Validez de la radiografía de tórax para diferenciar etiología bacteriana de viral en ni？os hospitalizados con neumonía

objective. to evaluate the accuracy of the chest radiograph to distinguish between bacterial and viral aetiology in children with pneumonia. population, material and methods. we analyzed the accuracy of chest radiograph for presuming aetiology using khamapirad score in 175 children aged under five years hospitalized for pneumonia in whom etiology was identified (viral= 127 and bacterial= 48). the interpretation of the chest radiograph was made independently by a pediatrician (p) and two pediatric radiologists (r1 - r2), and masked to etiology. kappa coefficient was calculated to within and between observed agreements. t test was applied to analyze the association between etiology (viral or bacterial) and the scores. optimum cut-off value for the score was identified (roc curve). results. we found and outstanding inter-observer agreement (p vs. r1 κ= 0.80; p vs. r2 κ= 0.83). the radiographic score was significantly higher in bacterial pneumonias than in viral pneumonias (3.7 ± 1.1 vs. -1.4 ± 1.2; p= 0.001) with a very good intra observed agreement (κ= 0, 93). a radiographic score ≥1 was identified as the optimum cut-off value to distinguish bacterial from viral pneumonia. at that level the score showed sensitivity: 100% (ic 95%: 90-100); specificity: 98% (ic 95%:93-99); positive predictive value: 96% (ic 95%:85-99); negative predictive value: 100% (ic 95%: 96-100). conclusions. the chest radiograph was accurate for distinguishing bacterial from viral pneumonia in children.