M. Radomski1, A. Zureikat1, S. M. Novak1, J. Steve1, J. Marsh1, K. K. Lee1, A. Tsung1, D. Bartlett1, H. J. Zeh1, M. E. Hogg1 1University Of Pittsburgh,Pittsburgh, PA, USA
Introduction: In this healthcare climate, much scrutiny is being paid to cost, readmission, and length of stay (LOS). Complex pancreatic surgeries have been associated with prolonged postoperative courses. Many studies have looked at overall data and low outliers for pancreaticoduodenectomy. However, little exists on high outliers and factors that contribute to prolonged stays. We sought to evaluate the contribution of social, preoperative, operative, and postoperative patient related factors to extended LOS in major pancreatic resections.
Methods: A retrospective review of a single institution’s pancreaticoduodenectomies (PD) and distal pancreatectomies (DP) was performed from 6/2009 to 3/2014 for all pathologies and technical approaches. Interquartile ranges (IQR) were calculated and the highest quartile was evaluated and compared to the lower three quartiles.
Results:A total of 350 PD (42% open) and 127 DP (21% open) patients were analyzed with a 3rd IQR of >14 and >8 days respectively. Social factors including race, distance, insurance status, and marital status were not significant for PD; however, divorce was associated with longer LOS in DP (p<0.0001). Pre-operative characteristics of higher age (p=0.0003), age adjusted Charlson Comorbidity Index (p=0.002), body mass index (p=0.01), and American Society of Anesthesia assessment (ASA, p=0.005) were associated with increased LOS in PD; whereas, only higher ASA (p=0.0188) was associated in DP. Albumin, Ca19-9, previous abdominal surgery, and neoadjuvant therapy were not significant for PD or DP. Increased operative time (p=0.009), blood loss (EBL, p=0.03), and transfusion (p=0.03) all were associated with longer LOS in PD; however, only EBL (p=0.03) was associated with longer LOS in DP. A trend toward more LOS outliers was seen in the open PD group (p=0.06) compared with robotic, but not in the DP group (p=0.63). For the PD group, pancreatic fistula (p<0.0001), delayed gastric emptying (p<0.0001), and pseudoaneurysm (p<0.0001) were associated with extended hospitalizations. No specific post-operative complications led to increased LOS in the DP group but the high LOS group was more likely to have had any complication (p=0.007) compared to the rest of the cohort. High LOS outliers were more likely to go to skilled nursing facilities, acute care facilities, and rehab than home in both PD (p=<0.0001) and DP (p=0.005) groups. The high LOS group had more readmission in the PD group (p=0.005) but not in the DP group (p=0.64).
Conclusion:More patient pre-, intra-, and post-operative factors lead to high LOS outliers seen in the PD group than the DP group; but the DP group also had a social factor associated with increased LOS. Pre- and Intraoperative factors are hard to change, but further subgroup analysis in the PD group looking at management of specific complications and physician related factors may help identify better or sooner management to decrease LOS for these factors.