Minimally Invasive Versus Open Pancreaticoduodenectomy for Cancer: Practice Patterns and Short-term Outcomes Among 7061 Patients.

Abstract

To describe national practice patterns regarding utilization of minimally invasive pancreaticoduodenectomy (MIPD) and compare short-term outcomes with those following open pancreaticoduodenectomy for cancer. There is increasing interest in use of MIPD; however, published data are limited to single institutional experiences. Adult patients undergoing pancreaticoduodenectomy were identified from the National Cancer Database, 2010-2011. Descriptive statistics and multivariable modeling were employed to characterize use of MIPD (laparoscopic or robotic) and compare short-term outcomes to those following open pancreaticoduodenectomy. A total of 7061 patients underwent pancreaticoduodenectomy: 983 had MIPD and 6078 had open procedures. The use of MIPD increased by 45% (179 cases) from 2010 to 2011. The majority of hospitals (92%) performing MIPD were low volume (≤ 10 cases/2 years). Factors independently associated with undergoing MIPD included fewer comorbidities, treatment at an academic institution, and a neuroendocrine tumor diagnosis (all P < 0.01). The unadjusted 30-day mortality rate was 5.1% for MIPD versus 3.1% after open surgery. For patients with adenocarcinoma, there were no differences between MIPD and open pancreaticoduodenectomy after multivariable adjustment in number of lymph nodes removed, rate of positive surgical margins, length of stay, or readmissions. However, 30-day mortality was higher for patients undergoing MIPD versus open surgery (odds ratio = 1.87, confidence interval: 1.25-2.80, P = 0.002). While there is increasing interest in employing MIPD for adenocarcinoma, its use is associated with increased 30-day mortality. The majority of hospitals performing MIPD were low volume. These results may suggest that MIPD is a complex procedure for which comprehensive protocols outlining criteria for implementation might be warranted to optimize patient safety.