Clinical characteristics analysis and predictive diagnostic model building of community-acquired pneumonia in adults requiring hospitalisation in Fujian Provincial Hospital

Abstract

Abstract Background We compared the clinical characteristics of infections caused by different pathogens and established a viral/bacterial infection prediction model to guide early clinical identification of pathogens among inpatients with community-acquired pneumonia (CAP). Methods Data were analysed to establish a prediction model for the early treatment of bacterial or viral infections. Basic data, clinical symptoms, laboratory examinations, and imaging of patients were collected and compared, and the virus/bacteria prediction equation was established. Results The proportion of patients with muscle soreness and headaches was significantly higher in the viral infection group than in the bacterial infection group. Procalcitonin (PCT) levels, erythrocyte sedimentation rate (ESR), and neutrophil alkaline phosphatase (NAP) scores were significantly higher in the bacterial infection group than in the viral infection group. The creatine kinase level was significantly higher in the viral infection group than in the bacterial infection group (P < 0.05). More patients in the atypical pathogen infection group (up to 52.0%) had real lung degeneration, and the difference was statistically significant compared with other groups (P < 0.005). Patchy shadows were more common in the viral infection group than in other groups (up to 92.5%). There were significant differences in the PCT levels and the presence of fever or muscle soreness between the groups. A binary logistic regression equation was obtained, which could predict the probability of viral infection (sensitivity 57.5%, specificity 67.7%, area under the ROC curve 0.651). Conclusions Adult CAP patients with viral infection are more likely to have headaches and muscle soreness than those with bacterial infection. An elevated PCT level, NAP score, and ESR indicated a high possibility of bacterial infection. Accordingly, a viral and bacterial infection prediction model was established.