RESUMO
BACKGROUND: A recent RCT has shown that routine intraoperative drain placement after pancreatoduodenectomy (PD) is not necessary. The aim was to confirm this in real-world conditions. METHODS: A propensity score-matched (PSM) analysis of patients with and without drainage after PD or distal pancreatectomy (DP) at a high-volume centre was undertaken. Main outcomes were complications and duration of hospital stay. Multivariable regression analysis was used to assessed which factors led to intraoperative drainage after the standard was changed from drain to no drain. RESULTS: Of 377 patients who underwent PD, 266 were included in PSM analysis. No drain was associated with fewer major complications (42 (31.6 per cent) versus 62 (46.6 per cent); P = 0.017), shorter duration of hospital stay (mean(s.d.) 14.7(8.5) versus 19.6(14.9) days; P = 0.001), and required fewer interventional drain placements (8.4 versus 19.8 per cent; P = 0.013). In PSM analysis after DP (112 patients), no drainage was associated with fewer clinically relevant postoperative pancreatic fistulas (9 versus 18 per cent; P = 0.016), fewer overall complications (mean(s.d.) comprehensive complication index score 15.9(15.4) versus 24.8(20.4); P = 0.012), and a shorter hospital stay (9.3(7.0) versus 13.5(9.9) days; P = 0.011). Multivisceral resection (OR 2.80, 95 per cent c.i. 1.10 to 7.59; P = 0.034) and longer operating times (OR 1.56, 1.04 to 2.36; P = 0.034) influenced the choice to place a drain after PD. Greater blood loss was associated with drainage after DP (OR 1.14, 1.02 to 1.30; P = 0.031). CONCLUSION: Standard pancreatic resections can be performed safely without drainage. Surgeons were more reluctant to omit drainage after complex pancreatic resections.
Pancreatic surgery has traditionally relied on the use of drains placed during surgery that should facilitate outflow of fluids from the operating site. This principle has recently been challenged by specially designed studies showing that patients who do not receive a drain may have fewer complications. The present study has demonstrated that these results also apply to routine clinical settings outside the constraints of surgical trials.