Warning indicators of COVID-19 severity: a retrospective observational study integrating modern biomarkers and traditional tongue features

ObjectiveThis study aims to identify early warning indicators of COVID-19 severity by integrating modern medical biomarkers with traditional Chinese medicine (TCM) tongue features.MethodsA retrospective observational study was conducted on 409 hospitalized COVID-19 patients from two centers in China...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhang Jing, Liu Yuntao, Zheng Danwen, Ye Gangfu, Chen Qiumin, Huang Jianshan, Wang Jiamei, Ma Zengming, Zhang Zhongde
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-04-01
Series:Frontiers in Medicine
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fmed.2025.1500605/full
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ObjectiveThis study aims to identify early warning indicators of COVID-19 severity by integrating modern medical biomarkers with traditional Chinese medicine (TCM) tongue features.MethodsA retrospective observational study was conducted on 409 hospitalized COVID-19 patients from two centers in China. Patients were stratified into severe (n = 50) and non-severe (n = 359) groups based on the 10th edition of China’s diagnostic guidelines. Data included demographics, clinical symptoms, tongue characteristics, and laboratory parameters. Univariate analyses (chi-square/Fisher’s exact tests) and stepwise logistic regression were performed to identify key predictors.ResultsAge (p < 0.001), fever (p < 0.001), elevated procalcitonin (PCT, p < 0.001), thick tongue fur (p = 0.003), and fat tongue shape (p = 0.002) were significant predictors of severity. The combined model integrating these factors demonstrated superior predictive performance (Nagelkerke R2 = 0.741).ConclusionIntegrating TCM tongue features (thick fur and fat shape) with clinical biomarkers (age, fever, and PCT) enhances early identification of severe COVID-19, particularly in resource-limited settings.
ISSN:2296-858X