Transection level matters: A prospective study on pancreatic fistula risk after left pancreatectomy.

BACKGROUND

Postoperative pancreatic fistula remains the main cause of morbidity after left pancreatectomy, but the impact of pancreatic transection level on postoperative outcomes remains debated.

METHODS

We retrospectively analyzed a prospectively maintained single-center database of consecutive patients undergoing left pancreatectomy for malignant, premalignant, or benign disease between 2016 and 2024. Transection level was classified as standard pancreatic body-tail resection, pancreatic neck-body-tail resection/extended pancreatic neck-body-tail resection, or pancreatic tail resection. The primary end point was clinically relevant postoperative pancreatic fistula. Secondary end points included postpancreatectomy fluid collections, organ/space surgical site infection, postpancreatectomy hemorrhage, major complications, mortality, and endocrine/exocrine insufficiency. Associations were assessed using multivariable logistic regression, propensity-score matching, sensitivity analyses, and concordance index comparison.

RESULTS

Among 971 patients, transection was pancreatic body-tail resection in 661 (68.1%), pancreatic neck-body-tail resection/extended pancreatic neck-body-tail resection in 116 (11.9%), and pancreatic tail resection in 194 (20.0%). Postoperative pancreatic fistula occurred in 205 patients (21.1%) and was more frequent after nonstandard transections (pancreatic body-tail resection, 17.3%; pancreatic neck-body-tail resection/extended pancreatic neck-body-tail resection, 31.9%; pancreatic tail resection, 27.8%; P < .001). In multivariable analysis, pancreatic neck-body-tail resection/extended pancreatic neck-body-tail resection (odds ratio, 1.87; 95% confidence interval, 1.05-3.28) and pancreatic tail resection (odds ratio, 2.07; 95% confidence interval, 1.19-3.58) were associated with postoperative pancreatic fistula versus pancreatic body-tail resection. The association persisted after propensity-score matching but was attenuated after adjustment for pancreatic thickness and stump management. Pancreatic neck-body-tail resection/extended pancreatic neck-body-tail resection was associated with postpancreatectomy hemorrhage, and both nonstandard transections were associated with organ/space surgical site infection.

CONCLUSION

Transection level was associated with postoperative pancreatic fistula, organ/space surgical site infection, and postpancreatectomy hemorrhage after left pancreatectomy, supporting its role as an intraoperative marker of anatomic and technical complexity.