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

Materials and Methods

Our study area was located in the southeastern part of Senegal (12°11'W, 12°33'N) in Kedougou, a department (the first-level administrative subdivision of the region) named after the town Kedougou, which is surrounded by an area of savannah and forest. In this paper the expression Kedougou area refers to the Savannah and forest galleries area, where most of our study took place. The population is essentially rural. The Kedougou area, which belongs to the Sudan-Guinean climate, is located at 1,200-1,300 mm isohyets; the area is part of the rainiest Senegalese region. The rainy season generally lasts from May/June to October/November, with maximum rainfall generally recorded in August or September. We conducted entomologic investigations in the Kedougou area in June, October, and November 1999. This schedule for site visits was based on our previous experience in the field and had two objectives: 1) to investigate vertical transmission of arboviruses with the emergence of nulliparous adult mosquitoes at the beginning of a rainy season after an arbovirus amplification, and 2) to isolate the maximum amount of virus at the end of the rainy season, the period of maximum arbovirus amplification. After results of virus isolation from mosquitoes showed evidence of DENV-2 circulation in 1999, we conducted intensive mosquito collections during the rainy season in June, August, September, October, and November 2000.

Because many arboviruses of medical and veterinary interest circulate in the study area, we used a variety of sampling methods to collect a wide range of mosquito species. Mosquitoes were then sampled by using human landing catches, CDC light traps with or without CO2, and animal bait traps. However, for collection of dengue vectors, we captured mosquitoes exclusively by human landing collections, using persons vaccinated against yellow fever virus and taking malaria chemoprophylaxis. These captures occurred between 5:30 p.m. and 8:30 p.m. in the forest gallery (located 10 km from Kedougou) and the villages of Ngari, Silling, Bandafassi, and Kénioto (Figure 1). The ecologic characteristics of the area have been described.[7,8,9,10,11] Twenty-four human volunteers participated each evening, including 18 persons in the forest gallery and 6 in one village. In the forest gallery, mosquito catches were performed at the ground level and in the canopy. Five platforms 6-9 m high served as capture sites in the canopy. Captured mosquitoes were frozen and then sorted on a chill table by using identification keys established by Edwards,[12] Ferrara et al.,[13] Huang,[14] and Jupp.[15] Mosquitoes were sorted into monospecific pools and frozen in liquid nitrogen for virus isolation attempts.

Map of the Kedougou area, Senegal, showing geographic position of villages and forest gallery where dengue virus serotype 2 vectors were collected.

To verify virus maintenance in the field by vertical transmission (after evidence of DENV-2 circulation was obtained in 1999), wild mosquito breeding sites were investigated. This investigation was undertaken during the dry season (outside of the adult mosquito's activity season), in February 2000, when all sylvatic mosquito breeding sites were dry. Tree holes were scraped with spoons and knives to collect Aedes eggs, known to be resistant to desiccation. Samples from these tree holes were stored in plastic boxes and flooded in the laboratory for egg hatching. Larvae were reared to the adult stage, then frozen, identified, and pooled by species for virus isolation attempts.

Virus isolations were performed on AP61 (Aedes pseudoscutellaris) mosquito cells lines, as described by Digoutte et al..[16] Virus was identified by using immunofluorescence with a specific immune ascitic fluid and confirmed by complement fixation or neutralization tests.

After initial results of virus isolation from mosquitoes showed evidence of DENV-2 circulation, serologic investigations on monkeys were undertaken January 31-February 6, 2000. Blood samples were collected from each wild monkey captured with flexible nets in the forest gallery. Briefly, animals were immobilized with ketamine HCl (Merial, Lyon, France), 10-15 mg/kg of body weight, injected intramuscularly. Blood was collected by femoral venipuncture into a heparinized vacuum tube, and gender was determined by inspection. Dental casts, morphologic measurements, and features relating to reproductive status (i.e., nipple length, scrotal pigmentation) were used to assign age classes of the animals. Each was tagged and released in the wild. In the field laboratory, heparinized blood was centrifuged for 15 minutes at 2,000 rpm, and the plasma was removed and stored in liquid nitrogen until testing. Serum samples were tested for DENV-2 immunoglobulin M (IgM)/IgG antibodies by using the enzyme-linked immunosorbent assay, as described.[17]

Rainfall fluctuations from 1972 to 1999 were retrospectively analyzed with respect to 1961-1990 mean rainfall (the normal level, as defined by the World Meteorological Organization) and correlated to DENV-2 emergence in the same period. Anomalies were calculated by subtracting the recorded seasonal rainfall during May to October in each year from the seasonal rainfall mean (or normal) from 1961 to 1990. Rainfall data were provided by Agence pour la Sécurité de la Navigation Aérienne, Dakar, Senegal. Only rainy season months were taken into account in the analysis.

Several entomologic indices were estimated: 1) the true infection rate (estimated number of dengue virus-positive mosquitoes per 100 mosquitoes tested) by using the methods of Chiang and Reeves[18] and Walter and others;[19] 2) and the entomologic inoculation rate (number of infected mosquito bites per human per evening). Rates obtained were compared by using the chi-square test with p <0.05 considered significant.