Rev. sci. tech. Off. int. Epiz., 2004, 23 (2), 613-623

Rift Valley feverG.H. Gerdes

Private Bag X05, Onderstepoort, 0110, South Africa

SummaryRift Valley fever (RVF) is an arthropod-borne viral disease of ruminants, camelsand humans. It is also a significant zoonosis which may be encountered as anuncomplicated influenza-like illness, but may also present as a haemorrhagicdisease with liver involvement; there may also be ocular or neurological lesions.In animals, RVF may be inapparent in non-pregnant adults, but outbreaks arecharacterised by the onset of abortions and high neonatal mortality. Jaundicehepatitis and death are seen in older animals. Outbreaks of RVF are associatedwith persistent heavy rainfall with sustained flooding and the appearance oflarge numbers of mosquitoes, the main vector. Localised heavy rainfall is seldomsufficient to create conditions for an outbreak; the simultaneous emergence oflarge numbers of first generation transovarially infected mosquitoes is alsorequired. After virus amplification in vertebrates, mosquitoes act as secondaryvectors to sustain the epidemic.

KeywordsAbortion – Flooding – Liver necrosis – Mosquito – Rift Valley fever – Transovarialtransmission – Zoonosis.

Introduction and historyA zoonosis may be defined as a disease that is naturallytransmitted between humans and animals. The mostserious zoonoses are viral, and Rift Valley fever (RVF) fallsinto this category. This arthropod-borne infection affects awide range of vertebrates, but clinical disease is limited todomestic ruminants and humans. The course of thedisease varies with host species and may be inapparent toperacute.

Rift Valley fever virus was first discovered andcharacterised in the Rift Valley of Kenya in 1931 (12), butoutbreaks probably occurred before this time.

Many sub-Saharan tropical and sub-tropical countries inAfrica have reported outbreaks of RVF and the disease isencountered in an enzootic or epizootic form along theeast and south coast of Africa and also in Madagascar. Thevirus has spread as far north as Egypt and more recently anoutbreak occurred in the Arabian Peninsula (Fig. 1). Largeoutbreaks have occurred in South Africa, which wereparticularly severe between 1950 and 1951 and even more

widespread between 1974 and 1975. Because clinicaldisease is generally not evident in indigenous breeds, RVFhas not been identified as a livestock problem in manyother countries in Africa, although large losses inindigenous livestock have occurred in Egypt and WestAfrica. There have been no Rift Valley fever epizootics insmaller countries such as Mali, Gabon, Congo, Chad,Botswana, Angola, Nigeria and Uganda, but evidence ofantibodies to the disease has been found in humans and/orlivestock. Large outbreaks in cattle, manifested byabortions, have been reported in Zimbabwe and Zambia.Monitoring of sentinel cattle in Zambia indicates an annualemergence of the virus after the seasonal rains, withantibody prevalence varying from 3%-8% to 20% (14).

The ability of RVF to move outside traditionally endemicareas, or even out of the continent of Africa, lies in the factthat there is an unusually large range of vectors capable oftransmitting the virus and a level of viraemia in ruminantsand humans that is sufficiently high to infect mosquitoes.Such a movement of the virus occurred following the

El Niño floods of 1997-1998 in the Horn of Africacountries, and there was an RVF epizootic in the ArabianPeninsula in 2000. Although it has not been proven thatthe virus was carried from infected to non-infected areas byviraemic animals, it is a possibility in this case, given theproximity of the Saudi and Somali port cities and theaccelerated trade which takes place just before the Hajjfestival. Viraemic animals may also have had a role to playin the spread of RVF from Northern Sudan into Egypt in1977. This type of transmission deserves furtherconsideration in a world where globalisation of trade andshort transport times are now the norm.

The importance of Rift Valleyfever for animal and publichealthThe words attributed to Louis Pasteur ‘the microbe isnothing; the terrain is everything’ are particularlyapplicable to RVF. Flooding in northeast Africa in the El Niño year of 1997-1998 gave rise to the most importantepidemic of RVF ever recorded. The greater than normalrainfall caused the combined flooding of two rivers inSomalia, and further flooding in Kenya, creating a hugeinland lake and setting up the conditions for an outbreakof RVF that affected 89,000 people in Kenya and Somaliaand caused 250 deaths.

Other pathogens, including haemorrhagic fever viruses,circulated in the hot, moist conditions, which weredescribed by the World Health Organization (WHO) as ‘aviral soup’. Bluetongue killed many sheep in Kenya, andthe final number of livestock deaths was greatly increasedbecause of the countless animals that drowned (32).

In addition to flooding, changing land use is another majorfactor in the emergence of disease, and in the case of RVFthis is reflected in the building of dams. Up to 1977, RVF had been seen largely as a veterinary problem inAfrica, but in Egypt in that year, six years after thecompletion of the Aswan dam, more than 200,000 peoplebecame sick, 598 deaths occurred and livestock lossesresulted in meat shortages (18). The mode of spread of thevirus, from an isolated focus in Northern Sudan to the Nilevalley, is unclear: it could have been spread by infectedstock or people, or transported by mosquitoes (1). Thedam had created many hectares of floodlands andincreased the breeding ground of mosquitoes, addingimpetus to the outbreak. Rift Valley fever was not enzooticin Egypt before 1977 and the virus disappeared in 1981,before re-emerging in 1993.

A second dam-linked incident occurred in 1987 in WestAfrica. One year after the construction of the Diama damalong the Senegal River, an outbreak of RVF was recordedin northern Senegal and Mauritania. Many thousands ofpeople became sick and about one fifth of those died (22).Abortion losses in livestock were also heavy. This incident,which was the result of the mass emergence of mosquitoesin northern Senegal, was interesting, given the fact that thefirst RVF virus isolate from mosquitoes in West Africa wasdiscovered as early as 1974 in southern Senegal and thatthe virus appears to be enzootic in the livestock there (45).

The East African outbreaks, particularly the earlier ones,were considered to be ‘veterinary/animal epidemics’. Forexample, in the 1930-1931 hepatitis outbreak in the RiftValley of Kenya there were heavy losses in sheep due todeath and/or abortion. In one documented example, 60 out of 80 lambs aged three to seven days died in thespace of 24 h and there were also reports of 3,500 lambsand 1,200 ewes being lost in a single month (12). In SouthAfrica between 1950 and 1951, 100,000 sheep died and500,000 aborted. Similar heavy losses occurred in cattle inZambia and Zimbabwe.

However, the Egyptian outbreak of 1977 and the WestAfrican outbreak of 1987 in Senegal and Mauritania were‘people epidemics’. The outbreak in the Nile valley anddelta caused nearly 600 human deaths. Among the200,000 people who became sick, retinitis was present ina large proportion of infected individuals, as well as livernecrosis, haemorrhagic fever, and encephalitis (24). Ocularinvolvement was characterised by retinal macular lesionsand was first noticed as an abrupt loss of visual clarity. The

Rev. sci. tech. Off. int. Epiz., 23 (2)614

Countries where outbreaks have occurredCountries where antibodies have been found

Fig. 1A map of Africa and the Arabian Peninsula illustrating endemicand epidemic areas of Rift Valley fever

habitats in an endemic cycle between mosquitoes and anunknown vertebrate host. Alternatively there may be virusactivity each year, with low-level transmission to livestockassociated with Aedine mosquitoes that breed in low-lyingdepressions which flood when abnormally heavy rainfallraises the water table sufficiently. (In sub-Saharan Africa,RVF is endemic because of transovarial transmission inAedes Neomelaniconion.)

Once there is evidence of past virus activity, countries arelikely to remain permanently infected. Some 20 countriesin Africa and Madagascar are infected and 23 species of mosquitoes are involved in the epizootic/enzootictransmission cycles of RVF. Countries in north and west Africa are arid and outbreaks occur independently ofrainfall, with mosquitoes that breed in large rivers and dams, not floodwater aedines. In Egypt, there are nomosquitoes capable of transovarial transmission and themajor vector is a Culex sp. which breeds in polluted waters (25).

In the rest of Africa, inter-epizootic survival depends ontransovarial transmission of the virus or venerealtransmission between mosquitoes with low-levelcirculation in livestock. In Egypt, new outbreaks are theresult of infected adult mosquitoes coming out ofhibernation, re-introduction of the virus via the transportof infected animals, or wind-borne transmission frominfected neighbouring countries. The latter is thought tohave occurred in 1977 in Egypt, when unusual southerlywinds were documented for a week at the end of

Rev. sci. tech. Off. int. Epiz., 23 (2) 615

heavy human involvement was due to direct contact withvectors through peridomestic mosquito bites. The manycases of liver involvement, which was seen for the first timein the 1975 outbreak in South Africa, were thought to bedue to the high incidence of schistosomosis amongEgyptians (31).

The human toll in Mauritania was also high, with over 200 deaths.

Aetiology and classificationThe Bunyaviridae is a large family of viruses and containsfive genera, four of which infect vertebrates, while theremaining genus, Tospovirus, contains a group of plantviruses. Three of the vertebrate-infecting genera,Bunyavirus, Phlebovirus and Nairovirus are associated witharthropods (and are further subdivided into complexes ofclosely related members), while the last genus, Hantavirus,has no known invertebrate association.

Rift Valley fever virus, in physical, chemical andmorphological terms, is a typical member of theBunyaviridae of the genus Phlebovirus. This genus alsoincludes the sandfly fevers. Rift Valley fever virus is anenveloped spherical virus of up to 120 nm in diameter,with short glycoprotein spikes projecting through abilayered lipid envelope (Fig. 2). The single strandedribonucleic acid (RNA) genome is divided into threesegments each in its own nucleocapsid. These segments arethe large (L), medium (M) and small (S) segmentsexpressing many copies of the N (nucleocapsid) and fewcopies of the L (transcriptase) structural proteins (46).

The S segment has bi-directional coding or ambisense RNAwhile the other two segments have negative sense RNA genomes (37).

Epidemiology and transmissionThe distribution of large RVF outbreaks is neither seasonalnor annual, but linked to the presence of water. Water isimportant to most blood-feeding insects, as they haveaquatic or semi-aquatic immature stages and thisautomatically limits the choice of breeding sites. In thesouth and east of Africa, outbreaks of RVF are associatedwith heavy rainfall, while in the drier north and west theyare linked to irrigated lands. Epizootics of RVF are cyclicalin nature and characterised by long inter-epizootic periods.These cycles may be comparatively short, i.e. five to fifteen years in wetter areas, or much longer, i.e. fromfifteen to thirty years or more where it is drier. In the inter-epizootic periods the virus may be present in forest-edge

Fig. 2Rift Valley fever virions in semi-purified enzyme-linkedimmunosorbent assay antigen

Rev. sci. tech. Off. int. Epiz., 23 (2)616

July/beginning of August, probably bringing large numbers of mosquitoes up from the north of Sudan. Thewaters of the Aswan dam had already significantlyincreased mosquito-breeding sites in Egypt.

Rainfall, and the presence or absence thereof, is animportant factor in the establishment of breeding sites, andit is above average and sustained rainfall which createsconditions for an outbreak of RVF. Persistent rainfall raisesthe water table and floods grassland and the shallowdepressions which characteristically dot the plateauregions of some African states. Eggs of the ground poolbreeding Aedes spp. are attached to vegetation at the edgeof these depressions, and when they flood large numbersof floodwater breeding mosquitoes emerge (13). Outbreaksmay be less frequent and less widespread in highlypopulated areas where there is intensive land use, as thisdecreases the presence of water and thus reduces thenumber of mosquito breeding sites.

Epidemics do not occur as a result of lateral spread from asingle source, but because of the intensification of vectoractivity due to the simultaneous emergence of a greatnumber of vectors. This would explain why localisedheavy rainfall does not produce an epidemic. Whenoutbreaks do occur there are often large distances betweenfoci of infection, and one theory being tested to account forthis is that mosquito eggs may be dispersed by wadingbirds (National Institute for Communicable Diseases,Johannesburg, South Africa, personal communication).Secondary, opportunistic mosquitoes then serve topropagate the epizootic.

There are many arthropods involved in the transmission ofRVF – whether by biological or mechanical means. Insouthern Africa alone, the virus has been isolated from wildmosquitoes of the Aedes, Culex, Anopheles, Mansonia andEretmapodites species. The floodwater-breeding aedines areof particular importance in the sub-genera Aedimorphusand Neomelaniconion. In Africa alone, 23 species ofmosquito from five genera are involved and, whilemosquitoes are the most important arthropod vectors,virus may be mechanically transmitted by Culicoides,Simulium, and even tick species. There is a complex cycleof enzootic maintenance and epizootic transmission of thevirus by different mosquito species. Transovarialtransmission within the Aedes species allows a singlegeneration of infected mosquitoes to emerge after flooding.These infect some livestock, and if the pools created by thefloodwater do not disappear then other mosquito speciesbreed rapidly and act as secondary vectors.

There are two different types of secondary cycles by whichthe virus is transmitted to humans: the sylvatic cycle andthe urban peridomestic cycle. In the sylvatic cycle, whichaffects people working in the livestock industry, humansbecome infected through contact with animals to whom

the virus has been transmitted by zoophilic mosquitoes. InSouth Africa between 1974 and 1976, 14.5% of peopleworking in this industry were seropositive, with seven reported deaths. Urban consumers of animalproducts are not affected by the sylvatic cycle as the virusdoes not survive below pH 6, and the pH of meat generallyfalls to 6 or less during processing (41). The second type ofsecondary cycle is the urban peridomestic cycle, in whichhumans become infected with RVF through the bites ofanthropophilic mosquitoes. The large number of humancases in Egypt in 1977 and 1978 occurred largely as aresult of this type of transmission; the antibody prevalencein people reached 29.6%, and there were 598 recordeddeaths. Human transmission takes place when the viruscomes into contact with abraded skin or mucousmembranes (aerosol with intranasal infection is also apossibility and is thought to have occurred in Saudi Arabiaduring the ritual slaughter of animals). There is alsocircumstantial evidence that infection can occur throughcontact with raw milk.

There appears to be little contact transmission betweenanimals in spite of the presence of virus in nasal dischargeand saliva. Laboratory infection has been reported andpeople at risk are now vaccinated and required to takeprotective measures.

Rift Valley fever transmission and the effects ofglobal warming At this point, it is also worth considering global warmingand its effect on extending the range of RVF vectors. Globalwarming is set to modify temperature, rainfall, wind andsea-levels and will therefore alter vector, vertebrate andvirus interactions. One American study on the effects ofglobal warming on mosquito-borne arboviruses reportedthat two American mosquito species had movednorthwards to slightly cooler areas to escape escalatingtemperatures (33). Various other studies have beenundertaken, including a WHO Task Group investigation,the results of which were released in a 1990 report on the‘potential health effects of climate change’ (43). Thesensitivity of diseases and their vectors in relation to globaltemperature changes was rated on a scale of zero to three.Among the dozen human illnesses evaluated none scoredzero or ‘no effect’. Malaria scored three, dengue two, and allother insect-borne viral diseases scored one.

The biology of infectionDiagnosing RVF can be a challenge because while someoutbreaks are obvious there are also cryptic cycles whichare not seen. Severe clinical disease in young animalsindicates that a breed is highly susceptible to the disease.

Table I lists the host range of vertebrates according to theirdegree of susceptibility to the virus, ranging from highlysusceptible to resistant (31).

Epidemics are characterised by sudden abortion storms,almost 100% mortality in newborns of less than one weekold and an influenza-like illness in humans (2).Complications may include hepatitis with generalisedhaemorrhages and cerebral or ocular infections in someindividuals (15). Camels, which do not exhibit diseasewill, however, abort, as do buffalo. Newborn kids andlambs are most susceptible and mortality may vary from70% to 100%. The disease follows a peracute course, withcollapse and death occurring 12 h after the onset of pyrexia(40°C to 42°C). After the incubation period, which can lastfrom between 12 h and 36 h, animals become listless andexhibit abdominal pain. Adult sheep and goats or olderlambs and kids are less susceptible and disease may varyfrom acute to inapparent, depending on the breed. There isan incubation period of 24 h to 72 h followed by anelevated temperature, lymphadenitis, and vomiting ordiarrhoea. The animal is listless, becomes recumbent andmay have haemorrhagic diarrhoea. Icterus can develop andabortion may occur. Mortality varies from 10% to 70%,although in adults with a more sub-acute form of thedisease it may only reach 20%. Calves less than ten daysold will also suffer a peracute form of RVF and die within24 h. Icterus is more commonly seen in older calves andadult cattle and death in adults is often less than 10% (21).The animal develops anorexia, diarrhoea and dysgalactiaand exhibits salivation and nasal discharge (17). Abortionmay occur as a result of infection of the foetus or as a resultof a febrile reaction. The foetus is often autolysed.

In indigenous breeds of African cattle, infection with RVFis often retrospectively diagnosed serologically and closerexamination may then reveal abortions and a drop in milkproduction.

Wild ruminants seroconvert after inapparent infection but,as mentioned above, buffalo abort.

Human infectionDisease in man is often an indicator of a larger problem,particularly in dry areas where livestock are not intensivelymanaged and abortions are not evident. Human exposureto the virus is often occupational, either through handlinginfected livestock or their products or by breathing inaerosols released at slaughter. Mosquito bites, as seen inEgypt, and the consumption of raw milk have beendocumented as routes of exposure. In a high densitypopulation area like the Nile delta, man may even beconsidered to be an amplifying host, because humans maydevelop viraemia, thus enabling biting mosquitoes totransmit the virus to additional hosts.

Human illness may take one of four forms, as follows:

a) an uncomplicated, febrile, influenza-like illness

b) a haemorrhagic fever with liver involvement,thrombocytopenia, icterus and bleeding tendencies

c) encephalitis following a febrile episode with confusionand coma. Death is infrequent but there may be someresidual damage

d) ocular involvement with reported blurred vision andloss of visual acuity due to retinal haemorrhage andmacular oedema. There may also be residual damage (31).

Exposure to the virus is followed by a short incubationperiod of three to seven days, chills, headache and abiphasic fever. The patient experiences muscle and jointpains, nausea and tenderness in the liver area. In mostcases, recovery is uneventful and complications are theexception rather than the norm.

Pathogenesis and lesionsIf an examination of blood is undertaken, the followingthree findings indicate the presence of RVF:

– profound leucopaenia

– elevated blood enzymes associated with severe liverdamage

– thrombocytopenia.

The main sites of viral replication are the liver and spleenand often the brain (3, 30). The brain is a good specimento submit in the case of an autolysed foetus. In extremelysusceptible animals, peracute liver damage is seen (16).Susceptibility decreases with age and adult sheep and goatsare often only moderately susceptible, as are cattle andhuman beings. This varies according to breed andindividual and a hepatic form of the disease may occurwith vasculitis and haemorrhagic manifestations. The RVFvirus replicates in many cells, and lesions in target organsare lytic in nature. Virus antigen has been demonstrated in

Rev. sci. tech. Off. int. Epiz., 23 (2) 617

Table IThe susceptibility of vertebrate hosts to Rift Valley fever virus

SusceptibleSusceptible (but infection

is inapparent)

Newborn lambs Calves Cattle Camels Birds

Newborn kids Sheep Goats Equines Reptiles

Puppies Buffalo Pigs Amphibians

Kittens Humans Dogs

Mice Cats

Hamsters Guinea pigs

Rabbits

Highlysusceptible

Moderatelysusceptible

Resistant

Rev. sci. tech. Off. int. Epiz., 23 (2)618

the walls of small vessels, such as the adrenocortical cellsand the glomeruli of the kidney, as well as in most areas ofthe spleen and all cells of the liver (42).

In an aborted foetus, virus may be found in visceral organsand the brain. Virus may be recovered from a placentawhich does not show any visible lesions. Hepatic necrosisis the most common lesion observed in foetuses at autopsyand severe lesions are seen in newborn lambs. The liver isenlarged and congested and bronze to yellow in colour. Itis soft and friable with a patchy congestion, haemorrhagesand scattered gray-white necrotic foci (Figs 3 and 4).Because of the peracute nature of the disease the necroticfoci become indistinct in the severely congesteddiscoloured liver (9, 11). Icterus is not seen in lambs,although in adult sheep with less severe lesions, this maybe a feature.

Calves and cattle display more localised liver lesions, necroticfoci are visible and lobulation is more distinct (10). There is ageneralised lymphadenopathy and often a severehaemorrhagic gastro-enteritis. The gut contents may be

chocolate brown in colour, particularly in newborn calves.There is a blood-stained accumulation of fluid in the bodycavities and widespread ecchymotic and petectialhaemorrhages on subcutaneous and serosal surfaces. The wallof the gallbladder is oedematous with haemorrhages (Fig. 5).

A severe lytic liver necrosis with dense aggregates ofnuclear and cellular debris, fibrin and inflammatory cells,is found in neonates and aborted foetuses of all species(Fig. 6). In adult animals necrosis is mostly multifocal andless diffuse.

DiagnosisSingle cases of RVF may be confused with many viraldiseases of sheep where there is sudden death, generalisedlymphadenopathy and haemorrhages throughout thecarcass. However, RVF should be considered when there isa sudden onset of abortions at all stages of pregnancy,sudden death in young animals following an acute febrile

Fig. 3Newborn lamb: a congested, discoloured liver with smallnecrotic foci

Fig. 4Lamb: the liver is mottled in appearance due to necrosis andhaemorrhage

Fig. 5Adult sheep: oedema and haemorrhages in the gallbladder wall

Fig. 6Lamb: an area of necrosis in the liver with nuclear and cellulardebris

disease and obvious liver involvement in all cases. Such anoutbreak may also be accompanied by severe flooding, thepresence of large aggregates of mosquitoes and aninfluenza-like illness in human beings exposed to infectedmaterial. The presence of haemorrhagic fever in someindividuals can create a sense of panic, sometimes leadingto a public health crisis.

Liver lesions are pathognomic for RVF, however, laboratoryconfirmation is advisable. In the case of live animals,specimens should include heparinised blood and serum,while the tissues of choice from dead animals are the liver,spleen, kidney, lymph nodes and heart blood. The brain ofan autolysed foetus should also be included. Samplesshould be securely packed and labelled to indicate thedangerous nature of the contents. Specimens should besent on ice at 4°C or, if there is likely to be a delay intransit, tissues may be sent in formalin for histopathologyand also placed in glycerol or saline. Paired serum samplesshould be considered due to the confusing clinicalpresentation of RVF in adult animals.

Viral antigen may be rapidly detected using a reversetranscriptase polymerase chain reaction technique (34, 36). Virus replicates rapidly in infant or unweanedmice, on hamster or monkey cell lines such as Vero, BHKor CER, and on mosquito cells. Primary ruminant cells ofcalves and lambs may be useful as well.

Antibody tests include immunodiffusion, complementfixation, haemagglutination inhibition (HAI), indirectimmunofluorescence (IFA) and virus neutralisation (8, 40). Immunoglobulin G (IgG) antibodies are producedearly in an infection, but of the available enzyme-linkedimmunosorbent assay (ELISA) tests, immunoglobulin M(IgM) ELISAs are favoured for a rapid diagnosis (29).

Although African phleboviruses other than RVF areunlikely to cause disease or cause confusion in thediagnostic interpretation, it should be noted that there maybe cross-reactivity associated with HAI tests. Screeningtests may be done with HAI, IFA and ELISAs, butneutralisation tests are considered to be specific.

Laboratory infections are common and appropriatemeasures should be taken to protect staff members,including mandatory immunisation.

Surveillance and controlRift Valley fever is intimately connected to sustained, heavyrainfall with flooding. Cycles are irregular and inter-epizootic periods may be long. Continuous control effortsmay be impractical, but it is possible to monitor enzooticvirus activity in livestock.

Routine screening of abortions, where practical, should beencouraged. Sentinel herd monitoring during the rainyseason will yield useful data. Herds should be strategicallyplaced near water sources which are known to flood andshould consist of young seronegative animals. Theseanimals should be regularly sampled for IgG/IgMantibodies. Data collected may alert authorities to thepresence of virus emergence/circulation in sufficient timeto contemplate control measures.

An even earlier warning system, five months in advance,using satellite weather data, was used to predict the RVFoutbreak in Mauritania that erupted in 1987 after thecompletion of the Diama Dam on the Senegal River. Thistype of remote sensing satellite warning system collects thefollowing data:

a) southern oscillation index (SOI) – measures thedifference in atmospheric pressure between Tahiti (eastPacific) and Darwin (west Pacific). Heavy rainfall isassociated with a negative SOI

b) sea surface temperature (SST) – of the Indian and Pacificoceans. Conditions are neutral if the SST remains between0.5°C above or below the norm for a given area of ocean.In the El Niño year of 1997-1998 the Pacific SST was 5°C above the norm and severe flooding occurred in theHorn of Africa

c) normalised differentiated vegetation index (NDVI) –measures the greenness and browness of the vegetation. Thegreen leaf biomass or photosynthetic capacity of a region isan indicator of the amount of moisture in the soil (23).

The correlation between increased vector activity andheavy rainfall predictions from weather satellite data hasbeen investigated and used successfully in a number ofimportant human and animal disease studies (6). Theanalysis of data sets may be as simple as assessinginformation about El Niño to predict a wet season or ascomplex as comparing land-based weather stationminimum and maximum temperatures and monthlyrainfall with satellite derived NDVI, middle InfraRedreflectance (MIR), land surface temperature (LST) and airtemperature (TAIR).

This data is a useful tool for predicting weather conditionsthat favour an RVF outbreak, but even if a season isparticularly wet, an outbreak will not occur unless there isalso active virus circulation and a large susceptiblevertebrate population.

Rift Valley fever remains difficult to predict but devastatingin its effect on both humans and animals.

Control measures include the following:

a) movement control with respect to trade and export

b) vector control with larvicides in vector breeding sitesrather than aerial spraying of adults

Rev. sci. tech. Off. int. Epiz., 23 (2) 619

c) vaccination of livestock – this is important for tworeasons, (i) livestock amplify the virus and mosquitoespreferentially feed on livestock (ii) the young/newborn aremost at risk of dying, but colostral immunity confersprotection against virulent RVF virus challenge.

The vaccine currently in use is a live, attenuated productincorporating the neurotropic attenuated mouse brainSmithburn strain, which was derived from the Entebbestrain in 1944, developed in 1949, and has been used inSouth Africa since 1952. Attenuation was achieved bypassage in suckling mice and embryonated eggs (39).Serial intra-cerebral passage enhances neurotropism andreduces hepatotropism. The vaccine is not used inpregnant animals except in the face of an outbreak,because a percentage of animals may abort (44). Thisvaccine is still neurotropic and affects the central nervoussystem of the foetus and induces foetal teratology andhydramnios (hydrops amnii) in the dam. Use of thisattenuated live vaccine in non-endemic areas may beconsidered during an outbreak, because the importance ofpreventing the negative effects of the virus in humans andlivestock may well outweigh the risk that the virus may re-assort or regain virulence (35). Vaccines underdevelopment include two live strains not yet available forfield use. One is derived from a mutagenised human fieldisolate from Egypt, MP 12; it is an efficient immunogen,but is still abortogenic. A mutagen attenuated RVF vaccineof MP 12 has however been found immunogenic and non-abortogenic in pregnant ewes and immunogenicand non-pathogenic in neonatal lambs in the face ofvirulent challenge exposure. This vaccine produces low-level transient viraemia in cows. The other vaccine, clone13, is also derived from a field strain of a naturally mildhuman case. It is a stable vaccine candidate with not muchlikelihood of reversion, as this clone has a large deletion of549 nucleotides in the non-structural protein gene (7, 20,26, 27, 28).

A vaccine for use in pregnant animals and in non-endemicareas is produced from a pathogenic field isolate. This virusis cell culture adapted and formalin inactivated. There aredisadvantages, which include the use of pathogenic strains,the necessity of two vaccinations and the costs involved (4, 5, 19). A human vaccine is produced in the UnitedStates for use in people at risk.

ConclusionRift Valley fever is a zoonosis and a haemorrhagic feverwhich can generate a sense of panic in the public and therehas been speculation as to where it might appear if thevirus were to ‘escape’ from Africa.

The virus could exit the country via either wind-bornespread of infected mosquitoes, as was probably the case in1977 when the virus spread from Sudan to Egypt, orvehicular transport of viraemic persons or animals, ashappened in 2000 when the virus spread from the Horn ofAfrica to Saudi Arabia. In both these situations, the viruswas deposited in an already flooded area with an extremelylarge local mosquito population and naive hosts. This isthe key to the generation of an outbreak – the obligatorypresence of flood conditions and large numbers ofmosquitoes.

The particular risk factors associated with the movement ofinfected animals are the high, intense viraemia and the widerange of vectors capable of transmitting the virus, includingmechanical vectors such as ticks and various biting flies suchas Glossina, Simulium, Stomoxys and Culicoides spp. Theviraemic period is about seven days, but the virus maypersist in the spleen of sheep for much longer. Although theincubation period of RVF is short, travel to a country with adam, large irrigation project or large delta system is aconcern. Potential risk areas include locations such as wherethe Tigris and Euphrates rivers empty into the Persian Gulfand the area where the Indus River delta flows into theArabian Sea. Should the virus spread to an arid or semi-aridone, a cryptic cycle might be the result, where an oddabortion or human case could pass unnoticed.

Given the right conditions, this transboundary disease hasthe potential to spread over large distances.

Acknowledgement

The author would like to thank the Department ofPathology at the University of Pretoria for permission touse their slide material.

Rev. sci. tech. Off. int. Epiz., 23 (2)620

Rev. sci. tech. Off. int. Epiz., 23 (2) 621

G.H. Gerdes

RésuméLa fièvre de la Vallée du Rift est une maladie virale transmise par les arthropodesqui touche les ruminants, les camélidés et les humains. C’est aussi une zoonoseimportante qui peut se manifester comme une maladie pseudo-grippale sanscomplication, mais peut aussi prendre la forme d’une maladie hémorragiqueavec atteinte hépatique ; elle peut également provoquer des lésions oculaires ouneurologiques. Chez les animaux, la fièvre de la Vallée du Rift peut ne pas semanifester chez les femelles adultes non gestantes, mais les foyers secaractérisent par l’apparition d’avortements et d’une forte mortalité néonatale.Chez les animaux plus âgés, on constate des ictères, des hépatites et des décès.Les foyers de fièvre de la Vallée du Rift sont associés à de fortes pluiespersistantes accompagnées d’inondations durables et à l’apparition de grandesquantités de moustiques, qui sont le principal vecteur. Il est rare que des fortespluies localisées suffisent à créer les conditions d’un foyer ; il faut aussil’émergence simultanée de nombreux moustiques de première générationinfectés par voie transovarienne. Après l’amplification du virus chez lesvertébrés, les moustiques agissent comme vecteur secondaire pour prolongerl’épidémie.

Mots-clésAvortement – Fièvre de la Vallée du Rift – Inondation – Moustique – Nécrose hépatique– Transmission transovarienne – Zoonose.

La fièvre de la Vallée du Rift

G.H. Gerdes

ResumenLa fiebre del Valle del Rift es una enfermedad vírica transmitida por artrópodosque afecta a rumiantes, camélidos y seres humanos. También es una importantezoonosis, que a veces se presenta como afección de tipo gripal sincomplicaciones, pero otras veces da lugar a una enfermedad hemorrágica conafectación hepática. Puede ocasionar asimismo lesiones oculares oneurológicas. En los animales, la fiebre del Valle del Rift puede ser asintomáticaen adultos no gestantes, pero los brotes se caracterizan por la aparición deabortos y una elevada tasa de mortalidad neonatal. En ejemplares de más edadse observan ictericia, hepatitis y muerte del animal. Los brotes de estaenfermedad vienen asociados a lluvias fuertes y persistentes, con inundacionescontinuas y la aparición de gran número de mosquitos, que son su principalvector. La presencia de fuertes lluvias localizadas no crea por sí sola lascondiciones necesarias para un brote, pues se requiere también la apariciónsimultánea y en gran número de una primera generación de mosquitosinfectados por vía transovárica. Tras la ampliación del virus en vertebrados, elinsecto actúa como vector secundario y perpetúa así la epidemia.

Palabras claveAborto – Fiebre del Valle del Rift – Inundación – Mosquito – Necrosis hepática –Transmisión transovárica – Zoonosis.

La fiebre del Valle del Rift

References1. Abd El-Rahim I.H.A., Abd El-Hakim U. & Hussein M.

(1999). – An epizootic of Rift Valley fever in Egypt in 1997.Rev. sci. tech. Off. int. Epiz., 18 (3), 741-748.

2. Abdel-Wahab K.S.E., El Baz L.M., El Tayeb E.M., Omar H.,Ossman M.A. & Yasin W. (1978). – Rift Valley fever virusinfections in Egypt: pathological and virological findings inman. Trans. roy. Soc. trop. Med. Hyg., 72 (4), 392-396.

3. Anderson G.W., Slone T.W. & Peters C.J. (1987). –Pathogenesis of Rift Valley fever virus (RVFV) in inbred rats.Microb. Pathogen., 2 (4), 283-293.

4. Barnard B.J.H. & Botha M.J. (1977). – An inactivated RiftValley fever vaccine. J. S. Afr. vet. med. Assoc., 48 (1), 45-48.

5. Barnard B.J.H. (1979). – Rift Valley fever vaccine – antibodyand immune response in cattle to a live and an inactivatedvaccine. J. S. Afr. vet. med. Assoc., 50 (3), 155-157.

6. Bayliss M., Mellor P.S. & Meiswinkel R. (1999). – Horsesickness and ENSO in South Africa. Nature, 397, 594.

7. Caplen H., Peters C.J. & Bishop D.H.L. (1985). – Mutagen-directed attenuation of Rift Valley fever as a method ofvaccine development. J. gen. Virol., 66, 2271-2277.

8. Clarke D.H. & Cassals J. (1958). – Techniques forhaemagglutination and haemagglutination inhibition witharthropod-borne viruses. Am. J. trop. Med. Hyg., 7, 561-573.

9. Coetzer J.A.W. (1977). – The pathology of Rift Valley fever. I.Lesions occurring in natural cases in newborn lambs.Onderstepoort J. vet. Res., 44, 205-212.

10. Coetzer J.A.W. (1982). – The pathology of Rift Valley fever. II.Lesions occurring in field cases in adult cattle, calves andaborted foetuses. Onderstepoort J. vet. Res., 49, 11-17.

11. Coetzer J.A.W. & Ishak K.G. (1982). – Sequentialdevelopment of liver lesions in newborn lambs infected withRift Valley fever virus. I. Macroscopic and microscopicpathology. Onderstepoort J. vet. Res., 49, 105-108.

12. Daubney R.J., Hudson J.R. & Garnham P.C. (1931). –Enzootic hepatitis of Rift Valley fever: an undescribed virusdisease of sheep, cattle and man from East Africa. J. Pathol.Bacteriol., 34, 543-579.

13. Davies F.G., Linthicum K.J. & James A.D. (1985). – Rainfalland epizootic Rift Valley fever. Bull. WHO, 63 (5), 941-943.

14. Davies F.G., Kilelu E., Linthicum K.J. & Pegram R.G. (1992).– Patterns of Rift Valley fever activity in Zambia. Epidemiol.Infect., 108 (1), 185-191.

15. Deutman A.F. & Klomp H.J. (1981). – Rift Valley feverretinitis. Am. J. Ophthamol., 92 (1), 38-42.

16. Easterday B.M., McGavran M.H., Rooney J.R. & Murhpy L.C.(1962). – The pathogenesis of Rift Valley fever in lambs. Am.J. vet. Res., 23, 470-478.

17. Erasmus B.J. & Coetzer J.A.W. (1981). – Symptomatologyand pathology of Rift Valley fever in domestic animals.Contrib. Epidemiol. Biostat., 3, 77-82.

18. Garrett L. (1995). – N’zara. In Coming plague: newlyemerging diseases in a world out of balance. Penguin Books,Middlesex, 192-221.

19. Harrington D.G., Lupton H.W., Crabbs C.L., Peters C.J.,Reynolds J.A. & Slone T.W. Jr. (1980). – Evaluation of aformalin-inactivated Rift Valley fever vaccine in sheep. Am. J. vet. Res., 41 (10), 1559-1564.

20. Hubbard K.A., Baskerville A. & Stephenson J.R. (1991). –Ability of a mutagenised virus variant to protect young lambsfrom Rift Valley fever. Am. J. vet. Res., 52, 50-55.

21. Jouan A., Coulibaly I., Adam F., Philippe B., Riou O.,Leguenno B., Christie R., Ouold Merzoug N., Ksiazek T. &Digoutte J.P. (1989). – Analytical study of a Rift Valley feverepidemic. Res. Virol., 140 (2), 175-186.

22. Karlen A. (1996). – A garden of germs. In Plague’s progress: asocial history of man and disease. Cassell Group, London,149-173.

23. Linthicum K.J., Anyamba A., Tucker C.J., Kelley P.W., Myers M.F. & Peters C.J. (1999). – Climate and satelliteindicators to forecast Rift Valley fever epidemics in Kenya.Science, 285, 397-400.

24. Meegan J.M., Hoogstraal H. & Moussa M.I. (1979). – Anepizootic of RVF in Egypt in 1971. Vet. Rec., 105, 124-125.

25. Mellor P.S. & Leake C.J. (2000). – Climatic and geographicinfluences on arboviral infections and vectors. In An updateon zoonoses (P.-P. Pastoret, ed.). Rev. sci. tech. Off. int. Epiz., 19 (1), 41-54.

26. Morrill J.C., Jennings G.B., Caplan H., Turell M.J., Johnson A.J. & Peters C.J. (1987). – Pathogenicity andimmunogenicity of a mutagen-attenuated Rift Valley fevervirus immunogen in pregnant ewes. Am. J. vet. Res., 48 (7),1042-1047.

27. Morrill J.C., Carpenter L., Taylor D., Ramsburg H.H., Quance J. & Peters C.J. (1991). – Further evaluation of amutagen-attenuated Rift Valley fever vaccine in sheep.Vaccine, 9 (1), 35-41.

28. Muller R., Saluzzo J.F., Lopez N., Dreier T., Turell M., Smith J. & Bouloy M. (1995). – Characterization of clone 13,a naturally attenuated avirulent isolate of Rift Valley fevervirus, which is altered in the small segment. Am. J. trop. Med.Hyg., 53 (4), 405-511.

29. Niklasson B., Peters C.J., Grandien M. & Wood O. (1984). –Detection of human immunoglobulins G and M antibodies toRift Valley fever virus by enzyme-linked immunosorbentassay. J. clin. Microbiol., 19 (2), 225-229.

30. Peters C.J. & Anderson G.W. (1981). – Pathogenesis of RiftValley fever. Contrib. Epidemiol. Biostat., 3, 21-41.

31. Peters C.J. & Meegan J.M. (1994). – Rift Valley fever. InHandbook of zoonoses. Section B: viral zoonoses, 2nd Ed.(G.W. Beran & J.H. Steele, eds). CRC Press, Boca Raton, 403-420.

Rev. sci. tech. Off. int. Epiz., 23 (2)622

32. Porter H.F. (1999). – Onward to Africa. In Forecast: disaster.The future of El Niño. Dell Publishing, New York, 127-144.

33. Reeves W.C., Hardy J.L., Reisen W.K. & Milby M.M. (1994).– The potential effect of global warming on mosquito-bornearboviruses. J. med. Entomol., 31 (3), 323-332.

34. Sall A.A., Zanotto de A. P.M., Zeller H.G., Digoutte J.P.,Thiongane Y. & Bouloy M. (1997). – Variability of the NS (S)protein among Rift Valley fever isolates. J. gen. Virol., 78,2853-2858.

35. Sall A.A., Zanotto de A. P.M., Sene O.K., Zeller H.G., Digoutte J.P., Thiongane Y. & Bouloy M. (1999). – Geneticreassortment of Rift Valley fever virus in Nature. J. Virol., 73(10), 8196-8200.

36. Sall A.A., Thonnon J., Sene O.K., Fall A., Ndiaye M., Baudez B., Mathiot C. & Bouloy M. (2001). – Single-tube andnested reverse transcription-polymerase chain reaction fordetection of Rift Valley fever virus in human and animal sera.J. virol. Meth., 91 (1), 85-92.

37. Schmaljohn C.S. & Hooper J.W. (2001). – Bunyaviridae: theviruses and their replication. In Fields virology, 4th Ed. (B.N. Fields, D.M. Knipe, P.M. Howley, D.E. Griffin, R.A.Lamb, M.A. Martin, B. Roizman & S.E. Straus, eds).Lippincott, Williams & Wilkins, Philadelphia, 1581-1602.

38. Sellers R.F., Pedgley D.E. & Tucker M.R. (1982). – Rift Valleyfever, Egypt 1977: disease spread by windborne insectvectors? Vet. Rec., 110 (4), 73-77.

39. Smithburn K.C. (1949). – Rift Valley fever: the neurotropicadaptation of the virus and the experimental use of thismodified virus as a vaccine. Br. J. experim. Pathol., 30, 1-16.

40. Swanepoel R., Struthers J.K., Erasmus M.J., Shepherd S.P.,McGillivray G.M., Erasmus B.J. & Barnard B.J.H. (1986). –Comparison of techniques for demonstrating antibodies toRift Valley fever virus. J. Hyg. (London), 97 (2), 317-329.

41. Swanepoel R. & Coetzer J.A.W. (1994). – Rift Valley fever. In Infectious diseases of livestock with special reference tosouthern Africa, Vol. I (J.A.W. Coetzer, G.R. Thomson & R.C. Tustin, eds). Oxford University Press, Cape Town, 688-717.

42. Van der Lugt J.J., Coetzer J.A.W. & Smit M.M. (1996). –Distribution of viral antigens in tissues of newborn lambsinfected with Rift Valley fever virus. Onderstepoort J. vet. Res.,63 (4), 341-347.

43. World Health Organization (WHO) (1990). – Potential healtheffects of climate change. A WHO Task Group Report.WHO/PEP/90/10. WHO, Geneva.

44. Yedloutschnig R.J., Dardiri A.H., Mebus C.A. & Walker J.S.(1981). – Abortion in vaccinated sheep and cattle afterchallenge with Rift Valley fever virus. Vet. Rec., 109 (17), 383-384.

45. Zeller H.., Fontenille D., Traoré-Lamizana M., Thiogane Y. &Digoutte J.P. (1997). – Enzootic activity of Rift Valley fevervirus in Senegal. Am. J. trop. Med. Hyg., 56 (3), 265-272.

46. Zeller H. & Bouloy M. (2000). – Infections by viruses of thefamilies Bunyaviridae and Filoviridae. In An update onzoonoses (P.-P. Pastoret, ed.). Rev. sci. tech. Off. int. Epiz., 19 (1), 79-91.

Rev. sci. tech. Off. int. Epiz., 23 (2) 623