Abstract
IMPORTANCE: Morbidity is still high in pancreatic surgery, driven mainly by gastrointestinal complications such as pancreatic fistula. Perioperative thoracic epidural analgesia (EDA) and patient-controlled intravenous analgesia (PCIA) are frequently used for pain control after pancreatic surgery. Evidence from a post hoc analysis suggests that PCIA is associated with fewer gastrointestinal complications. OBJECTIVE: To determine whether postoperative PCIA decreases the occurrence of gastrointestinal complications after pancreatic surgery compared with EDA. DESIGN, SETTING, AND PARTICIPANTS: In this adaptive, pragmatic, international, multicenter, superiority randomized clinical trial conducted from June 30, 2015, to October 1, 2017, 371 patients at 9 European pancreatic surgery centers who were scheduled for elective pancreatoduodenectomy were randomized to receive PCIA (n = 185) or EDA (n = 186); 248 patients (124 in each group) were analyzed. Data were analyzed from February 22 to April 25, 2019, using modified intention to treat and per protocol. INTERVENTIONS: Patients in the PCIA group received general anesthesia and postoperative PCIA with intravenous opioids with the help of a patient-controlled analgesia device. In the EDA group, patients received general anesthesia and intraoperative and postoperative EDA. MAIN OUTCOMES AND MEASURES: The primary end point was a composite of pancreatic fistula, bile leakage, delayed gastric emptying, gastrointestinal bleeding, or postoperative ileus within 30 days after surgery. Secondary end points included 30-day mortality, other complications, postoperative pain levels, intraoperative or postoperative use of vasopressor therapy, and fluid substitution. RESULTS: Among the 248 patients analyzed (147 men; mean [SD] age, 64.9 [10.7] years), the primary composite end point did not differ between the PCIA group (61 [49.2%]) and EDA group (57 [46.0%]) (odds ratio, 1.17; 95% CI, 0.71-1.95 P = .54). Neither individual components of the primary end point nor 30-day mortality, postoperative pain levels, or intraoperative and postoperative substitution of fluids differed significantly between groups. Patients receiving EDA gained more weight by postoperative day 4 than patients receiving PCIA (mean [SD], 4.6 [3.8] vs 3.4 [3.6] kg; P = .03) and received more vasopressors (46 [37.1%] vs 31 [25.0%]; P = .04). Failure of EDA occurred in 23 patients (18.5%). CONCLUSIONS AND RELEVANCE: This study found that the choice between PCIA and EDA for pain control after pancreatic surgery should not be based on concerns regarding gastrointestinal complications because the 2 procedures are comparable with regard to effectiveness and safety. However, EDA was associated with several shortcomings. TRIAL REGISTRATION: German Clinical Trials Register: DRKS00007784.