Repeat Serum Sodium Values over 145 Predict Diabetes Insipidus after Endoscopic Transphenoidal Pituitary Tumor Resection

Background: Central diabetes insipidus (DI) is a common complication following pituitary surgery caused by damage or irritation to the posterior pituitary, pituitary stalk, or hypothalamic osmoregulatory centers. Most cases of postoperative DI are reported as transient, resolving within 1 week of surgery, though a small proportion of patients develop nontransient DI. Although there are generally accepted risk factors to DI development after surgery, such as Rathke’s cleft cyst, craniopharyngioma, craniotomy, and tumor size, other risk factors may contribute and have not been clearly defined. Elevated serum sodium values are a known characteristic of DI; however, few studies have looked at the association between the development of this condition and postoperative sodium values.

Objective: This study was aimed to determine if DI is associated with sodium values within the first 48 hours after endoscopic transsphenoidal pituitary surgery and whether other perioperative variables may be associated with the development of DI.

Methods: A total of 347 patients undergoing 366 surgeries were included for analysis. Medical records were reviewed for demographics, repeat surgery, tumor pathology, extent of resection, intraoperative or postoperative CSF leak, hospital stay, follow-up, and maximum serum sodium concentration within the first 48 hours postoperatively. Bivariate associations between each explanatory variable of interest and postoperative DI was examined using Wald χ ² tests. Multivariable logistic regression models were used to evaluate the association between postoperative 48-hour serum sodium concentrations and DI, after covariate adjustment

Results: Of 360 resections, 52 patients (14.4%) showed evidence of DI postoperatively. Transient postoperative DI (lasting less than 7 days) occurred in 23 patients (6.4%), and nontransient DI occurred in 29 patients (8.1%). Of those with nontransient DI, 10 patients were deamino-8-D-arginine vasopressin (ddAVP)-independent within 6 months, 1 patient after 1 year, and 18 patients remained ddAVP-dependent with permanent DI (5.0%). Odds of DI were higher with higher recorded 48-hour maximum serum sodium measurements (OR = 1.86; 95% CI = 1.59–2.17; indicating differential odds with unit increases in serum sodium measurements). Odds of DI were also higher with higher frequencies of serum sodium concentration measurements over 145 (OR = 9.22; 95% CI = 5.11–16.63). Two serum sodium values above 145 predicted DI with 99.03% specificity and 50% sensitivity and had a 98.09% PPV and 66.45% NPV. In bivariate analyses, statistically significant associations were observed between patient age and DI (OR = 0.98, 95% CI = 0.96–0.99 for unit increases in age). There were no significant associations observed between sex, CSF leaks, extent of resection, or prior surgery and DI.

Conclusion: Higher maximum recorded values of serum sodium, and multiple serum sodium over 145, within 48 hours of postsurgery were associated with higher odds of DI. While elevated serum sodium is a known feature of diabetes insipidus, the positive and negative predictive values presented are useful resources for clinicians in medication decision making, safe discharge, and patient counseling.