Mapping the viruses belonging to the order Bunyavirales in China

Abstract

Background: Viral pathogens belonging to the order Bunyavirales pose a continuous background threat to global health, but the fact remains that they are usually neglected and their distribution is still ambiguously known. We aim to map the geographical distribution of Bunyavirales viruses and assess the environmental suitability and transmission risk of major Bunyavirales viruses in China.

Methods: We assembled data on all Bunyavirales viruses detected in humans, animals and vectors from multiple sources, to update distribution maps of them across China. In addition, we predicted environmental suitability at the 10 km × 10 km pixel level by applying boosted regression tree models for two important Bunyavirales viruses, including Crimean-Congo hemorrhagic fever virus (CCHFV) and Rift Valley fever virus (RVFV). Based on model-projected risks and air travel volume, the imported risk of RVFV was also estimated from its endemic areas to the cities in China.

Results: Here we mapped all 89 species of Bunyavirales viruses in China from January 1951 to June 2021. Nineteen viruses were shown to infect humans, including ten species first reported as human infections. A total of 447,848 cases infected with Bunyavirales viruses were reported, and hantaviruses, Dabie bandavirus and Crimean-Congo hemorrhagic fever virus (CCHFV) had the severest disease burden. Model-predicted maps showed that Xinjiang and southwestern Yunnan had the highest environmental suitability for CCHFV occurrence, mainly related to Hyalomma asiaticum presence, while southern China had the highest environmental suitability for Rift Valley fever virus (RVFV) transmission all year round, mainly driven by livestock density, mean precipitation in the previous month. We further identified three cities including Guangzhou, Beijing and Shanghai, with the highest imported risk of RVFV potentially from Egypt, South Africa, Saudi Arabia and Kenya.

Conclusions: A variety of Bunyavirales viruses are widely distributed in China, and the two major neglected Bunyavirales viruses including CCHFV and RVFV, both have the potential for outbreaks in local areas of China. Our study can help to promote the understanding of risk distribution and disease burden of Bunyavirales viruses in China, and the risk maps of CCHFV and RVFV occurrence are crucial to the targeted surveillance and control, especially in seasons and locations at high risk.

Keywords: Bunyavirales; Crimean-Congo hemorrhagic fever virus; Ecological niche model; Rift Valley fever virus; Risk assessment.

Fig. 1

The flow diagram of data collection for Bunyavirales viruses in China. *China Information System for Disease Control and Prevention (CISDCP); the report data of SFTS and HFRS from CISDCP including laboratory-confirmed and clinically diagnosed cases, were both annual data at the county level. A record represents the number of SFTS or HFRS cases reported in a year at the county level. ^#Geographic location duplicates included between point-level records and point-level records, between polygon-level records and polygon-level records, between polygon-level records and point-level records (the record about polygon corresponding to the point-level record was removed). The full name of each pathogen was shown in Table 1

Fig. 2

Geographical distribution of Bunyavirales viruses detected in humans and vectors. A Distribution of human cases infected with Bunyavirales viruses in China. *The RVFV case was imported. The number of human cases of unclassified hantavirus and SFTSV was calculated from the data of China Information System for Disease Control and Prevention. ^‡Unclassified hantavirus was a general designation for the Orthohantavirus members of the Hantaviridae family including HTNV and SEOV in China, shown as polygon data, and the numbers represents different shades of gray. B The distribution of seroprevalence (infection rate of specific antibody/antigen) for Bunyavirales viruses in humans. ^†Human beings infected with EBIV, GTV, SSHV, BATV, PUUV, WENV and WUXV were only reported with serological evidences. ^#Positive rate was not calculated for the data with detections less than 10 or not provided. C The distribution of Bunyavirales virus species determined in vectors. D The distribution of vector species that carrying Bunyavirales viruses. Bunyavirales viruses in vectors were all detected by molecular methods or pathogen isolation. Locations with positive records were positioned at the center of either provinces or cities depending on data availability, among which the data at the province level were circled with dashed lines. The full name of each pathogen was shown in Table 1

Fig. 3

The relationship matrix of Bunyavirales virus species and involved vectors. The viruses that could infect humans with any detection method, including molecular, serological and pathogen isolation method, were marked by purple fonts. ^†This represented infection of human-infected viruses in humans detected only by serological methods. *Hantavirus referred to unclassified hantavirus. Names of vectors were marked in blue if there was any proof through literature or search engines that they bite humans. The red square indicated that viruses can infect humans and can be carried by human-biting vectors, and it turned into blue if the vector cannot bite humans. Light yellow squares indicated viruses were not pathogenic to humans carried by human-biting vectors, and it turned into grey if carried by vectors of non-biting human. Bunyavirales viruses in vectors were all detected by molecular methods or pathogen isolation. The full name of each pathogen was shown in Table 1

Fig. 4

Recorded and predicted risk distribution of CCHFV presence in China. A Records of CCHFV were all from literature review. An occurrence record was defined as one or more con-firmed infection(s) with CCHFV at a unique location (the same administrative district or 10 km × 10 km pixel for points) regardless of the type of hosts, detection methods or time points with positive detection. The CCHFV occurrence records detected by serological methods were marked by green, otherwise it turned into red or blue if the records were detected by molecular methods and virus isolation. The coordinates of centroid points were both displayed for administrative district or 10 km × 10 km pixel level records. B Predicted risk distribution of CCHFV after averaged 100 boot-strapping BRT models. Source data are provided in Additional file 2

Fig. 5

Average number of suitable months per year and annual imported risk for RVFV. A The average number of suitable months per year across years 2000–2020 was shown. Places in red were suitable for more months per year, on average. The annual imported risk index was calculated by adding the monthly imported risk index at the city level. The annual city imported risk level was classified into four categories: very low (imported risk index ≤ 10^–2); low (imported risk index between 10^–2 and 10^–1); medium (imported risk index between 10^–1 and 1); high (imported risk index > 1). B The annual imported risk was from the infected countries in RVFV epidemic area to the provinces in China. The annual imported risk index at the provincial level was calculated by adding the annual imported risk index of cities in the same province. The following are the full names of RVFV infected countries: SZ (Swaziland); ZA (South Africa); NA (Namibia); BW (Botswana); UG (Uganda); SS (South Sudan); SO (Somalia); RW (Rwanda); MZ (Mozambique); MG (Madagascar); KE (Kenya); BI (Burundi); CG (Republic of Congo); CD (Democratic Republic of the Congo); TD (Chad); CF (Central African Republic); AO (Angola); SN (Senegal); NE (Niger); MR (Mauritania); ML (Mali); GN (Guinea); GM (Gambia); SD (Sudan); EG (Egypt); YE (Yemen); SA (Saudi Arabia). Source data are provided in Additional file 2