Background: Many children undergo chest radiography (CXR) in their evaluation of a febrile illness. Pneumonia without signs of respiratory distress or ausculatory findings has been previously described (termed occult pneumonia [OP]).
Objective: The objectives of this study were to determine the incidence of OP among children who have CXR performed and to identify clinical predictors of OP.
Methods: A prospective observational study of children undergoing CXR for possible pneumonia was conducted. Standardized data forms were completed before the CXR. Univariate analysis and recursive partitioning were used to identify predictors of OP.
Results: Of 1866 patients enrolled, 308 had no evidence of respiratory distress or lower respiratory tract findings and were studied for OP. Twenty-one patients had radiographic OP (6.8%; 95% confidence interval [CI], 4.0%-10.6%). Age, height of fever, duration or quality of cough, and pulse oximetry were not associated with OP. A decision rule based on fever for 1 day or longer or with a combination of fever for less than 1 day but worsening cough identifies patients at greater risk for OP (likelihood ratio, 1.47; 95% CI, 1.21-1.77). No patient with fever for less than 1 day and without any cough or without worsening cough had pneumonia (likelihood ratio, 0.40; 95% CI, 0.19-0.84).
Conclusions: Occult pneumonia was identified in 1 of 15 patients undergoing CXR without respiratory distress or ausculatory findings. Obtaining a CXR for the detection of OP in children without cough and with fever for less than 1 day in duration should be discouraged.