Over 100 Years of Rift Valley Fever: A Patchwork of Data on Pathogen Spread and Spillover

Abstract

During the past 100 years, Rift Valley fever virus (RVFV), a mosquito-borne virus, has caused potentially lethal disease in livestock, and has been associated with significant economic losses and trade bans. Spillover to humans occurs and can be fatal. Here, we combined data on RVF disease in humans (22 countries) and animals (37 countries) from 1931 to 2020 with seroprevalence studies from 1950 to 2020 (n = 228) from publicly available databases and publications to draw a more complete picture of the past and current RVFV epidemiology. RVFV has spread from its original locus in Kenya throughout Africa and into the Arabian Peninsula. Throughout the study period seroprevalence increased in both humans and animals, suggesting potentially increased RVFV exposure. In 24 countries, animals or humans tested positive for RVFV antibodies even though outbreaks had never been reported there, suggesting RVFV transmission may well go unnoticed. Among ruminants, sheep were the most likely to be exposed during RVF outbreaks, but not during periods of cryptic spread. We discuss critical data gaps and highlight the need for detailed study descriptions, and long-term studies using a one health approach to further convert the patchwork of data to the tale of RFV epidemiology.

Keywords: ProMED; Rift Valley fever phlebovirus; bunyavirales; mosquito-borne disease; notifiable disease.

Figure 1

Flowchart of the source and data acquisition. A total of 275 sources ((A) two OIE databases and ProMED, and (B–E): 272 publications) were used to inform the RVF case and RVFV seroprevalence dataset for humans and animals. RVF reports were extracted from databases (A) and supplemented with historical review papers (B). Seroprevalence data was first extracted from publications identified through cross-referencing review papers (C) and the search strategy of Clark et al., 2018 was repeated to identify and add publications from 2017 to 2020 (D). In step (E), a search was conducted to explore if publications were available for African, Southern and Western Asian countries without RVF reports and RVFV seroprevalence studies to confirm absence of RVFV evidence. Data was included until 31 December 2020; the last update was conducted on 13 January 2021 by repeating step A and D. ADM1: Administrative Level 1, e.g., districts, provinces.

Figure 2

Regional variation in RVF reports and seroprevalence studies over time. (A) Number of included publications by year. Vertical lines represent the first year during which OIE and ProMED reports were included. (B) For each country, the years during which Rift Valley fever cases were reported, and the years during which Rift Valley fever virus seroprevalence data were collected are marked for animals and humans. When results from multiple years were reported as one, e.g., 1991–2000, we marked the first year in the figure. Serological data also include studies where no RVFV antibodies were detected; note that this is not a RVFV detection chart.

Figure 3

RVFV reporting and activity. (A,B) The year of the first (A) and most recent (B) RVFV publication by country. The studies on the Canary Islands, Comoros Islands, and Mayotte are not shown. No publications were found for the countries in grey. (C,D) Number of years with human (C) or animal (D) RVF case records per country. (E,F) Percent of individuals positive for RVFV antibodies of all individuals sampled per country for human samples (E) and animal samples (F). All samples, except IgM-only records, were included. Gray indicates that no RVF case reports or seroprevalence studies were found for this country. It should be noted that the presence of individuals with a positive RVFV antibody test does not ascertain local circulation of the virus and should be interpreted with caution, particularly for countries with no confirmed cases.

Figure 4

Regional human, individuals selected by random sampling, (A,B) and ruminant (C,D) RVFV seroprevalence in the absence of recent human and/or animal RVF cases or within the same year, or year post, reported RVF cases. (A,C) Sample size and seroprevalence per country and year. The grey, dashed vertical lines represent the mean seropositivity (A,C) and equal odds (B,D). Horizontal lines represent confidence intervals around the point estimate (B,D). Southern Asia is represented by studies from Iran and India; Western Asia is represented by Iraq, Saudi Arabia, Turkey and Yemen.

Figure 5

Rift Valley fever virus seroprevalence when RVF cases had never been reported in the country (human and/or animal). The outlier with no known RVF cases represents samples from Tunisiacollected between 2017 and 2018 (34.5% positive of 470 camels) [57], and two outliers currently known to report RVF cases represent samples collected in South Sudan from 1979 to 1983 (33.7% positive of 92 ruminants, cattle and goats) [58], and Cameroon in 1968 (33.6% of 122 sheep) [59].

Figure 6

Animal and human RVF seroprevalence data from the same year and country. Point estimates represent mean seropositivity (percent of individuals testing positive of the total number of individuals tested) by country and year. The point estimates are color-coded by region. Lines represent the range in seroprevalence estimates reported when multiple publications were included in the prevalence estimate.