Impact of empirical antibiotic regimens on mortality in neutropenic patients with bloodstream infection presenting with septic shock

Abstract

Objectives: We analyzed risk factors for mortality in febrile neutropenic patients with bloodstream infections (BSI) presenting with septic shock and assessed the impact of empirical antibiotic regimens. Methods: Multicenter retrospective study (2010-2019) of two prospective cohorts comparing BSI episodes in patients with or without septic shock. Multivariate analysis was performed to identify independent risk factors for mortality in episodes with septic shock. Results: Of 1563 patients with BSI, 257 (16%) presented with septic shock. Those patients with septic shock had higher mortality than those without septic shock (55% vs 15%, p<0.001). Gram-negative bacilli caused 81% of episodes with septic shock; gram-positive cocci, 22%; and Candida species 5%. Inappropriate empirical antibiotic treatment (IEAT) was administered in 17.5% of septic shock episodes. Empirical β-lactam combined with other active antibiotics was associated with the lowest mortality observed. When amikacin was the only active antibiotic, mortality was 90%. Addition of empirical specific gram-positive coverage had no impact on mortality. Mortality was higher when IEAT was administered (76% vs 51%, p=0.002). Age >70 years (OR 2.3, 95% CI 1.2-4.7), IEAT for Candida spp. or gram-negative bacilli (OR 3.8, 1.3-11.1), acute kidney injury (OR 2.6, 1.4-4.9) and amikacin as the only active antibiotic (OR 15.2, 1.7-134.5) were independent risk factors for mortality, while combination of β-lactam and amikacin was protective (OR 0.32, 0.18-0.57). Conclusions: Septic shock in febrile neutropenic patients with BSI is associated with extremely high mortality, especially when IEAT is administered. Combination therapy including an active β-lactam and amikacin results in the best outcomes.