Identifying risk factors for postoperative diabetes insipidus in more than 2500 patients undergoing transsphenoidal...

Diabetes insipidus (DI) following transsphenoidal surgery can adversely impact quality of life and be difficult to manage. This study sought to characterize pre- and perioperative risk factors that may predispose patients to DI after pituitary surgery.

A retrospective review of patients treated at a single institution from 2007 to 2019 was conducted. DI was defined as postoperative sodium > 145 mEq/L and urine output > 300 ml/hr and/or postoperative desmopressin (ddAVP) use. DI was further characterized as transient or permanent. Uni- and multivariate analyses were performed to determine variables associated with postoperative DI.

The authors identified 2529 patients who underwent transsphenoidal surgery at their institution. Overall, DI was observed in 270 (10.7%) of the 2529 patients, with 114 (4.5%) having permanent DI and 156 (6.2%) with transient symptoms. By pathology type, DI occurred in 31 (46.3%) of 67 craniopharyngiomas, 10 (14.3%) of 70 apoplexies, 46 (14.3%) of 322 Rathke's cleft cysts, 77 (7.7%) of 1004 nonfunctioning pituitary adenomas (NFPAs), and 62 (7.6%) of 811 functioning pituitary adenomas (FPAs). Final lesion pathology significantly affected DI rates (p < 0.001). Multivariate analysis across pathologies showed that younger age (odds ratio [OR] 0.97, p < 0.001), intraoperative CSF encounter (OR 2.74, p < 0.001), craniopharyngioma diagnosis (OR 8.22, p = 0.007), and postoperative hyponatremia (OR 1.50, p = 0.049) increased the risk of DI. Because surgery for each pathology created specific risk factors for DI, the analysis was then limited to the 1815 pituitary adenomas (PAs) in the series, comprising 1004 NFPAs and 811 FPAs. For PAs, younger age (PA: OR 0.97, p < 0.001; NFPA: OR 0.97, p < 0.001; FPA: OR 0.97, p = 0.028) and intraoperative CSF encounter (PA: OR 2.99, p < 0.001; NFPA: OR 2.93, p < 0.001; FPA: OR 3.06, p < 0.001) increased DI rates in multivariate analysis. Among all PAs, patients with DI experienced peak sodium levels later than those without DI (postoperative day 11 vs 2). Increasing tumor diameter increased the risk of DI in FPAs (OR 1.52, p = 0.008), but not in NFPAs (p = 0.564).

In more than 2500 patients treated at a single institution, intraoperative CSF encounter, craniopharyngioma diagnosis, and young age all increased the risk of postoperative DI. Patients with postoperative hyponatremia exhibited higher rates of DI, suggesting possible bi- or triphasic patterns to DI. Greater vigilance should be maintained in patients meeting these criteria following transsphenoidal surgery to ensure early recognition and treatment of DI.