Population pharmacokinetic modelling and simulation
Two previous published population pharmacokinetic models have been published for hydroxychloroquine (HCQ) in patients with RA [3] and lupus erythematosus [4] using whole-blood HCQ concentrations. Four additional models describe plasma concentrations or merged blood and plasma concentrations [13-15]. Blood concentrations are known to be more reproducible (because higher, with lesser analytical sensitivity issues). We therefore used the model by Caramichael et al [3] for subsequent simulation after independent validation. The authors developed a one-compartment model with first-order elimination and absorption and an absorption lag time. Nine of the patients received oral dose and intravenous infusion for the bioavailability study. Patients received Plaquenil® tablets, each tablet having 200 mg of racemic HCQ sulfate equivalent to 155 mg of racemic HCQ base. The only covariate retained in the model is the methotrexate, a drug most commonly used in RA.
This model was externally validated using digitised blood concentrations obtained from Morito et al. paper. In this study, HCQ was used to treat patients with cutaneous lupus erythematosus [4]. The 90 patients received one of these three dosing regimens of HCQ sulfate, depending on their ideal body weight: 200 mg daily (n=20), 200 mg and 400 mg every other day (n=55) or 400 mg daily (n=15). The steady-date blood concentration data (3 samples/patient) were digitised from Figure 1 in the paper by Morito et al [4], using MATLAB R2016b software (The MathWorks Inc, Natick City, MA, USA). Monte Carlo simulations were performed using NONMEM software, version 7.3 (Icon Development Solutions, Ellicott City, MD, USA).
This model was also used for simulations of the serum concentrations serum PK data including data from 20 patients from Gautret et al. study [6] with a serum/whole blood correction ratio of 0.53 [16]. Gautret et al. study patients were confirmed COVID-19 and received 200 mg every 8 hours during 10 days. It is assumed that they were trough concentrations, measured before the first dose of the indicated day. NONMEM software was used for this purpose.
Subsequently, we used this model to perform simulations of blood concentrations of HCQ base for different dosing protocols for treatment of COVID-19. Table 1 includes the relevant information on the simulations performed. The figures were generated using MATLAB software.