Perioperative nasogastric decompression strategies in laparoscopic pancreaticoduodenectomy: a propensity score–matched retrospective cohort study

In this study, we aimed to assess whether postoperative nasogastric tube (NGT) maintenance was associated with better outcomes than postoperative omission and complete perioperative omission and whether complete omission was safe and feasible compared with postoperative omission in laparoscopic pancreaticoduodenectomy (LPD). This single-centre retrospective cohort included 957 adults undergoing LPD from April 2016 to April 2025. Patients were classified according to NGT management as routine placement (RP), postoperative omission (PO), or routine omission (RO). Primary outcomes were clinically relevant delayed gastric emptying (CR-DGE) and Comprehensive Complication Index (CCI). Pairwise 1:1 propensity score matching was performed. Among 957 patients, 324 were RP, 102 PO, and 531 RO. In RP, median initial NGT duration was 3 days (IQR 2–5), with 10.5% retained beyond 7 days. After matching, RO had lower CR-DGE (16.95% vs. 37.29%; P < 0.001) and CCI (23.01 vs. 32.38; P < 0.001), earlier semi-solid diet tolerance (9.04 vs. 10.88 days; P = 0.010), and shorter length of stay (20.53 vs. 22.90 days; P = 0.043) than RP. Restricting CR-DGE to the oral-intake criterion attenuated but did not eliminate the difference (11.02% vs. 25.42%; P = 0.009). PO had lower CCI than RP (27.15 vs. 34.62; P = 0.018), while CR-DGE did not differ significantly (28.12% vs. 42.19%; P = 0.110). CR-DGE and CCI did not differ significantly between RO and PO; rescue NGT insertion was less frequent with RO (9.38% vs. 28.12%; P = 0.006), but pleural effusion was more frequent (82.81% vs. 62.50%; P = 0.007). Postoperative NGT retention was associated with higher CCI than both omission strategies and more CR-DGE than RO. RO and PO did not differ significantly in CR-DGE or CCI, while RO had fewer rescue NGT insertions.