Weighting the impact of virulence on the outcome of Pseudomonas aeruginosa bloodstream infections

Abstract Objectives We assessed the association between the lethality of Pseudomonas aeruginosa in a Caenorhabditis elegans model and outcomes of P. aeruginosa bloodstream infections. Methods A total of 593 P. aeruginosa bloodstream isolates recovered from a prospective Spanish multicenter study were analyzed. Clinical variables, susceptibility profiles and Type III Secretion System (TTSS) genotypes ( exoU / exoS genes) were available from previous studies. A C. elegans virulence score (CEVS) was used, classifying the isolates into high (CEVS 4-5), intermediate (CEVS 3) and low-virulence (CEVS 1-2). The main outcome analyzed was 30-days mortality. Results Up to 75% (446/593) of the isolates showed a high-virulence phenotype, and 17% (101/593) a low-virulence one. No association between virulence phenotype and the main outcome variable (30-days) mortality was evidenced (29/101(28.7%) vs 127/446 (28.5%), p =1. However, an inverse association between C. elegans virulence and MDR and XDR profiles was documented (O.R=0.655 (0.571-0.751)) and O.R=0.523 (0.436-0.627), p ˂0.001, respectively), whereas the exoU genotype was significantly more frequent among isolates showing high-virulence (10/101 (9.9%)vs 112/446 (25.1%), p ˂0.001). Moreover, although significance was not reached, strains showing a high-virulence phenotype tended to be associated with community acquired infections (1/101 (1%) vs 25/446 (5.6%), p =0.065), whereas low-virulence phenotypes tended to be associated with a higher illness severity [such as higher median Pitt Score (2(1-4) vs 1(0-3), p =0.036) or initial multiorgan disfuction (17/101(16.8%) vs 41/446 (9.2%), p =0.024)], with some underlying conditions (such as chronic renal failure 24/101 (23.8%)vs 59/446 (13.2%), p =0.013) and with the respiratory source of infections (17/101 (16.8%) vs 45/446 (10.1%), p =0.058). Conclusions Our results indicate that P. aeruginosa virulence phenotype in C. elegans model correlates with virulence genotype (TTSS) and resistance profile, but it is a poor prognostic marker of mortality in bloodstream infections.