• Rahul S Koti, Kurinchi S Gurusamy, Giuseppe Fusai, and Brian R Davidson.
• Department of Surgery, Royal Free Hospital and University College School of Medicine, Royal Free Hospital, London, UK.
• HPB (Oxford). 2010 Apr 1;12(3):155-65.

BackgroundThe use of synthetic analogues of somatostatin following pancreatic surgery is controversial. The aim of this meta-analysis is to determine whether prophylactic somatostatin analogues (SAs) should be used routinely in pancreatic surgery.MethodsRandomized controlled trials were identified from the Cochrane Library Trials Register, MEDLINE, EMBASE, Science Citation Index Expanded and reference lists. Data were extracted from these trials by two independent reviewers. The risk ratio (RR), mean difference (MD) and standardized mean difference (SMD) were calculated with 95% confidence intervals (95% CIs) based on intention-to-treat or available case analysis.ResultsSeventeen trials involving 2143 patients were identified. The overall number of patients with postoperative complications was lower in the SA group (RR 0.71, 95% CI 0.62-0.82), but there was no difference between the groups in perioperative mortality (RR 1.04, 95% CI 0.68-1.59), re-operation rate (RR 1.15, 95% CI 0.56-2.36) or hospital stay (MD -1.04 days, 95% CI -2.54 to 0.46). The incidence of pancreatic fistula was lower in the SA group (RR 0.64, 95% CI 0.53-0.78). The proportion of these fistulas that were clinically significant is not clear. Analysis of results of trials that clearly distinguished clinically significant fistulas revealed no difference between the two groups (RR 0.69, 95% CI 0.34-1.41). Subgroup analysis revealed a shorter hospital stay in the SA group than among controls for patients with malignant aetiology (MD -7.57 days, 95% CI -11.29 to -3.84).ConclusionsSomatostatin analogues reduce perioperative complications but do not reduce perioperative mortality. However, they do shorten hospital stay in patients undergoing pancreatic surgery for malignancy. Further adequately powered trials of low risk of bias are necessary.

22 keywords...

hide…