An unusually long Rift valley fever inter-epizootic period in Zambia: Evidence for enzootic virus circulation and risk for disease outbreak

Abstract

Rift valley fever (RVF) is a mosquito-borne disease of animals and humans. Although RVF outbreaks are usually reported at 5-15-year intervals in sub-Saharan Africa, Zambia has experienced an unusually long inter-epizootic/-epidemic period of more than three decades. However, serological evidence of RVF virus (RVFV) infection in domestic ruminants during this period underscores the need for comprehensive investigation of the mechanisms of virus perpetuation and disease emergence. Mosquitoes (n = 16,778) captured from eight of the ten provinces of Zambia between April 2014 and May 2019 were pooled (n = 961) and screened for RVFV genome by a pan-phlebo RT-PCR assay. Aedes mosquito pools (n = 85) were further screened by nested RT-PCR assay. Sera from sheep (n = 13), goats (n = 259) and wild ungulates (n = 285) were screened for RVFV antibodies by ELISA while genome detection in pooled sera (n = 276) from domestic (n = 248) and wild ungulates (n = 37) was performed by real-time RT-PCR assay. To examine the association between the long inter-epizootic period and climatic variables, we examined El Niño-Southern Oscillation indices, precipitation anomalies, and normalized difference vegetation index. We then derived RVF risk maps by exploring climatic variables that would favor emergence of primary RVFV vectors. While no RVFV genome could be detected in pooled mosquito and serum samples, seroprevalence was significantly high (OR = 8.13, 95% CI [4.63-14.25]) in wild ungulates (33.7%; 96/285) compared to domestic ruminants (5.6%; 16/272). Retrospective analysis of RVF epizootics in Zambia showed a positive correlation between anomalous precipitation (La Niña) and disease emergence. On risk mapping, whilst northern and eastern parts of the country were at high risk, domestic ruminant population density was low (< 21 animals/km2) in these areas compared to low risk areas (>21 animals/km2). Besides evidence of silent circulation of RVFV and the risk of disease emergence in some areas, wildlife may play a role in the maintenance of RVFV in Zambia.

Fig 1. Agroecological regions and location of sample collection areas by district.

Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license of Global Administrative Areas (GADM), copyright 2018.

Fig 2. El Niño-Southern Oscillation indices.

(A) Equatorial pacific (Niño 3.4) sea surface temperature anomalies (1973–2019). (B) Southern Oscillation Index (SOI) anomalies (1973–2019). (C) Outgoing long wave radiation anomalies (1974–2019). (D) Precipitation anomalies for Zambia (1981–2019) indicating wet (0–2) to extremely wet (>2) conditions.

Fig 3. December-January-February (DJF) precipitation, rainfall anomaly index (RAI) and Standardized precipitation evapotranspiration index (SPEI) time-series analysis.

(A,D,G) December-January-February precipitation during RVF outbreaks. (B,E,H) Rainfall anomaly Index indicating positive rainfall anomalies during RVF outbreaks. (C,F,I) Standardized precipitation evapotranspiration index (1973–1990) showing near normal, (-0.99–0.99), very wet (1.50–1.99) and extremely wet (>2.0) conditions during RVF outbreaks.

Fig 4. Mean normalized difference vegetation index (NDVI) for November and March.

NDVI were computed as means for the period 2000–2019. (A & C) NDVI for November at the onset of the rainy season. (B & D) NDVI for March at the end of the rainy season showing anomalous (NDVI > 0.76) vegetation growth. Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license of Global Administrative Areas (GADM), copyright 2018.

Fig 5. Riparian mean normalized difference vegetation index (NDVI) for the November and March.

NDVI were calculated as means for the period 2000–2019. (A) Riparian NDVI for November. (B) Riparian NDVI for March. (C) Riparian NDVI > 0.76 for November showing RVF high risk areas. (D) Riparian NDVI > 0.76 for March showing RVF high risk areas. Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license of Global Administrative Areas (GADM), copyright 2018.

Fig 6. December-January-February (DJF) precipitation and soil moisture content.

DJF precipitation and soil moisture content were computed as means with respect to the 1998–2019 and 2000–2019 climatological means, respectively. (A) Mean DJF precipitation showing high rainfall variability. (B) Mean DJF soil moisture content indicating areas that are at high risk of floods during seasons of above normal rainfall. (C) Correlation between increased riparian NDVI and mean DJF precipitation. (D) Correlation between increased riparian NDVI and mean DJF soil moisture content. Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license of Global Administrative Areas (GADM), copyright 2018.

Fig 7. Ruminant population density and RVF risk map.

(A) Domestic ruminant population density map. (B) Location of RVF outbreaks, past and present seropositive results. (C & D) RVF high risk areas in at the onset (November) and end of the rain season (March). Shapefile republished from DIVA-GIS database (https://www.diva-gis.org/) under a CC BY license of Global Administrative Areas (GADM), copyright 2018.

Fig 8. Permanent and ephemeral water bodies in Chililabombwe and Monze districts.

Waterbodies were mapped as cumulative totals for March and October for the period 2017–2020. (A & B) Dambos in the Wet (March) and Dry Season (October) in Chililabombwe District on the Copperbelt Province. (C & D) Dambos in the wet (March) and dry (October) season in Monze District in Southern Province. Black and Blue arrows indicate permanent and ephemeral water bodies, respectively. Base map republished from OpenStreetMap (https://www.openstreetmap.org/copyright) under a CC BY license.