Bioanalytical Methods and Data Handling
A fully validated liquid chromatography-tandem mass spectrometry
(LC-MS/MS) method determining yimitasvir concentration in phase 1 and
phase 1b studies had been reported elsewhere [7, 8].
Plasma concentration of yimitasvir in phase 2 study was analyzed using a
validated LC-MS/MS method equipped with a Shimadzu LC-30AD liquid
chromatography-SCIEX API6500 mass spectrometer. Chromatographic
retention and separation were achieved on a XBridge Peptide BEH C18, 50
× 2.1 mm, 3.5 μm column. Gradient elution was used with 2 mM ammonium
acetate in water with 0.1% formic acid as mobile phase A and
acetonitrile:methanol (30:70, v:v) as mobile phase B at a flow rate of
0.5 ml min-1. The column temperature was maintained at
45 °C. Quantitation was accomplished in positive mode with
precursor-to-production pairs of m/z 428.5→315.3 for yimitasvir and m/z
432.5→319.5 for the internal standard d8-DAG (isotope-labelled
yimitasvir), respectively.
The calibration curve was linear in the range of 5.00-5000 ng
ml-1 for both methods, and the lower limit of
quantitation (LLOQ) was 5.00 ng ml-1. Accuracy and
precision were within the acceptable criteria of ±15% for quality
control (QC) samples and of ±20% for LLOQs. Both methods were fully
validated in accordance with National Medical Products Administration
(NMPA) of China, U.S. Food and Drug Administration (U.S. FDA) and
European
Medicines Agency (EMA) guidelines. These two bioanalytical methods were
not cross-validated.
Pharmacokinetic data with an absolute value of conditional weighted
residual (CWRES) |CWRES| > 6 in the
structural models were regarded as outliers. Outliers were omitted
because these observations had a potential to negatively influence the
convergence and/or poor estimation precision of parameters. If outliers
were removed in the process of model development, the final model was
re-run with or without the outliers to assess the potential influence on
parameter estimates. If the frequency of LLOQ data was less than 10%,
PK samples below LLOQ were excluded from model development. Otherwise,
Beal’s M3 method was used for handling the LLOQ data [11]. For
covariates with missing values in less than 10% of subjects, continuous
covariates were imputed as the population median, while a new category
of ‘missing’ is generated for categorical covariates. No formal
covariate screen procedure would be conducted if the covariates missed
in more than 10% of the subjects.