Piperacillin-Tazobactam vs. Cefepime for Kidney Injury and Neurologic Dysfunction (ACORN Trial)
Piperacillin-tazobactam, marketed as Zosyn, has been one of the most-prescribed broad spectrum antibiotics for many years, used for its bactericidal activity against most gram-negative and anaerobic bacteria, including Pseudomonas. Cefepime has similar antibacterial activity, and both are widely used in the ICU as empiric coverage in patients with potentially serious gram-negative infections.
Pip-tazo has been repeatedly associated with an increased risk for acute kidney injury over the past several years, based on retrospective observational studies with risk for confounding. Answering this research question has been made more difficult by pip-tazo’s possible status as a pseudo-nephrotoxin (elevating creatinine levels in the blood without harming the kidneys) and frequent co-administration with vancomycin, a known nephrotoxin. Over the years, dozens of copycat studies (which are relatively easy for academics to churn out) created a steady, low-quality drumbeat of negativity toward Zosyn.
For its part, cefepime has been strongly suspected of causing neurologic dysfunction, based on many case reports and the fact that it crosses the blood-brain barrier in significant concentrations. Cefepime has not been associated with kidney dysfunction.
The rumors of Zosyn-induced kidney injury and cefepime-induced neurotoxicity persisted mainly because no one had performed a large randomized trial. Until now.
Investigators randomized 2511 patients in EDs and ICUs at Vanderbilt University, whose attendings intended to prescribe antipseudomonal coverage, to receive either pip-tazo or cefepime (open-label). About three-quarters of patients were also prescribed vancomycin (at similar rates between groups). Patients received antibiotics a median 3 days.
At 14 days, there were no differences between groups in the rates of acute kidney injury, need for renal replacement therapy, or death.
Those receiving cefepime were adjudicated to have greater rates of neurologic dysfunction (a positive CAM-ICU test or RASS score -4 or -5). At 14 days, patients receiving cefepime had 11.9 days free of delirium and coma vs. 12.2 days for those prescribed pip-tazo. (Higher = better for days free of bad things.)
I say “adjudicated” because while the delirium assessments performed at Vanderbilt (where CAM-ICU was pioneered) may be the best anywhere, there’s still some subjective interpretation involved (did he squeeze my hand? I think he did), and this was an open-label trial, in which cefepime was already suspected of causing neurologic dysfunction, all of which together create some potential for unmeasured bias.
The absolute rates of neurologic dysfunction (as defined above) were 3.5% higher in the cefepime group (20.8% vs 17.3%). If broadly true, that would signify a number needed to harm of about 29 to induce delirium or coma with cefepime, relative to Zosyn.
Cefepime crosses the blood-brain barrier, and dozens of case reports have associated the drug with seizures, encephalopathy, delirium, and coma, with a possibly increased incidence of neurologic adverse effects among patients with renal failure or sepsis.
Zosyn seems to have been cleared of all renal wrongdoing and to have won this round against cefepime. It will be interesting to see if and how cefepime’s potential neurotoxicity is further tested, at Vanderbilt or elsewhere. Read in JAMA