Mismatch Rate of Empirical Antimicrobial Treatment in Fracture-Related Infections

Abstract

Objectives: To evaluate the current standard of care regarding empirical antimicrobial therapy in fracture-related infections (FRIs).

Design: Retrospective cohort study.

Setting: Level I Trauma Center.

Patient selection criteria: Adult patients treated for FRI with surgical debridement and empirical antibiotics between September 1, 2014, and August 31, 2022. Patients were excluded if less than 5 tissue samples for culture were taken, culture results were negative, or there was an antibiotic-free window of less than 3 days before debridement.

Outcome measures and comparisons: FRI microbial etiology, antimicrobial resistance patterns (standardized antimicrobial panels were tested for each pathogen), the mismatch rate between empirical antimicrobial therapy and antibiotic resistance of causative microorganism(s), and mismatching risk factors.

Results: In total, 75 patients were included [79% (59/75) men, mean age 51 years]. The most prevalent microorganisms were Staphylococcus aureus (52%, 39/75) and Staphylococcus epidermidis (41%, 31/75). The most frequently used empirical antibiotic was clindamycin (59%, 44/75), followed by combinations of gram-positive and gram-negative covering antibiotics (15%, 11/75). The overall mismatch rate was 51% (38/75) [95% confidence interval (CI), 0.39-0.62] and did not differ between extremities [upper: 31% (4/13) (95% CI, 0.09-0.61), lower: 55% (33/60) (95% CI, 0.42-0.68, P = 0.11)]. Mismatching empirical therapy occurred mostly in infections caused by S. epidermidis and gram-negative bacteria. Combination therapy of vancomycin with ceftazidime produced the lowest theoretical mismatch rate (8%, 6/71). Polymicrobial infections were an independent risk factor for mismatching (OR: 8.38, 95% CI, 2.53-27.75, P < 0.001).

Conclusions: In patients with FRI, a mismatching of empirical antibiotic therapy occurred in half of patients, mainly due to lack of coverage for S. epidermidis , gram-negative bacteria, and polymicrobial infections. Empirical therapy with vancomycin and ceftazidime produced the lowest theoretical mismatch rates. This study showed the need for the consideration of gram-negative coverage in addition to standard broad gram-positive coverage. Future studies should investigate the effect of the proposed empirical therapy on long-term outcomes.

Level of evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.