Utility of serum amyloid A for predicting bacterial pneumonia with coronavirus disease 2019 pneumonia

Park, Hee Geun; Kwak, Sae Jeung; Yun, Sung Hyun

J Korean Soc Emerg Med > Volume 37(3); 2026 > Article

Original Article

Journal of The Korean Society of Emergency Medicine 2026;37(3): 133-142.

코로나바이러스감염증-19 폐렴 환자에서 세균성 폐렴 예측을 위한 혈청 아밀로이드 A의 유용성

박희근

, 곽세정

, 윤성현

가톨릭관동대학교 국제성모병원 응급의학과

Utility of serum amyloid A for predicting bacterial pneumonia with coronavirus disease 2019 pneumonia

Hee Geun Park

, Sae Jeung Kwak

, Sung Hyun Yun

Department of Emergency Medicine, Catholic Kwandong University International ST. Mary’s Hospital, Incheon, Korea

Correspondence

Sung Hyun Yun ,Tel: 032-290-3030, Fax: 032-290-2886, Email: jimin0415@ish.ac.kr,

Received: May 17, 2025; Revised: November 3, 2025 Accepted: December 3, 2025. Published online: June 30, 2026.

ABSTRACT

Objective:
This study evaluated the role of serum amyloid A (SAA) and other inflammatory biomarkers in the early prediction of bacterial pneumonia coinfection among patients with coronavirus disease 2019 (COVID-19) pneumonia.
Methods:
Three hundred and two adult patients diagnosed with COVID-19 pneumonia from a single center emergency department between 2020 and 2024 were analyzed retrospectively. The patients were stratified into COVID-19 monoinfection (n=249) and bacterial coinfection (n=53) groups based on sputum culture results. The clinical characteristics, inflammatory markers-including white blood cell (WBC) count, neutrophils, lymphocytes, neutrophil-to-lymphocyte ratio (NLR), erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and SAA-and the clinical outcomes were assessed. Procalcitonin was excluded because of the limited availability in the emergency setting. Logistic regression and receiver operating characteristic curve analyses were performed to identify the independent predictors and optimal cut-off values.
Results:
The bacterial coinfection group showed significantly higher levels of WBC, neutrophils, NLR, ESR, CRP, and SAA (all P<0.001) and lower lymphocyte counts (P=0.038). SAA was the sole independent predictor in multivariate analysis (odds ratio, 1.15; 95% confidence interval, 1.10-1.21; P<0.001). ROC analysis showed excellent diagnostic performance of SAA with an area under the curve, sensitivity, and specificity of 0.907, 96.2%, and 80.7%, respectively. Commonly isolated pathogens included Streptococcus pneumoniae, Klebsiella pneumoniae, Staphylococcus aureus, and Pseudomonas aeruginosa. A coinfection was associated with increased mechanical ventilation, intensive care unit admission, mortality, and prolonged hospitalization.
Conclusion:
SAA showed superior diagnostic utility over traditional biomarkers for the early identification of bacterial pneumonia coinfection in COVID-19 patients, facilitating prompt management and potentially improving clinical outcomes.

Key words: COVID-19; Bacterial pneumonia; Serum amyloid A protein