Article — Artikel

Rabies in the Eastern Cape Province of South Africa –Where are we going wrong?

S J van Sitterta*, J Raathb, G W Akolc, J M Miyend, B Mlahlwaa and C T Sabetad

INTRODUCTIONRabies is a fatal disease caused by a virus

of the family Rhabdoviridae, genusLyssavirus. The severe nature of the esti-mated 55 000 human deaths a year causedby the disease, mainly in Africa and Asia24,underline its impact. Lyssavirus currentlyincludes 11 genotypes recognised by theInternational Committee on Virus Taxon-omy19, 4 of which have been isolated inSouth Africa to date (Rabies virus, Mokolavirus, Lagos bat virus and Duvenhagevirus)31. For the Rabies virus genotype, thedomestic dog is the major reservoir andvector and is responsible for most humancases worldwide31. However, differentRabies virus biotypes within particularspecies and geographical areas canexist40. For example, in southern Africa 2biotypes (canine and mongoose) are

recognised. Mongoose rabies is adaptedto and circulates within herpestid species,particularly the yellow mongoose(Cynictis penicillata). The cosmopolitancanid biotype, on the other hand, circu-lates within members of the familyCanidae: dogs (Canis familiaris), blackbacked jackals (C. mesomelas) andbat-eared foxes (Otocyon megalotis)31,32,35.While mongoose rabies is considered tobe indigenous to the area, canine rabiesappears to have been introduced duringmodern times31.

The 1st confirmed diagnosis of rabies inSouth Africa was in 1893 following an out-break of the disease in the Eastern Capefrom an imported dog6,40. This outbreakwas brought under control, however, andthe spatial spread of rabies was thenclosely associated with wildlife speciesparticularly the yellow mongoose. By1940, canine rabies had begun to spreadsouth of the Zambezi River so that by 1950it appeared in South Africa in the north-ern Limpopo Province, where it is stillpresent6,26,40. Subsequent to the 1950Limpopo outbreak, canine rabies spreadto Mozambique from where it enteredKwaZulu-Natal in 196125,40. Although theKwaZulu-Natal outbreak was broughtunder control, the disease reappeared in1976 following an influx of refugees fromMozambique40. This outbreak could not

be brought under control and marked thestart of the incessant southward spread ofcanine rabies. In 1986, the disease reachedthe Eastern Cape, when it was confirmedin the northern areas of Transkei (Malutiand Umzimvubu local municipalities ofthe Eastern Cape Province)2. Over thenext 4 years, canine rabies continued tospread throughout Transkei and by theearly 1990s had reached East London.Currently, rabies is a re-emerging publichealth problem in the Eastern Cape,evident by the fact that from 2008 to 2009more human cases were reported thanfrom any other province5.

Although rabies has been confirmed inwild and domestic animal species in theEastern Cape, canine rabies was the maindriver of the observed temporal trendsover the period 1986–2009 (Fig. 1). Themongoose biotype in the Eastern Cape ismaintained especially in the Karoo andadjoining areas (Cacadu, Western ChrisHani and Ukhahlamba District munici-palities), but is absent in the formerTranskei area32. Among domestic animalsin the Eastern Cape, rabies has been diag-nosed predominantly in dogs (52 %) fol-lowed by cattle (34 %), goats (6 %), sheep(4 %), domestic cats (2 %) and otherdomestic animals (2 %) over the period1986–2009 (Eastern Cape Department ofAgriculture, Forestry and Fisheries,unpubl. data). The high incidence ofrabies in cattle and other dead-end hostscompared to the rest of South Africa18

seems to indicate that surveillance cover-age is still insufficient, and that theburden of rabies in the province may beunderestimated.

Control of rabies through targeted vac-cination of at least 70 % of the dog popula-tion is the standard method of controllingrabies and has been used successfully indeveloped countries10,42, although a recip-rocal increase (whether real or relative) ofthe disease in wildlife species may occur32.Vaccination of dogs and cats has beenimplemented in the Eastern Cape Prov-ince on an annual basis since 1986, but hasneither curbed the spread nor preventedthe rise in the number of canine rabiescases and dead-end host species, proba-bly because less than 70 % of the popula-tion was reached2. The problem iscompounded by the lack of information

0038-2809 Jl S.Afr.vet.Ass. (2010) 81(4): 207–215 207

aDepartment of Agriculture, Forestry and Fisheries,Eastern Cape Province, Veterinary Services-AnimalHealth (Emalahleni LM), Private Bag X7093, Queens-town, 5320 South Africa.

bDepartment of Agriculture Forestry and Fisheries,Eastern Cape Province, GIS unit, PO Box 131, Cradock,5880 South Africa.

cDepartment of Agriculture, Forestry and Fisheries, East-ern Cape Province, Centre of Veterinary Excellence,Dohne ADI, Private Bag X14, Stutterheim, 4930 SouthAfrica.

dAgricultural Research Council – Onderstepoort Veteri-nary Research Institute, OIE Rabies Reference Labora-tory, Private Bag X05, Onderstepoort, 0110 South Africa.

*Author for correspondence. Present address: PO Box124, Malelane, 1320 South Africa.E-mail: sybrandvs@yahoo.com

Received: May 2010. Accepted: October 2010.

ABSTRACTRabies is a growing problem in the Eastern Cape Province of South Africa. This study inves-tigated dog ecology, vaccination coverage and rabies neutralising antibody levels in203 randomly selected dogs within a local municipality in the former Transkei area.Responses to vaccination were also evaluated in 80 of these dogs. The population wasremarkably uniform in size, breed and condition. Slightly over 1/5th of the populationwas between 6 weeks and 1 year of age, while very few dogs reached 10 years or older.According to owner responses, the Animal Health Technicians achieved a total vaccinationcoverage of 65 % of owned dogs over several years, but only 56 % within the previous 12months. Only 32 % of dogs had adequate circulating rabies virus neutralisation antibodies(≥0.5IU/ ). After vaccination, 83 % had seroconverted to this level. The magnitude of sero-conversion was independent of body condition or age. This study proposes a differentapproach to vaccination strategies than those currently employed in certain areas of theprovince.

Keywords: dog, Eastern Cape, ecology, lyssavirus, rabies, seroconversion, South Africa,Transkei, vaccination.

Van Sittert S J, Raath J, Akol G W, Miyen J M, Mlahlwa B, Sabeta C T Rabies in the EasternCape Province of South Africa – where are we going wrong? Journal of the South AfricanVeterinary Association (2010) 81(4): 207–215 (En.). Department of Agriculture, Forestry andFisheries, Eastern Cape Province, Veterinary Services-Animal Health (Emalahleni LM),Private Bag X7093, Queenstown, 5320 South Africa.

on the ecology of the dog population inthis province. Indeed, information on dogpopulation ecology appears to be verylimited in most parts of Africa33. In orderto develop an effective approach to rabiescontrol it is essential to understand thedog ecology and establish monitoringmeasures for assessing the effectivenessof the vaccination campaigns given theshortcomings highlighted above. Thispaper outlines preliminary findings of astudy aimed at identifying some of thekey ecological factors that appear to influ-ence the successful immunisation of dogsagainst rabies in a communal farmingarea in the province. The study furtherinvestigated factors that may influencethe ability of dogs to respond effectivelyto rabies vaccines.

MATERIALS AND METHODS

Study areaThis study was conducted in the

Emalahleni Local Municipality, which isadministered from the central town ofLady Frere (27�14’E; 31�42’S), and is partof the greater Chris Hani District Munici-pality in the Eastern Cape Province ofSouth Africa (Fig. 2). The majority(3317 km2) of the 3551 km2 area compriseswhat was formerly Transkei whereraising livestock on communal grazing isthe main farming activity. In the north-ernmost area the people farm livestock ona commercial level. Apart from 3 smalltowns, the mountainous area consists ofrural villages of varying population sizes.The housing in the district is generallytraditional round huts and more modernbrick houses, especially in the towns. Thetotal human population in the Emala-hleni municipality was estimated at

125 288 in 2007 (Statistics South Africa,pers. comm.). No formal census data ofdog numbers in Emalahleni are available,and thus the dog population could onlybe estimated as ranging from 8348 to25 058, based on a human to dog ratio of1:5 to 1:1523. During 2008–2009 10 cases oflaboratory confirmed rabies (8 dogs, 1 bulland 1 goat) were reported in Emalahleni.These cases were separated temporallythroughout the 2 years and spatiallyacross the district, and it was not estab-lished if it comprised 1 or many out-breaks. Every year a rabies vaccinationcampaign is launched throughout themunicipality, where dogs and cats arevaccinated free of charge through a streetto street or farm to farm vaccinationstrategy. This proactive vaccinationcampaign is conducted by animal healthtechnicians (AHTs) in addition to addi-tional vaccinations in the face of an out-break.

Sampling methodsThe dog population density is propor-

tional to the human population densityand is often remarkably similar acrossdifferent African communities9,20,28. As asampling strategy, 203 GPS coordinateswere randomly generated throughoutthe municipality. In order to reflect theexpected spread of the canine population,however, the coordinates were weighted/focused towards areas of higher humanpopulation density. This was carried outby selecting 1.4 %, 2.9 %, 7.4 %, 14.7 %,29.4 % and 44.1 % of the GPS coordinatesto fall in human population densities of0–5, 5–10, 10–25, 25–50, 50–100 and >100people per square kilometre respectively(Fig. 3). ArcGIS® software (ESRI Inc.,Redlands, USA) was used to randomly

allocate biased coordinates throughoutthe municipality based on data from theSouth African Geo-referenced Informa-tion System (www.agis.agric.za). Thesecoordinates were located with a Garmin®

nüvi® 220W series GPS tracking device(Garmin Distribution Africa, Honeydew,South Africa). The closest household toeach set of coordinates that owned a dogor dogs was used as a sampling point.

Sample collection and storageThe most recent vaccination campaign

preceding the study was from October toDecember 2008. Sampling took placefrom November 2008 to May 2009.Sampling that coincided with the 2008campaign (i.e. November and December2008) was done within areas that werealready covered during the 2008 cam-paign. At each sampling point, 1 dog ofat least 6 weeks of age was randomlyselected. With the owner’s consent, bloodwas collected from the cephalic, lateralsaphenous or jugular veins. The bloodwas allowed to clot before being centri-fuged and the serum stored at –20 °Cwithin 12 hours. On the day of bloodcollection, the dog was injected subcuta-neously with 1 m of the same vaccine usedduring the 2008 rabies campaign, namelyRabdomun® (Intervet/Schering-Plough,Isando, RSA (batch A667B2)). Each m ofvaccine had an antigenic value of ≥5.0, asdetermined by the modified NIH test,and at least 107,3 MLD50 Flurry LEP strainvirus grown on Baby Hamster Kidney(BHK) 21 cell culture, inactivated with anaziridine compound and absorbed intoaluminium hydroxide (package insert). Aquestionnaire administered in isiXhosa,Afrikaans or English was completedthrough an owner interview for each dog

208 0038-2809 Tydskr.S.Afr.vet.Ver. (2010) 81(4): 207–215

Fig. 1: Trends of rabies in domestic (including all canids) and wildlife (predominantly herpestids) species in the Eastern Cape Province duringthe period 1984–2009. Rabies cases reported as domestic prior to 1986 (typing results not available) were cases that occurred in 2 bat-earedfoxes (1984), 1 bovine (1984), 1 canine (1984), 2 feline (1984 and 1985) and 1 caprine (1985), all of which occurred in the western part ofthe province that is predominantly endemic to herpestid rabies. (Eastern Cape Department of Agriculture, Forestry and Fisheries, VeterinaryServices, unpubl. data).

in the study sample. The questionnairenoted the GPS coordinates of the house-hold, the name of owner and the name ofthe locality, i.e. village/farm. Data re-corded about the dog included name, age,sex (male/female/sterilised/lactating),breed, size, condition, rabies vaccinationhistory, function (pet/security/hunt-ing/herding) and any noteworthy externalcharacteristics (external parasites/inju-ries/external tumours). The dog in ques-tion was classified as confined (i.e.permanently confined) or roaming freely(including semi restriction, i.e. access tothe general dog population some of thetime without owner supervision). Rabiesvaccination history was ascertained byasking if the dog was previously vacci-nated (yes/no/unsure) and if previouslyvaccinated, when last (less than 12months ago/more than 12 months ago/unsure). In most cases, however, the vac-cination history could not be verified. Thecondition of the dog was subjectivelyclassified as ‘poor’, ‘thin’, ‘good’ or ‘over-weight’: dogs with mange, hair loss, highexternal parasite burden or which wereemaciated were placed into the ‘poor’ cat-egory; ‘thin’ dogs had a low muscle andfat mass, a well kept coat with no or fewexternal parasites; ‘good’ condition dogshad an adequate muscle and fat mass, a

well kept coat with no or few externalparasites and overweight dogs had a wellkept coat with no or few external parasitesand a fat mass which made palpation ofbony prominences on the spine and pel-vic area difficult. Interviews and condi-tion scoring were conducted by thelocal state veterinarian. It should beemphasised that parasite burden wasjudged only on the presence of adult ticksand fleas seen or felt on the dog’s headand external pinnae, dorsal and lateralbody wall or tail. No attempt was made tocollect or count these parasites. Any dogadditions or losses to the household in thepast 2 years were recorded as well as thecurrent number of dogs living at thehouse. A post-vaccination blood samplewas taken between 30 and 60 days(mean = 52 days) after the 1st sample from80 of the sampled dogs. The selection ofdogs for the post-vaccination sample wasdone randomly.

Determination of rabies antibody titresRabies antibody titres were assessed

using a standard 48 hour fluorescent anti-body virus neutralisation test (FAVNT) asdescribed previously7. Briefly, 3-fold serumdilutions were incubated with a 100TCID50/ of challenge rabies virus (ATCCVR959, CVS-II) and any un-neutralised

challenge virus allowed to grow on sus-ceptible Baby Hamster Kidney cells(BHK C13-ATCC: CL-10, DiagnosticHybrids, USA). Virus growth on BHKcells was detected by acetone-fixation ofthe monolayer and stained with fluores-cein isothiocyanate (FITC)-labelled anti-rabies monoclonal hyperimmune serum(Onderstepoort Veterinary Institute,Pretoria, South Africa). The microtitreplates were examined under a fluores-cence microscope to detect infected cellsand 50 % endpoint titres calculated usingthe Spearman-Kärber method21,38,41. TheFAVN tests were performed at theOnderstepoort Veterinary Institute, anOIE rabies reference laboratory.

Data analysisData were analysed with SPSS® statistics

17.0 (SPSS Inc, Chicago, USA). Proportionsof responses were evaluated through achi-square test for an r × k contingencytable. Where proportions of previouslyvaccinated dogs were evaluated, ‘unsure’responses (n = 5) were omitted. Serumtitre results in International Units permillilitre (IU/ ) were assessed for normal-ity through normal probability plots.Since the distribution of serum titres wereskewed to the right, data were log trans-formed. The difference between the 1st

0038-2809 Jl S.Afr.vet.Ass. (2010) 81(4): 207–215 209

Fig. 2: Location of the study area within the Eastern Cape, South Africa. The map indicates the former Transkei (North) and Ciskei (South)areas, as well as the Emalahleni local municipality, which formed part of the westernmost part of the former Transkei area.

and 2nd (30–60 days post-vaccination)serum rabies antibody titre were evalu-ated through a paired t-test. A repeatedmeasures analysis of variance (or mixeddesign ANOVA) with ‘age group’ and‘condition score’ as between subject fac-tors, was used to evaluate the interactioneffects of these factors with the repeatedserum titre measurements. For purposesof this mixed design ANOVA, the agegroup factor was pooled into the follow-ing categories: less than 1 year, 1–3 yearsand more than 3 years. The body condi-tion factor was grouped into the follow-ing categories: poor, thin and good/overweight. The level of significance wasset at P < 0.05.

RESULTS

Description of the canine populationThe majority (96 %) of the dogs were

classified as a medium-sized dog, whilelarge dogs represented 3 % and smalldogs 1 %. The breed most represented inthe study could be classified as Africanis,or the local landrace17, comprising 93 %,while cross-breeds represented 4 % andother dog breeds 3 % of the population.The male:female ratio was 1.7:1. Ownersvery often did not know the exact age oftheir dogs, but were able to classify them

into the age groups specified in the ques-tionnaire. The dogs in the sample popula-tion had the following age groupdistribution: 6–12 weeks = 3 %, 3–12months = 18 %, 1–3 years = 43 %, 3–10years = 35 % and more than 10 years =1 %. The dogs were generally in a ‘good’condition with 3 %, 76 %, 14 % and 7 % ofthe sample population classified into‘overweight’, ‘good’, ‘thin’ and ‘poor’categories respectively.

Nearly a quarter of the households(23 %) reported that they use dogs forhunting/herding animals in addition tokeeping them as pets or for securityreasons. Most of the dogs in the study(77 %) were not confined to the house androamed freely. The mode of dogs perhousehold that owned dogs was 2 (mini-mum = 1; 1st quartile = 2; median = 3;3rd quartile = 4; maximum = 10). Themedian number of dogs entering thehousehold (born/bought/adopted)during the 2 years prior to the study was1 (range 0 to 18) and for dogs leaving thehousehold (died/sold/given away) themedian was 0.5 (range 0–18). However, bymany owners’ admission this was an esti-mate and they could not exactly recall thedog turnover rate at their house over2 years.

Animal health technician efficacyAnimal health technicians are responsi-

ble for vaccinating as many dogs and catsas possible during the annual vaccinationcampaigns and in Emalahlenimunicipality there are currently 8 AHTs.The effectiveness of the previous years’vaccination coverages was evaluatedwith the question: ‘Has the dog sampledbeen vaccinated before?’ (Table 1). Of the203 respondents, 65 % reported ‘yes’, 2 %were unsure, and 33 % reported ‘no’. Ofthe dogs that had been vaccinated previ-ously, 64 % had been vaccinated oncebefore, 20 % twice, 10 % 3 times and 1 %4 times (5 % were unsure as to how manytimes the dog was vaccinated previously).Slightly more than half of the dogs (56 %)had been vaccinated within the 12months before being sampled, while 5 %of owners were not sure when last theirdogs had been vaccinated. The effect ofthe population density group the ownerresided in was not significant with regardto the reported vaccination status of his/her dog (χ2

(5) = 2.97, P = 0.71). IndividualAHT efficacies were compared by group-ing the responses in terms of the geo-graphical areas for which each AHT wasresponsible. Nine geographical areaswere identified, of which 8 AHT each has

210 0038-2809 Tydskr.S.Afr.vet.Ver. (2010) 81(4): 207–215

Fig. 3: Emalahleni Local Municipality indicating population density and the sampling points.

had his/her own area, while a single areawas shared between 2 AHTs. The reportedvaccination status of dogs was signifi-cantly associated with the technicianresponsible for its vaccination (χ2

(8) =18.08, P = 0.02). Twenty-three per cent ofowners reported that they used their dogsfor hunting or herding as opposed tousing them solely as pets or for security.Of the dogs used for hunting/herding,72 % were reported as previously vacci-nated while 62 % of dogs which are usedpurely as pets or for security had beenvaccinated before according to theirowners. The relationship between a dog’suse and its reported vaccination statuswas not significant, however (χ2

(1) = 1.05,P = 0.31). There were also no significantrelationships between a dog’s reportedvaccination status and its sex (χ2

(1) = 1.05,P = 0.31) or body condition (χ2

(3) = 5.37,P = 0.15). Conversely, a dog’s reportedvaccination status was significantly asso-ciation with its age (χ2

(4) = 23.91, P < 0.001)(Table 1).

Serological description of the dogssampled

Rabies-neutralising antibody titreswere determined in 203 dogs. The geo-metric mean rabies titre was 0.38 IU/ (1stquartile = 0.15 IU/ (lowest measuredpoint); median = 0.21 IU/ ; 3rd quartile =0.65 IU/ ). Of the 203 dogs sampled, 51 %had titres above 0.2 IU/ , while only 32 %had titres of ≥0.5 IU/ – an animal with atitre of 0.5 IU/ and above is highly likelyto be protected against rabies virus infec-tion in the face of challenge3, 42. When con-sidering only those dogs whose ownersreported previous vaccination, 43 % hadrabies-neutralising antibody titres of0.5 IU/ and above. On the other hand,when considering only dogs with adequate

neutralising antibody titres, 11 % had notreceived a previous vaccination accord-ing to their owners. There was a highlysignificant relationship between a dog’sreported vaccination status and whetheror not it had an adequate rabies-neutralis-ing antibody titre (χ2

(1) = 21.3, P < 0.001).There was no significant relationshipbetween a dog’s body condition andwhether or not it had a titre of ≥0.5 IU/(χ2

(3) = 3.85, P = 0.28), but there was suffi-cient evidence to conclude that titres of0.5 IU/ and above were associated withincreasing age (χ2

(4) = 26.64, P < 0.001).

Response to vaccinationEighty of the 203 dogs originally sampled

were retested 30–60 days later to deter-mine the response to vaccination. Of the80 dogs that were resampled, 83 % hadtitres that indicated successful sero-conversion. Of those dogs that previouslyhad rabies antibody titres below 0.5 IU/ ,64 % seroconverted to titres above 0.5 IU/(range: 0.5–23.38 IU/ ). The geometricmean titre of the retested dogs was1.44 IU/ (1st quartile = 0.64 IU/ ;median = 1.30 IU/ ; 3rd quartile =2.61 IU/ ). A paired-samples t-test indi-cated that the resampled dogs had asignificantly greater titre 30–60 days afterthe rabies vaccination (Fig. 4a, t(79) = 11.3,P < 0.001) (1st quartile of difference =0.26 IU/ ; median = 0.7 IU/ ; 3d quartile ofdifferences = 2.14 IU/ ). There seemed tobe minor differences in the magnitude ofresponse to vaccination of the differentfactor groups (Fig. 4b,c,d), althoughneither the ‘age group × vaccinationinteraction’ (F(2,72) = 0.65, P = 0.53) nor the‘condition × vaccination interaction’(F(2,72) = 1.52, P = 0.225) was significant.However, there was a significant maineffect of ‘age group’ (F(2,72) = 5.74, P =

0.005. This effect tells us that at least 1 ofthe age groups differed in their serumtitres.

DISCUSSIONThe estimation of the canine population

size in Emalahleni could be narroweddown further to c. 16 500, based on 2008’svaccination figures in the municipality(unpubl. records) and vaccination cover-age of about 56 % a year (see above). Thecanine population in the study area wasremarkably homogeneous in size, breed,and condition. Interestingly, however,only a small proportion of dogs was as-signed to the ‘poor condition’ category(7 %). The proportion of dogs classified asin good condition (76 %) is slightly higherthan similar studies in rural southernAfrica (56 %30 and 67 %34) and rural Zam-bia (60 %), but similar to urban dogs inZambia (74 %)11. It could be argued that ahigh proportion of these dogs, by virtueof not having similar access to veterinaryservices or balanced diets as dogs in moredeveloped areas of the country, should beprone to high parasite burdens and/ormalnutrition, and thus be overrepre-sented in the ‘poor condition’ category.This contrary finding might be ascribed tothe subjective nature of the assessment,but may also be ascribed to an inherentresistance to parasites and disease of thelocal land-race, and the ability to look foradditional sources of food in the ruralenvironment17. It is important to remem-ber, however, that the rural ‘village’scenario should be distinguished fromthat which occurs in townships/shantytowns in South Africa, as the findingabout dog condition and breed might notbe the same in both27. Furthermore, if amore objective approach is taken throughmeasurement of clinical parameters,counting internal and external parasitesand serological screening for variousdiseases, it may be that a larger propor-tion of animals will be identified as inneed of veterinary attention30,34.

As in similar studies in Africa23,34,37, therewere more males than females. This couldbe ascribed to better care of male animalsdue to preference for using male animalsduring hunting23, although in Emalahlenionly 23 % of owners reported using theiranimals for either hunting or herding. Assterilisation services are either too expen-sive or difficult to arrange (only 7 (3 %) ofthe owners reported their dog as steri-lised, of which 6 were males), it might alsobe that owners prefer to keep males ratherthan females as a form of populationcontrol at home.

Slightly over 1/5th (21 %) of the dogs inthe study sample were between 6 weeksand 1 year of age, an observation similar

0038-2809 Jl S.Afr.vet.Ass. (2010) 81(4): 207–215 211

Table 1: Summary of the dogs sampled in the Emalahleni municipality. Dogs were dividedinto groups based on sex,condition,age and rabies vaccination status. In each group,dogswere further divided into previous rabies vaccination status.

Previous vaccination status Total

Classification Yes No Unsure

Total no. of dogs in study 203

Sex: Male 86 39 2 127Female 45 28 3 76

Condition: Poor 7 6 1 14Thin 21 6 2 29Good 98 55 1 154Overweight 6 0 2 6

Age group: 6 weeks to 3 months 0 5 0 53–12 months 18 18 1 371–3 years 54 32 1 873–10 years 56 12 3 71More than 10 years 3 0 0 3

Previously vaccinated 131 67 5 203

to a study in Tunisia37, where 28 % of thedogs were less than 1 year old (includinganimals younger than 6 weeks). On theother hand, a study in a district in Kenyafound that more than 50 % of the dogpopulation was less than 1 year old23 . Theproportion of dogs aged less than 1 year iscrucial in rabies eradication as this agegroup is very often unvaccinated39 orshows a smaller proportion of sero-conversion (≥0.5 IU/ ) than older dogs37.Moreover, only a small proportion (1 %)of dogs were reported to reach 10 years ormore. This is similar to a previous studyin which the authors found no dog overthe age of 8 years in a town in formerBophuthatswana (in present-day NorthWest Province, South Africa)34. This indi-cates that although many dogs may notseem to be in a poor condition based onthis study’s condition scoring system,more dogs may be in need of veterinaryattention or improved care in order toimprove their longevity. Further studieson the causes of dog mortality withinthese communities are thus needed.

The finding that 77 % of the owned dogpopulation is free-roaming is in accor-dance with observations in other ruralAfrican communities17,23. It is important tonote that the feral (or ownerless) dogproportion is often very small in thesecommunities, seldom exceeding 5 %20.Thus eradication of unrestricted or free-roaming animals as an adjunct in thecontrol of rabies is not practical in thiskind of community because of the diffi-culty in distinguishing between ownedand feral animals. Even if feral animalswere to be eliminated from the popula-tion, the majority of owned dogs wouldstill roam freely. Furthermore, it has beenfound that dog control in the form ofquarantine, movement control and dogremoval had been counterproductive insome cases in eastern and southern Africain recent years, provoking public antipathytowards rabies control campaigns33.Therefore, although free-roaming dogsare an important factor in the epidemiol-ogy of rabies outbreaks, we can expectthis trend to continue within many Afri-can communities. The experience of fieldwork in Emalahleni has shown, however,

212 0038-2809 Tydskr.S.Afr.vet.Ver. (2010) 81(4): 207–215

Fig. 4: Box and whisker plots of antibodytitres (logarithmically transformed) at the be-ginning of the study (n = 203) and 30–60days post-vaccination (n = 80). The dashedline denotes the antibody titre which is con-sidered as successful seroconversion(0.5 IU/ ). a, Antibody titres pooled beforeand after vaccination; b, antibody titresgrouped according to age categories; c,antibody titres grouped according to condi-tion;d, antibody titres grouped according toprevious vaccination status reported byowners.

that dogs tended not to wander far fromtheir homes unless accompanied by theirowners, and were generally available forvaccination when requested, especiallywhen extension work and public educa-tion has been thorough.

Although the figures for this study werehigher than the 29 % reported in a similarstudy in Kenya23 it is of concern that lessthan 70 % of dogs had been reported asvaccinated previously, as it has beenshown that 70 % vaccination coveragewould prevent outbreaks of the disease96.5 % of the time10. Even more concern-ing is the proportion of dogs reported bytheir owners to have been vaccinatedwithin the previous 12 months (56 %). Inthe Emalahleni municipality a dog’s re-ported vaccination status was signifi-cantly associated with the technicianresponsible for the area in which the dogresided, indicating differing campaignefficacies between AHTs. Furthermore,there was no significant relationship be-tween a dog’s vaccination status and thedog’s condition, sex or use in hunting orherding. It thus seems that a dog is vacci-nated ultimately because an animalhealth technician was present within hisstreet or village during a campaign, andnot whether the dog was important to theowner or not, or if the dog received ade-quate care. In other words, it is unlikelythat a dog owner will travel with his/herdogs to the local state veterinary office fora free rabies vaccination, irrespective ofthe function or importance of the dog tothe owner, or the care level that the dogreceives.

The immune mechanisms throughwhich an animal is protected against chal-lenge from rabies virus still needs elucida-tion, but is likely to involve both cellularand humoral immune responses14. Never-theless, the correlation between highlevels of neutralising antibody and pro-tection against challenge is high, as hasbeen reported in an extensive review onthe subject3. The World Health Organisa-tion has subsequently recommended that0.5 IU/ be regarded as the minimum neu-tralising antibody titre for a dog to be con-sidered adequately protected against thedisease42.

The annual vaccination campaigns inEmalahleni resulted in only 32 % of dogssampled in this study having adequateprotective antibodies, which, interest-ingly, is the same proportion found indogs in a similar study in urban Tunisia37.Rabies-neutralising antibody titres tendto wane rapidly after primary vaccina-tion3. Furthermore, the vaccine type andmanufacturer can also have an influenceon the peak antibody levels reached andthe longevity of the antibody response. In

this regard a study that compared 2different commercial vaccines in labora-tory dogs it was found that after a singlevaccination the percentages of dogs withtitres above 0.5 IU/ after 1 week were93 % and 100 %, and after 1 month 93 %and 67 %, respectively29. Even thoughthere is large individual variation in theneutralising antibody responses, laboratorydogs do show better antibody responsesthan pet dogs and stray dogs3. In thisstudy, 43 % of dogs whose owners re-ported previous vaccination had anti-body titres of 0.5 IU/ or above. Thisresponse is lower than the percentage inother communities: a study in Spainfound it to be 58 %12; in urban Bolivia itwas also found to be 58 %39; in rural andperi-urban Kenya it was 48 %23 and inurban Botswana 52 %36. Considering thatthe owner’s responses regarding thevaccination status of their dogs could notbe verified, it becomes difficult to evaluatethe low proportion of dogs with adequateneutralising antibody titres in Emalahleni,but it may, amongst others, be attribut-able to: low vaccination coverage byAHTs in the preceding rabies campaign, arapid drop in antibody titre after beingvaccinated in the preceding campaignand/or incorrect handling and adminis-tration of vaccines during campaigns. Asthis study found that the magnitude ofseroconversion in Emalahleni was inde-pendent of body condition and age theycould be eliminated as contributingfactors to such a low level of animalsprotected in the sample. Titres of 0.5 IU/and above were associated with increas-ing age, however, which could indicatethat the older dogs received more than1 vaccination and therefore maintainedantibody levels above the threshold levellonger than younger dogs. As the extentto which maternally derived rabies anti-bodies interferes with vaccination in pupsis controversial, we concur with otherauthors23,37 that vaccination of dogs youn-ger than 3 months of age should beconsidered to increase the proportion ofdogs immunised and to increase theproportion of dogs that will receive 2ndand 3rd booster vaccinations. Impor-tantly, however, owner education shouldbe thorough so as to prevent owners fromassociating the usually high neonatal andjuvenile morbidity and mortality rate inthese rural settings23 with rabies vaccina-tion.

Similar to a previous finding23, 11 % ofdogs had rabies-neutralising antibodytitres of ≥0.5 IU/ despite owners report-ing no previous vaccination. As all thesedogs were older than 1 year, this cannotbe ascribed to maternally derived anti-bodies. This finding could thus either be

due to owner ignorance of the vaccina-tion status of his/her dog, a possible falsepositive reaction of the FAVN test, oreven exposure to lyssavirusses in nature.Indeed, there are reported cases of rabiesantibody titres being detected in previ-ously unvaccinated but clinically healthyanimals where there was a long incuba-tion period after infection, where animalswere clinically unaffected carriers ofrabies virus or had recovered from clinicalinfection1,4,13–16. If clinically healthy dogswere able to maintain and spread theinfection, this might play a significant rolein the epidemiology of rabies in the area.

Comparison of antibody titres 30–60days post-vaccination in this study (n =80) showed that the majority (83 %) ofdogs seroconverted adequately. In a simi-lar study in urban Tunisia3 7, sero-conversion varied between 72 % 1 monthafter vaccination to 36 % 12 months aftervaccination. On the other hand, theproportion of dogs that seroconverted tolevels of ≥ 0.5 IU/ after vaccination in thisstudy is slightly less than the 85 % and>90 % of dogs responding adequately to1 or more than 1 rabies vaccination,respectively, in a study of sera from petdogs vaccinated in veterinary practices inFrance8. This could indicate that the dogsin Emalahleni do not seroconvert asefficiently as dogs that have better accessto veterinary care. The geometric meantitre 30–60 days post-vaccination withinthis study (1.44 IU/ ) compares favourablywith other studies29,37,39. However, asthe number of vaccinations a dog hadreceived before the study could not beverified, further interpretation of thisresult was not attempted.

Vaccination campaigns tend to getcoverage in the major and minor media inthe area, and considerable effort is madeto design and distribute informationabout the disease in the province. Fur-thermore, the vaccination campaign islaunched at major community meetingsacross the province every year. Accordingto this evaluation of technician efficacy abreakdown in the campaign occurs after-wards, at field level, with the delivery ofvaccine within communities. This is couldbe attributed to difficulty in supervisingthe process and lack of motivation ofAHTs to complete the campaign in veryoften demanding environments, as wellas a relatively young dog population.Ideally dogs in the area should be vacci-nated twice a year, as this could alsotheoretically lower the proportion of dogsthat need to be vaccinated in order toblock the spread of disease22. Unfortu-nately, however, the Eastern Cape rabieseradication campaign competes withother important disease control pro-

0038-2809 Jl S.Afr.vet.Ass. (2010) 81(4): 207–215 213

grammes like anthrax, sheep scab, bovinebrucellosis and tuberculosis, Newcastledisease and other emergency disease out-breaks (for example the recent classicalswine fever and Rift Valley fever out-breaks). Consequently, twice a year vacci-nation would be impractical andunsustainable with the current number ofAHTs available. There is thus an urgentneed to find a balance between the ‘goldstandard’ in rabies control and what is re-peatedly practical in Emalahleni.

In response to this specific challenge,the 2009 rabies campaign in Emalahleniwas approached by pooling AHTs in anarea rather than scattering their man-power throughout the district. All theAHTs were thus concentrated in a villagefor a shorter time than would be the casewith individual efforts before moving onto the next village/area. This approachhad several advantages: greater visibilityof the campaign, improved supervisionduring the campaign and an increase inproportion of animals vaccinated to 72 %(based on our canine population esti-mate). Disadvantages included greatertransport costs (as AHTs now had to travelto areas other than those for which theyhad been responsible), and a longer pe-riod to conclude the campaign in the mu-nicipality (4 months instead of 3 months).Although certain models do predict 70 %vaccination coverage to be adequate inblocking the spread of disease10,22, theyoften assume a that a single injection willconfer lifelong immunity, and that eachvaccinated animal has seroconvertedsuccessfully. As we have seen in thisstudy, 83 % of animals vaccinated duringthe study seroconverted successfully af-ter 30–60 days. Thus, models should alsobe developed to test whether a 70 % vacci-nated proportion would be adequate inblocking the spread of infection given thesuccessful seroconversion proportionabove, also taking into account the num-ber of dogs receiving 2 or more vaccina-tions.

CONCLUSIONRabies is an emerging public health

problem in the Eastern Cape Province,notwithstanding the fact that every year avaccination campaign is launched to vac-cinate as many dogs and cats as possible.This study highlights some of the areaswhere this worrying trend might origi-nate. Problems identified during previ-ous dog ecology studies in Africa wereconfirmed in Emalahleni, and we canassume that these trends will continueowing to various socio-economic factorsas listed previously40. However, in SouthAfrica, we are fortunate to have adequategovernment funding in terms of rabies

disease surveillance and provision ofvaccines. It is thus important that duringplanning of pro-active rabies vaccina-tion campaigns, work and supervision atfield level is not neglected. This can beachieved by active involvement of super-visors during the campaign who willexperience first-hand the the difficultiesmet during vaccine administration atground level, and attending to staff short-ages of AHTs and state veterinarians.

This study ’s preliminary findingsprompt similar investigation into otherrural and urban municipalities in theEastern Cape and in the rest of SouthAfrica, as findings about dog ecology andvaccination efficacy may differ regionally.Importantly, more studies should focuson the evaluation of different rabiescontrol and vaccination strategies, andwhy current methods are failing in SouthAfrica. Rabies vaccination strategies in theface of an outbreak also deserve specialattention.

ACKNOWLEDGEMENTSThe study was partially funded by the

OIE Rabies Reference Laboratory costcentre 04/03/P001. Vehicles and supplieswere provided by the Eastern CapeDepartment of Agriculture, Forestry andFisheries. The authors thank the follow-ing persons: Ms Almir Karstens and MsKatherina Fourie for their assistance at theQueenstown Veterinary Laboratory, DrCebisa Mnqeta for his support andinterest and Mr Simon Mandla, whoseexcellent dog (and people) handling skillsduring field work made this project a loteasier. Enkosi!

REFERENCES1. Aghomo H O, Oduye O O, Tomori O, Ibe M

1989 Isolation of rabies virus from clinicallyhealthy and previously unvaccinated dogs.Bulletin of Animal Health and Production inAfrica 37: 131–135

2. Akol G, Lwanga-Iga I, Amaral L A,Kroll-Lwanga-Iga S 1993 Rabies control inTranskei South Africa. Proceedings of theSouthern and Eastern African Rabies GroupInternational Symposium, Pietermaritzburg,South Africa, 29–30 April 1993: 81–91

3. Aubert T F A 1992 Practical significance ofrabies antibodies in cats and dogs. Revuescientifique et technique, Office Internationaldes Êpizooties 11: 735–760

4. Bigler W J, McLean R G, Trevino H A 1973Epizootiologic aspects of raccoon rabies inFlorida. American Journal of Epidemiology 98:326–335

5. Blumberg L, Leman P, Paweska J, Weyer J2010 Human rabies in South Africa, 2009.Communicable Diseases Surveillance Bulletinof the National Institute of CommunicableDiseases 8: 16–17

6. Brown K In press. Rabid epidemiologies:the emergence and resurgence of rabies intwentieth century South Africa. Journal ofthe History of Biology. DOI 10.1007/s10739-010-9241-9

7. Cliquet F, Aubert M, Sagné L 1998 Develop-ment of a fluorescent antibody virus neutra-lisation test (FAVN test) for the quantitationof rabies-neutralising antibody. Journal ofImmunological Methods 212: 79–87

8. Cliquet F, Verdier Y, Sagné L, Aubert M,Schereffer J L, Selve M, Wasniewski M,Servat A 2003 Neutralising antibody titra-tion in 25 000 sera of dogs and cats vaccinatedagainst rabies in France, in the frameworkof the new regulations that offer an alterna-tive to quarantine. Revue scientifique et tech-nique, Office International des Êpizooties 22:857–866

9. Cohen C, Sartorius B, Sabeta C, Zulu G,Paweska J, Mogoswane M, Sutton C, Nel LH, Swanepoel R, Leman P A, Grobbelaar AA, Dyason E, Blumberg L 2007 Epidemiol-ogy and molecular virus characterization ofreemerging rabies, South Africa. EmergingInfectious Diseases 13: 1879–1886

10. Coleman P G , Dye C 1996 Immunizationcoverage required to prevent outbreaks ofdog rabies. Vaccine 14: 185–186

11. De Balogh K K, Wandeler A I, Meslin F X1993 A dog ecology study in an urban and asemi-rural area of Zambia. OnderstepoortJournal of Veterinary Research 60: 437–443

12. Delgado S, Cármenes P 1997 Immuneresponse following a vaccination campaignagainst rabies in dogs from northwesternSpain. Preventive Veterinary Medicine 31:267–261

13. Fekadu M 1991 Latency and aborted rabies.In Baer G M (eds) The natural history of rabies.CRC Press, Boca Raton, Florida: 191–198

14. Fekadu M 1993 Canine rabies. OnderstepoortJournal of Veterinary Research 60: 421–427

15. Fekadu M 1988 Pathogenesis of rabies virusinfection in dogs. Reviews of InfectiousDiseases 10 Suppl 4: S678–683

16. Fekadu M 1972 Atypical rabies in dogs inEthiopia. Ethiopian Medical Journal 10: 79–86

17. Gallant J 2002 The story of the African dog.University of Natal Press, Pietermaritzburg

18. Gummow B, Roefs Y A A, de Klerk G 2010Rabies in South Africa between 1993 and2005 – what has been achieved? Journal ofthe South African Veterinary Association 81:16–21

19. International Committee on Taxonomy ofViruses 2010 ICTV master species list 2009version 5. Online at: http://talk.ictvonline.org/files/ictv_documents/m/msl/1231.aspx(accessed 5 August 2010)

20. Kaare M, Lembo T, Hampson K, Ernest E,Estes A, Mentzel C, Cleaveland S 2009Rabies control in rural Africa: evaluatingstrategies for effective domestic dog vacci-nation. Vaccine 27: 152–160

21. Kärber G 1931 Beitrag zur kollektivenBehandlung pharmakologischer Reihen-versuche. Naunyn-Schmiedebergs Archiv fürExperimentelle Pathologie und Pharmakologie162: 480–483

22. Kitala P M, McDermott J J, Coleman P G,Dye C 2002 Comparison of vaccinationstrategies for the control of dog rabies inMachakos District, Kenya. Epidemiology andInfection 129: 215–222

23. Kitala P M, McDermott J J, Kyule M,Gathuma J, Perry B, Wandeler A 2001 Dogecology and demography information tosupport the planning of rabies control inMachakos District, Kenya. Acta Tropica 78:217–230

24. Knobel D L, Cleaveland S, Coleman P G,Fèvre E M, Meltzer M I, Miranda M E G,Shaw A, Zinsstag J , Mesl in F 2005

214 0038-2809 Tydskr.S.Afr.vet.Ver. (2010) 81(4): 207–215

Re-evaluating the burden of rabies in Africaand Asia. Bulletin of the World Health Organi-zation 83: 360–368

25. Mansvelt P R 1962 Rabies in South Africa.Field control of the disease. Journal of theSouth African Veterinary Medical Association33: 313–319

26. Mansvelt P R 1956 Rabies in the northernTransvaal (1950) outbreak. Journal of theSouth African Veterinary Medical Association27: 167–178

27. McCrindle C M E, Gallant J, Cornelius S T,Schoeman H S 1999 Changing roles of dogsin urban African society: A south Africanperspective. Anthrozoos 12: 157–161

28. Meltzer M I , Rupprecht C E 1998 A review ofthe economics of the prevention and controlof rabies: Part 2: Rabies in dogs, livestock andwildlife. PharmacoEconomics 14: 481–498

29. Minke J M, Bouvet J Cliquet F, WasniewskiM, Guiot A L, Lemaitre L, Cariou C, CozetteV, Vergne L, Guigal PM 2009 Comparison ofantibody responses after vaccination withtwo inactivated rabies vaccines. VeterinaryMicrobiology 133: 283–286

30. Minnaar W N, Krecek R C 2001 Veterinaryneeds of dogs in two resource limited com-munities in the Gauteng and North West

Provinces in South Africa. Journal of theSouth African Veterinary Association 72:209–213

31. Nel L H , Markotter W 2007 Lyssaviruses.Critical Reviews in Microbiology 33: 301–324

32. Nel L H, Sabeta C T, Von Teichman B, JafthaJ B, Rupprecht C E, Bingham J 2005 Mon-goose rabies in southern Africa: A re-evaluation based on molecular epidemiol-ogy. Virus Research 109: 165–173

33. Perry B D 1993 Dog ecology in eastern andsouthern Africa: implications for rabiescontrol. Onderstepoort Journal of VeterinaryResearch 60: 429–436

34. Rautenbach G H, Boomker J, De Villiers I L1991 A descriptive study of the caninepopulation in a rural town in SouthernAfrica. Journal of the South African VeterinaryAssociation 62: 158–162

35. Sabeta C T, Bingham J, Nel L H 2003Molecular epidemiology of canid rabies inZimbabwe and South Africa. Virus Research91: 203–211

36. Sebunya T K, Ndabambi N, Mpuchane S2007 A serosurvey of rabies antibodies indogs in Gaborone, Botswana. Journal ofAnimal and Veterinary Advances 6: 549–552

37. Seghaier C, Cliquet F, Hammami S, Aouina

T, Tlatli A, Aubert M 1999 Rabies mass vacci-nation campaigns in Tunisia: Are vacci-nated dogs correctly immunized? AmericanJournal of Tropical Medicine and Hygiene 61:879–884

38. Spearman C 1908 The method of ‘right andwrong’ cases (‘constant stimuli’) withoutGauss’s formulae. British Journal of Psychol-ogy 227–242

39. Suzuki K, González E T, Ascarrunz G, LozaA, Pérez M, Ruiz G, Rojas L, Mancilla K,Pereira J A C, Guzman J A, Pecoraro M R2008 Antibody response to an anti-rabiesvaccine in a dog population under fieldconditions in Bolivia. Zoonoses and PublicHealth 55: 414–420

40. Swanepoel R 2004 Rabies. In Coetzer J A W,Tustin R C (eds) Infectious diseases of livestock(2nd edn) Vol. 2. Oxford University PressSouthern Africa, Cape Town: 1123–1182

41. Thrusfield M 2007 Diagnostic testing. InThrusfield M (ed.) Veterinary epidemiology(3rd edn). Blackwell Science, Oxford:307–308

42. World Health Organization 2005 WHO ex-pert consultation on rabies (2004: Geneva,Switzerland). WHO Technical Report Series931: viii + 88

0038-2809 Jl S.Afr.vet.Ass. (2010) 81(4): 207–215 215