Distinguishing Between Bacterial and Viral Community-Acquired Pneumonia in Hospitalized Adults: The Importance of...

Our study aims to compare patients with identified viral and bacterial pathogens and develop a predictive model for determining community-acquired pneumonia (CAP) etiology.

In this retrospective case-control study, hospitalized adults with confirmed bacterial or viral CAP were compared at a 1:3 ratio. Patients were matched by gender, hospitalization department, age, and Charlson comorbidity index. The etiology of CAP was established through blood and sputum culture, polymerase chain reaction of respiratory specimens, and urinary antigen tests. Both objective variables and subjective symptoms were assessed using multivariate conditional logistic regression.

Our study included 400 patients. Using the best-developed model, we found that patients with bacterial CAP were more likely to exhibit unilateral infiltration on visualization (odds ratio [OR] 20.47; 95% confidence intervals [CI; 7.51-55.81], p < 0.001), experience chills (OR 20.69; 95% CI [6.08-70.38], p < 0.001), have a higher heart rate (OR 1.04; 95% CI [1.01-1.06], p = 0.001), and demonstrate an elevated neutrophils/lymphocytes ratio (OR 1.58; 95% CI [1.16-2.14], p = 0.003). They also showed a decreased level of consciousness (OR 3.16; [1.16-8.61], p = 0.024), and were more likely to receive vasopressors within the first 24 h of admission (OR 8.28; 95% CI [1.75-39.22], p = 0.008). Conversely, they were less likely to have general weakness (OR 0.20; 95% CI [0.09-0.47], p < 0.001) and tended to have lower total serum protein levels (OR 0.06; 95% CI [0.01-0.45], p = 0.006). This model showed sensitivity of 90%, specificity of 91%, and positive predictive value of 97%.

The developed model offers a valuable tool for assessing the etiology of CAP upon admission.