The Outcomes of Non-Trauma Splenectomy from Nationwide Inpatient Sample

A Y Zemlyak, MD, V B Tsirline, MD, S El Djouzi, MD, A L Walters, MS, A E Lincourt, PhD, MBA, R F Sing, DO, B T Heniford, MD

Carolinas Medical Center

Introduction
Due to the impact of LeapFrog and other outcomes measurement organizations and studies, a strong push toward regionalization for many solid organ operations began in the early 2000’s. This study examines the effects of regionalization for medically indicated, non-trauma splenectomies (NTS) in USA.
Methods
Nationwide Inpatient Sample data were analyzed for NTS based on ICD-9-CM codes for 1998-99 (1990’s) and 2008-2009 (2000’s). Demographics, co-morbidities and complications were compared by hospital status (teaching vs non-teaching) and location (urban vs rural) using standard statistical methods.
Results
NIS recorded 7,062 cases performed in 1990’s and 5,530 in 2000’s. Teaching hospitals accounted for 55.5% of NTS in 1990’s vs 61.5% in 2000’s (p<0.001). In 1990’s, 86.9% of cases were performed in urban hospitals vs 91.3% in 2000’s (p<0.001). In both decades patients were older in non-teaching versus teaching hospitals (47.6 vs 43, p<0.0001 and 51.4 vs 46.2, p<0.0001). The Charlson Co-morbidity Index (CCI) scores did not differ between hospitals in either decade. Non-teaching hospitals had more medical morbidities: CHF (1990’s: 6.05% vs 3.91%, p<0.0001 and 2000’s: 5.84% vs 3.36%, p<0.0001) and COPD (8.6% vs 6.56%, p=0.001 and 13.8% vs 10.2%, p<0.0001). In 1990’s in-hospital mortality was higher in teaching (7.31% vs 5.16%, p=0.0002) and urban (6.64% vs 4.45%, p=0.01) hospitals; there was no difference in the 2000’s. In 1990’s, teaching hospitals had more patients with severe liver disease and HIV (2.04% vs 1.37%, p=0.032 and 1.15% vs 0.64%, p=0.025). The same pattern was observed in urban vs rural hospitals (1.89% vs 0.76%, p=0.014 and 1.04% vs 0.11, p=0.002); there was no difference in 2000’s. In both decades in-hospital mortality correlated strongly with the presence of severe liver disease (p<0.0001), and with HIV in 2000’s (p=0.0003). Surgical complication rate (including bleeding/hematoma, abscess, pancreatic fistula and re-operation) was below 0.5% in both decades and not different per the hospital setting.
Conclusion
Although non-trauma splenectomies are performed increasingly in urban and teaching hospitals, the surgical outcomes are not affected by regionalization. The spectrum of presenting co-morbidities differs per hospital setting, however. Inpatient mortality appears to be primarily related to severe liver disease despite the decade of care.

Session: Podium Presentation

Program Number: S027