Evaluation of factors affecting therapeutic levetiracetam drug levels in intensive care unit patients: a retrospective study

Levetiracetam (LEV) is a widely used anti-seizure medication (ASM) with linear pharmacokinetics and a wide therapeutic range. Although therapeutic drug monitoring (TDM) for LEV is not routinely recommended, critically ill patients in the intensive care unit (ICU) may experience altered pharmacokinetics. This study aimed to evaluate the factors influencing LEV serum levels in ICU patients and to identify predictors of therapeutic drug concentrations. This retrospective observational study included ICU patients aged ≥ 18 years who received LEV therapy and underwent TDM between January 2021 and July 2024 at a tertiary care hospital. LEV serum levels were categorized as subtherapeutic (< 12 mg/L), therapeutic (12–46 mg/L), or supratherapeutic (> 46 mg/L). Data on demographics, clinical parameters, renal function (estimated glomerular filtration rate [eGFR] calculated with the CKD-EPI 2009 equation and used as a surrogate for creatinine clearance), and concurrent ASM use were analyzed. Predictors of log-transformed LEV concentration were assessed using multiple linear regression. A total of 66 patients were included, with 59.1% achieving therapeutic LEV levels. Subtherapeutic and supratherapeutic levels were observed in 25.8% and 15.2% of patients, respectively. Advanced age and renal dysfunction (eGFR < 30 mL/min/1.73 m²) were significantly associated with supratherapeutic levels ( p = 0.044 and p < 0.001, respectively). In multiple linear regression on log-transformed concentrations, renal dysfunction (+ 150%), in-hospital death, treatment rather than prophylactic indication, and daily LEV dose were independent predictors of serum LEV concentration (adjusted R² = 0.325, p < 0.001). LEV concentration did not discriminate seizure occurrence (area under the receiver operating characteristic curve 0.568, 95% CI 0.415–0.722). LEV concentrations showed substantial variability in ICU patients and were independently associated with renal function, LEV dose, indication, and discharge status. However, serum LEV concentration did not discriminate seizure occurrence. These findings support individualized interpretation of LEV TDM results, particularly at the extremes of renal function, but do not demonstrate that achieving the predefined reference range improves seizure control or clinical outcomes.