Neutrophil-lymphocyte ratio as a point-of-care marker for predicting bacterial etiology in pediatric community-acquired Pneumonia: A comparative analysis with C -reactive protein

Objectives: To evaluate the role of NLR in predicting bacterial etiology in pediatric community acquired pneumonia and compare its diagnostic performance with CRP. Methods: This retrospective analytical study was conducted in a tertiary care center in South India. The clinico-laboratory data of children with pneumonia were reviewed. The diagnostic performance was evaluated using appropriate statistical methods. Results: Among 514 children with community-acquired pneumonia, 398 were included. Bacterial pneumonia was diagnosed in 250 (62.8 %) children. The median NLR among children with bacterial pneumonia was 3.3 (2.18, 5.7) compared to 1.3 (0.685, 2.595) among children with non-bacterial pneumonia [p < 0.001]. A CRP cut-off value of 31.5 mg/L had a sensitivity of 85.2 % and a specificity of 93.2 %, respectively [AUC 0.923; 95 % CI (0.895, 0.952); p < 0.001]. The positive and negative predictive values of CRP to predict bacterial pneumonia were 95.52 % and 78.86 %, respectively, while the positive and negative likelihood ratios were 12.61 (6.92, 22.99) and 0.16 (0.12, 0.21), respectively. An NLR cut-off of 1.74 had a sensitivity of 87.6 % and a specificity of 62.8 %, respectively [AUC 0.78, 95 % CI (0.729, 0.831); p < 0.001]. The positive and negative predictive values of NLR to predict bacterial pneumonia were 79.93 % and 75 %, respectively, while the positive and negative likelihood ratios were 2.36 (1.90, 2.92) and 0.19 (0.14, 0.28), respectively. Bayesian reasoning on positive NLR shifted the post-test probability of bacterial pneumonia to 79.9 %. Conclusion: NLR has moderate specificity and should be viewed primarily as a sensitive, accessible, and rapid screening or triage tool, especially useful in resource-limited settings.