The Effect of C-Reactive Protein on Chest X-Ray Interpretation

Background: Chest X-Ray (CXR) and blood-tests such as blood-count and inflammatory markers are often used in the ED to diagnose pneumonia and to discern between viral and bacterial etiologies. We suspected that cross-interaction exists between these factors, specifically - between C-Reactive Protein (CRP) and CXR interpretation, introducing a bias into the diagnosis.

Methods: An internet-based survey was conducted among 337 pediatricians in Israel. Each participant was given two short vignettes and a CXR of a child with suspected pneumonia. Participants were randomly assigned to one of three groups: Controls - where no lab tests were given and two intervention groups where one case was given with a high CRP (15.1 mg/dL) and the other with low CRP (2.3 md/dL), each group with a different order of CRP values. Between subjects and within subjects analyses were conducted to study the effect of CRP on CXR interpretation.

Results: Overall, 301 clinicians (89.3%) completed the study. the percentage of subjects that answered that “the CXR fits a case of bacterial pneumonia” was 39% when CRP was low or when no CRP was given, vs. 62% when high CRP was given (p = 0.002). The effect on antibiotic treatment was even more pronounced, with 86% of the subjects stating that they will prescribe antibiotics when CRP was high vs. only 53% when low-concentration of CRP was given or 56% when CRP was not given at all (p < 0.001).

Conclusion: Exposure to CRP leads to biases in the interpretation of CXR.