Good afternoon, everyone, and thank you for joining us for the AANS-CNS Virtual Joint Tumor Section. My name is Rishika Shoshi, and I'm a rising fourth-year medical student at the University of California, San Diego School of Medicine. I recently completed a research year at the University of California, San Francisco, in Dr. Manish Aghi's lab. As part of our clinical research efforts, we curated a registry of over 2,500 patients receiving transplant oil surgery. Today, I will be talking about our study to identify risk factors for postoperative diabetes insipidus in over 2,500 patients undergoing transplant oil surgery at a single institution. I have no disclosures. Diabetes insipidus, and more specifically, central diabetes insipidus, is the most common source of postoperative morbidity following transplant oil resection of cellar and paracellar lesions. The rates of DI following transplant oil surgery vary widely, with various studies indicating overall DI rates from 10% to 30%. While the overall rate of DI remains relatively high, most studies indicate the rate of permanent DI to be closer to 2% to 4%. Clinically, DI presents with polyuria and polydipsia, and can impose a significant burden on patient quality of life. Generally, DI is thought to arise from manipulation of the infundibulum, or pituitary stalk, during surgery. While similar studies have been published, there remains variability in identified risk factors due to limited sample sizes. Our goal was to identify risk factors that may predispose patients to developing DI postoperatively in a large cohort from a single institution. For our study, we conducted a retrospective review of all patients who underwent transplant oil surgery from 2007 to 2019 at UCSF. In total, we identified 2,563 patients. To define diabetes insipidus in our cohort, we limited our definition to two primary criteria, the first being a postoperative sodium of 145 mEq and a urine output of 300 cc per hour, or patients who received desmopressin, or DDAVP, postoperatively. Patients who had pre-existing DI at baseline were excluded from our study and analysis. To identify risk factors associated with DI, we first conducted univariate analysis on a multitude of patient, surgical, and pathological characteristics, followed by subsequent multivariate analysis to identify factors independently associated with DI. In our patient cohort, a total of 251 patients overall exhibited clinical signs for DI postoperatively, corresponding to 9.8% of our total cohort. Of these 251 patients, 157, or 6.1% of the total, had transient symptoms, while 94 patients, or 3.7%, had permanent DI. Permanent DI was defined as the presence of symptoms at the last follow-up. The majority of cases collected from our review were conducted microscopically, with endoscopic technique being used more widely over the last 6-7 years. Pathologies included in our study were craniopharyngioma, Rathke's cleft cyst, and pituitary adenomas, both functioning and non-functioning, which were present in about a 50-50 split within the pituitary adenoma subgroup. We next looked at the rates for DI in each specific lesion type, and it is clear that rates of DI, both transient and permanent, were significantly higher in patients with craniopharyngiomas, which is consistent with what has been presented previously in the literature. We also observed higher rates of DI in Rathke's cleft cyst patients when compared to pituitary adenomas. Within the pituitary adenomas, both non-functioning and functioning adenomas once again exhibited very similar rates. Each of these differences was statistically significant. We next conducted our multivariate analysis in four different groups. First, we included all pathologies in our analysis. Here, we identified four risk factors as being independently associated with development of DI postoperatively. We observed that younger patients exhibited higher risk, as did patients with craniopharyngiomas and patients with intraoperative CSF encounters. Patients who underwent microscopic surgery, on the other hand, were associated with a lower risk of DI. The identification of craniopharyngioma and CSF encounters intraoperatively as being associated with higher risk are once again consistent with similar studies that have previously been published. We next ran a multivariate analysis looking only at pituitary adenomas. In this cohort, we identified sex as an independent risk factor, with male patients being associated with higher risks of DI. Additionally, age, microscopic approach, and intraoperative CSF encounters were once again significantly associated with DI. As was the case previously, younger patients were associated with higher risk, as were those who had intraoperative CSF encounters during their surgery. For our final two groups, we looked individually at non-functioning and functioning adenomas to see if there were differences in identifying risk factors for DI. In both cohorts, we similarly identified age, microscopic approach, and intraoperative CSF encounter as independently associated risk factors. Once again, these trends were the same as previously described, with younger patients at higher risk of DI, as are patients with CSF encounters intraoperatively. In summary, our study, which was a large retrospective review of transfinoyl cases conducted at a single institution, corroborated previous results that craniopharyngiomas are associated with a higher risk of developing DI postoperatively. In addition, we identified age as a significant risk factor, showing younger patients as exhibiting higher risk of developing DI. When we considered just pituitary adenomas, we identified male sex as associated with higher risk of developing DI as well. Finally, across all four cohorts, we identified intraoperative CSF encounters as a significant risk factor for developing DI postoperatively. This is consistent with some previously published studies, and can possibly be attributed to greater stock manipulation from operations that require entering the supercellar cistern. We observed that microscopic approach was associated with lower risk of DI in our four groups as well, and this could be due to sample size bias, as there are far less endoscopic cases in our cohort versus microscopic approaches. Additionally, we noted that endoscopic approach was generally being utilized for larger or more complex tumors, which often included cases with greater cavernous sinus invasion. I would like to thank the AANS-CNS Joint Section on Tumors for allowing us to give these virtual presentations, and most importantly, I would like to thank my mentor Dr. Aghi for a wonderful and enriching research year, as well as the entire Department of Neurological Surgery at UCSF. Thank you.

postoperative diabetes insipidus

transsphenoidal surgery

risk factors

craniopharyngiomas

microscopic approach