Impact of dynamic risk stratification for postoperative pancreatic fistula using serum pancreatic amylase on the day after pancreatoduodenectomy.

Kato, Tomotaka; Murase, Yoshiki; Baba, Yasutaka; Takase, Kenichiro; Watanabe, Yuichiro; Okada, Katsuya; Aikawa, Masayasu; Okamoto, Kojun et al. · Surgery · 2025

retrospective_cohort · Level III

Where this comes from

Abstract

An appropriate combination of various biomarkers, including serum pancreatic amylase, drain amylase, and C-reactive protein levels, is pivotal for assessing the dynamic changes in patient status after pancreatoduodenectomy. This study aimed to establish a dynamic risk stratification for clinically relevant postoperative pancreatic fistula using clinically applicable parameters on postoperative days 1 to 3. Data from 234 patients who underwent pancreatoduodenectomy were analyzed. Multivariate analyses were conducted using biomarkers available on postoperative days 1 to 3. Clinically relevant postoperative pancreatic fistula, grade C postoperative pancreatic fistula, and postpancreatectomy hemorrhage were observed in 56 (24%), 12 (5%), and 24 (10%) patients, respectively. Although the prognostic value of drain amylase and C-reactive protein improved over time, serum pancreatic amylase levels on postoperative days 1 and 2 were equally predictive of clinically relevant postoperative pancreatic fistula. Patients with high levels of serum pancreatic amylase had significantly high fistula risk scores. Multivariate analysis revealed that serum pancreatic amylase ≥75 U/L and drain amylase ≥2,000 U/L on postoperative day 1 were independent risk factors (odds ratio: 18.0 and 3.45, respectively). Adding C-reactive protein on postoperative day 2 to the stratification on postoperative day 1 improved the diagnostic accuracy (odds ratio: 3.19). Finally, risk assessment on postoperative day 3, using serum pancreatic amylase ≥75 U/L on postoperative day 1, drain amylase ≥5,000 U/L on postoperative day 3, and C-reactive protein ≥22 mg/L on postoperative day 3, identified 75% of patients at highest risk for clinically relevant postoperative pancreatic fistula (odds ratio: 19.7, 5.41, and 5.01, respectively), with an accuracy of 87%. Dynamic risk stratification using serum pancreatic amylase on postoperative day 1 gradually increased the accuracy of identifying high-risk patients from postoperative days 1 to 3. Using suitable biomarkers, including those at each perioperative stage, can facilitate the early identification of high-risk patients.

Medical subject headings