S214. HEART RATE VARIABILITY PREDICTS CLOZAPINE PLASMA LEVEL IN PATIENTS WITH SCHIZOPHRENIA IN A POLYPHARMACY CONDITION

Abstract Background Clozapine (CLZ) is most specific suitable treatment of refractory schizophrenia; however, it might lead a variety of adverse reactions that cause drug non-compliance. Although it is necessary to conduct therapeutic drug monitoring through measuring psychoactive drug plasma concentration measurement, it has limitation in that it takes several days and it is invasive. Previous studies have suggested that the heart rate variability (HRV) might serves as non-invasive and fast marker to monitoring pharmacodynamic intra-individual effects of the drug. This study aimed to investigate whether and which HRV parameters predict CLZ plasma concentration in natural clinical and polypharmacy conditions. Methods Total 85 patients (Male = 49 and Female = 36; Mean age = 35.95±9.45 years old; Mean CLZ intake = 6.21±4.86 years old) with schizophrenia who had received clozapine without the change of clozapine dosage at least 7 days before blood sampling for TDM and electrocardiogram (ECG) measuring participated in this study. ECG data were collected during resting-state. Linear and nonlinear HRV analysis were ran. Results Based on significant results from correlational analysis with the CLZ plasma level, univerate generalized linear regression model (GLM) with gamma distribution were conducted with adjusting the effect of age and sex. As results, for linear parameters of HRV, pNN5, pNN10, and VLF were inversely associated with the CLZ plasma level (Ps<0.05). For nonlinear HRV, DFA2 and crossover were positively and CSE30 were inversely associated with the CLZ concentration (Ps<0.05). After that, step-wise multiple GLM was conducted and the result yield that pNN5 and crossover are best predictors of CLZ plasma level (β=-0.01, t=-3.85, P<0.001 and β=0.50, t=5.02, P<0.001, respectively). The final model with pNN5 and predicted 35.78% of CLZ concentration. Among dependent variables in this model, crossover were positively correlated with CLZ dose (r=0.48, P<0.001). Discussion HRV, especially for crossover, could be a candidate of a surrogate marker of CLZ concentration for drug-monitoring in patients with treatment-resistant schizophrenia in a polypharmacy condition. Findings of this study would helpful to make a clinical decision of CLZ dosage adjustment on site without the delay.