What is required in terms of mass drug administration to interrupt the transmission of schistosome parasites in regions of endemic infection?

Abstract

Background: Schistosomiasis is endemic in 54 countries, but has one of the lowest coverages by mass drug administration of all helminth diseases. However, with increasing drug availability through donation, the World Health Organisation has set a goal of increasing coverage to 75 % of at-risk children in endemic countries and elimination in some regions. In this paper, we assess the impact on schistosomiasis of the WHO goals in terms of control and elimination.

Methods: We use an age-structured deterministic model of schistosome transmission in a human community and the effect of mass drug administration. The model is fitted to baseline data from a longitudinal re-infection study in Kenya and validated against the subsequent re-infection data. We examine the impact on host worm burden of the current treatment trend, extrapolated to meet the WHO goals, and its sensitivity to uncertainty in important parameters. We assess the feasibility of achieving elimination.

Results: Model results show that the current treatment trend, extrapolated to the WHO goals, is able to greatly reduce host worm burdens. If coverage is continued at the same level beyond 2020, elimination is possible for low to moderate transmission settings, where transmission intensity is defined by the basic reproduction number, R0. Low levels of adult coverage have a significant impact on worm burden in all settings. Model validation against the re-infection survey demonstrates that the age-structured model is able to match post-treatment data well in terms of egg output, but that some details of re-infection among school children and young adults are not currently well represented.

Conclusions: Our work suggests that the current WHO treatment goals should be successful in bringing about a major reduction in schistosome infection in treated communities. If continued over a 15 year period, they are likely to result in elimination, at least in areas with lower transmission.

Fig. 1

Distribution of schistosomiasis worldwide in 2013 as reported by the WHO [1]

Fig. 2

National coverage in endemic regions as reported by the WHO. The age-specific (SAC and adults) coverages, were estimated from the national figures by assuming that SAC account for 70 % of all those treated [4], and are 46 % of the current Population requiring PC (based on PCT data bank [22]). The trends in SAC and adult coverage are represented by the orange and grey lines, respectively. Solid lines represent reported figures and broken lines the project trends in coverage to reach 75 % coverage for SAC in 2020 and maintain the current ratio of SAC/adult coverage. In model projections, coverage levels are assumed to be constant after 2020

Fig. 3

Age mean intensity (epg) profiles for S.mansoni in two Kenyan villages from Fulford et al. (1995) pre (solid line) and post (dashed line) MDA with praziquantel. Note the large impact that MDA has in these settings. The village in Graph B was used in parameter estimation employing likelihood methods to fit the model to the observed mean epg counts

Fig. 4

Maximum likelihood fit of the deterministic model to age cross sectional data of the mean intensity of infection (epg counts) from Iietune village in Kenya (Fulford et al., 1995) as plotted in Fig. 3. The solid line is the best fit and the crosses are the raw data. The estimated parameter values are recorded in Table 1

Fig. 5

Changes in the mean worm burden in the three age groups and in the overall population mean between 2001 and 2020 with coverage as defined in Fig. 2. Percentages indicate the worm burden reductions in the respective age groups

Fig. 6

Predicted time series for mean intensities of infection (epg) in pre-SAC, SAC, adults and overall from 2003 to 2020, with the coverage trends defined in Fig. 2. After 2020 the coverage is assumed to remain constant at 75 % in Sac and 27.4 % in adults

Fig. 7

‘Elimination surface’ for S mansoni calculated for the parameter assignments listed in Table 1. Levels of coverage above the surface are predicted to reduce worm burden in the population within 15 years to such a level that it cannot subsequently recover

Fig. 8

Validation of model behaviour against baseline and reinfection data from Iietune village, Machakos district, Kenya. Panel A shows the fitted equilibrium age profile from the model against baseline data from 1983 and the response of the model in 1987 after two rounds of treatment (1983, 1985). Panel B shows the details of the model age profile against the re-infection survey data