Meta-analysis of randomized controlled trials comparing minimal-invasive versus open pancreatic surgery

Abstract

Abstract Objective To compare perioperative outcomes of minimal-invasive pancreatic surgery (MIPS) to open pancreatic surgery (OPS) using data obtained from randomised controlled trials (RCTs). Methods Literature search was done in CENTRAL, Medline and Web of Science; all currently available RCTs comparing MIPS and OPS in adult patients requiring elective distal pancreatectomy (DP) or partial pancreatoduodenectomy (PD) were included in the analysis. Outcomes were mortality, general and pancreatic surgery specific morbidity and length of hospital stay (LOS). Subgroup analysis between PD and DP was performed. A random-effects model for Mantel-Haenzsel and inverse variance analysis was used. Cochrane RoB 2.0 tool and GRADE approach was used for assessment of risk of bias and certainty of evidence. Results A total of 166 patients for DP (80 MIPS vs. 86 OPS) in two RCTs and 856 for PD (429 MIPS vs. 427 OPS) in five RCTs were analysed. Meta-analysis showed a reduction of LOS by 1.2 days (-2 to -0.3, p<0.01) and intraoperative blood loss by 143 ml (-186 to -101, p<0.01) in MIPS. In the subgroup analysis reduction of LOS was only present for minimal-invasive DP (-2 days, -2.3 to -1.7, p<0.01). A minimal-invasive approach showed reduction of surgical site infections (odds ratio 0.4, 0.1 to 0.96, p=0.04) and intraoperative blood loss (-141 ml, -181 to -101, p<0.01) only in PD. Duration of surgery was about 96 min longer in minimal-invasive PD (56 to 137 min, p<0.01), but not in DP. Analysis showed no significant difference between MIPS and OPS, as well as between subgroups, in mortality, complications according to Clavien-Dindo >=3 and pancreatic surgery specific complications like postoperative pancreatic fistula and delayed gastric emptying. Certainty of evidence according to the GRADE approach was moderate to low. Conclusion This meta-analysis of level 1 evidence shows reduced LOS and intraoperative blood loss in MIPS compared to OPS. Morbidity and mortality did not differ between MIPS and OPS. Longer duration of surgery in minimal-invasive PD hints that minimal-invasive technique in PD is more challenging than in DP. However, patients undergoing minimal-invasive PD seem to benefit from lower surgical site infections and less intraoperative blood loss.