Costs of bloodstream infections caused by Escherichia coli and influence of extended-spectrum-beta-lactamase production and inadequate initial antibiotic therapy

Abstract

Escherichia coli is the leading cause of bloodstream infections (BSIs) caused by Gram-negative bacteria. The increasing prevalence of antibiotic-resistant E. coli strains, particularly those producing extended-spectrum β-lactamases (ESBLs), increases the odds that empirically prescribed antimicrobial therapy for these infections will be inadequate, but the economic impact of this risk has not been fully evaluated. In the present retrospective 1-year analysis of 134 consecutive E. coli BSIs in our hospital, we explored the clinical and economic impacts of (i) inadequate initial antimicrobial treatment (IIAT) (i.e., empirical treatment with drugs to which the isolate had displayed in vitro resistance) of these infections and (ii) ESBL production by the bloodstream isolate. Cost data were obtained from the hospital accounting system. Compared with the 107 (79.8%) adequately treated patients, the 27 (20.1%) who received IIAT had a higher proportion of ESBL BSIs (74.0% versus 15.8%), longer (+6 days) and more costly (+EUR 4,322.00) post-BSI-onset hospital stays, and higher 21-day mortality rates (40.7% versus 5.6%). Compared with the 97 non-ESBL infections, the 37 (27.6%) ESBL BSIs were also associated with longer (+7 days) and more costly (+EUR 5,026.00) post-BSI-onset hospital stays and increased 21-day mortality (29.7% versus 6.1%). These findings confirm that the hospital costs and mortality associated with E. coli BSIs are significantly increased by ESBL production and by IIAT.

FIG. 1.

Cost of antimicrobial drug therapy in subgroups defined by the infecting pathogen's ESBL status and by the adequacy of the initial antibiotic therapy. The mean ± SD cost of antimicrobial treatment was EUR 763 ± 437 for ESBL BSIs versus EUR 474 ± 270 for non-ESBL BSIs (P < 0.001), and it was EUR 715 ± 437 for patients with IIAT versus EUR 513 ± 311 for those with AIAT (P = 0.02).

FIG. 2.

Box plot showing percent differences between the costs of ESBL and non-ESBL E. coli BSIs and between the costs of IIAT and AIAT E. coli BSIs. The results are based on a Monte Carlo simulation with 1,000 iterations. Boxes represent interquartile ranges (lower border, 25th percentile; upper border, 75th percentile), and the horizontal lines within the boxes indicate the medians (50th percentile). Whiskers indicate minimum and maximum values.

FIG. 3.

Kaplan-Meier curves showing the impact on 21-day mortality of ESBL production (solid line) versus non-ESBL production (dashed line) by the bloodstream isolate (P < 0.001) (a) and inadequate (solid line) versus adequate (dashed line) initial antibiotic therapy (P < 0.001) (b).