The equivalence of intravenous push (IVP) and piggyback (IVPB) administration has not been evaluated in the critically ill population for most medications, but it is especially relevant for antibiotics, such as cefepime, that exhibit time-dependent bactericidal activity. A single center, retrospective, observational pre/post-protocol change study included critically ill adults who received cefepime as empiric therapy between August 2015 and 2021. The primary outcome was treatment failure, which was defined as a composite of escalation of antibiotic regimen or all-cause mortality. Secondary outcomes included adverse drug events, days of cefepime therapy, total days of antibiotic therapy, and ICU and hospital length of stay. Outcomes were compared using Chi-squared, Mann Whitney U, and binary logistic regression as appropriate. A total of 285 patients were included: 87 IVPB and 198 IVP. Treatment failure occurred in 18% (n = 16) of the IVPB group and 27% (n = 54) of the IVP group (p = 0.109). There were no significant differences in secondary outcomes. Longer duration of antibiotics (odds ratio [OR] 1.057, 95% confidence interval [CI] 1.013-1.103), SOFA score (OR 1.269, 95% CI 1.154-1.397) and IVP administration of cefepime (OR 2.370, 95% CI 1.143-4.914) were independently associated with treatment failure. Critically ill patients who received IVP cefepime were more likely to experience treatment failure in an adjusted analysis. The current practice of IVP cefepime should be reevaluated, as it may not provide similar clinical outcomes in the critically ill population.
Keywords: antibacterial agents; cefepime; critical illness; drug administration routes; sepsis; treatment failure.