Purpose: To describe the prevalence of patients admitted to hospital with a diagnosis of community-acquired pneumonia who have normal chest radiographs; the extent to which patients actually had pneumonia on radiographs; and to compare presentation and outcomes in patients with a lower respiratory tract infection and those whose clinical diagnosis of pneumonia was confirmed by radiography.
Methods: We studied a population-based cohort of 2706 adults who were admitted with suspected pneumonia and managed using a clinical pathway. We stratified patients by presence or absence of radiograph-confirmed pneumonia, and compared their characteristics and in-hospital mortality. We also performed an independent review of a 10% sample of "normal" chest radiographs and classified them according to the presence or absence of pneumonia.
Results: One third (n=911) of patients admitted with pneumonia had their initial radiograph reported as "no pneumonia." Independent review found that only 7% (6/92) of radiographs developed an opacity that confirmed pneumonia. Characteristics were similar among admitted patients irrespective of radiographic findings, although patients without pneumonia on radiograph were older (mean [+/- SD] age, 73 +/- 15 years vs. 68 +/- 19 years, P <0.001) and had greater pneumonia-specific severity-of-illness scores (104 +/- 32 vs. 99 +/- 37, P=0.004). Patients without radiographic confirmation of pneumonia had similar rates of positive sputum cultures (32% [87/271] vs. 30% [208/706], P=0.42) and blood cultures (6% [35/576] vs. 8% [100/1241], P=0.13), but microbiology results differed, with a shift away from Streptococcus pneumoniae towards other streptococci species and gram-negative aerobic bacilli. In-hospital mortality was similar for both groups of patients (8% [64/911] in the unconfirmed pneumonia group vs. 10% [165/1795] in the confirmed group, adjusted P=0.09).
Conclusion: One third of patients suspected of having pneumonia and admitted to hospital did not have pneumonia, but had serious lower respiratory tract infections with substantial rates of bacteremia and mortality. The absence of radiographic findings should not supercede clinical judgment and empiric treatments in these patients.