This study employed longitudinal data-independent acquisition drain-fluid proteomics to profile postoperative drainage in two surgical cohorts with the objective of defining procedure-specific signatures of anastomotic leakage and identifying candidate biomarkers beyond routinely measured proteins. The cohorts comprised 33 pancreatic surgery patients with 201 drain-fluid samples and 34 colorectal surgery patients with 143 drain-fluid samples. Sampling covered an observation window of up to 10 days postoperatively.
The approach combined evaluation of predefined, established biomarkers across postoperative time with an unrestricted discovery analysis to detect additional proteins and temporal patterns associated with leakage. The authors emphasize that the data are derived from a preprint and have not yet been peer reviewed.
In the pancreatic surgery cohort, established enzyme markers—particularly AMY2A (amylase isoform) and PNLIP (pancreatic lipase)—showed strong early discrimination between patients who developed leakage and those who did not. The discrimination by these enzyme markers was most pronounced in the early postoperative period (post-operative day [POD] 1–3).
The pattern observed suggests that enzyme enrichment in drain fluid is an early indicator of pancreatic anastomotic disruption, with subsequent proteomic changes unfolding over the first postoperative week.
Applying the same predefined marker strategy to the colorectal surgery cohort produced less consistent results. Established enzyme markers that performed well in pancreatic patients did not robustly discriminate leak from non-leak in colorectal patients across the postoperative timeline reported.
This inconsistency indicates that biomarker dynamics and the biological processes underlying leakage differ by surgical procedure, and that strategies effective in one context (pancreatic surgery) may not translate directly to another (colorectal surgery).
Discovery analysis of longitudinal proteomic data revealed divergent temporal patterns between the two procedures. In pancreatic leakage, the dominant early event (POD 1–3) was enzyme enrichment, followed by increases in proteins associated with inflammation, antimicrobial activity, oxidative stress, and matrix remodeling. These successive changes outline a trajectory from enzymatic leakage to secondary tissue and immune responses.
In contrast, colorectal leakage showed its strongest proteomic divergence around POD3 and was characterized primarily by reduced abundance of proteins linked to wound healing. The depleted proteins included factors involved in endothelial integrity, extracellular matrix organization, complement regulation, and coagulation balance. Rather than an early enzyme-dominant signal, colorectal leakage appeared to reflect a failure or alteration in reparative and hemostatic processes that becomes most apparent around the third postoperative day.
From the procedure-specific discovery patterns, the authors nominated novel candidate markers for future validation. For pancreatic surgery, PRSS2 and AZU1 emerged as additional candidates arising from the early enzyme and immune-associated signatures. For colorectal surgery, FBLN1 and CDH5 were identified among proteins linked to wound healing and endothelial function that showed reduced abundance in leaking patients.
The report frames these candidates as hypotheses for follow-up validation studies to establish diagnostic performance and clinical utility; no diagnostic performance metrics or validation outcomes are reported in the source material.
The full proteomic datasets supporting the analysis are available on the PRIDE database (PXD082170 and PXD082236). Additional information and data are available upon reasonable request to the corresponding author.
Ethics approvals for sample collection and proteomic analysis were reported from the ethics committee of the University Hospital in Pilsen (Ref. No.: 509/2020; proteomic analysis Ref. No.: 50/24) and from the cantonal ethics committee BASEC (collection BASEC no. Req-2022-01026; proteomic analysis BASEC no. Req-2026-01175). The authors state that all necessary patient consent was obtained and appropriate institutional forms archived.
Competing interests declared that two authors (Alexander Jessernig and Inge K. Herrmann) are inventors on a patent for anastomotic leak detection filed by ETH Zurich and Empa (EP23203384). Other authors declared no conflicts.
The authors conclude that postoperative anastomotic leakage follows procedure-specific proteomic trajectories and that longitudinal drain-fluid proteomics provides a biologically informed framework for biomarker discovery beyond routinely measured proteins. Key practical observations from the source material are that enzyme markers such as AMY2A and PNLIP are promising early discriminators for pancreatic leakage, while colorectal leakage may require alternative marker strategies focused on wound-healing and endothelial integrity pathways.
The article emphasizes that the nominated candidates (PRSS2, AZU1, FBLN1, CDH5) require follow-up validation studies to determine diagnostic potential. Because the report is a medRxiv preprint and not peer reviewed, the authors note it should not be used to guide clinical practice until validated and externally confirmed.