Relevance of Crackles to Clinician Diagnosis of Pneumonia

BACKGROUND AND OBJECTIVES Clinicians often use the presence of auscultatory crackles to guide antibiotic treatment for pneumonia, despite evidence that crackles do not strongly predict pediatric pneumonia. We evaluated the association of crackles with both clinician-diagnosed pneumonia and radiographic pneumonia. We hypothesized that physicians would overestimate the prevalence of radiographic pneumonia in children with focal crackles. METHODS We performed a prospective cohort study of children aged 3 months to 18 years undergoing chest radiography to evaluate for pneumonia in an urban tertiary-care pediatric emergency department. Using multivariable regression, we assessed the odds of both clinician-diagnosed pneumonia (based on a binary question posed to physicians) and radiographic pneumonia by presence of crackles. We compared the proportion of children with clinician-diagnosed pneumonia and radiographic pneumonia across combinations of examination findings. RESULTS Our cohort comprised 1594 children, with a median age of 3.3 years (IQR, 1.5–6.1). Among 260 children with focal crackles, 59 (22.7%) had radiographic pneumonia. Compared with children without any crackles, the adjusted odds ratio (aOR) of clinician-diagnosed pneumonia in children with focal crackles was 6.8 (95% CI, 4.8–9.8), whereas the aOR of radiographic pneumonia was 1.7 (95% CI, 1.2–2.4). In children with focal crackles alone or combined with other specified examination findings, physicians consistently overestimated the likelihood of pneumonia as compared with chest radiography. CONCLUSIONS Physicians overestimate the association between crackles and radiographic pneumonia. The use of clinical risk scores may aid in objectively quantifying the risk of radiographic pneumonia in children with respiratory illness.