Background/objectives: Acute pancreatitis is complicated by local and systemic complications in 20-30% of the patients. Accurate prediction of severity may be important for clinical decision making. Our aim is to identify and compare the accuracy of laboratory biomarkers that predict severity and complications in adult patients.
Methods: Medline, EMBASE, Web of Science and Cochrane Library (1993 to August 2020) were searched for studies with an unselected population of patients with acute pancreatitis, that contains accuracy data for ≥1 laboratory biomarker(s) and/or APACHE-II score for the prediction of a patient outcomes of interest during the first 48 h of admission. The primary outcome is moderate severe or severe acute pancreatitis (MSAP/SAP). Secondary outcomes are severe acute pancreatitis, pancreatic necrosis and organ failure. Risk of bias was assed using QUADAS-2. Biomarkers extracted from ≥3 unique sources, were analyzed using hierarchical summary receiver operating characteristic (HSROC) and bivariate model analysis.
Results: In total, 181 studies were included in the qualitative analysis reporting on 29 biomarkers. For the primary outcome at admission, summary sensitivities and specificities were, respectively, 87% (95% CI 69-95%) and 88% (95% CI 80-93%) for IL-6 at a threshold of >50 pg/ml, 72% (95% CI 64-79%) and 76% (95% CI 67-84%) for an APACHE-II score of ≥8, and 53% (95% CI 35-71%) and 82% (95% CI 74-88%) for CRP >150 mg/l. HSROC curve analysis confirmed these results.
Conclusion: This study indicates superiority of IL-6 for the early prediction of MSAP/SAP and may be used for to guide clinical decision making.
Keywords: Acute pancreatitis; Biomarkers; Prediction studies.
Copyright © 2020 IAP and EPC. Published by Elsevier B.V. All rights reserved.