Abstract Rationale Anti-anaerobic antibiotics are frequently included in empiric regimens for suspected sepsis. Prior studies have suggested an association between receipt of anti-anaerobic antibiotics and mortality. However, whether anaerobic antibiotic coverage itself contributes to differences in patient outcomes remains unclear. Methods We performed a secondary analysis of the ACORN trial, a randomized trial that compared cefepime versus piperacillin-tazobactam in adults hospitalized with acute infection. To evaluate the effect of empiric anaerobic coverage, a propensity score was developed using the likelihood of receiving metronidazole in patients randomized to the cefepime arm based on the following baseline covariates: age, sex, race, location at enrollment, presence of sepsis, source of infection, Sequential Organ Failure Assessment (SOFA) score, receipt of vancomycin, transplant status, neutropenia, chronic kidney disease (CKD), acute kidney injury, coma, delirium, Charlson Comorbidity Index, lactate, white blood cell (WBC) count, receipt of continuous sedation, and receipt of nephrotoxic or neurotoxic medications. We then examined the effect of anaerobic coverage on the primary outcome of 14-day in-hospital mortality using a logistic regression model including the propensity score and age, sex, baseline vasopressor use, mechanical ventilation, SOFA score, source of infection, and location. Sensitivity analysis included a comparison of 14-day mortality using an unadjusted chi-square analysis in the subgroup of patients excluding those with sites of infection that routinely warrant anaerobic coverage (intra-abdominal, cellulitis, abscess, necrotizing fasciitis). Odds ratios greater than 1 represent higher mortality in the cefepime group. Results Among 2,511 patients (1,297 piperacillin-tazobactam; 1,214 cefepime), median age was 58 years, 42.7% were female, and 54.2% had confirmed sepsis. In the propensity score-adjusted model, trial group (cefepime vs piperacillin-tazobactam) was not significantly associated with mortality (OR 1.28; 95% CI, 0.90 - 1.83; p = 0.17). Independent predictors of mortality included older age, vasopressor use, higher SOFA score, mechanical ventilation, and ICU compared to ED location on enrollment (all p 0.01, see Table). In the subgroup of patients excluding infections warranting anaerobic coverage (n = 771 piperacillin-tazobactam; n = 697 cefepime), mortality was 8.0% with piperacillin-tazobactam versus 10.0% with cefepime (p = 0.18). Conclusions In this secondary analysis of the ACORN trial, empiric anaerobic coverage was not associated with excess mortality. Among patients without indications for anaerobic therapy, mortality was similar between cefepime and piperacillin-tazobactam treated groups. These findings argue against intrinsic harm from anaerobic coverage in empiric treatment of sepsis. This abstract is funded by: None
Magner et al. (Fri,) studied this question.