Pesticide pollution in freshwater paves the way for schistosomiasis transmission

Abstract

Schistosomiasis is a severe neglected tropical disease caused by trematodes and transmitted by freshwater snails. Snails are known to be highly tolerant to agricultural pesticides. However, little attention has been paid to the ecological consequences of pesticide pollution in areas endemic for schistosomiasis, where people live in close contact with non-sanitized freshwaters. In complementary laboratory and field studies on Kenyan inland areas along Lake Victoria, we show that pesticide pollution is a major driver in increasing the occurrence of host snails and thus the risk of schistosomiasis transmission. In the laboratory, snails showed higher insecticide tolerance to commonly found pesticides than associated invertebrates, in particular to the neonicotinoid Imidacloprid and the organophosphate Diazinon. In the field, we demonstrated at 48 sites that snails were present exclusively in habitats characterized by pesticide pollution and eutrophication. Our analysis revealed that insensitive snails dominated over their less tolerant competitors. The study shows for the first time that in the field, pesticide concentrations considered "safe" in environmental risk assessment have indirect effects on human health. Thus we conclude there is a need for rethinking the environmental risk of low pesticide concentrations and of integrating agricultural mitigation measures in the control of schistosomiasis.

Figure 1

Dominance of host snails of the pathogens of human schistosomiasis among the study sites. Sites were depicted by the shapes of the icons as either reservoirs (triangles), streams/channels (circles) or rice fields (squares). The dominance of host snails (number of transmitting planorbid snails/total number of individuals per site) is represented by the shade of the icon. Maps created using DIVA-GIS 7.5.0. https://diva-gis.org/.

Figure 2

Species sensitivity distribution (SSD) of freshwater macroinvertebrates from the study region to common agricultural insecticides. Data points show the acute LC_5024h for various species. The SSD curves were fitted using a quasibinomial GLM with logit-link; means ± 95% confidence intervals are shown. (a) Sensitivity distribution to the neonicotinoid insecticide imidacloprid. χ² = 230.69, res. df = 11, p < 0.001, McKelvey-Zavoina’s pseudo-R² = 0.29. For Melanoides sp., Bulinus africanus and Biomphalaria pfeifferi the LC₅₀ exceeded the highest test concentration and was extrapolated from non-linear regression (Melanoides sp.) or estimated. (b) Sensitivity distribution to the organophosphorus insecticide Diazinon. χ² = 115.89, res. df = 8, p < 0.001, McKelvey-Zavoina’s pseudo-R² = 0.40. (a,b) The acute LC_5048h of the most sensitive standard reference taxa (Chironomus riparius and Daphnia magna) was added for comparison^, and used for the calculation of toxic units (see text).

Figure 3

Pesticide pollution increases the incidence (probability of occurrence) of snails that act as hosts of schistosomiasis. Binomial GLM with complementary log-log link function; χ² = 7.60, res. df = 46, p = 0.006, McFadden’s pseudo-R² = 0.16. Means ± 95% confidence intervals are shown. Pesticide pollution was quantified as log₁₀ of the maximum ratio of a pesticide concentration measured in a grab sample of water vs. the acute LC₅₀of that pesticide for a standard reference organism (TU_max). TU_max of marginally polluted sites (n = 4) was set to a minimum of TU −5.

Figure 4

Ranking the relevance of environmental variables in driving the abundance of host snails. We combined all environmental variables that on their own showed a significant effect on the incidence or on the population density of host snails and combined them in a hurdle model. The model was subjected to hierarchical partitioning to identify the independent contribution of each environmental variable to the goodness-of-fit (quantified as log-likelihood of the hurdle model). Each step, an environmental variable was either included both in the zero and the count part of the model or excluded completely.

Figure 5

Principal component analysis of the environmental variables that drive the abundance of Schistosoma hosts. The 1^st principal component explains 33.0% of the variation and is associated with pesticide pollution, turbidity and the dominance of potential competitor species of the host snails. The 2^nd principal component explains 29.4% of the variation and is associated with the species richness, dissolved oxygen and again with the dominance of competitors. Colors indicate the number of host snails collected.

Figure 6

Pesticide pollution favors tolerant snails over less tolerant competitors. (a) No significant change in the community composition of grazers, predators and other taxa with pesticide pollution (PERMANOVA; n = 48, F = 0.80, res. df = 46, p = 0.502, R² = 0.02). For the graph, the range of TU_max values was evenly split in three categories, and for each pollution category the mean proportion of each guild on the macroinvertebrate community is shown. Because some taxa belong to more than one guild, we calculated proportions as the individuals in a guild divided by the summed-up individuals in all guilds (≠ the total individual number) so that the proportions sum up to 1. (b) Within the guild of grazers, the dominance of snails increases with pesticide pollution (n = 47, χ² = 13.82, res. df = 45, p < 0.001, McKelvey-Zavoina’s pseudo-R² = 0.37). One site was omitted because no grazers were found. Quasi-binomial GLM with logitlink function; means ± 95% confidence intervals are shown.