Risk prediction for linezolid-induced thrombocytopenia and exposure-guided safety optimization

Linezolid (LZD)-associated thrombocytopenia (TP), a type of major hematotoxicity that can lead to the discontinuation of LZD treatment in patients with methicillin-resistant Staphylococcus aureus infections, may be prevented through therapeutic drug monitoring (TDM) combined with the assessment of individual pharmacokinetics and risk factors within a model-informed precision-dosing framework. Exposure-related risk factors including renal impairment, advanced age, pediatric status, metabolite accumulation, and concomitant medications with the potential for drug-drug interactions alter systemic pharmacokinetics and increase the probability of TP. By contrast, exposure-independent factors such as low-baseline platelet count, prolonged treatment duration, and infection severity further affect susceptibility to platelet decline. Population pharmacokinetic/pharmacodynamic models and machine learning-based approaches have been developed to predict TP using early treatment drug concentrations and platelet dynamics. Collectively, these findings support a shift from fixed dosing to exposure-guided individualized therapy. Early TDM with risk stratification enables proactive dose adjustment for the maintenance of concentrations below toxicity-associated thresholds while preserving antimicrobial efficacy. Therefore, model-informed precision-dosing represents a rational and clinically applicable strategy to prevent LZD-induced TP and allow the safe continuation of therapy.