Readmissions following pancreaticoduodenectomy for pancreas cancer: a population-based appraisal

Abstract

Procedure complexity and volume-outcome relationships have led to increased regionalization of pancreaticoduodenectomy (PD) for pancreas cancer. Knowledge regarding outcomes after PD comes from single-institutional series, which may be limited if a significant number of patients follow up at other hospitals. Thus, readmission data may be underreported. This study utilizes a population-based data set to examine readmission data following PD. California Cancer Registry (1994-2003) was linked to the California's Office of Statewide Health Planning and Development (OSHPD) database; patients with pancreatic adenocarcinoma who had undergone PD, excluding perioperative (30-day) mortality, were identified. All hospital readmissions within 1 year following PD were analyzed with respect to timing, location, and reason for readmission. Our cohort included 2,023 patients who underwent PD for pancreas cancer. Fifty-nine percent were readmitted within 1 year following PD and 47% were readmitted to a secondary hospital. Readmission was associated with worse median survival compared with those not readmitted (10.5 versus 22 months, p<0.0001). Multivariate analysis revealed that increasing T-stage, age, and comorbidities were associated with increased likelihood of readmission. Diagnoses associated with high rates of readmission included progression of disease (24%), surgery-related complications (14%), and infection (13%). Diabetes (1.4%) and pain (1.5%) were associated with low rates of readmission. We found a readmission rate of 59%, which is much higher than previously reported by single institutional series. Concordantly, nearly half of patients readmitted were readmitted to a secondary hospital. Common reasons for readmission included progression of disease, surgical complications, and infection. These findings should assist in both anticipating and facilitating postoperative care as well as managing patient expectations. This study utilizes a novel population-based database to evaluate incidence, timing, location, and reasons for readmission within 1 year following pancreaticoduodenectomy. Fifty-nine percent of patients were readmitted within 1 year after pancreaticoduodenectomy and 47% were readmitted to a secondary hospital.