Screening of asymptomatic children for tuberculosis is a lateral chest radiograph routinely indicated?

Abstract

Rationale and objectives: The aim of this study was to determine whether a lateral chest radiograph provides additional diagnostic information to a posteroanterior (PA) radiograph in the screening of asymptomatic children with positive purified protein derivative (PPD) skin tests in a nonendemic area.

Materials and methods: This was an Institutional Review Board-approved, Health Insurance Portability and Accountability Act-compliant, retrospective study of 605 consecutive pediatric patients (294 males, 311 females; mean age, 10.8 ± 5.2 years) with positive PPD skin test results, who underwent PA and lateral chest radiographs between July 2003 and May 2009 at a tertiary care pediatric hospital in a nonendemic area for tuberculosis (TB). Two pediatric radiologists independently reviewed each chest radiograph for evidence of abnormalities that may be indicative of acute or chronic TB infection. The reviewers first analyzed the PA radiograph alone and subsequently evaluated the PA and the lateral radiograph together to determine whether any observed abnormality was identified only on the lateral radiograph. When an abnormality was detected on both PA and lateral radiographs, the reviewers determined whether the abnormality on the lateral radiograph changed the reviewer's decision based on the PA radiograph alone. Assessment of nonconcordance between PA and lateral chest radiographs for each reviewer was evaluated by the McNemar test of matched binary pairs. Agreement between reviewers for detecting abnormalities on radiographs was evaluated by using the kappa (κ) statistic.

Results: The frequency of an abnormal chest radiograph related to TB was 1.8% (11/605). The PA radiograph showed abnormalities in all 11 (100%) children with radiographic abnormalities. Lateral radiographs showed abnormalities related to TB in 2 (18.2%) of 11 cases found to be abnormal on PA radiographs. Nine (81.8%) of 11 abnormalities on PA radiographs were not detected on the lateral chest radiographs. There was statistical evidence of nonconcordance between PA and lateral chest radiographs in detecting TB-related abnormalities for reviewer 1 (P < .001) and reviewer 2 (P = .004). In cases with abnormalities observed on both PA and lateral radiographs, there were no cases in which information obtained from the lateral chest radiograph resulted in a change in interpretation based on the PA radiograph alone. A high level of agreement was observed between the two independent reviewers in detecting TB-related abnormalities on PA radiographs (κ = 0.84, P < .001).

Conclusions: A PA radiograph alone is sufficient for TB screening of asymptomatic pediatric patients with positive PPD skin test results in an area non-endemic for TB.