K. E. Cunningham2, M. S. Zenati2, J. Petrie3, J. Steve2, M. E. Hogg2, H. J. Zeh2, A. H. Zureikat2 2University Of Pittsburgh Medical Center,Surgical Oncology,Pittsburgh, PA, USA 3University Of Pittsburgh Medical Center,Finance,Pittsburgh, PA, USA
Introduction: Immediate post-operative admission to the ICU following pancreaticoduodenectomy (PD) has been standard of care for many institutions. Over the last decade, minimally invasive pancreaticoduodenectomy has emerged as safe and feasible. The aim of this project was to examine if omission of an immediate post-operative ICU admission would be safe and result in decreased length of stay (LOS) and reduced cost for patients undergoing robotic pancreaticoduodenectomy (RPD).
Methods: From December 2014 to June 2015, a non-ICU admission policy on post-operative day zero (POD0) was implemented for all patients undergoing RPD. Prior to this date, all RPDs were routinely admitted to the ICU on POD0. Using a prospectively maintained database, we compared the outcomes of the non-ICU RPD cohort to patients routinely admitted to the ICU post-operatively prior to implementation of this policy (January 2014-November 2014). All cases were analyzed on an intent-to-treat basis, thereby minimizing selection bias.
Results: The ICU cohort (n=49, average age 65.6 ±12, 51% females) and non-ICU cohort (n=34, average age 66.4 ± 9, 38% females) were comparable with no statistically significant differences with respect to age, sex, BMI, CCI and ASA score, pre-operative tumor size, diagnosis, receipt of neoadjuvant therapy, operative time, and estimated blood loss. Seven patients (21%) from the non-ICU group were directly admitted to the ICU post-operatively based on unpredicted changes in intra-operative clinical status necessitating ICU care. They were included in the non-ICU group and their outcomes were analyzed on an intent-to-treat basis. The rates of Clavien complications, pancreatic leak, reoperation, readmission and mortality demonstrated no statistically significant difference between both groups (Table 1). A statistically significant trend toward reduced total hospital LOS in the non-ICU group was noted (median 6.95 days versus 7.7 days, P=0.083). This reduced LOS and avoidance of routine post-operative ICU admission translated into a cost reduction from $25,812 (IQR $19,875 -$29,853) in the ICU group to $19,739 (IQR $17,964 -$25,521) in the non-ICU group, P=0.01. The reduction in cost remained statistically significant even after adjusting for all related demographics and perioperative characteristics.
Conclusion: A standard policy of omitting a post-operative ICU admission on POD0 following RPD is safe and can result in reduced length of stay and overall savings in total hospital cost.
