Population pharmacokinetic model of ivermectin in mass drug administration against lymphatic filariasis

Abstract

Background: Ivermectin (IVM) is a broad-spectrum anthelmintic drug used to treat diseases caused by filarial worms, such as onchocerciasis and lymphatic filariasis (LF). IVM is part of a triple-drug therapy used by the Mass Drug Administration (MDA) as a preventive strategy to eradicate LF in sub-Saharan Africa. The drug shows high variability in drug exposure in previous pharmacokinetic studies. This study aims to build a population pharmacokinetic (PopPK) model to identify and quantify the possible sources of the variability of IVM exposure after a single-oral dose in LF-infected subjects and healthy individuals.

Methodology / principal findings: In this analysis, 724 samples were collected from treatment-naïve Wuchereria bancrofti-infected (n = 32) and uninfected (n = 24) adults living in Côte d'Ivoire who had received one dose of IVM as a part of triple-drug therapy. PopPK analysis was conducted using Phoenix NLME 8.3 software. The Monte Carlo simulation based on the final model was performed to simulate drug exposure among different dosing groups (200 μg/kg, 18 mg, and 36 mg). A two-compartment model with zero-order dose input into the absorption compartment with a lag time function followed by first-order absorption and linear elimination best described the IVM's pharmacokinetic (PK) parameters. The final model identifies that the PK parameters of IVM are not affected by LF infection. Sex was a significant covariate on the peripheral volume of distribution (Vp/F, 53% lower in men than in women). IVM drug exposure shows linear pharmacokinetic behavior among the simulated dosing groups with similar drug exposure based on sex.

Conclusion/significance: We have developed a PopPk model to describe and identify possible sources of the variability of IVM exposure. To our knowledge, this is the first PopPK study of IVM in patients with LF.

Trial registration: NCT02845713; NCT03664063.

Fig 1. IVM plasma concentration vs. time after 200 μg/kg dosing (n = 56).

Fig 2. Schematic description of the selected population structural model.

Tlag, lag time function; TK0, zero dose input into absorption compartment; Ka, first–order absorption rate constant; VC/F, central volume of distribution; Vp/F, peripheral volume of distribution; CL/F, clearance; Q/F, intercompartmental clearance.

Fig 3

The goodness of fit plots of final model Observed plasma concentrations of IVM versus individual predicted concentration (A) and population predicted concentration (B), conditional weighted residuals (CWRES) versus time after dose (C) and population predicted concentration (D).). Black lines in (A) and (B) are the line of identity. The blue line in (C) and (D) represent the locally weighted scatterplot smoothing line (LOWESS); dashed black lines are margins (y = 2) of outliers.

Fig 4. ETA box plots for sex impact on Vp/F.

Box plots show the impact of sex on Eta values of Vp/F in the base model (A) and after incorporating the sex effect in the final model (B). (0 = Female, 1 = male).

Fig 5. Visual predictive check (VPC) of the final model over the time from 0 and 168 hr (Left) and over the time from 0 and 72 hr for external validation (right), following IVM oral administration (n = 1000).

A solid red line represents the 50^th percentile of observed data (blue dots). Dashed red lines represent the 5^th and 95^th percentiles of observed data. Shaded areas (blue and red) represent a 95% prediction interval of the 5^th, 50^th, and 95^th simulated data.

Fig 6. Comparison of model–based simulated IVM exposure (Cmax, AUClast) across simulated dosing groups (18 mg, 200 μg/kg, and 36 mg) (upper panels) and simulated exposure stratified by sex (lower panels).

The box plot shows the median, 25th, and 75th quartiles.