Resolution of Chest X-ray Abnormalities for Pneumonia

Abstract & Commentary

By Dean L. Winslow, MD, FACP, FIDSA Chief, Division of AIDS Medicine, Santa Clara Valley Medical Center; Clinical Professor of Medicine, Stanford University School of Medicine Dr. Winslow is a consultant for Bayer Diagnostics, and is on the speaker's bureau for GlaxoSmithKline and Pfizer. This article originally appeared in the December 2007 issue of Infectious Disease Alert. It was edited by Stan Deresinski, MD, and peer reviewed by Connie Price, MD.

Synopsis: Two hundred eighty-eight patients hospitalized with severe community-acquired pneumonia (CAP) were followed for 28 days in a prospective multicenter study. At day 7, 25% of patients had resolution of CXR abnormalities and 56% had improvement. At day 28, 53% of patients had resolution of CXR abnormalities and 78% had clinical cure. By multivariate analysis, delayed resolution of CXR abnormalities by day 7 was associated with multilobar disease, dullness to percussion by physical exam, elevated CRP (> 200 mg/L), and tachypnea (respiratory rate > 25/min.) on admission.

Source: Bruns AH, et al. Patterns of resolution of chest radiograph abnormalities in adults hospitalized with severe community-acquired pneumonia. Clin Infect Dis. 2007;45:983-991.

This interesting study from the netherlands prospectively evaluated 288 consecutive patients with severe CAP (ATS pneumonia severity index > 90) admitted to the hospital, on which clinical data and CXR's were available at admission, day 7, and day 28. Mean age of patients was 69.7 years, and 53.5% had comorbid conditions, including CHF, underlying neoplasm, cerebrovascular diseases, and renal disease. Of these patients, 21.5% had microbiologically-documented infection with pneumococcus, 9.7% had infection with an atypical pathogen, 51.4% had pneumonia of unknown etiology, 3.8% had infection with multiple pathogens, and 17.4% had infection with other pathogens, including gram-negative enteric organisms, or Pseudomonas, S. aureus, H. influenzae, or M. catarrhalis. Twenty (6.9%) patients died.

Univariate analysis for delayed resolution of CXR abnormalities at day 28 showed the following parameters to be correlated: higher PSI, S. pneumoniae infection, multilobar pneumonia, PCO2 < 30 mm Hg, CRP > 200 mg/L, and BUN > 10uM on admission. By multivariate analysis, delayed resolution of CXR abnormalities by day 28 was associated only with CRP > 200 mg/L on admission.


My experience in several large teaching hospitals over the last 20 years is that patients admitted with pneumonia often are subjected to routine daily CXR's during the first few days in the hospital, and generally every 3 days or so until hospital discharge, despite the presence of clinical improvement. This large, prospective, multicenter study conducted in immunocompetent adult patients hospitalized with CAP shows that only one-quarter of the patients resolved their CXR abnormalities by day 7 and approximately one-half did so by day 28. This study clearly suggests that frequent CXR's obtained prior to hospital discharge, in patients who are clinically improving, are unnecessary and unlikely to be useful. The study also suggests that one of the "old saws" many of us were taught during our Internal Medicine training (without any literature support) in the 1970s, which recommended deferring aggressive work-up of persistent radiographic abnormalities following CAP unless those abnormalities persisted beyond 6 weeks, was correct. While not specifically addressed in this study, it is likely that an interval as long as 8-12 weeks, following an episode of CAP, seems to be reasonable before performing follow-up radiography (including thoracic CT scans) or bronchoscopy to exclude noninfectious causes of persistent CXR abnormalities.