Pleural effusion is a common finding among patients presenting with respiratory symptoms. The value of the bedside examination to detect pleural effusion is unclear.
To systematically review the evidence regarding the accuracy of the physical examination in assessing the probability of a pleural effusion.
We searched MEDLINE (1950-October 2008) and EMBASE (1980-October 2008) using Ovid to identify English-language studies conducted in a clinical setting. Additional studies were identified by searching the bibliographies of retrieved articles and contacting experts in the field.
We included prospective studies of diagnostic accuracy that compared at least 1 physical examination maneuver with radiographic confirmation of pleural effusion.
Three authors independently appraised study quality and extracted relevant data. Data regarding participant recruitment, reference standard, diagnostic test(s), and test accuracy were extracted. Disagreements were resolved by consensus.
We identified 310 unique citations, but only 5 prospectively conducted studies met inclusion criteria (N = 934 patients). A random-effects model was used for quantitative synthesis. Of the 8 physical examination maneuvers evaluated in the included studies (conventional percussion, auscultatory percussion, breath sounds, chest expansion, tactile vocal fremitus, vocal resonance, crackles, and pleural friction rub), dullness to conventional percussion was most accurate for diagnosing pleural effusion (summary positive likelihood ratio, 8.7; 95% confidence interval, 2.2-33.8), while the absence of reduced tactile vocal fremitus made pleural effusion less likely (negative likelihood ratio, 0.21; 95% confidence interval, 0.12-0.37).
Based on the limited number of studies, dullness to percussion and tactile fremitus are the most useful findings for pleural effusion. Dull chest percussion makes the probability of a pleural effusion much more likely but requires a chest radiograph to confirm the diagnosis. When the pretest probability of pleural effusion is low, the absence of reduced tactile vocal fremitus makes pleural effusion less likely so that a chest radiograph might not be necessary depending on the overall clinical situation.