RESUMO
Background: Effective treatment using antibiotic vancomycin requires close monitoring of serum drug levels due to its narrow therapeutic index. In the current practice, physicians use various dosing algorithms for dosage titration, but these algorithms reported low success in achieving therapeutic targets. We explored using artificial intelligent to assist vancomycin dosage titration. Methods: We used a novel method to generate the label for each record and only included records with appropriate label data to generate a clean cohort with 2,282 patients and 7,912 injection records. Among them, 64% of patients were used to train two machine learning models, one for initial dose recommendation and another for subsequent dose recommendation. The model performance was evaluated using two metrics: PAR, a pharmacology meaningful metric defined by us, and Mean Absolute Error (MAE), a commonly used regression metric. Results: In our 3-year data, only a small portion (34.1%) of current injection doses could reach the desired vancomycin trough level (14-20 mcg/ml). Both PAR and MAE of our machine learning models were better than the classical pharmacokinetic models. Our model also showed better performance than the other previously developed machine learning models in our test data. Conclusion: We developed machine learning models to recommend vancomycin dosage. Our results show that the new AI-assisted dosage titration approach has the potential to improve the traditional approaches. This is especially useful to guide decision making for inexperienced doctors in making consistent and safe dosing recommendations for high-risk medications like vancomycin.
RESUMO
BACKGROUND/OBJECTIVES: The pharmacokinetics (PK) of drugs is dramatically altered in critical illness. Augmented renal clearance (ARC), a phenomenon characterized by creatinine clearance (CrCl) greater than 130 ml/min/1.73m2, is commonly described in critically ill patients. Levetiracetam, an antiepileptic drug commonly prescribed for seizure prophylaxis in the neurosurgical ICU, undergoes predominant elimination via the kidneys. Hence, we hypothesize that current dosing practice of intravenous (IV) levetiracetam 500 mg twice daily is inadequate for critically ill patients due to enhanced drug elimination. The objectives of our study were to describe the population PK of levetiractam using a nonparametric approach to design an optimal dosing regimen for critically ill neurosurgical patients. METHODS: This was a prospective, observational, population PK study. Serial blood samples were obtained from neurosurgical ICU patients who received at least one dose of IV levetiracetam. We used uHPLC to analyze these samples and Pmetrics™ software to perform PK analysis. RESULTS: Twenty subjects were included, with a median age of 54 years and CrCl of 104 ml/min. A two-compartmental model with linear elimination adequately described the profile of levetiracetam. Mean clearance (CL) was 3.55 L/h and volume of distribution (V) was 18.8 L. No covariates were included in the final model. Monte Carlo simulations showed a low probability of target attainment (PTA, trough at steady state of ≥6 mg/L) with a standard dose of 500 mg twice daily. A dose of at least 1000 mg twice daily was required to achieve 80% PTA. Two subjects, both with subtherapeutic trough levels, developed early onset seizures. CONCLUSION: Our study examined the population PK of levetiracetam in a critically ill neurosurgical population. We found that this population displayed higher clearance and required higher doses to achieve target levels.