Total pancreatectomy: a national study.

BACKGROUND: Total pancreatectomy (TP) is performed for various indications. Historically, morbidity and mortality have been high. Recent series reporting improved peri-operative mortality have renewed interest in TP. We performed a national review of TP including indication, patient/hospital characteristics, complications and peri-operative mortality.
METHODS: The Nationwide Inpatient Sample (NIS) was queried to identify TPs performed during 1998 to 2006. Univariate analyses were used to compare patient/hospital characteristics. Multivariable logistic regression was performed to identify predictors of in-hospital mortality. Post-operative complications/disposition were assessed.
RESULTS: From 1998 to 2006, 4013 weighted patient-discharges occurred for TP. Fifty-three per cent were male; mean age 58 years. INDICATION: neoplastic disease 67.8%. Post-operative complications occurred in 28%. Univariate analyses: TPs increased significantly (1998, n = 384 vs. 2006 n = 494, P < 0.01). 77.1% of TPs occurred in teaching hospitals (P < 0.0001), 86.4% in hospitals performing <five pancreatectomies/year (P < 0.0001). In-patient mortality was 8.5% with a significant decrease (12.4% 1998-2000 vs. 5.9% 2002-2006, P < 0.01). Multivariable analyses: advanced age [referent < or = 50 years; > or = 70 Adjusted odds ratio (AOR) 3.4, 95% confidence interval (CI) 1.33-8.67], select patient comorbidities and year (referent = 2004-2006; 1998-2000 AOR 2.70; 95% CI 1.41-5.14) independently predicted in-patient mortality whereas hospital surgical volume did not. DISCUSSION: TP is increasingly performed nationwide with a concomitant decrease in peri-operative mortality. Patient characteristics, rather than hospital volume, predicted increased mortality.