Amplification of the Sylvatic Cycle of Dengue Virus Type 2, Senegal, 1999-2000: Entomologic Findings and Epidemiologic Considerations

Mawlouth Diallo, Yamar Ba, Amadou A. Sall, Ousmane M. Diop, Jacques A. Ndione, Mireille Mondo, Lang Girault, Christian Mathiot


Emerging Infectious Diseases. 2003;9(3) 

In This Article


After 8 years of silence, DENV-2 reemerged in a sylvatic cycle in the Kedougou area in 1999. With new isolations, the number of DENV-2 strains isolated from mosquitoes in Senegal reached 350. Except for Ae. aegypti and Ae. vittatus, which were associated with DENV-2 for the first time in Senegal, all strains were obtained from vectors previously described for the DENV-2 sylvatic cycle in West Africa and belonging to the subgenus Diceromyia (Ae. furcifer and Ae. taylori) and Stegomyia (Ae. luteocephalus).[20,21,22] Population density of vector species and their spatial dynamics permitted us to assess a potential role in the transmission cycle for each infected mosquito species. Ae. taylori and Ae. luteocephalus, because of their scarcity in villages, may have a role limited to the forest gallery. However, Ae. furcifer may contribute to both sylvatic transmission and virus dissemination from the forest zone to human habitats since this species is the only infected one abundant in the domestic environment. Ae. aegypti, the principal DENV-2 epidemic vector worldwide, while very scarce in our catches, was more abundant in the forest gallery than in the villages. Ae. aegypti's low rate of biting humans may reflect the zoophilic tendency of this species, which probably exists in the area only in its sylvatic form, Ae. aegypti formosus,[11] which uses tree holes as breeding sites. (By contrast, the domestic form, Ae. aegypti aegypti, preferentially colonizes artificial water containers.) However, the low entomologic inoculation rates of Ae. aegyptiin sylvatic areas suggests its limited role in sylvatic amplification cycles of DENV-2 despite its high infection rates. Although Ae. vittatus was infected with DENV-2 in our study, its role in dengue transmission has never been demonstrated. Further studies are needed to determine its involvement in the sylvatic cycle.

Sylvatic amplifications of DENV-2 have previously been observed in Senegal in 1974, 1980-1982, and 1989-1990,[6,7,8] defining periods of occurrence with silent intervals of 5-8 years.[23] This periodicity of occurrence shows the similarity between DENV-2 and yellow fever virus, which share the same vectors, vertebrate hosts, and ecologic niche.[8] Moreover, this pattern of occurrence shows that DENV-2 is maintained in the Kedougou area by a still-unknown mechanism, although two have been hypothesized: vertical transmission of the virus or its maintenance through a secondary cycle. Vertical transmission is the most probable hypothesis since it has been supported by laboratory and field evidence for Ae. aegypti[24,25] and Ae. taylori.[6] The isolation from male Ae. furcifer during this investigation is the first field evidence of vertical transmission of DENV-2 by this species in Kedougou and reinforces this hypothesis. Our investigation of tree holes also supports the hypothesis of a secondary transmission cycle involving mosquito species other than those identified as vectors. These investigations showed high representation of species considered scarce or absent, which was the case for Ae. longipalpis belonging to the Finlaya subgenus, a forest/enzootic vector of dengue in Asia; Ae. bromeliae, one of the major yellow fever virus vectors in East Africa; and Ae. stockesi.[1,14,26] The role of these species may have been overlooked because of biased sampling methods for mosquito captures. Further research about the existence of a secondary cycle is needed.

In DENV-2 sylvatic circulation, vertebrate hosts may serve either as amplifiers or reservoirs. Unfortunately, data obtained in this study from monkey serum samples were not conclusive about their role during amplification because of IgG cross-reactions between flaviviruses.[27] Thus, the role of monkeys in sylvatic cycle of DENV-2 remains unresolved. Extensive investigations are needed to assess the role of vertebrate hosts in the amplification and maintenance of DENV-2 in natural conditions through herd immunity, turnover of susceptible vertebrates hosts, or persistent infection.[28,29]

Analysis of rainfall amount and distribution did not show any correlation with sylvatic DENV-2 emergence unlike the situation with other arboviruses such as Rift Valley fever.[30,31] Rainfall has an impact on two epidemiologic parameters important in arbovirus transmission: vector density, which controls the transmission level, and adult mosquito longevity, which makes transmission possible and durable. For the sylvatic DENV-2 vectors in the Kedougou area, rainfall seems to interfere, particularly with the virus amplification period (generally October-November), coinciding with the end of the rainy season. Indeed, although vector population densities are already high at the beginning of the rainy season, the virus emergence and the maximum amplification period occur only at the end of the rainy season. Probably the greater longevity of the female vectors at the end of the rainy season, attributable to low intensity of precipitation, allows them to achieve a complete extrinsic incubation of DENV-2.

DENV-2 isolations from Ae. furcifer captured in peridomestic habitats demonstrate that the virus circulates in all villages where mosquito catches were undertaken. However, no DENV-2 clinical case was recorded in the region. This finding suggests that DENV-2 is confined to the forest or, if human-mosquito-human transmission occurs, the level is low and leads to nonsymptomatic cases. Similar situations were observed during the 1981 and 1990 DENV-2 epizootics, when only a few sporadic human cases were recorded.[32,33] The same situation occurred in Burkina Faso and Côte d'Ivoire,[21,22,34]), where sylvatic transmission of DENV-2 was observed without an epidemic. Only one epidemic, which occurred in Burkina Faso in 1982, has been reported in West Africa; that epidemic was suspected of being caused by the introduction of a virus strain from the Seychelles Islands.[35] Therefore, the risk of a DENV-2 epidemic in Senegal is very low if only the sylvatic cycle is taken into account. However, a risk exists since importation of an epidemic strain in an urban area cannot be excluded, as witnessed by the epidemic in Burkina Faso in 1982.