Association Between the Sequence of β-Lactam and Vancomycin Administration and Mortality in Patients With Suspected Sepsis.

Abstract

Timely antibiotic initiation is critical to sepsis management, but there are limited data on the impact of giving β-lactams first versus vancomycin first among patients prescribed both agents. We retrospectively analyzed all adults admitted to 5 US hospitals from 2015-2022 with suspected sepsis (blood culture collected, antibiotics administered, and organ dysfunction) treated with vancomycin and a broad-spectrum β-lactam within 24 hours of arrival. We estimated associations between β-lactam- versus vancomycin-first strategies and in-hospital mortality using inverse probability weighting (IPW) to adjust for potential confounders. Among 25 391 patients with suspected sepsis, 21 449 (84.4%) received β-lactams first and 3942 (15.6%) received vancomycin first. Compared with the β-lactam-first group, patients administered vancomycin first tended to be less severely ill, had more skin/musculoskeletal infections (20.0% vs 7.8%), and received β-lactams a median of 3.5 hours later relative to emergency department arrival. On IPW analysis, the β-lactam-first strategy was associated with lower mortality (adjusted odds ratio [aOR]: .89; 95% CI: .80-.99). Point estimates were directionally similar but nonsignificant in a sensitivity analysis using propensity score matching rather than IPW (aOR: .94; 95% CI: .82-1.07) and in subgroups of patients with positive blood cultures, methicillin-resistant Staphylococcus aureus cultures, and those administered antipseudomonal β-lactams. Among patients with suspected sepsis prescribed vancomycin and β-lactam therapy, β-lactam administration before vancomycin was associated with a modest reduction in in-hospital mortality. These findings support prioritizing β-lactam therapy in most patients with sepsis but merit confirmation in randomized trials given the risk of residual confounding in observational analyses.