Indices of onchocerciasis transmission by different members of the Simulium damnosum complex conflict with the paradigm of forest and savanna parasite strains

Abstract

Onchocerciasis in savanna zones is generally more severe than in the forest and pathologies also differ geographically, differences often ascribed to the existence of two or more strains and incompatibilities between vectors and strains. However, flies in the forest transmit more infective larvae than their savanna counterparts, even in sympatry, contradicting expectations based on the forest and savanna strains paradigm. We analysed data on the numbers of Onchocerca volvulus larvae of different stages found in 10 different taxonomic categories of the Simulium damnosum complex derived from more than 48,800 dissections of flies from Sierra Leone in the west of Africa to Uganda in the east. The samples were collected before widespread ivermectin distribution and thus provide a baseline for evaluating control measures. Savanna species contained fewer larvae per infected or per infective fly than the forest species, even when biting and parous rates were accounted for. The highest transmission indices were found in the forest-dwelling Pra form of Simulium sanctipauli (616 L3/1000 parous flies) and the lowest in the savanna-inhabiting species S. damnosum/S. sirbanum (135) and S. kilibanum (65). Frequency distributions of numbers of L1–2 and L3 larvae found in parous S. damnosum/S. sirbanum, S. kilibanum, S. squamosum, S. yahense, S. sanctipauli, S. leonense and S. soubrense all conformed to the negative binomial distribution, with the mainly savanna-dwelling species (S. damnosum/S. sirbanum) having less overdispersed distributions than the mainly forest-dwelling species. These infection patterns were maintained even when forest and savanna forms were sympatric and biting the same human population. Furthermore, for the first time, levels of blindness were positively correlated with infection intensities of the forest vector S. yahense, consistent with relations previously reported for savanna zones. Another novel result was that conversion rates of L1–2 larvae to L3s were equivalent for both forest and savanna vectors. We suggest that either a multiplicity of factors are contributing to the observed disease patterns or that many parasite strains exist within a continuum.