia test, peak cortisol < 18 µg/dL in the short ACTH test, morning cortisol < 8 µg/dL, or random cortisol < 5 µg/dL with a low to normal serum ACTH level.

The primary outcome was the incidence of 3-month postoperative hypocortisolism.

Sample size calculation

In two previous studies of NFPA patients with preoperative normal HPA axis undergoing endoscopic transsphenoidal tumor removal surgery, the incidence of 3-month postoperative hypocortisolism was approximately 20% [15, 19]. To reproduce this proportion with a 99% confidence interval (CI) and margin of error of 0.05, at least 425 patients were required for this study.

Statistical analysis

Statistical analyses were conducted using SPSS software (version 25.0; IBM Corp., Armonk, NY, USA). Categorical variables are presented as the number of patients (proportions) and continuous variables as the mean (standard deviation) or median (interquartile range). The chi-squared test or Fisher’s exact test was used to compare categorical variables. For continuous variables, the Shapiro–Wilk test was first used to determine the normality of the data distribution, and Student’s t-test and the Mann–Whitney U test were used to compare normally distributed and skewed variables, respectively. Variables with a P value < 0.075 in univariate logistic regression analysis, and well-known factors for postoperative hypocortisolism were included in the multivariate logistic regression to identify risk factors for postoperative hypocortisolism. A P value < 0.05 was considered statistically significant.

Results

A total of 432 NFPA patients with preoperative normal HPA axis underwent endoscopic transsphenoidal tumor removal surgery during the study period; 3 patients were excluded due to missing data for preoperative ACTH or cortisol level. Ultimately, 429 patients were included in the data analysis.

The demographic and perioperative data are summarized in Table 1. A total of 74 and 355 patients received intraoperative sevoflurane-based inhalational and propofol-based intravenous anesthesia, respectively. Six variables (age, preoperative steroid administration, tumor volume, immunohistochemical staining, optic chiasm compression, and cavernous sinus invasion) were matched with a ratio of 1:1 in the two groups using propensity score matching. After checking that there was no variable showing a significant imbalance between two groups (P = 0.695 in overall balance test), 73 patients were chosen in each group and included in data analysis. After propensity score matching, there were no significant differences in demographics, oncologic and perioperative data between the sevoflurane-based inhalation and propofol-based intravenous anesthesia group.

Table 1 Comparisons of demographic and perioperative data between two anesthetic techniques
Full size table

Three-month postoperative hypocortisolism was observed in 69 (16.1%) and 25 (17.1%) patients before and after propensity score matching, respectively. The sevoflurane group had a higher incidence of postoperative hypocortisolism before (n = 20 [27.0%] vs. n = 49 [13.8%], P = 0.008) and after (n = 20 [27.4%] vs. n = 5 [6.8%], P = 0.002) propensity score matching, respectively.

Patients with postoperative hypocortisolism had a lower portion of patients with body mass index (BMI) ≥ 25 (n = 6 [24.0%] vs. n = 60 [49.6%], P = 0.034) and higher proportion of musculoskeletal disease (n = 4 [16.0%] vs. n = 5 [4.1%], P = 0.047), compared with those with postoperative normal HPA axis Table 2. In addition, patients with postoperative hypocortisolism had a larger tumor volume (12.0 [6.0–16.0] vs. 6.8 [4.4–10.8] cm3, P = 0.019) and longer surgical duration (120.0 [92.5–159.5] vs. 95.0 [75.0–130.0] min, P = 0.044). Also, patients with postoperative hypocortisolism had a higher proportion of postoperative permanent DI (5 [20.0%] vs. 8 [6.6%], P = 0.048) and postoperative steroid administration (11 [44.0%] vs. 24 [19.8%], P = 0.020) than those with postoperative normal HPA axis.

Table 2 Comparisons of demographic and perioperative data between patients with 3-month postoperative hypocortisolism and those with 3-month postoperative normal hypothalamic–pituitary–adrenal (HPA) axis
Full size table

The results of univariate and multivariate logistic regression analyses are shown in Table 3. The sevoflurane-based inhalation anesthetic technique (5.37 [1.80–15.98], P = 0.003) and postoperative steroid administration (2.89 [1.06–7.92], P = 0.039) were significant predictors of postoperative hypocortisolism.

Table 3 Predictive factors for 3-month postoperative hypocortisolism in patients undergoing endoscopic transsphenoidal surgery for non-functional pituitary adenomas with preoperative normal hypothalamic–pituitary–adrenal axis
Full size table

Discussion

Postoperative hypocortisolism in patients undergoing transsphenoidal surgery for NFPA is a major concern for both neurosurgeons and neuroanesthesiologists. Postoperative hypocortisolism is a potentially life-threatening condition associated with significantly increased mortality and poor quality of life in NFPA patients [9, 26]. Thus, identification of predictive factors for 3-month postoperative hypocortisolism is clinically relevant in NFPA patients undergoing endoscopic transsphenoidal surgery with preoperative normal HPA axis.

The present study showed that the sevoflurane anesthetic technique was a significant predictor of 3-month postoperative hypocortisolism in NFPA patients with preoperative normal HPA axis. Sevoflurane anesthesia had a 5.4-fold higher incidence of postoperative hypocortisolism than propofol anesthesia. By contrast, patients who received propofol anesthesia more frequently suffered from postoperative permanent diabetes insipidus (DI). Postoperative permanent DI is known to be associated with irreversible hypothalamus and/or pituitary stalk injuries during surgery [5]. Thus, postoperative permanent DI can be associated with postoperative hypocortisolism. Indeed, our result showed that postoperative permanent DI was significantly associated with postoperative hypocortisolism in univariate logistic regression analysis, although it was not significant in multivariate analysis. Taken together, our findings suggest that propofol seems to have a significant protective effect on long-term steroid dependency. Unfortunately, we were not able to explain the exact pathophysiological mechanism on this finding that the functional outcomes of ACTH at 3 months postoperatively differed between the two different anesthetic techniques. In part, this finding can be explained by the difference in anti-inflammatory and antioxidant properties between propofol and sevoflurane. In a previous study, propofol had greater anti-inflammatory effects than sevoflurane in patients undergoing craniotomy [23]. Propofol exerts anti-inflammatory effects by upregulating annexin A1 (a regulatory protein of pro-inflammatory mediators) and inhibiting the release of inflammatory factors [28, 36]. Greater susceptibility to inflammation is associated with impaired HPA axis function at the pituitary gland level [35, 41]. The anti-inflammatory effects of propofol might contribute to preservation of the function of the HPA axis, leading to a lower rate of 3-month postoperative hypocortisolism in intravenous anesthesia group patients. In addition, propofol, compared with sevoflurane, has greater or similar antioxidant effects [12, 14, 17]. In a previous experimental study, antioxidant-induced activation of the HPA axis via downregulation of glucocorticoid receptors in the pituitary gland was reported [30]. Thus, the antioxidant effects of propofol might reduce the incidence of 3-month postoperative hypocortisolism by activating the HPA axis.

In a previous study, no significant difference in pituitary hormonal outcomes at 3 months postoperatively in NFPA patients was observed between sevoflurane-based inhalation and propofol-based intravenous anesthetic techniques [25]. By contrast, this study showed a relevant association between inhalation anesthesia and 3-month postoperative hypocortisolism. This discrepancy may be due to differences in the preoperative ACTH functional status. In the previous study, NFPA patients with preoperative hypocortisolism and normal HPA axis (20% and 80%, respectively) were included; however, in the present study, only NFPA patients with preoperative normal HPA axis were evaluated. Patients with preoperative hypocortisolism tends to have persistent postoperative hypocortisolism. Notably, 37–50% of NFPA patients with preoperative hypocortisolism undergoing endoscopic transsphenoidal surgery show 3-month postoperative hypocortisolism due to impaired HPA axis function [15, 25]. Based on the results of this study, propofol-based intravenous anesthesia may be more beneficial than sevoflurane-based inhalation anesthesia in reducing the rate of 3-month postoperative hypocortisolism in NFPA patients with preoperative normal HPA axis undergoing endoscopic transsphenoidal surgery. However, to verify the beneficial effects of propofol anesthesia on postoperative hypocortisolism, a large-scale prospective study is necessary.

In the present study, postoperative steroid administration was also predictive of 3-month postoperative hypocortisolism with an odds ratio of 2.9. Postoperative steroid was administrated to patients with postoperative morning cortisol < 8 µg/dL. A low level of early postoperative morning cortisol is known to be associated with 3-month postoperative hypocortisolism. A previous study showed that 2-day postoperative morning cortisol level was a significant predictor of postoperative hypocortisolism in pituitary adenoma patients undergoing transsphenoidal resection [29]. Postoperative hyponatremia is caused by transient endocrine function disorder such as syndrome of inadequate secretion of antidiuretic hormone (ADH) [40]. Glucocorticoid deficiency can result in excessive secretion of ADH by glucocorticoid negative feedback [32]. In this study, patients with postoperative hyponatremia were treated with oral hydrocortisone. A previous study demonstrated that postoperative hyponatremia was associated with postoperative hypocortisolism in pituitary adenoma patients [11]. Taken together, NFPA patients who require steroid supplements after transsphenoidal surgery may be at risk of postoperative hypocortisolism.

Age was a significant predictor of 3-month postoperative hypocortisolism before propensity score matching in this study. Old age is associated with changes of endocrine functions [39]. In the elderly, decreased functional activity of the HPA axis, increased adrenal sensitivity to ACTH, and reduced levels of cortisol regulatory hormones (e.g., dehydroepiandrosterone) result in dysregulation of cortisol and thus a high serum cortisol level [1, 10]. In addition, the duration of stress response induced by surgery is likely to be prolonged in older patients [37]. A sustained high serum cortisol level induces HPA axis dysfunction via a negative feedback mechanism as well as atrophy of the normal adrenal cortex [33]. This may in part explain the high incidence of postoperative hypocortisolism in old patients in the present study. Similarly, in a previous study, old age was associated with new-onset pituitary failure in patients undergoing endonasal transsphenoidal pituitary adenoma removal [7].

Long surgical time was also predictive of 3-month postoperative hypocortisolism in the present study before propensity score matching. Prolonged surgical time can cause more severe inflammatory and neuroendocrine stress response. In addition, suprasellar extension or tumor adhesion may extend the surgical time, which can in turn cause pituitary gland injury and postoperative hypocortisolism.

Musculoskeletal disease was a significant predictor of 3-month postoperative hypocortisolism before propensity score matching. Musculoskeletal diseases are common in old patients. In addition, arthralgia can be caused by hypocortisolism [16, 34]. Impairment of HPA axis causes a high pain sensitivity due to inadequate stress response [4]. Thus, musculoskeletal disease may be associated with postoperative hypocortisolism in NFPA patients.

The present study had several limitations. First, some potential biases (i.e., selection bias) associated with the retrospective design could have affected the results. Also, only the association, not causality, between sevoflurane-based inhalation anesthesia and postoperative hypocortisolism was implied because this was a retrospective and observational study. Second, the presence or absence of postoperative hypocortisolism was evaluated only at 3 months postoperatively; thus, further research is necessary to determine the association of anesthetics with long-term postoperative hypocortisolism in NFPA patients undergoing endoscopic transsphenoidal surgery. Third, the Nagelkerke R2 values were relatively low, which indicates that other key predictors of postoperative hypocortisolism may have been omitted from the analysis. Fourth, because the number of patients in the sevoflurane anesthesia group was relatively small, the study may have been underpowered. However, changes in the proportion of two anesthetic techniques used and incidence of 3-month postoperative hypocortisolism over time were not significantly different.

Conclusions

In conclusion, sevoflurane anesthesia and postoperative steroid administration were associated with 3-month postoperative hypocortisolism in patients with preoperative normal HPA axis undergoing endoscopic transsphenoidal surgery for NFPA removal. A large-scale prospective study is necessary to confirm the negative association between sevoflurane-based anesthesia and postoperative ACTH functional outcome. Also, even though patients in the propofol group more frequently suffered from postoperative permanent DI, they showed a low incidence of 3-month postoperative hypocortisolism, suggesting that propofol may have a significant protective effect on postoperative hypocortisolism.