Effect of facility volume on giant pituitary adenoma neurosurgical outcomes

Abstract

OBJECTIVE Giant pituitary adenomas (PAs), defined as 4 cm or greater at their maximum diameter, are commonly treated with neurosurgical intervention as the first-line therapy. However, existing studies are from high-volume institutions whose outcomes may not be representative of many cancer centers. In the present study, the authors use a large cancer registry to evaluate demographics, national treatment trends, and outcomes by facility volume to address knowledge gaps for this uncommon tumor. METHODS The National Cancer Database was queried for adult patients with PAs who had undergone resection from 2004 to 2016. Univariate and multivariate logistic regression modeling was used to evaluate the prognostic impact of covariates on short-term outcomes including 30-day readmission (30R), 30-day mortality (30M), 90-day mortality (90M), and prolonged length of inpatient hospital stay (LOS). Propensity score matching was used for validation. RESULTS Among the 39,030 patients who met the study inclusion criteria, 3696 giant PAs were identified. These tumors had higher rates of subtotal resection (55% vs 24%, p < 0.001), adjunctive radiotherapy (15% vs 5%, p < 0.001), and hormonal therapy (8% vs 4%, p < 0.001) than nongiant PAs. The giant PAs also had worse 30M (0.6% vs 3.1%, p < 0.001), 90M (1.0% vs 5.0%, p < 0.001), 30R (4.0% vs 6.3%, p < 0.001), and LOS (22.2% vs 42.1%, p < 0.001). On multivariate analysis for giant PA, decreased tumor size, younger age, race other than African American, lower comorbidity score, and high-volume facility (HVF; defined as ≥ 2.5 giant PA cases per year) were statistically significant predictors of favorable outcomes. Specifically, 30M, 90M, 30R, and LOS were decreased by 50%, 43%, 55%, and 32%, respectively, when giant PAs were treated at HVFs (each p < 0.05). HVFs more often used the endoscopic approach (71% vs 46%, p < 0.001) and less adjuvant radiotherapy (11% vs 16%, p < 0.001). Propensity score matching validated 30M, 30R, and LOS outcome differences in a cohort of 1056 patients. CONCLUSIONS This study provides evidence of superior outcomes when giant PAs are treated at HVFs. These results likely reflect the relation between physician experience and outcomes for these uncommon tumors, which suggests the need for institutional collaboration as a potential goal in their surgical management.