1400. Retractor Related Liver Injuries After Pancreaticoduodenectomy: A Retrospective Analysis of a Single Institution Experience
Authors * Denotes Presenting Author
  1. William Law *; Dartmouth Hitchcock Medical Center
  2. Qingyuan Song; Dartmouth Hitchcock Medical Center
  3. Kerrington Smith; Dartmouth Hitchcock Medical Center
  4. Robert Percarpio; Dartmouth Hitchcock Medical Center
  5. Nancy McNulty; Dartmouth Hitchcock Medical Center
To determine the risk factors, incidence, severity, reporting rates, and outcomes of retractor related liver injuries (RRLI) after pancreaticoduodenectomy.

Materials and Methods:
The records of all pancreaticoduodenectomies performed at a single institution from 1/2014 to 12/2019 were analyzed. Inclusion criteria were preoperative CT or MRI within 1 year of surgery and postoperative CT or MRI within 30 days of surgery. Data was collected via review of the electronic medical record. Preoperative imaging was reviewed to identify hepatic steatosis. Postoperative imaging was reviewed by two abdominal radiologists with 6 years (RP) and 19 years (NM) experience and liver injuries were graded using American Association for Surgery of Trauma (AAST) guidelines.

Of 230 pancreaticoduodenectomies performed, 109 pancreaticoduodenectomy patients were included in the study. There were 100 open, 6 robotic, and 3 combined cases. RRLI occurred in 24 (22%) of patients. Of these RRLI, 20 (83.3%) were in the open group and 4 (16.7%) were in the robotic group. Types included: laceration 9/24 (37.5%), subcapsular hematoma 2/24 (8.3%), and intraparenchymal hematoma 13/24 (54.2%). Liver injury grade distribution was: I – 3/24 (12.5%), II – 18/24 (75%), III – 3/24 (12.5%), IV – 0/24 (0%), and V – 0/24 (0%). Only 15/24 (62.5%) were reported on initial CT interpretation. Body mass index and hepatic steatosis were not significant predisposing factors for liver injury. A lower platelet count was seen in patients with injuries (median 219 versus 250, p=0.023). The operations were significantly longer in the injury group (mean 545 minutes versus 496 minutes, p=0.019) and postoperative AST (U/L) and ALT (U/L) were significantly elevated [median AST 221 versus 70.5 (p<0.001) and ALT 236 versus 66 (p<0.001)]. No significant difference in hospital length of stay or post-operative pain scores were noted. Out of all the patients that met inclusion criteria there was one death within 30 days from the noninjury group.

RRLI have been reported after open and laparoscopic abdominal surgeries.1–3 Our study is the first to fully characterize a large group of patients for RRLI after pancreaticoduodenectomy. RRLI occur relatively frequently; however, are often unrecognized by both surgical and radiology teams. The majority of injuries were AAST grade I or II and no changes in patient outcomes were identified in our study. Raising awareness of the frequency and types of liver injury in pancreaticoduodenectomy patients is important for surgical and radiology teams and further study may aid in prevention and improved outcomes.