Facteurs prédictifs de bactériémie chez le sujet âgé : création d’un score de probabilité de bactériémie aux urgences

Abstract : Objectives: the aim of this study was to identify potentials predictors of bacteremia in elderly and to create an easily reproducible predictive bacteremia score for the emergency department. Methods: we conducted a monocentric retrospective study at the “Côte Basque” hospital in France from December 2015 to September 2017 included. All patients aged 65 years or older with definitive infectious diagnosis were included. Socio-demographic, clinical and biological informations were collected in the emergency department using the medical information system. A multivariate logistic regression analysis was performed to identify factors associated with the presence of bacteremia. A probability score of bacteremia was then created and tested within the same cohort to determine its internal validity. Results: among the 266 patients studied (median age 78 years, 55,6 % of men), 101 (38 %) had bacteremia and 165 (62 %) had a bacterial, viral or fungal infection without bacteremia. The most common bacteria found in bacteremic patients were Staphylococcus aureus (23,8 %) and Escherichia coli (21,8 %). The overall mortality rate was 6 % (11 % for bacteremic patients). Five parameters were significantly associated (p<0.05) with the presence of bacteremia after logistic regression: a history of valvulopathy, > 38.5 fever, a Quick-SOFA > 2, a CRP > 70 mg/L and a neutrophil-to-lymphocyte ratio (NLR) > 5. A score between 5 and 8 was associated with a bacteremia probability between 20 % and 25 %, a score between 9 and 12 was associated with a probability between 50 % and 60 %, and a score above 13 was associated with a bacteremia risk of more than 75 %. The area under the ROC curve (AUC) found after modeling the study population was 0.764, greater than the AUC of the CRP, NLR and Quick-SOFA. Conclusion: bacteremia in the elderly can be predicted by calculating an easily reproducible and inexpensive clinical-biological probability score. A validation study could confirm and refine this score performance.
Document type :
Master Thesis
Complete list of metadatas

Cited literature [161 references]  Display  Hide  Download

Contributor : Bu Carreire Université de Bordeaux <>
Submitted on : Tuesday, March 12, 2019 - 3:04:40 PM
Last modification on : Saturday, March 30, 2019 - 2:02:16 AM


Files produced by the author(s)


  • HAL Id : dumas-02065215, version 1



Romain Boissin. Facteurs prédictifs de bactériémie chez le sujet âgé : création d’un score de probabilité de bactériémie aux urgences. Médecine humaine et pathologie. 2018. ⟨dumas-02065215⟩



Record views


Files downloads