e18627 Background: HSCT patients are highly susceptible to severe infections. Conventional antimicrobial susceptibility testing (AST) (phenotypic and molecular), does not reflect in vivo infection conditions and evaluates antibiotic activity against a single “primary” pathogen at serum-based concentrations. These approaches fail to account for polymicrobial biofilms, bacterial-fungal interactions, and limited antibiotic penetration into infected tissues, resulting in ineffective antibiotic selection. AtbFinder is a novel, culture-independent, non–MIC-based phenotypic test that selects antibiotics under conditions mirroring each patient’s infection environment. It preserves polymicrobial biofilms from clinical samples and tests antibiotics at clinically achievable, site-specific concentrations, including tissue-penetration-adjusted levels for lungs. We compared clinical outcomes following AtbFinder-guided versus standard AST-guided antibiotic selection in HSCT patients with severe infections. Methods: This prospective, open-label, IIT pilot study included 12 HSCT febrile patients (6 sepsis, 6 respiratory infections). All patients had failed two antibiotic regimens selected by standard AST within the preceding 30 days. 64% had mixed bacterial–fungal infections, and 45% harbored ESKAPE pathogens. No antibiotics were administered for 5d prior to enrollment. AtbFinder evaluated up to 184 antibiotics per patient within 6–20 hours. Clinical outcomes were compared with matched retrospective controls treated with antibiotics selected with standard AST at the same hospitals. Controls were selected from patients managed during the same time period immediately prior to AtbFinder implementation and were matched by HSCT status, infection type, and clinical severity. Results were benchmarked against broth microdilution and PCR-based resistance gene detection. Results: At 30 days following completion of AtbFinder-guided therapy, all patients achieved clinical and laboratory-confirmed infection resolution, compared with 40% of matched controls (p<0.001). Use of broad-spectrum antibiotics decreased by 65% (p<0.01). ICU length of stay was reduced by 74%, and 90-day hospital readmission was 0% in the AtbFinder group versus 45% in controls. Among patients with ESKAPE pathogens, AtbFinder-guided antibiotic selection demonstrated comparable efficacy, whereas the control group showed a 3.5-fold lower response rate (p<0.01). Molecular analyses confirmed preservation of patient-specific polymicrobial biofilm communities within the AtbFinder assay. Conclusions: In this prospective pilot study, antibiotic selection guided by a polymicrobial biofilm–preserving phenotypic assay was associated with improved clinical outcomes, reduced broad-spectrum antibiotic use, and shorter ICU stays in HSCT patients with hard-to-treat and antibiotic-resistant infections.
Tetz et al. (Thu,) studied this question.