(M1130-09-50) Volumetric Absorptive Microsampling Method as a Better Blood Sampling Tool for the Therapeutic Drug Monitoring of Everolimus in Liver Transplant Recipients

Purpose: Therapeutic drug monitoring has been routinely performed to maintain the appropriate whole blood concentrations of everolimus with venous blood collection in patients with organ transplantation. A minimally invasive method of volumetric absorptive microsampling (VAMS) has been developed as an alternative one to the invasive venous blood sampling. This study aimed to compare the whole blood concentrations of everolimus obtained from venous blood samples (C_IV) with those from VAMS samples (C_VAMS) from patients with liver transplantation to investigate the applicability of VAMS to the monitoring of everolimus. In addition, evaluated were the effects of hematocrit and blood collection time on the applicability.

Methods: This study recruited 45 adult patients with liver transplantation who had taken the stable doses of everolimus for more than a week before the enrollment. From all study participants, venous blood samples were collected into two vacutainer tubes containing EDTA by phlebotomy and two VAMS samples by finger prick. One venous blood sample from each participant was used for the measurement of hematocrit immediately after blood collection, while the other was stored at -80 ± 10 °C until the measurement of C_IV. All VAMS samples were placed in aluminum foil pouch with a desiccant gel immediately after collection and stored at 20 ± 5 °C with humidity < 40% until the measurement of C_VAMS. C_IV and C_VAMS were measured using a validated ultra-performance liquid chromatographic method with tandem mass spectrometric detection (UPLC-MS/MS). The comparison between C_VAMS and C_IV was performed by using a Deming regression, a linear regression and a Bland-Altman analysis. The proportional and constant biases between C_VAMS and C_IV were quantified using the slope and intercept of the Deming regression line, respectively. A value of the coefficient of determination (r²) between C_VAMS and C_IV was obtained using the linear regression. In the Bland-Altman analysis, the difference between C_VAMS and C_IV (i.e., C_VAMS - C_IV) was plotted against the average value of C_VAMS and C_IV (i.e., (C_VAMS + C_IV)/2). The 95% limit of agreement (LoA) for the mean difference between C_VAMS and C_IV was obtained from the standard deviation (SD) of the C_VAMS - C_IV in the analysis. In addition, linear regression methods were used to assess the impact of hematocrit and the blood sampling time points on C_VAMS - C_IV.

Results: The final analysis included the data obtained from 32 participants. The data from thirteen participants were excluded as their VAMS samples were contaminated by touching the storage container (n=6), inspected as oversampled (n=6), and determined below the lower limit of quantification (n=1). The daily doses of everolimus in the evaluable participants were in the range of 1.0 to 2.5 mg with the median of 2.0 mg. The sampling time points after the last administration of everolimus were 0.2 to 5.85 hours with the median of 3.1 hours. The values of hematocrit were 29.8% to 49.3% with the median of 39.1%. The calibration curves of the UPLC-MS/MS assay were linear (r² ≥ 0.995) over the concentration range of 1.00 to 250 ng/mL. The accuracy and precision of the assay assessed using quality control samples were in the range of -3.5% to 4.7% and 1.8% to 7.9%, respectively. The relationship determined in Deming regression between C_VAMS and C_IV was C_IV = 1.04*C_VAMS + 0.55. Neither proportional nor constant bias was observed between C_VAMS and C_IV as the 95% confidence interval (95% CI) for the slope (0.91 to 1.18) and the intercept (-0.05 to 1.16) do not include 1 and 0, respectively. C_VAMS and C_IV were strongly associated (r² = 0.92, p < 0.0001) when determined in the linear regression. The Bland-Altman plot demonstrated a random scattering of the C_VAMS-C_IV points with the mean difference ± SD of -0.79 ± 0.90 ng/mL between C_VAMS and C_IV with their 95% LoA ranging between -2.55 and 0.97 ng/mL. Hematocrit and blood sampling time points did not have meaningful effect on the relationship between C_VAMS and C_IV as the r² values of the linear regression expressing the impact of hematocrit and time points on C_VAMS - C_IV were 0.006 (p = 0.68) and 0.003 (p = 0.77), respectively.

Conclusion: The results demonstrate a good agreement between C_VAMS and C_IV implying the interchangeability between VAMS and venous blood sampling. There was no need to consider the impacts of hematocrit or sampling time points on the interchangeability between C_VAMS and C_IV. Hence, the minimally invasive VAMS appears to be a better sampling method for the therapeutic drug monitoring of everolimus in patients with liver transplantation.

Acknowledgements: This work was funded by SoVanGen Co., Ltd and National Research Foundation of Korea (2018R1D1A1B07044329).