Hospital procedural volume is inversely linearly associated with complications in endoscopic endonasal transsphenoidal resection of pituitary tumors

OBJECTIVE Prior studies have shown that high-volume centers have improved outcomes in patients undergoing transsphenoidal resection for pituitary tumors. However, those investigations have not examined this association specifically for the endoscopic approach and are limited by characterizing the volume outcome association using arbitrary dichotomous volume cutoffs. The objective of the current study was to delineate the continuous volume-outcome relationship adjusted for risk factors in patients with benign pituitary tumors undergoing endoscopic endonasal transsphenoidal surgery (ETSS) and systemically identify volume cutoffs after which there is no significant increase in complication risk. METHODS In this retrospective analysis, the Nationwide Readmissions Database (NRD) was queried for patients with benign pituitary tumors who underwent ETSS from 2016 to 2018. ICD-10 coding was used for cohort selection. Patient and hospital characteristics were extracted from standard NRD-collected variables. The association of institutional procedural volume and outcomes (major complications, sellar tumor–specific complications, and discharge disposition) were evaluated using multivariable analysis. RESULTS A total of 14,947 patients (median age 56 years) with benign pituitary tumors who underwent ETSS were identified. Most patients received treatment at institutions with at least 13 cases per year (top 75th percentile). The multivariable analysis of volume as a continuous variable demonstrated that risk of major complications (e.g., sepsis) decreased at a steady rate (OR 0.984, 95% CI 0.977–0.992; p < 0.0001) per 1 procedure increase at institutions with a procedural volume of 1–57 cases per year. From 58 cases per year, there was no longer a decrease in risk (OR 1.001, 95% CI 0.996–1.006; p = 0.68). The risk of sellar tumor–specific complications (e.g., endocrinopathies and cranial nerve palsies) decreased throughout the entire volume range (OR 0.997, 95% CI 0.996–0.998; p < 0.0001). Furthermore, there was no linear response in discharge disposition, but the highest quartile was associated with the least likelihood of nonroutine discharge. CONCLUSIONS A multivariable analysis with institutional case volume as a continuous variable exhibited a linear association with risk of major and sellar tumor–specific complications specific to this patient population. Future studies are needed to further characterize the factors that contribute to this additive relationship.