Drain fluid biomarkers for the diagnosis of clinically relevant postoperative pancreatic fistula: a diagnostic accuracy systematic review and meta-analysis

Abstract

Background: Pancreatectomy is the only curative treatment available for pancreatic cancer and a necessity for patients with challenging pancreatic pathology. To optimize outcomes, postsurgical complications such as clinically relevant postoperative pancreatic fistula (CR-POPF) should be minimized. Central to this is the ability to predict and diagnose CR-POPF, potentially through drain fluid biomarkers. This study aimed to assess the utility of drain fluid biomarkers for predicting CR-POPF by conducting a diagnostic test accuracy systematic review and meta-analysis. Methods: Five databases were searched for relevant and original papers published from January 2000 to December 2021, with citation chaining capturing additional studies. The QUADAS-2 tool was used to assess the risk of bias and concerns regarding applicability of the selected studies. Results: Seventy-eight papers were included in the meta-analysis, encompassing six drain biomarkers and 30 758 patients with a CR-POPF prevalence of 17.42%. The pooled sensitivity and specificity for 15 cut-offs were determined. Potential triage tests (negative predictive value >90%) were identified for the ruling out of CR-POPF and included postoperative day 1 (POD1) drain amylase in pancreatoduodenectomy (PD) patients (300 U/l) and in mixed surgical cohorts (2500 U/l), POD3 drain amylase in PD patients (1000–1010 U/l) and drain lipase in mixed surgery groups (180 U/l). Notably, drain POD3 lipase had a higher sensitivity than POD3 amylase, while POD3 amylase had a higher specificity than POD1. Conclusions: The current findings using the pooled cut-offs will offer options for clinicians seeking to identify patients for quicker recovery. Improving the reporting of future diagnostic test studies will further clarify the diagnostic utility of drain fluid biomarkers, facilitating their inclusion in multivariable risk-stratification models and the improvement of pancreatectomy outcomes.