The center volume-outcome effect in pancreas transplantation: a national analysis.

BACKGROUND: Although increased hospital volume has been correlated with improved outcomes in certain surgical procedures, the effect of center volume on pancreas transplantation (PT) is less understood. Our study aims to establish whether a volume-outcome effect exists for PT.

METHODS: Through an established linkage between the University HealthSystem Consortium and the Scientific Registry of Transplant Recipients (SRTR) databases, we performed a retrospective cohort analysis of adult PT recipients between 2009 and 2012. Surgical volume was divided equally into low volume (LV), middle volume (MV), and high volume (HV) tertiles for each year that was studied. Hospital outcomes were measured through University HealthSystem Consortium, and long-term outcomes were measured through Scientific Registry of Transplant Recipients. Statistical analysis was performed using regression analyses and the Kaplan-Meier method. Median follow-up period was 2 y.

RESULTS: Among the 2309 PT recipients included, 815 (35.3%) were performed at LV centers, 755 (32.7%) at MV centers, and 739 (32.0%) at HV centers. Compared with MV and LV centers, organs transplanted at HV centers were more frequently donation after cardiac death (5.1% versus 2.4% versus 3.3%, P = 0.01) and from older donors (2.8% [>50 y] versus 0.8% versus 0.1%, P < 0.001). In addition, HV recipients were older (31.5% [>50 y] versus 20.9% versus 19.7%, P < 0.001) and had worse functional status (39.5% dependent versus 9.7% versus 9.9%, P < 0.001). Patient and graft survival were similar across hospital volume tertiles. Center volume was not predictive of readmission rates, total length of stay, intensive care unit length of stay, or total direct cost on multivariate analysis (all P > 0.05).

CONCLUSIONS: Short- and long-term outcomes after PT are not affected by hospital volume. Although LV centers confine their cases to low-risk patients, HV centers transplant a higher percentage of high-risk donor and recipient combinations with equivalent outcomes.