The First Decade of Laparoscopic Pancreaticoduodenectomy in the United States: Costs and Outcomes Using the Nationwide Inpatient Sample

Thuy B Tran, MD, Monica M Dua, MD, David J Worhunksy, MD, George A Poultsides, MD, Jeffrey A Norton, MD, Brendan C Visser, MD. Stanford University

INTRODUCTION: Minimally invasive pancreaticoduodenectomy (PD) remains an uncommon procedure, and the safety and efficacy remain uncertain beyond single institution case series. An analysis on national costs and outcomes of minimally invasive PD has yet to be studied. Fears persist that longer operative times and instrument costs make LPD cost-ineffective. The aim of this study is to compare outcomes and costs between laparoscopic (LPD) and open PD (OPD) using a large population-based database.

METHODS: The Nationwide Inpatient Sample (NIS) database (a sample of approximately 20% of all hospital discharges) was queried to identify patients who underwent PD from 2000 to 2010. Total hospital charges were adjusted to 2010 dollars. Patient demographics, comorbidities, hospital characteristics, inflation-adjusted total charges, and complications were evaluated using univariate and multivariate analyses. Hospitals were categorized as high volume hospitals (HVH) if more than 20 PD were performed annually, while those performing less than 20 PD were defined as low volume hospitals (LVH).

RESULTS: Of the 15,574 PD cases performed during the study period, 681 LPD were performed (4.4%). Thus, approximately 3,400 LPD were performed over the decade. Patients undergoing OPD and LPD were very similar overall. Compared to OPD, patients who underwent LPD were slightly older (65 vs. 67 years; p=0.001), and more likely to have been treated at teaching hospitals (91% vs. 80.9%, p<0.001) and at HVH (56.6 vs. 66.1%; p<0.001). But cardiac, pulmonary, and renal comorbidities were similar between the groups. Higher rates of complications were observed in OPD than LPD (46% vs. 39.4%; p=0.001), though mortality was similar (5% for OPD vs 3.8% for LPD, p=0.27). Inflation-adjusted hospital charges were similar between OPD and LPD ($87,577 vs. $81,833, p=0.199; Figure 1). However, hospital stay was a day longer in OPD compared to LPD group (12 vs 11 days, p<0.001).

Stratifying outcomes by hospital volume, LPD at LVH resulted in longer hospital stay (13 d at LVH vs 9 d at HVH, p<0.001) and significantly higher charges ($106,367 at LVH vs $76,572 at HVH, p=0.016; Figure 2). On multivariate analysis, after controlling for demographics and comorbidities, teaching status and HVH correlated with low risk of morbidity and mortality. Blood transfusion and LVH were independent predictors of mortality.

CONCLUSIONS: Despite the relative rarity of the procedure, laparoscopic pancreaticoduodenectomy has slightly lower hospital stay, lower morbidity, and similar hospital charges compared to the traditional open version. The costs associated with operating room time and instruments are likely offset by the shorter hospital stay. LPD performed at high volume hospitals are associated with shorter stay, a lower rate of complications, and lower total hospital charges compared to low volume hospitals. Laparoscopic pancreaticoduodenectomy at high volume hospitals is justified.



« Return to SAGES 2015 abstract archive