O21 Progress towards Therapeutic Drug Monitoring via breath analysis

Background Therapeutic Drug Monitoring (TDM) is essential aspect for the clinical management of patients. However, despite the clear advantages of TDM, it faces several challenges to being more widely used in the clinic. Specially challenging is the case of TDM in children, as they experience rapid physiologic developments, leading to great pharmacokinetic and pharmacodynamic variability. Breath analysis provides a patient-friendly approach to support TDM. To explore this possibility, we are currently running a pilot study, whereby we analyze the exhaled breath of pediatric patients receiving anti-seizure or chemotherapy drugs requiring TDM. Methods We analyzed by secondary electrospray ionization-high resolution mass spectrometry (SESI-HRMS) exhaled breath of pediatric patients under therapy for Valproic acid (VPA; n = 27), Lamotrigine (n = 19), Levetiracetam (n = 15), Oxcarbazepine (n = 11) and Methotrexate (n = 4). Systemic blood concentrations were measured simultaneously to the breath test. In the case of VPA, we constructed a regression model to predict systemic blood concentrations based on the signal intensity of breath mass spectral features. For the rest of the drugs listed, due to its current limited size, we conducted preliminary data visualization approaches. Results We found that exhaled metabolites of VPA allow to predict free VPA blood concentrations with a root-mean-square error of 1.5 mg/L for concentrations in the range 0–12 mg/L. This prediction is accomplished within 20 minutes, comprising the breath test and data analysis. For Levetiracetam, Lamotrigine and Oxcarbazepine the data analysis is still ongoing. For MTX we found breath metabolites clearly altered as a result of the drug administration. However, a great inter-individual variability was also observed. Conclusion We conclude that breath analysis may support current TDM approaches. This will lead to new opportunities to guide the dose of drugs with a high level of accuracy, in real-time and non-invasively. Disclosure(s) PS gratefully acknowledges the financial support of the Fondation Botnar (Switzerland) and the Swiss National Science Foundation (320030_173168 and PCEGP3_181300). PS is part of the board of directors of Deep Breath Initiative AG. JU was supported by research fellowship of the University Children’s Hospital Basel. VS and JVDA were supported by the Eckenstein-Geigy Foundation. *AD and PS: shared senior authorship