Flaviviruses are small, ~50 nm in diameter, enveloped, positive-sense and single-stranded RNA viruses. Flaviviruses are arthropod-borne viruses (arboviruses) belonging to the Flaviviridae family in the genus Orthoflavivirus. Over 70 viral species exist within this genus, and they are classified based on their vectors, that is, tick-borne, mosquito-borne, insect-specific or no known vector¹⁴ as indicated in Table 1. This genus includes medically or veterinary important mosquito-borne flaviviruses such as DENV, ZIKV, YFV, Japanese encephalitis virus (JEV), WNV, USUV and WSLV; tick-borne flaviviruses such as tick-borne encephalitis virus (TBEV), Kyasanur forest virus (KFDV) and Powassan virus (POWV) and as well as lesser known flaviviruses that include BANV, BAGV and SPOV. Flaviviruses are further classified into serocomplexes distinguishable by neutralising antibody reactivity and hence indicate antigenic relationships and serological cross-reactivity based on shared antigen. The amino acid sequence similarity of the envelope (E) protein is approximately 70%–80% for virus species within a serocomplex and 40%–50% between serocomplexes.¹⁵ Serocomplexes with known medical significance include the Dengue virus group, yellow fever virus group, Japanese encephalitis group, Spondweni virus group, Ntaya virus group and mammalian tick-borne viruses (see Figure 1 and Table 1). Figure 1 indicates the relationships of representative members of each complex, based on analysis of the complete polyprotein, indicating antigen relationships. Table 1 shows the geographic distribution of selected mosquito-borne members of each serocomplex relevant to South Africa, selected medically significant tick-borne flaviviruses, an example of an insect-specific flavivirus and one with no known vector.

Flaviviruses are also grouped by their clinical manifestations; however, this depends on the virus and the affected individual. Approximately 80%–85% of flavivirus infections are asymptomatic. Some flaviviruses are neurotropic, causing severe neurological syndromes including encephalitis, meningitis and acute flaccid paralysis, for instance, WNV, USUV, JEV and TBEV. Other flaviviruses cause visceral disease, which results in haemorrhagic syndromes, liver failure and vascular compromise, for instance, YFV = virus and POWV.

West Nile virus was initially detected in 1937 in Uganda in a febrile patient. Thereafter, epidemics have been reported in Europe, Africa, North America, the Middle East and West Asia. West Nile virus is a member of the JEV serocomplex with nine phylogenetic lineages based on genetic sequence variation, WNV lineage 1 to lineage 9 (WNV-L1 – L9). WNV-L1 is divided into clades A, B and C; Clade A includes strains from Africa, Europe, America and the Middle East, and Clade B, also known as Kunjin virus, was isolated in Australia and Clade C includes strains detected in India.⁹¹ WNV-L1 and L2 were detected in South Africa and are of public health concern; they are the most pathogenic, widely distributed and implicated in several outbreaks worldwide.^91,92 In addition to WNV-L1 and L2, L7 also known as Koutango virus (KOUTV) and L8 circulate in Africa.⁹³

There are other lineages that have not yet been associated with human and/or animal disease and not isolated in Africa: these include WNV-L3 (Rabensburg virus), detected in the Czech Republic; WNV-L4, isolated and reported in Russia; WNV-L5, isolated in India; WNV-L6, isolated in Spain and WNV-L9, isolated in Austria.^94,95,96 Viruses from these lineages have not yet been associated with human or animal disease.

West Nile virus is endemic to South Africa, particularly in the Highveld and Karoo regions, where its primary vector, Cx univittatus, and avian hosts such as corvids, passerine and raptors are prevalent.^6,29 Other vectors of WNV include Cx. theileri, Cx. pipiens, Cx. quinquefasciatus and Aedes spp. mosquitoes.^26,29,97 Epidemics of WNV are typically triggered by flooding and elevated temperatures, which foster an ideal environment for vector breeding. The largest WNV epidemic in South Africa occurred in 1974, resulting in over 10 000 human cases across a 2500-km² area of the Karoo and Northern Cape Provinces.²⁸ This was followed by a smaller epidemic between 1983 and 1984, coinciding with the Sindbis (SINV) virus outbreak, in the Witwatersrand-Pretoria region, now known as Gauteng.⁹⁸ SINV is an arbovirus from the family Togaviridae. In all recorded epidemics, WNV infections were self-limiting and primarily associated with mild febrile illness; symptoms included rash, myalgia and arthralgia, with no reported fatalities. Seroprevalence studies revealed that 55% of humans and 53% of wild birds in the affected areas tested positive for WNV, with some regions reporting seroprevalence as high as 80% to 85%.²⁹ Prior to these epidemics, sporadic cases were documented in the former Transvaal (Gauteng) and Orange Free State (now referred to as Free State), and serological surveys confirmed widespread WNV distribution among humans and various animal species, including cattle, horses and wildlife.^99,100,101

West Nile virus outbreaks were recorded on the Highveld in 1974, where over 10 000 febrile cases were reported; in 1984 and 2004, where hundreds of people were affected. In addition to outbreaks, sporadic WNV cases are diagnosed by the National Institute for Communicable Diseases (NICD). Cases are usually confirmed by detecting immunoglobulin M (IgM) antibody; however, serological cross-reactivity must be considered when interpreting results. Cases have been recorded from different provinces in South Africa, for example, in 2017, a veterinarian from Mpumalanga province who frequently performed bird-tagging and animal autopsies, tested positive for WNV-specific immunoglobulin M (IgM) and immunoglobulin G (IgG) antibodies by enzyme-linked immunosorbent assay (ELISA).¹⁰² On a separate occasion, a patient from a farm in the Northern Cape province tested positive for WNV.¹⁰³

In addition to the acute cases detected by NICD, seroprevalence studies and vector surveillance have assisted with identifying WNV in South Africa. A seroprevalence study detected WNV antibodies in 7.9% of veterinarians in South Africa, their distribution corresponding with reported cases in WNV-positive animals.¹⁰⁴ A passive entomological survey conducted from 2011 to 2018 detected WNV in Cx. univittatus, Cx. pipiens, Cx. theileri, Cx. poicilipes, Cx. simpsoni, Cx. bitaeniorhynchus, An. gambiae and Aedes mosquitoes from Limpopo, Mpumalanga and Gauteng provinces. The study indicated the emergence of potential new mosquito vectors in conservation and peri-urban areas.⁸ Vector competency studies are required to confirm the role of these mosquitoes in the transmission and spread of WNV in South Africa.

West Nile virus is neurotropic and affects humans, both domesticated and wild animals and avian species. In humans, approximately 80% of WNV infections are asymptomatic, while 20%–25% develop West Nile fever, a self-limiting condition characterised by high fever, headache, body aches, maculopapular rash, nausea and vomiting.¹⁰⁵ Only about 1% of cases lead to serious neurological complications, including encephalitis, meningitis, myelitis and acute flaccid paralysis.^92,105 Disturbingly, WNV has been detected in 3.5% of unresolved cases of human neurological disease in Gauteng provincial hospitals.¹⁰⁶ In a separate study, WNV was identified as the causative agent in 8% to 11% of hospitalised patients from Mpumalanga and Gauteng provinces presenting with acute febrile and/or neurological symptoms.¹⁰⁷ These findings indicate that WNV is overlooked in neurological cases and should be included in diagnostic tests. Horses are highly susceptible domestic animals and are sentinels for WNV. Clinical signs in horses can range from asymptomatic to severe neurological symptoms due to virus-induced encephalitis.^23,108 In birds, WNV infections are typically asymptomatic, but in the United States of America (USA), species such as blue jays (Cyanocitta cristata) and crows (Corvus brachyrhynchos) are highly susceptible, often experiencing fatal neurologic disease.¹⁰⁹

West Nile virus is a zoonotic virus maintained in a sylvatic cycle involving various avian and mosquito species. Migratory birds such as corvids, passerines and raptors are primary reservoirs and amplifying hosts, responsible for long-distance dispersal of WNV. Cx. univittatus is the primary vector^109,110 and can transmit WNV to humans, horses and other species through bites. These hosts are considered ‘dead-end’ as the levels of viraemia are below the threshold level required to transmit to arthropods. Migratory birds such as barn swallows, European rollers and bee-eaters play a crucial role in WNV’s geographic distribution. Remarkably, ticks have been found naturally infected with WNV; however, their competency as vectors remains poorly understood.¹¹¹ Human-to-human transmission of WNV has been recorded through organ transplants,¹¹² blood transfusions^113,114 and breast milk,¹¹⁵ with one report of mother-to-child transmission.¹¹⁶

Wesselbron virus was initially isolated from infected sheep in 1955 in the small rural town of Wesselsbron, in the northern district of the Free State, South Africa. This initial outbreak was marked by significant mortality in newborn lambs and abortions in pregnant ewes.¹¹⁷ Wesselsbron virus causes abortion and the death of newborn offspring in cattle, sheep and goats. In 1956, WSLV antibodies were detected in sheep during an Rift Valley fever (RVF) outbreak in Kroonstad, northeast of Wesselsbron. Since then, WSLV has been identified in various vertebrates, including cows, sheep and humans.

Wesselsbron virus is primarily transmitted by infected Aedes mosquitoes, with Culex, Anopheles and Mansonia also implicated. Reservoirs include domesticated ruminants, wild birds and rodents.¹¹⁸ The levels of viraemia in sheep and cattle can reach threshold levels that are sufficient to infect mosquitoes.¹¹⁸ Humans can contract WSLV through mosquito bites or handling infected tissues. There are no reports of human-to-human transmission.

Initially, seroprevalence studies indicated that WSLV circulated predominantly in the more tropical parts of South Africa, such as KwaZulu-Natal. Cattle, sheep, goats and donkeys were infected due to the ecological niche of mosquito vectors such as Ae. circumluteolus, Ae. mcintoshi and Ochlerotatus juppi.^71,119 In contrast, more temperate plateau regions such as the Free State, Karoo and Gauteng province had a lower seroprevalence. Similarly, serological studies on humans from KwaZulu-Natal, the southern Cape, revealed WSLV prevalence rates of 32% and 0.7%, respectively. The high variance in the WSLV prevalence rates in two places is attributed to factors such as vector distribution, human activities and environmental conditions. A seroprevalence study, using an indirect IgG ELISA, from the Free State recorded a high seroprevalence rate of 27.6% (54/196) in humans, 19.3% (396/2052) in cattle and 1.1% (1/88) in sheep. Although neutralisation assays were not carried out for accurate confirmation of results, the ELISA used a recombinant ED III antigen, which can differentiate between flaviviruses.¹⁰¹ The high serodetection rate may have been due to including samples collected after the 2010–2011 RVF outbreak, when there was excessive rainfall that likely favoured increased mosquito populations and potential for other arboviruses, such as WSLV, to occur. Differences in sample collection times and geographic location are possible explanations for the high variance observed in seroprevalences between species. During the 2010–2011 RVF outbreak, two human cases of Wesselsbron disease were reported in South Africa.¹²⁰ Like WNV, WSLV is implicated in neurological equine disease.¹⁰⁸

The actual burden of WSLV is unknown due to serological cross reactivity with other flaviviruses and similar clinical presentation in livestock to RVFV, therefore misdiagnosed if not specifically tested. WSLV and RVFV often occur concurrently and are transmitted by similar vectors.¹¹⁷ A wide range of domestic animals, such as cattle, sheep, camels, pigs, goats, donkeys and horses, are susceptible to WSLV infection, with WSL disease being particularly significant in sheep in South Africa.¹²⁰ In humans, WSLV infection typically manifests as a sudden onset of influenza-like symptoms, including headache, fever, myalgia, rigours and arthralgia.¹¹

Usutu virus (USUV), SPOV, BAGV and BANV are lesser-known flaviviruses previously detected in South Africa, but their current status is unknown.

Flaviviruses are arboviruses with three overlapping transmission cycles: sylvatic, rural and urban, with different vectors and hosts involved in each cycle and location.¹²⁹Arboviruses are transmitted to humans, birds, livestock and NHPs, which act as reservoir or amplifying host through bites of hematophagous vector arthropods such as ticks and mosquitoes. The sylvatic transmission cycle is maintained between the arthropod vector (mosquitoes or ticks) and a reservoir or amplifying host such as NHPs, rodents, birds and bats. In a rural transmission cycle also known as the emergence zone, human activities such as hunting, farming and herding encroach on the sylvatic transmission cycle with humans infected through arthropod bites. The urban transmission cycle involves ‘domesticated’ mosquitoes, the mosquitoes that breed on stagnant water around human settlements as vectors and humans as hosts.¹²⁹

Many arboviruses, including flaviviruses such as YFV, ZIKV and WNV, have escaped from their sylvatic cycle and spread globally as a result of anthropogenic factors such as climate change, urbanisation, population growth, global travel and most significantly, viral mutations allowing adaptation to new vectors.^130,131 Increased ambient temperatures and flooding as a result of climate change create favourable environments for flavivirus vectors, thereby heightening the risk of vector–human interactions and subsequent infection surges.^132,133 Additionally, migration and/or displacement of people and livestock from one place to another (change in land use) due to climate disasters such as flooding and drought can trigger outbreaks of infectious diseases.

The rising human population, particularly in densely populated cities like Johannesburg, exacerbates the emergence of flaviviruses. Overburdened sewage systems in these urban areas frequently burst, resulting in standing water in the streets, ideal breeding grounds for mosquitoes. In addition, authorities are expanding urban areas into previously protected wild habitats, increasing opportunities for arthropods to interact with wildlife and humans, further elevating the risk of transmission.

Moreover, the international movement of people, livestock and cargo facilitates the dispersal of vectors and pathogens to new regions, leading to the emergence of infections in previously unaffected areas. Changes in viral genetics are particularly concerning, as they could result in the development of new variant strains with heightened virulence and viraemia levels in vertebrates. For example, WNV-L2 was considered less virulent than WNV-L1 before the occurrence of the six amino acid substitutions at E(V159I), NS1 (L338T), NS2A (A126S), NS3 (N421S), NS4B (L20P) and NS5 (Y254F) proteins. Amino acid substitutions resulted in increased virulence, causing serious disease in South Africa and other countries among humans, horses and birds.¹³⁴ Genetic mutations also enable viruses to adapt and live in new environments and infect different hosts and vectors, enhancing vector competence and transmission rates. Lastly, the development of pesticide resistance among vectors poses a significant challenge, complicating vector management and contributing to the emergence of flavivirus outbreaks. For instance, insecticide resistance has been recorded in some flavivirus vectors such as Cx. quinquefasciatus, Cx. pipiens, Ae. albopictus and Ae. aegypt.^{135,136,137,138,139,140,141,142,143}

In South Africa, arthropod survey studies collect mosquitoes and ticks in January to April and November to December (highly dependent on rains for the year), when it is wet and warm, tick and mosquito season. Arboviral infections such as WNV are recorded by the NICD at this time. During the mosquito and tick season, an arboviral infection should be suspected when a patient presents with flu-like symptoms such as fever, headache, muscle pains and even seizures and neurological signs. The gold standard for accurately diagnosing flavivirus infections is viral isolation in culture. However, this method requires skilled personnel to maintain the virus viability without contamination, is time consuming and necessitates specialised biosafety and containment facilities that are often unavailable in resource-limited settings. Nucleic acid amplification tests (NAAT) have emerged as the most rapid and accurate diagnostic tools, offering the advantage of detecting both viable and non-viable infectious agents.¹⁴⁴ Detection of viral nucleic acids in vectors and tissues from infected humans or animals such as the spleen, brain, serum, liver and cerebrospinal fluid (CSF) is carried out by reverse transcription-polymerase chain reaction (RT-PCR) followed by sequence analysis.^{145,146,147,148}

A significant challenge with nucleic acid detection and viral isolation for diagnostic purposes is that viraemia may be very low, as severe symptoms often develop only after viraemia has declined to undetectable levels.¹⁴⁹ Consequently, serological testing of serum and/or CSF for antibodies remains crucial in surveillance studies.¹⁵⁰ However, results must be interpreted cautiously due to potential cross-reactions stemming from antigenic similarities among different flaviviruses. Consequently, a positive viral-specific IgG and IgM test should be confirmed with a neutralising antibody test to prevent misdiagnosis due to serological cross-reactivity. In South Africa, certain private and public laboratories offer both serological and PCR tests for flaviviruses such as ZIKV, DENV, YFV and WNV, at the discretion of the examining doctor. There are various commercially available ELISA kits (Table 2) to detect flaviviruses; however, neutralisation tests are required to accurately differentiate between flaviviruses. These kits are quite expensive and laboratories from low-income countries frequently cannot afford them, and many rely on in-house assays. In South Africa, the NICD offers a diagnostic service based on NAAT and serology, for selected arboviral infections known in the country or with potential to occur in travellers from endemic regions.

Additionally, post-mortem histopathological examinations and immunohistochemistry using formalin-fixed tissue samples of animal tissues such as brain, spleen and liver from suspected cases can be used for detection, with confirmation by nucleic acid amplification followed by nucleotide sequencing.

Preventing flavivirus infections in humans and animals primarily relies on effective mosquito vector control programmes. In addition, mosquito bite prevention measures, including repellents, sleeping under nets, putting animals in enclosed structures during the night and avoiding low-lying wetland areas, can significantly reduce their interaction with vectors. Veterinary and laboratory personnel handling infected or potentially infected tissues and materials must wear personal protective equipment, such as gloves, masks and face shields, to prevent contact with the virus and should avoid procedures that aerosolise the virus. In some countries, blood donors returning from WNV-afflicted areas are excluded to avoid transmission of WNV during blood transfusion. Although rare, human-to-human transmission has been recorded for WNV in the USA.¹⁵¹

Continuous use of repellents and insecticides has raised challenges such as resistance and environmental harm. Therefore, biological control measures such as gene editing to obtain sterile male mosquitoes are being evaluated in Tanzania, while Wolbachia-based approaches have been adopted by countries such as Indonesia, Vietnam, Malaysia, Thailand and Taiwan.¹⁵² Wolbachia pipientis, an obligate intracellular mosquito parasite, uses two mechanisms to reduce the transmission of arboviruses: it induces cytoplasmic incompatibility and produces an infertile progeny from an uninfected female and a Wolbachia-infected male, thus reducing the vector population. Wolbachia infection also blocks the replication of pathogens such as viruses in the mosquito, reducing the efficiency of vectors.

Flavivirus vaccine development is a critical concern, especially for ZIKV and DENV, due to the complexity of the immune response and the severity of the disease outcomes. At present, Q-denga (TAK-003), a live attenuated vaccine containing DENV 1–4 serotypes for children between 9 and 16 years, is available for DENV. There are also highly effective and safe vaccines against some flaviviruses, such as YFV and JEV, for humans. The live attenuated 17D-204 YF vaccine developed by weakening the Asibi yellow fever strain through serial passaging of cell cultures has been used since 1938. The vaccine provides lifelong protection for 80%–90% people within 10 days of vaccination. Even though there are antigenic similarities in flaviviruses, the 17D-204 YF vaccine does not offer reliable cross-protection against ZIKV, DENV and WNV. However, some flavivirus vaccines provide some cross-protection to related viruses and/or within the same serocomplex in other models. For instance, WNV vaccination offers protection against USUV disease in mice¹⁵³ and also, JEV and Saint Louis encephalitis virus (SLEV) vaccination provided protection to lethal WNV challenge in a hamster model.¹⁵⁴

Currently, there are no licensed vaccines for humans against WNV, WSLV, USUV, BAGV, SPOV or BANV. However, a number of WNV vaccines have gone through human clinical trials, such as VRC-WNVDNA020-00-VP, HydroVax-00, Chimerivax-WNV02 and rWN/DEN4Δ30 and a DNA vaccine encoding WNV prM/E. Among these, ChimeriVax-WNV02 is the most promising candidate assessed in phase II clinical trials.¹⁵⁵ Currently, there is an FLAVIVACCINE project that aims to develop a broad-spectrum, mosquito saliva-targeted vaccine that protects against multiple flaviviruses.

For veterinary use, a live, attenuated WSLV vaccine exists for non-pregnant animals, providing lifelong immunity.¹¹⁷ WNV vaccines, including West Nile-Innovator DNA vaccine (Fort Dodge Animal Health, commercialised by Pfizer) and Recombiteck Equine West Nile Virus Vaccine, produced by Merial-Sanofi Aventis,^156,157 available for use in South Africa.

Currently, there are no specific antivirals against flaviviruses. In general, antivirals are designed to interfere with critical stages of the virus life cycle, that is, replication, protein synthesis, assembly, entry and virus egress. Research in this field is ongoing, with a lack of funding to assess antivirals through all the clinical stages a major obstacle. A number of compounds have been evaluated for the antiviral efficacy only in in vitro, in vivo and first-stage clinical trials. Notably, ZIKV-Ig (ZIKV), TY014 (YFV) and Tyzivumab (ZIKV) were assessed in first-stage clinical trials.¹⁵⁸ Some Food and Drug Administration (FDA)-approved drugs, such as Dasabuvir, Efavirenz and Tipranavir inhibit replication of multiple flaviviruses in vitro.¹⁵⁹ Dasabuvir, an antiviral drug against the hepatitis C virus (HCV), demonstrated strong and antiviral effects against ZIKV, WNV and TBEV in Vero cells. Tipranavir, an antiviral for treating human immunodeficiency virus (HIV) infection, had antiviral effects against ZIKV and TBEV in Vero cells. Another antiviral agent for treating HIV, efavirenz, demonstrated antiviral effects against TBEV, WNV, ZIKV, DENV and YFV in Vero cells.¹⁵⁹ Interestingly, targeted therapies, including monoclonal antibodies directed against the flavivirus surface E glycoprotein and non-structural protein 1 (NS1), have been shown to be effective against WNV in mice.¹⁶⁰

To effectively address the public health threats posed by flaviviruses in South Africa, several key recommendations should be implemented. Initially, it is essential to establish and maintain robust One Health surveillance programmes that monitor the prevalence and spread of flaviviruses in both human and animal populations.¹⁶¹ Such surveillance should include syndromic screening for lesser-known flaviviruses such as BANV, BAGV, WSLV and USUV. In addition, investing in the continuous development and validation of molecular diagnostic tools tailored to detect local strains of flaviviruses is crucial. This investment will help reduce misdiagnosis and improve the accuracy of surveillance efforts. Public awareness campaigns should also be implemented to educate the public and healthcare professionals about flavivirus transmission, symptoms and preventive measures. Emphasising the importance of avoiding arthropod bites and using protective clothing can significantly reduce infection rates.

Furthermore, prioritising research and development of effective vaccines for humans against flaviviruses is vital, particularly for those that currently lack licensed vaccines. Investigating the potential for cross-protection among flavivirus vaccines, as observed with USUV and WNV, should also be a focus. A One Health approach with collaboration between public health and veterinary services is essential to ensure a comprehensive approach to flavivirus management. This includes implementing vaccination programmes for animals and monitoring zoonotic transmission. Additionally, clinical guidelines to diagnose and manage flavivirus infections should be regularly reviewed and updated to incorporate the latest research findings and technological advancements.