Abstract
Background: Legionella pneumonia remains a diagnostic challenge. The legionella urinary antigen test (LUT) primarily detects Legionella pneumophila serogroup 1, accounting for 64% of Danish cases, and is often the only legionella test performed. We aimed to identify variables predictive of a positive or negative test result and to explore how the LUT was used in clinical practice. Methods: The study was an audit-based cohort study. LUT-positive patients were compared with three randomly selected age- and gender-matched LUT-negative referent patients admitted at a Danish university hospital during 2003–2013. Data were extracted from charts and databases. Positive and negative likelihood ratios (LR+ and LR–) were calculated. For CURB-65 and sepsis, sensitivity analyses were made due to incomplete data. Results: In all, 25 cases were compared with 75 referents. Factors associated with LUT positivity included recent travel outside Scandinavia (LR + 5.3), Na+ < 130 mEq/L (LR + 4.3), confusion (LR + 4.2), C-reactive protein (CRP) > 200 mg/L (LR + 3.5), temperature > 39°C (LR + 3.5), and CURB-65 score ≥ 3 (LR + 3.0–15.0, depending on the model). Decreasing the likelihood of LUT positivity were CRP < 200 mg/L (LR– 0.1), absence of sepsis (LR– 0.1–0.2, depending on the model), absence of tachycardia (heart rate < 90) (LR– 0.2) and normal pulmonary auscultation (LR– 0.3). Additional legionella tests were performed in 60% of the cases and 13% of the referents. Conclusion: Classical features of severe pneumonia are associated with a positive LUT. The LUT is often used inappropriately and should be accompanied by PCR analysis.
Keywords::
Acknowledgments
We are indebted to Steen Hoffmann and Søren Uldum at Statens Serum Institut for providing data on LUTs and PCR analyses performed there.
Declaration of interest: The authors report no conflict of interests. No funding was received for the study.