Modelling the impact of ivermectin on River Blindness and its burden of morbidity and mortality in African Savannah: EpiOncho projections

Abstract

Background: The African Programme for Onchocerciasis Control (APOC) has refocused its goals on the elimination of infection where possible, seemingly achievable by 15-17 years of annual mass distribution of ivermectin in some African foci. Previously, APOC had focused on the elimination of onchocerciasis as a public health problem. Timeframes have been set by the World Health Organization, the London Declaration on Neglected Tropical Diseases and the World Bank to achieve these goals by 2020-2025.

Methods: A novel mathematical model of the dynamics of onchocercal disease is presented which links documented associations between Onchocerca volvulus infection and the prevalence and incidence of morbidity and mortality to model outputs from our host age- and sex-structured onchocerciasis transmission framework (EpiOncho). The model is calibrated for African savannah settings, and used to assess the impact of long-term annual mass administration of ivermectin on infection and ocular and skin disease and to explore how this depends on epidemiological and programmatic variables.

Results: Current onchocerciasis disease projections, which do not account for excess mortality of sighted individuals with heavy microfilarial loads, underestimate disease burden. Long-term annual ivermectin treatment is highly effective at reducing both the morbidity and mortality associated with onchocerciasis, and this result is not greatly influenced by treatment coverage and compliance. By contrast, impact on microfilarial prevalence and intensity is highly dependent on baseline endemicity, treatment coverage and systematic non-compliance.

Conclusions: The goals of eliminating morbidity and infection with ivermectin alone are distinctly influenced by epidemiological and programmatic factors. Whilst the former goal is most certainly achievable, reaching the latter will strongly depend on initial endemicity (the higher the endemicity, the greater the magnitude of inter-treatment transmission), advising caution when generalising the applicability of successful elimination outcomes to other areas. The proportion of systematic non-compliers will become far more influential in terms of overall success in achieving elimination goals.

Figure 1

Schematic representation of the disease model. Prevalence of troublesome itch was estimated based on a relationship with the prevalence of adult female worms, previously derived using the ONCHOSIM model [25,26]. Incidence of blindness was estimated as a function of microfilarial load (lagged by two years) based on a log-linear Poisson model [46]. The number of individuals with visual impairment was estimated using a published ratio between the prevalence of visual impairment and that of blindness [47]. Excess mortality due to onchocerciasis was assumed to occur via mortality among individuals suffering onchocerciasis-related vision loss (blindness and visual impairment) [10,48], plus an independent (from the former) risk of mortality among sighted individuals with high microfilarial load (lagged by two years) [8,9]. Further descriptions are provided in the main text and Additional file 1: Text S1.

Figure 2

Relationship between the level of endemicity and pre-control disease burden associated with onchocerciasis in savannah areas of Africa. Total disability adjusted life-years (DALY) associated with onchocerciasis (black); years of life with disability (YLD) associated with onchocerciasis-related blindness (dark blue); YLD associated with onchocerciasis-related visual impairment (red); YLD associated with onchocerciasis-related troublesome itch (green); years of life lost (YLL) associated with vision loss (light blue); YLL associated with high microfilarial load (purple).

Figure 3

Impact of annual ivermectin distribution on the morbidity associated with onchocerciasis in savannah areas of Africa. (a) Prevalence of blindness due to onchocerciasis (across all ages). (b) Prevalence of visual impairment due to onchocerciasis (across all ages). (c) Prevalence of troublesome itch due to onchocerciasis (across all ages). Red, blue and green lines correspond, respectively, to a baseline endemicity of 80%, 60% and 40% microfilarial prevalence. Results shown assume a therapeutic coverage of 80%, 0.1% of systematic non-compliance, perennial transmission, and a 7% cumulative reduction in microfilarial production by female adult worms per ivermectin dose. The commencement of the intervention at year 1 is represented by the vertical dashed lines. Delays in the decrease of blindness and visual impairment are due to a two-year lag between vision loss in the present and microfilarial infection in the past. The initial sharp decline in troublesome itch is due to the assumed therapeutic effect of ivermectin followed by a more gradual decrease as adult worm prevalence declines.

Figure 4

Impact of annual ivermectin distribution on the intensity (a) and prevalence (b) of microfilarial infection. Red, blue and green lines correspond, respectively, to a baseline endemicity of 80%, 60% and 40% microfilarial prevalence. Microfilarial intensity is quantified as the mean microfilarial load per mg of skin in those aged ≥ 20 years. The dashed horizontal lines illustrate the upper and lower bounds (5% and 1% prevalence) of the current operational thresholds for cessation of treatment, namely an observed microfilarial prevalence below 5% in all surveyed villages and 1% in 90% of the surveyed villages [57]. Assumptions are as in legend of Figure  3. The inset in Figure 4 (a) zooms in microfilarial infection intensity (in the ≥ 20 yr of age) for the last four years of the simulated intervention programme.

Figure 5

Impact of annual ivermectin distribution on incidence of blindness due to onchocerciasis in savannah areas of Africa. Red, blue and green lines correspond to, respectively, a baseline endemicity of 80%, 60% and 40% microfilarial prevalence. The commencement of the intervention at year 1 is represented by the vertical dashed line. The initially delayed decrease is due to a two-year lag between blindness incidence in the present and microfilarial load in the past. Results shown assume a therapeutic coverage of 80%, 0.1% of systematic non-compliance, perennial transmission, and a 7% cumulative reduction in microfilarial production by female adult worms per ivermectin dose.

Figure 6

Impact of annual ivermectin distribution on the excess mortality associated with onchocerciasis in savannah areas of Africa. (a) Associated with a high microfilarial load. (b) Associated with vision loss (blindness/visual impairment). (c) Total excess death associated with onchocerciasis. Red, blue and green lines correspond, respectively, to a baseline endemicity of 80%, 60% and 40% microfilarial prevalence. Assumptions are as in Figure  3. The commencement of the intervention at year 1 is represented by the vertical dashed lines. The initially delayed decrease of excess mortality is due to a two-year lag between incidence of death in the present and microfilarial load in the past.