Impact of Molecular Adsorbent Recirculating System Therapy on Piperacillin Concentration in Critically Ill Patients: An Observational Cohort Study

OBJECTIVES: To assess the impact of Molecular Adsorbent Recirculating System (MARS) therapy on total plasma piperacillin concentrations in critically ill patients with acute liver failure or acute-on-chronic liver failure. DESIGN: Retrospective, single-center observational cohort study. SETTING: University ICU. PATIENTS: Consecutive adult patients treated with MARS therapy and piperacillin-tazobactam between March 2023 and June 2025. INTERVENTIONS: MARS therapy was delivered as 8-hour sessions for up to three consecutive days. Piperacillin-tazobactam was administered by continuous infusion following a loading dose. MEASUREMENTS AND MAIN RESULTS: Venous blood samples were collected immediately before and after each MARS session. Total piperacillin concentrations were quantified using ultrahigh-performance liquid chromatography. The primary endpoint was the relative reduction in plasma piperacillin concentration induced by MARS therapy. Thirty patients (121 piperacillin measurements across 77 sessions) were analyzed. Seventy-three percent had cirrhosis (median Model for End-Stage Liver Disease Score 31 [23-38]). The mean relative reduction in piperacillin concentration per session was 34% (± 7.9%; p < 0.05). In univariable mixed-effects analysis, MARS session (β -47.49 mg/L; p < 0.001) and creatinine clearance (β -0.37 mg/L per mL/min; p < 0.001) were associated with lower concentrations. In multivariable modeling, higher creatinine clearance was independently associated with greater intra-session piperacillin loss ( p < 0.001), whereas higher bilirubin levels were associated with reduced drug loss ( p = 0.01). Overall, 31.6% of post-session concentrations were less than 80 mg/L, and 38% of patients had at least one concentration outside the predefined target range. CONCLUSIONS: MARS therapy was associated with a one-third reduction in total plasma piperacillin concentrations per session. These findings support systematic therapeutic drug monitoring and individualized antibiotic dose adjustment during MARS therapy.