How does onchocerciasis-related skin and eye disease in Africa depend on cumulative exposure to infection and mass treatment?

Abstract

Background: Onchocerciasis (river-blindness) in Africa is targeted for elimination through mass drug administration (MDA) with ivermectin. Onchocerciasis may cause various types of skin and eye disease. Predicting the impact of MDA on onchocercal morbidity is useful for future policy development. Here, we introduce a new disease module within the established ONCHOSIM model to predict trends over time in prevalence of onchocercal morbidity.

Methods: We developed novel generic model concepts for development of symptoms due to cumulative exposure to dead microfilariae, accommodating both reversible (acute) and irreversible (chronic) symptoms. The model was calibrated to reproduce pre-control age patterns and associations between prevalences of infection, eye disease, and various types of skin disease as observed in a large set of population-based studies. We then used the new disease module to predict the impact of MDA on morbidity prevalence over a 30-year time frame for various scenarios.

Results: ONCHOSIM reproduced observed age-patterns in disease and community-level associations between infection and disease reasonably well. For highly endemic settings with 30 years of annual MDA at 60% coverage, the model predicted a 70% to 89% reduction in prevalence of chronic morbidity. This relative decline was similar with higher MDA coverage and only somewhat higher for settings with lower pre-control endemicity. The decline in prevalence was lowest for mild depigmentation and visual impairment. The prevalence of acute clinical manifestations (severe itch, reactive skin disease) declined by 95% to 100% after 30 years of annual MDA, regardless of pre-control endemicity.

Conclusion: We present generic model concepts for predicting trends in acute and chronic symptoms due to history of exposure to parasitic worm infections, and apply this to onchocerciasis. Our predictions suggest that onchocercal morbidity, in particular chronic manifestations, will remain a public health concern in many epidemiological settings in Africa, even after 30 years of MDA.

Fig 1. Comparison of ONCHOSIM model predictions with multi-country data by age patterns for different onchocercal skin disease subtypes and endemicity strata.

The data points [42] consist of three onchocerciasis endemic countries with individual data from 4,810 persons (≥5 years) living in 26 villages. Each panel represents a different disease manifestation or disease stage. The lines represent the model predictions by ONCHOSIM by pre-control age patterns in prevalence of onchocercal skin disease. Please note the different scales for the y-axes in the panels.

Fig 2

Pre-control prevalence of onchocercal eye disease (OED) by mf prevalence in the total population: model predictions versus observed data (data points) for A) savanna areas, and B) forest areas. The red and blue lines represent the model-predicted pre-control prevalence of blindness and vision loss (any OED, i.e. blindness + visual impairment), respectively. In panel 2A the vertical dotted line is the mf prevalence at 73% at which the prevalence of visual impairment is assumed to be 1.8 times the prevalence of blindness [48]. In panel 2B the blue triangles represent data points for the data on any OED [39]. Please note the different infection metrics for measurement of onchocerciasis prevalence (mf prevalence versus community microfilarial load, CMFL).

Fig 3. The model-predicted association between prevalence of nodules and prevalence of subtypes of OSD at the community level.

The data [42,58,59,61] consist of a combined 19,781 persons living in 64 villages in Africa. Data were collected by means of cross-sectional dermatological surveys of individuals aged ≥5 years. Each panel represents the prevalence of a different disease manifestation or disease stage (y-axis). The black lines represent the pre-control model predictions by ONCHOSIM for prevalence of onchocercal skin disease by prevalence of infection. The model-predictions for mf prevalence rates below 15% are not shown due to unstable runs. Please note the different scales for the y-axes in the panels. The data from Murdoch et al. 2002 [42] (blue coloured bullets) were used for the quantification of the model, and the remaining data sources [58,59,61] were used for external validation of the model. The data from Murdoch et al. 2017 [61] only contains data on the prevalence of troublesome itch rather than severe itch. Troublesome itch was defined as any form of itching with or without insomnia, whereas severe itching was defined as itching with insomnia.

Fig 4. Predicted impact of annual mass drug administration (MDA) on the prevalence of morbidity due to onchocerciasis.

Coloured lines represent different levels of treatment coverage. The lines that start at different points on the y-axis represent various pre-control endemicity levels (from upper to lower lines: very hyperendemic, hyperendemic, mesoendemic). Different panels represent the various subtypes of onchocercal skin disease (OSD) and onchocercal eye disease (OSD). The predicted trends are based on the average of 750 simulations. Please note the different scales for the y-axes in the panels.