Background: This intercontinental study aimed to study gram-negative rod (GNR) resistance in hematopoietic stem cell transplantation (HSCT).

Methods: GNR bacteremias occurring during 6 months post-HSCT (February 2014-May 2015) were prospectively collected, and analyzed for rates and risk factors for resistance to fluoroquinolones, noncarbapenem anti-Pseudomonas β-lactams (noncarbapenems), carbapenems, and multidrug resistance.

Results: Sixty-five HSCT centers from 25 countries in Europe, Australia, and Asia reported data on 655 GNR episodes and 704 pathogens in 591 patients (Enterobacteriaceae, 73%; nonfermentative rods, 24%; and 3% others). Half of GNRs were fluoroquinolone and noncarbapenem resistant; 18.5% carbapenem resistant; 35.2% multidrug resistant. The total resistance rates were higher in allogeneic HSCT (allo-HSCT) vs autologous HSCT (auto-HSCT) patients (P < .001) but similar in community-acquired infections. Noncarbapenem resistance and multidrug resistance were higher in auto-HSCT patients in centers providing vs not providing fluoroquinolone prophylaxis (P < .01). Resistance rates were higher in southeast vs northwest Europe and similar in children and adults, excluding higher fluoroquinolone- and β-lactam/β-lactamase inhibitor resistance rates in allo-HSCT adults. Non-Klebsiella Enterobacteriaceae were rarely carbapenem resistant. Multivariable analysis revealed resistance risk factors in allo-HSCT patients: fluoroquinolone resistance: adult, prolonged neutropenia, breakthrough on fluoroquinolones; noncarbapenem resistance: hospital-acquired infection, breakthrough on noncarbapenems or other antibiotics (excluding fluoroquinolones, noncarbapenems, carbapenems), donor type; carbapenem resistance: breakthrough on carbapenem, longer hospitalization, intensive care unit, previous other antibiotic therapy; multidrug resistance: longer hospitalization, breakthrough on β-lactam/β-lactamase inhibitors, and carbapenems. Inappropriate empiric therapy and mortality were significantly more common in infections caused by resistant bacteria.

Conclusions: Our data question the recommendation for fluoroquinolone prophylaxis and call for reassessment of local empiric antibiotic protocols. Knowledge of pathogen-specific resistance enables early appropriate empiric therapy. Monitoring of resistance is crucial.

Clinical Trials Registration: NCT02257931.