A simple scoring system for predicting the risk of delayed hyponatremia after endoscopic transsphenoidal surgery for pituitary adenomas.

Purpose To identify high-risk patients for delayed postoperative hyponatremia (DPH) early, we constructed a simple and effective scoring system


Introduction
Pituitary adenoma is the most common benign tumor in the sellar region.Although surgical treatment has become increasingly mature, postoperative complications are still inevitable, which is an important factor affecting the postoperative recovery of pituitary adenoma patients.Delayed postoperative hyponatremia (DPH)is de ned as the occurrence of hyponatremia after the third postoperative day (POD3) following transsphenoidal surgery (TSS) [1,2], with an incidence rate of 9-23% [3].Symptoms include headache, nausea, vomiting, altered mental state, and seizures, and can even lead to death [4].
Therefore, timely treatment is necessary as it does not resolve on its own.A common cause of DPH is syndrome of inappropriate antidiuretic hormone secretion (SIADH), while cerebral salt wasting syndrome is rare [5].In patients with DPH, the lowest blood sodium level is typically observed around the 7th day after surgery, and most patients experience symptoms of hyponatremia between the 4th and 7th days after surgery [6].In recent years, due to the widespread use of endoscopic techniques, the length of hospital stay for patients with pituitary adenoma has greatly reduced, and most patients being discharged within 3 days after surgery [7].This may lead to an increase in the number of patients experiencing DPH after discharge.Bohl et al. found that DPH accounted for 56% of patients who were readmitted within 30 days after TSS [8].The treatment of delayed hyponatremia is challenging because there is limited understanding of the predictive factors associated with its occurrence.Once it occurs, it can prolong hospitalization, increase the economic burden on patients, and increase mortality.Currently, there is still a lack of a simple and effective assessment method for predicting the occurrence of DPH.

Patient cohort
Institutional review board approval was obtained for this retrospective study.All data were anonymous, and patient consent was waived.The study retrospectively analyzed 141 patients who underwent endoscopic TSS for pituitary adenoma and received histopathological con rmation between January 2019 and December 2022.All eligible patients were operated on by two experienced senior surgeons.

Group design
We divided the patients into two groups, DPH group and non-DPH (NDPH) group, based on the blood sodium concentration after the third postoperative day (POD3) following endoscopic transsphenoidal surgery.DPH is de ned as serum sodium level <135mEq/L after the third day of surgery.

Data collection
The data were collected retrospectively by reviewing patient electronic medical records.The tumor diameter and volume were measured using preoperative magnetic resonance imaging.The tumor's anterior-posterior, transverse, and vertical diameters were de ned as the longest axis along these three directions.The tumor volume was estimated by multiplying the three diameters and dividing the result by two [9].Cavernous sinus invasion was classi ed according to the Knosp classi cation, and an invasive tumor was de ned as a Knosp grade ≥3 [10].Low potassium levels on the rst or second day after surgery were de ned as a serum potassium level below the lower limit of normal (3.5mEq/L).The difference in blood sodium levels before and on the rst day after surgery is de ned as the value obtained by subtracting the preoperative serum sodium level from the serum sodium level on the rst day after surgery.The difference in serum sodium levels between postoperative day 1 and day 2 is de ned as the value of serum sodium level on day 1 minus the serum sodium level on day 2, abbreviated as serum sodium (POD1 -POD2).In the 2017 World Health Organization classi cation, a Ki-67 index ≥3% was considered an invasive pituitary adenoma [11].Postoperative diabetes insipidus was de ned as polyuria (urine output ≥300ml/h for at least 2 hours) after TSS surgery.All patients in this study received routine iodine-soaked gauze lling in the nasal cavity after surgery, which was later removed, resulting in a longer hospital stay and longer monitoring of serum sodium levels.After discharge, the patients were regularly followed up in the neurosurgery or endocrinology clinic for hormone and electrolyte monitoring.For patients with delayed hyponatremia, the time of onset was de ned as the time of the rst recorded low serum sodium level, although some patients may have developed hyponatremia before the test was performed.This could potentially result in the actual onset of DPH occurring earlier than the time of detection.

Statistical analysis
We used SPSS version 26.0 (IBM Inc, Armonk, NY, USA).Normality was tested using the Kolmogorov-Smirnov test.Independent samples t-tests were used for continuous variables with a normal distribution, while Mann-Whitney U tests were used for non-normally distributed variables.For categorical variables, chi-square tests were used, except for variables with values of ≤5 in any cell of the contingency table, where Fisher's exact test was used.A multivariate logistic regression analysis was conducted for factors that had statistical signi cance in the univariate analysis, where P<0.05 was considered statistically signi cant.All patients were scored according to the scoring system, to nd the best critical value for predicting scores, the receiver operating characteristic (ROC) curve was drawn and the area under the curve (AUC) calculated.The univariable analysis revealed that age and Na (POD1 -POD2) were predictive factors for DPH (Table 1).ROC curve analysis determined the optimal cutoff values for age and Na (POD1 -POD2) to be 48 years and 1 mEq/L, respectively (Fig. 1).Subsequent group analyses were conducted using these cutoff values.

Results
Patients' Characteristics Among the 141 patients included in the study, 59 (41.8%) were male and 82 (58.2%) were female, with a median age of 47 years (range, 19-76 years).36 patients (25.5%) developed delayed hyponatremia after surgery, of which 22 patients (61.1%) exhibited clinical symptoms such as nausea, vomiting, headache, and dizziness.The mean time to onset of delayed hyponatremia was 7.8 days postoperatively with a median of 7 days (range, 4-16 days).

Factors related to DPH.
Through independent sample t-tests, non-parametric Chi-square tests, and Mann-Whitney U tests, there were statistically signi cant differences (P<0.05) between the DPH group and non-DPH group in terms of age, BMI, tumor volume, tumor diameter, the maximum tumor diameter, compression of the optic chiasm, Knosp grade ≥3, intraoperative cerebrospinal uid leakage, hypokalemia on the rst or second day after surgery, hyponatremia on the second day after surgery, serum sodium (POD1 -POD2), and postoperative diabetes insipidus.However, there were no signi cant differences between the DPH group and non-DPH group regarding gender, preoperative comorbidities (hypertension, diabetes, Cushing's disease), preoperative and postoperative thyroid function, tumor cysts or hematoma, and immunohistochemical type (Table 1).
The SKAPD score The SKAPD score (Sodium, Knosp grade, Age, Potassium, Diabetes insipidus) is assigned based on the odds ratio values for serum sodium (POD1 -POD2) ≥1 mEq/L, Knosp grade ≥3, age ≥48 years old, hypokalemia on the rst or second day after surgery, and postoperative diabetes insipidus (Table 3).The SKAPD score was calculated for each patient, and a receiver operating characteristic (ROC) curve was plotted.The results showed that the optimal cut-off value for the SKAPD score was 10, with a sensitivity of 77.8%, speci city of 79.1%, and AUC of 0.856 (95%CI, 0.787-0.925),indicating a moderate to good accuracy in predicting DPH (Fig 2 .).

Discussion
This clinical study investigated the incidence and predictive factors of delayed hyponatremia following endoscopic transsphenoidal surgery (TSS) and constructed the SKAPD score based on these factors.The results showed that the incidence of delayed hyponatremia was 25.5%.The independent factors predicting delayed hyponatremia after surgery were age ≥ 48 years old, Knosp grade ≥ 3, hypokalemia on the rst or second day after surgery, serum sodium (POD1 -POD2) ≥ 1 mEq/L, and postoperative diabetes insipidus.We found that early prediction of the occurrence of DPH can be achieved based on the SKAPD score.The incidence of DPH in our study cohort was 25.5% compared to other studies, and the higher incidence may be due to the fact that our patients received routine nasal lling iodoform gauze strips after surgery, which subsequently required removal.Patients are hospitalized for longer periods of time and are more closely monitored for serum sodium levels.
Previous studies have reported a relationship between age and delayed hyponatremia after surgery.For example, Lee et al. reported that 50 years old was the cut-off age for predicting delayed hyponatremia [12], while Tomita et al. reported 55 years old [13], and Kinoshita et al. reported 60 years old as the cut-off age [14].Our ndings on predictive factors and critical values are similar to previous reports.The optimal critical value for age prediction was found to be 48 years old.We believe that older patients are more likely to develop DPH.This could be due to poorer compensatory ability of pituitary function in elderly patients, leading to postoperative pituitary dysfunction and electrolyte imbalance.In addition, some studies have reported that due to gender differences in the response to antidiuretic hormone, females are more susceptible to delayed hyponatremia after surgery than males [15].However, our study found that gender was not a predictor of delayed hyponatremia after surgery.[12].For patients with Knosp grade ≥ 3 tumors, larger and longer surgical procedures are typically required, which may result in greater manipulation of the pituitary stalk and subsequent release of ADH, ultimately leading to uid overload and hyponatremia.Increasing the surgical experience of neurosurgeons can reduce the occurrence rate of complications after pituitary adenoma surgery [17].Therefore, it is necessary to enhance the surgical training and learning of neurosurgeons to improve their surgical skills and expertise in this eld.
Currently, some studies have analyzed the relationship between postoperative hypokalaemia and delayed hyponatremia, but there is no evidence of a correlation between postoperative hypokalaemia and delayed hyponatremia.The de nition of postoperative hypokalaemia used in these studies was a potassium level < 3.5mEq/L after surgery.In this clinical study, only potassium levels on the rst and second postoperative days were collected, and hypokalaemia occurring on or after the third day was excluded.This not only reduces interference from inadequate potassium intake but also allows for earlier prediction of the risk of developing delayed postoperative hyponatremia after discharge.Patients with preoperative hypokalaemia are corrected before surgical treatment.Seow et al. suggest that hypokalaemia may be due to high levels of adrenocorticotropic hormone secreted by pituitary adenoma [18].You et al. believe that intraoperative hypokalaemia is a cause of postoperative hypokalaemia.Firstly, the use of glucocorticoids and mineralocorticoids during the operation can cause a decrease in blood potassium levels.Secondly, destruction of pituitary cells during surgery can release hormones, and the tumor tissue is squeezed out, leading to further hormone secretion [19].The most common cause of delayed hyponatremia is syndrome of inappropriate antidiuretic hormone secretion (SIADH) [5], and early release of ADH may also cause hypokalaemia.To further understand the impact of early postoperative hypokalaemia, ACTH and potassium levels should be monitored during the operation, and changes in antidiuretic hormone levels should be monitored early after surgery.
Krogh et al. reported that patients with a decrease in blood sodium concentration on the rst day after surgery had an increased risk of developing hyponatremia [20].Other studies have also indicated that a signi cant drop in blood sodium concentration 2 days after TSS compared to preoperative levels is an independent predictor of delayed hyponatremia, suggesting that abnormal secretion of ADH may occur on the second day after TSS [21].Our results indicate that a signi cant drop in blood sodium concentration on the second day after surgery is a risk factor for delayed hyponatremia.We believe that patients with sodium (POD1-POD2) ≥ 1mEq/L are more likely to have a tendency to decrease sodium levels after surgery and eventually develop delayed hyponatremia.Therefore, when patients develop Na (POD1 -POD2) ≥ 1 mEq/L, we can consider attempting prophylactic sodium supplementation through oral administration.Tanaka et al. reported that postoperative transient diabetes insipidus (DI) may be a risk factor for delayed hyponatremia [22].In our study, 3 cases of permanent DI were observed among the subjects, but they were not differentiated as transient or permanent DI in detail.Zada et al. mentioned that patients who develop DI after surgery have a 48% increased risk of developing hyponatremia [23], which may be associated with injury to the pituitary stalk during the surgical procedure.It has been reported that hyponatremia after DI is caused by SIADH, due to the irregular release of arginine vasopressin from the neurohypophysis without neural control [24].
Due to the variability of previous research results, the etiology of delayed hyponatremia is clearly complex and multifactorial.To our knowledge, this is the rst established scoring system that can predict the risk of delayed hyponatremia following Transsphenoidal Surgery (TSS).The SKAPD score has an Area Under the Curve (AUC) of 0.856 in predicting the occurrence of DPH, indicating a good predictive performance.
The majority of factors included in this scoring system are preoperative and early postoperative factors.
With the help of the SKAPD score, neurosurgeons can accurately identify potential patients at risk of delayed hyponatremia early on.Accurate prediction of delayed hyponatremia will help improve resource allocation and patient safety.High-risk patients who are prone to developing delayed hyponatremia may be considered for appropriate extension of hospitalization or more frequent outpatient follow-up visits after discharge, depending on the individual situation.

Limitations
This is a single-center study, which may introduce center bias into our results.In addition, hyponatremia can only be detected in patients who are hospitalized or followed up in outpatient clinics.The SKAPD score was generated retrospectively based on the cohort we expected to collect and has not been prospectively or externally validated.Our study included all patients who developed delayed postoperative hyponatremia (DPH), without speci cally focusing on those who developed symptomatic hyponatremia or required treatment or readmission.

Conclusions
The independent predictive factors for DPH after endoscopic TSS in patients with pituitary adenoma are age ≥ 48 years, Knosp grade ≥ 3, serum sodium (POD1 -POD2) ≥ 1 mEq/L, postoperative hypokalemia on the rst or second day, and postoperative DI.The accuracy of the SKAPD score in predicting DPH is moderate to good, but still requires prospective external validation. Tables The result of receiver operating characteristic (ROC) curve analysis showed that the optimal cut-off values of age and the difference in blood sodium levels between the rst and second day after surgery for predicting delayed hyponatremia were 48 years old and 1 mEq/L, respectively.
The result of the ROC curve analysis showed that the optimal cut-off value for the SKAPD score was 10, with a sensitivity of 77.8% and a speci city of 79.1%.The area under the curve (AUC) was 0.856 (95% CI, 0.787-0.925).
Yoon et al. suggested that longer surgical operation time increases the risk of developing delayed hyponatremia [16].Lee et al. reported a higher risk of postoperative hyponatremia in patients with pituitary adenoma, which require more surgical manipulations and can potentially interfere with the pituitary stalk and hypothalamus

Table 1
Clinical characteristics and auxiliary examinations.

Table 2
Multivariable logistic regression analysis of the risk of postoperative-delayed hyponatremia onset.CI, con dence interval; POD, postoperative day; DI, diabetes insipidus.Signi cant P values are shown in bold.