<i>Legionella</i> 5S rRNA PCR melting temperature analysis discriminates high-risk species associated with disease severity

ABSTRACT Legionella species are important causes of community-acquired pneumonia. The diagnostic performance and clinical implications of a fluorescence resonance energy transfer (FRET)-based real-time PCR assay targeting the Legionella 5S rRNA gene were evaluated, with emphasis on melting temperature (Tm) analysis to differentiate detected species and associated clinical outcomes. A retrospective multicenter cohort study of adults with laboratory-confirmed Legionella infection between 2019 and 2025 was conducted. PCR crossing point (Cp) and Tm data were analyzed and compared with species identified in culture from respiratory specimens. Associations between Tm and clinical outcomes were assessed using multivariable logistic regression. Among 51,489 PCR tests performed, 823 (1.6%) were positive. Among 376 Legionella infections, PCR demonstrated the highest diagnostic yield. Among patients undergoing all three diagnostic tests, the addition of PCR increased detection from 49% with urine antigen and culture combined to 100% ( P &lt; 0.001). Of 189 PCR-positive patient specimens with available Tm data, Tm showed a bimodal distribution, being either &lt;66°C or ≥66°C. A Tm cutoff of ≥66°C discriminated L. pneumophila / L. longbeachae from other Legionella species with excellent performance (AUC, 1.00). Overall, 29.6% of infections had a Tm &lt;66°C, consistent with non- pneumophila/ non- longbeachae Legionella species. Infections due to L. pneumophila / L. longbeachae , which correspond to high Tm values in the study assay, were independently associated with ICU admission (adjusted odds ratio, 2.85; 95% CI, 1.33–6.11; P = 0.007), but not 90-day mortality. Real-time PCR targeting the Legionella 5S rRNA gene provides superior detection compared to culture and/or urine antigen testing and enables higher-risk species discrimination through Tm analysis. IMPORTANCE Nearly one-third of Legionella infections in this cohort were attributable to non- pneumophila species, highlighting limitations of diagnostic strategies reliant on Legionella urine antigen testing or L. pneumophila PCR. The findings suggest that species-level inference can be derived from melting-temperature data generated during routine PCR testing. Incorporating validated melting-temperature interpretation into reporting workflows may improve epidemiologic surveillance and clinical understanding of characteristics of Legionella pneumonia.