Background Prophylactic somatostatin to reduce the incidence of clinically relevant postoperative pancreatic fistula after pancreaticoduodenectomy remains controversial. We assessed the preventive efficacy of somatostatin on clinically relevant postoperative pancreatic fistula in intermediate-risk patients who underwent pancreaticoduodenectomy at pancreatic centres in China. Methods In this multicentre, prospective, randomised controlled trial, we used the updated postoperative pancreatic fistula classification criteria and cases were confirmed by an independent data monitoring committee to improve comparability between centres. The primary endpoint was the rate of clinically relevant postoperative pancreatic fistula within 30 days after pancreaticoduodenectomy. Results Eligible patients (randomised, n = 205; final analysis, n = 199) were randomised to receive postoperative intravenous somatostatin (250 mu g/h over 120 h; n = 99) or conventional therapy (n = 100). The primary endpoint was significantly lower in the somatostatin vs control group (n = 13 vs n = 25; 13% vs 25%, P = 0.032). There were no significant differences for biochemical leak (P = 0.289), biliary fistula (P = 0.986), abdominal infection (P = 0.829), chylous fistula (P = 0.748), late postoperative haemorrhage (P = 0.237), mean length of hospital stay (P = 0.512), medical costs (P = 0.917), reoperation rate (P > 0.99), or 30 days' readmission rate (P = 0.361). The somatostatin group had a higher rate of delayed gastric emptying vs control (n = 33 vs n = 21; 33% vs 21%, P = 0.050). Conclusions Prophylactic somatostatin treatment reduced clinically relevant postoperative pancreatic fistula in intermediate-risk patients after pancreaticoduodenectomy.