Macrofilaricidal Efficacy of Repeated Doses of Ivermectin for the Treatment of River Blindness

Abstract

Background: Mass drug administration (MDA) with ivermectin is the cornerstone of efforts to eliminate human onchocerciasis by 2020 or 2025. The feasibility of elimination crucially depends on the effects of multiple ivermectin doses on Onchocerca volvulus. A single ivermectin (standard) dose clears the skin-dwelling microfilarial progeny of adult worms (macrofilariae) and temporarily impedes the release of such progeny by female macrofilariae, but a macrofilaricidal effect has been deemed minimal. Multiple doses of ivermectin may cumulatively and permanently reduce the fertility and shorten the lifespan of adult females. However, rigorous quantification of these effects necessitates interrogating longitudinal data on macrofilariae with suitably powerful analytical techniques.

Methods: Using a novel mathematical modeling approach, we analyzed, at an individual participant level, longitudinal data on viability and fertility of female worms from the single most comprehensive multiple-dose clinical trial of ivermectin, comparing 3-monthly with annual treatments administered for 3 years in Cameroon.

Results: Multiple doses of ivermectin have a partial macrofilaricidal and a modest permanent sterilizing effect after 4 or more consecutive treatments, even at routine MDA doses (150 µg/kg) and frequencies (annual). The life expectancy of adult O. volvulus is reduced by approximately 50% and 70% after 3 years of annual or 3-monthly (quarterly) exposures to ivermectin.

Conclusions: Our quantification of macrofilaricidal and sterilizing effects of ivermectin should be incorporated into transmission models to inform onchocerciasis elimination efforts in Africa and residual foci in Latin America. It also provides a framework to assess macrofilaricidal candidate drugs currently under development.

Keywords: ivermectin; macrofilaricide; multiple dose; onchocerciasis; river blindness.

Figure 1.

Schematic representation of the study design.

Figure 2.

Summary of estimated parameters that define the antifilarial activity of multiple doses of ivermectin under different assumptions on the transmission dynamics during the trial. In each panel, the data points and vertical lines indicate the means and 95% Bayesian credible intervals (BCIs) of the posterior distributions estimated from variants A (constant short-term and long-term transmission), B (seasonal short-term transmission; constant long-term transmission), C (constant short-term transmission; declining long-term transmission), and D (seasonal short-term transmission; declining long-term transmission) of the best fitting model 3 (see Supplementary Table 3 for deviance information criteria). The horizontal dotted lines in each panel indicate the null (zero) effect of each parameter. Hence, parameters with BCIs that cross the dotted line can be interpreted as not statistically significantly different from zero. For example, in panels (A) and (D) none of the BCIs include 0, indicating statistically significant (permanent) macrofilaricidal and sterilizing activity of multiple-dose ivermectin regimens. By contrast, in panels (B) and (E) all BCIs cross 0, indicating no statistically significant effect of dose on either macrofilaricidal or permanent sterilizing activity. The parameter posteriors estimated from the model variants are generally similar and therefore robust to different assumptions on the transmission dynamics during the trial.

Figure 3.

Dynamics of Onchocerca volvulus viability and fertility through multiple rounds of ivermectin treatment. Each panel depicts either the percentage of live female O. volvulus (panels A and B) or the percentage of fertile (live) female O. volvulus (panels C and D) exposed to ivermectin treatments given annually (panels A and C) or 3-monthly (panels B and D) at standard (150 µg/kg) and high doses (800 µg/kg). The thick gray and black solid lines indicate, respectively, the marginal mean (averaged across all hosts) dynamics for standard or high-dose regimens. The thin gray lines indicate the individual host dynamics mediated by the estimated random effects terms within the statistical model (see Supplementary Materials for statistical details). For clarity, no graphical distinction is given among individual hosts given either standard or high-dose regimens. In each panel, the triangular and circular markers denote, respectively, the observed percentages of live or fertile (live) females extracted from participants given either standard or high-dose regimens. Vertical error bars denote 95% confidence intervals that were calculated using a nonparametric weighted bootstrapping procedure to account for the variable number of female O. volvulus extracted from each host. Horizontal error bars (often narrower than the data point) indicate the range of nodulectomy dates at each sampling round.

Figure 4.

The life expectancy of adult Onchocerca volvulus multiply exposed to ivermectin. The dashed and dotted lines in panel A show the estimated life expectancy (in years) of adult female O. volvulus exposed to annual and 3-monthly treatments with ivermectin at a standard dose (150 µg/kg) (there is no substantive difference in the corresponding estimates for a high dose of 800 µg/kg dose, Figures 2 and 3). The solid line in this panel (A) depicts the estimated life expectancy of hypothetically unexposed worms. In panel (B), the dashed and dotted lines express the estimated life expectancies as percentage reductions compared to hypothetically unexposed worms. Estimates are plotted against the date of infection, that is, the time of establishment of female worms. Treatment dates with ivermectin are represented by the triangles: 3-monthly treatments on the top row and annual treatments below. Hence, worms that establish at an earlier date are exposed to more treatments and incur greater reductions in life expectancy. In each panel, gray shaded areas represent 95% Bayesian credible intervals. Life expectancies were calculated from the estimated parameter posteriors of model 3A (the best fitting model, see Supplementary Table 3 for deviance information criteria). Details of these calculations are given in the Supplementary Materials.