All-cause and Infection-attributable Mortality Amongst Adults With Bloodstream Infection-a Population-based Study

Jonathan Underwood; Rowena Griffiths; David Gillespie; Ashley Akbari; Haroon Ahmed

doi:10.1093/ofid/ofae126

All-cause and Infection-attributable Mortality Amongst Adults With Bloodstream Infection-a Population-based Study

Open Forum Infect Dis. 2024 Mar 6;11(5):ofae126. doi: 10.1093/ofid/ofae126. eCollection 2024 May.

Authors

Jonathan Underwood^{1

2}, Rowena Griffiths³, David Gillespie⁴, Ashley Akbari³, Haroon Ahmed⁵

Affiliations

¹ Division of Infection and Immunity, Cardiff University, Cardiff, UK.
² Department of Infectious Diseases, Cardiff and Vale University Health Board, Cardiff, UK.
³ Population Data Science, Swansea University, Swansea, UK.
⁴ Centre for Trials Research, Cardiff University School of Medicine, Cardiff, UK.
⁵ Division of Population Medicine, Cardiff University School of Medicine, Cardiff, UK.

Abstract

Background: Bloodstream infections (BSIs) are common, life-threatening infections. However, it remains unclear whether deaths following BSIs are primarily from uncontrolled infection or underlying comorbidities. We aimed to determine the overall mortality, infection-attributable mortality, and causes of death for four leading BSI pathogens.

Methods: This retrospective cohort study was conducted within the Secure Anonymized Information Linkage Databank, containing anonymized population-scale electronic health record data for Wales, UK. We included adults with Escherichia coli, Klebsiella spp, Pseudomonas aeruginosa, and Staphylococcus aureus BSI between 2010 and 2022 using linked data from Public Health Wales and the Office for National Statistics. Thirty-day all-cause and sepsis-specific mortality, as a proxy for infection-attributable mortality, were compared using Cox proportional hazards and competing risk regression, respectively.

Results: We identified 35 691 adults with BSI (59.6% E coli). Adjusted analyses revealed that all organisms had a higher 30-day mortality versus E coli with Pseudomonas aeruginosa the highest (hazard ratio, 1.96 [1.76-2.17], P < .001). Cancer was the leading cause of death following BSIs for all organisms, particularly deaths occurring between 30 and 90 days (35.9%). A total of 25.5% of deaths within 30 days involved sepsis. Methicillin-resistant Staphylococcus aureus was associated with the highest sepsis mortality versus E coli (hazard ratio, 2.56 [2.10-3.12], P < .001). Peak C-reactive protein was positively associated with increased sepsis mortality (P < .001).

Conclusions: This population-level study challenges the assumption that most deaths following BSIs are directly attributable to uncontrolled infection, particularly subacutely more than 30 days from BSI. Our findings underscore the need for reevaluating clinical trial design and developing better preventive strategies for BSIs.

Keywords: bloodstream infection; cause of death; inflammation; mortality; sepsis.