Risk Factors for Infection-Attributable Mortality in Patients With Bacteremia: A Competing Risk Analysis.

BACKGROUND

Identifying risk factors for mortality in patients with bacteremia (SAB) is crucial due to its high fatality. However, data on risk factors for infection-attributable deaths considering competing risk events such as non-infection-attributable deaths remain limited. We performed a competing risk analysis to elucidate risk factors associated with 30-day infection-attributable mortality in a large cohort of patients with SAB.

METHODS

This retrospective cohort study included adult patients diagnosed with SAB at a tertiary hospital from August 2008 to December 2019. Competing risk analysis was performed using Fine and Gray models to estimate subdistribution hazard ratios (sHRs) for 30-day infection-attributable death.

RESULTS

Among 1936 patients, 444 (22.9%) died within 30 days. Of these, 338 (76.1%) were infection-attributable and 106 (23.9%) were non-infection-attributable deaths. The multivariable Fine and Gray model identified significant risk factors for 30-day infection-attributable death (sHRs with 95% confidence intervals): an increase in age by 10 years (1.14 [1.02-1.26]), presence of malignancy (1.54 [1.17-2.02]), liver cirrhosis (2.15 [1.56-2.97]), corticosteroid use (1.61 [1.19-2.17]), septic shock (3.28 [1.98-5.42]), elevated C-reactive protein (1.60 [1.19-2.14]), pneumonia (1.81 [1.21-2.72]), persistent bacteremia (1.73 [1.31-2.30]), and failure to remove the eradicable focus (2.40 [1.38-4.19]) or absence of an eradicable focus (1.49 [1.08-2.04]). Except for age and malignancy, these factors were not significantly associated with non-infection-related death.

CONCLUSIONS

Specific risk factors for infection-attributable death in patients with SAB were identified, distinct from those for nonattributable death. These findings can aid in the early identification of patients at risk for SAB-attributable mortality.