INTRODUCTION: Chest X-ray (CR) utility to predict etiology in community-acquired pneumonia (CAP), in children has been controversial. Nevertheless, some authors propose the use of well standardized radiological patterns. OBJECTIVE: To evaluate usefulness of modified Swischuk's radiological patterns (RaP) to determine probable etiology in children hospitalized with CAP. PATIENTS AND METHODS: Eighty children were studied using 6 standardized RaP. The RaP were blindly and individually analyzed by 9 pneumologists, who registered the results (stage 1). Thereafter, (stage 2) a second evaluation adding clinical information was performed. Then, the patients complementary examinations and clinical evolution were included (stage 3). The reference standard (RS) was generated from a blind consensus. Every result was compared with the RS using Student test. RESULTS: According to RS, children were classified as having a virus-like pneumonia in 63% of cases, bacterial in 13%, mixed in 16%, atypical in 5% and ADV in 3%. The agreement of stage 1 and 2 with RS was 64 and 77%, respectively. Virus and bacterial RaP agreement increased from 66 to 82 % (p < 0.001) and from 82 to 90% (p < 0.05), respectively after incorporating clinical parameters. CONCLUSION: Modified Swischuk's RaP used in association with clinical elements allows a quite satisfactory approach to etiologic diagnosis of CAP.
INTRODUCTION: Chest X-ray (CR) utility to predict etiology in community-acquired pneumonia (CAP), in children has been controversial. Nevertheless, some authors propose the use of well standardized radiological patterns. OBJECTIVE: To evaluate usefulness of modified Swischuk's radiological patterns (RaP) to determine probable etiology in children hospitalized with CAP. PATIENTS AND METHODS: Eighty children were studied using 6 standardized RaP. The RaP were blindly and individually analyzed by 9 pneumologists, who registered the results (stage 1). Thereafter, (stage 2) a second evaluation adding clinical information was performed. Then, the patients complementary examinations and clinical evolution were included (stage 3). The reference standard (RS) was generated from a blind consensus. Every result was compared with the RS using Student test. RESULTS: According to RS, children were classified as having a virus-like pneumonia in 63% of cases, bacterial in 13%, mixed in 16%, atypical in 5% and ADV in 3%. The agreement of stage 1 and 2 with RS was 64 and 77%, respectively. Virus and bacterial RaP agreement increased from 66 to 82 % (p < 0.001) and from 82 to 90% (p < 0.05), respectively after incorporating clinical parameters. CONCLUSION: Modified Swischuk's RaP used in association with clinical elements allows a quite satisfactory approach to etiologic diagnosis of CAP.