Prognostic Factors and Optimal Empiric Therapy for Extended-Spectrum Beta-Lactamase (ESBL)-Producing Enterobacterales Bacteremia

Background Bloodstream infections (BSIs) caused by extended-spectrum β-lactamase-producingEnterobacterales (ESBL-E) are associated with increased morbidity and mortality, yet the optimal empiric treatment strategy remains uncertain. We aimed to identify prognostic factors for early clinical response and 30-day mortality in patients with ESBL-E bacteremia and to evaluate the impact of the appropriateness of empiric antimicrobial therapy. Methods We conducted a retrospective cohort study at a 916-bed tertiary care hospital in Tokyo, Japan, including all patients with ESBL-E bacteremia between April 2018 and March 2023. Clinical, microbiological, and therapeutic data were extracted from electronic medical records. Independent variables were selected based on prior literature, including age, sex, Charlson Comorbidity Index (CCI), quick sequential organ failure assessment (qSOFA) score, infection setting, invasive device use, C-reactive protein (CRP) level, and appropriateness of empiric antimicrobial therapy. Logistic regression was used to identify predictors of clinical improvement within 72 hours and 30-day all-cause mortality. Results A total of 138 patients were included, with a median age of 72.7 years; 83% of infections were caused by <i>Escherichia coli</i>. The urinary tract was the most common source of infection (58%). Approximately half of the infections were hospital-acquired (49%), and 46% were community-acquired. Appropriate empiric therapy, defined according to Clinical and Laboratory Standards Institute (CLSI) criteria, was administered in 60% of cases. Logistic regression analysis revealed that a qSOFA score ≥2 was significantly associated with reduced clinical improvement at 72 hours (odds ratio (OR), 0.29; 95% confidence interval (CI), 0.11-0.76) and with increased 30-day mortality (OR, 9.42; 95%CI, 2.27-39.11). Conversely, community-acquired infection was independently associated with reduced mortality risk (OR, 0.25; 95%CI, 0.06-0.94). Appropriateness of empiric antimicrobial therapy was not significantly associated with either early clinical response or 30-day mortality in multivariable analysis. In subgroup analysis of patients with a qSOFA score ≥2, inappropriate empiric therapy tended to be associated with higher 30-day mortality (OR, 7.22; 95%CI, 0.74-107.88), though this did not reach statistical significance. Conclusion Among patients with ESBL-E bacteremia, a qSOFA score ≥2 at the time of blood culture collection was the strongest independent predictor of both poor early clinical response and higher 30-day mortality, regardless of empiric antimicrobial appropriateness. These findings suggest that patient severity at presentation may outweigh the impact of initial antimicrobial selection in determining outcomes. However, potential for clinical deterioration with inappropriate therapy in high-risk patients underscores the need for timely, effective empiric coverage in those with suspected sepsis and elevated qSOFA scores. Risk stratification using qSOFA may support tailored empiric therapy decisions, helping to balance effective patient care with antimicrobial stewardship. Prospective, multicenter studies are warranted to validate these results and refine treatment strategies in the era of increasing antimicrobial resistance.