ORIGINAL ARTICLES

Borrelia burgdorferi Serosurvey in Wild Deerin England and Wales

Silvia Alonso,1 Francisco J. Marquez,2 and Laia Solano-Gallego3

Abstract

Lyme disease is the most common vector-borne disease in the United Kingdom and its incidence has beenincreasing in recent years. However, limited information is available on its epidemiology and dynamics in the U.K.A survey in wild deer to investigate the presence of antibodies reactive to Borrellia burgdorferi was conducted toobtain initial information on the distribution pattern of the spirochete in England and Wales. Samples from roedeer (n = 604) and red deer (n = 80) were collected in eight different locations. An ELISA protocol was developed toidentify antibodies reactive to B. burgdorferi s.l. Seropositivity was investigated by location of sampling, over time,and in relation to the level of deer tick infestation. Twenty-three percent of animals had a positive serology.Seroprevalence varied according to location with the southern forestry districts showing higher seroprevalencerates. One northern location showed an unexpectedly high proportion of positive deer. Variations in the pro-portion of positive animals were also observed over time. Tick load was higher from spring through autumn, andits relation to seroprevalence was compatible with higher tick infectivity during the spring and summer months.This study represents the first assessment of distribution of Borrelia antibodies in deer in the U.K. and identifiesareas that are potential hot spots for human Lyme borreliosis. Targeted epidemiological studies should be con-ducted to evaluate the actual disease risk for humans.

Key Words: Borrelia burgdorferi—Antibody testing—England—Wales—Wild deer.

Introduction

Lyme disease is currently the most common tick-bornedisease in humans in Europe (Smith et al. 2000; Smith

and Takkinen 2006), with trends of increasing incidence re-ported in recent years in almost all countries (Smith et al.2000). In the United Kingdom (U.K.), although the number ofcases has remained historically low, its incidence has experi-enced a fivefold increase in 10 years, and more than 1000locally-acquired cases were serologically confirmed in 2008(Department for Environment Food and Rural Affairs 2008;Health Protection Scotland 2010). There is no reason to believethat this increasing trend will reverse in the near future. Thegreater public access to the countryside, where reservoirs ofthe pathogen are present, is suspected to be one of the mostimportant risk factors, together with a greater awareness ofthe disease and increased surveillance (Department for En-vironment Food and Rural Affairs, 2006; Gilbert 2010).However, due to the relatively recent emergence of this dis-

ease, limited information is available about its epidemiologyand dynamics in the U.K.

In western Europe, Borrelia burgdorferi sensu latu infection istransmitted by ticks of the species Ixodes ricinus. Several ani-mal species are reservoirs of the pathogen, making its epide-miology and transmission cycle complex (Barbour 1998; DeMeneghi 2006; Linard et al. 2007; Jaenson et al. 2009). Borreliainfection causes disease in humans and in certain animalssuch as ruminants, horses, and dogs (Williams et al. 2002).Various Borrelia burgdorferi genospecies are known to causeLyme disease, five of which have so far been identified aspathogenic in Europe ( Jaenson et al. 2009) and three others,although they have been found in human cases, are of un-known pathogenicity at present (European Concerted Actionon Lyme Borrelliosis). Ixodes ricinus ticks are commonly foundin the British Isles (Pietzsch et al. 2005; Scharlemann et al.2008; Jameson and Medlock 2011). Small mammals and birdsare the most important reservoir for B. burgdorferi infection.However, other vertebrates such as deer, although

1Veterinary Epidemiology and Public Health Group, and 3Department of Pathology and Infectious Diseases, Royal Veterinary College,Hertfordshire, U.K.

2Biologıa Animal, Biologıa Vegetal y Ecologıa. Universidad de Jaen. Campus Las Lagunillas, Jaen, Spain.

VECTOR-BORNE AND ZOONOTIC DISEASESVolume 12, Number 6, 2012ª Mary Ann Liebert, Inc.DOI: 10.1089/vbz.2011.0635

448

incompetent reservoirs for the spirochete, act as feedingsource for ticks, potentially contributing to the disseminationof this infection (Simpson 2002; Williams et al. 2002). A rap-idly increasing wild deer population in England, estimated tobe 320,000 animals in 2004 (Alastair Ward, personal com-munication) is often mentioned as a major factor in increasedtick abundance (Bohm et al. 2007).

Previous studies conducted in various European countrieshave found a high proportion of wild deer with specificantibodies against B. burgdorferi antigen, demonstrating thatalthough unable to infect vectors, they are usually in contactwith infected ticks (Martinez et al. 1999; Travnicek et al.2003; Skarphedinsson et al. 2005). Therefore, it has beensuggested that deer could act as sentinels for the presence ofthe pathogen in a given area (Skarphedinsson et al. 2005). Todate, partly due to the limited interest in the disease and thehistorically low number of human cases, there are no sys-tematic studies on the distribution of antibodies against B.burgdorferi in deer in England and Wales (Lovett et al. 2008).Therefore, the aims of this study were to conduct a generalassessment of the pattern of B. burgdorferi presence in variouslocations in England and Wales using antibodies in deer as aproxy, to evaluate associated risk factors for higher ser-oprevalence, and to contribute to our understanding of thisinfection in the U.K.

Materials and Methods

Samples and data collection

Blood samples were collected from wild red deer (Cervuselaphus) and roe deer (Capreolus capreolus) at the time ofculling in eight different locations across England and Walesfrom April 2009 to March 2010. Seven forestry commissions(FC) in England and one in Wales participated in the study,representing diverse eco-geographical areas (Fig. 1). A fewadditional samples (n = 12) were received from non-FCstalkers. Convenience sampling was performed by in-structed rangers. Blood samples were collected in plain tubesand were kept refrigerated ( + 4�C) until postage to the lab-oratory. Rangers recorded the degree of tick infestationof each deer sampled based on three categories: none (novisible ticks in the animal body), low (only few ticks on thelegs), and high (ticks present on both the lower and the upperparts of the animal’s body). Up to 5 ticks were collected fromeach sampled deer infested with ticks. Collection of ticks wasnot systematic (i.e., not proportional to the tick infestationlevel or representative of the range of ticks on the animal).The ticks were morphologically identified as previouslydescribed (Hillyard 1996).

Information on the characteristics of each of the sampleddeer was obtained from the centralized database of the for-estry commission, where information on each culled deer inthe U.K. is kept. This included the animal identificationnumber (National TAG number), the date and location ofculling (FC name and number), the species, gross weight, andapproximate age ( < 1 year, 1–2 years, and ‡ 3 years), as well asthe National Grid Reference coordinates for the point wherethe animal was culled.

On arrival to the laboratory, the samples were centrifugedand sera was separated and stored at - 20�C until analysis.

Sampling frequency varied over the year, as did the relativeproportion of deer species and the sex of sampled animals.

These variations are determined by the structure of the Na-tional culling program, with roe deer males being culled fromApril to October each year, red deer males culled from Augustto April each year, and females of both species being culledfrom November until the beginning of the calving season atthe end of March.

Antibody testing

An ELISA protocol for antibodies against Leishmania spp.(Riera et al. 1999) was adapted to detect B. burgdorferi anti-bodies in sera from deer, taking into consideration ELISAprotocols previously described by other authors (Magnarelliet al. 1993; Bhide et al. 2004). Briefly, microtiter plates werecoated with 0.1 mL of B. burgdorferi (30-AL40; FitzgeraldIndustries International, Concord, MA) whole antigen so-lution (10 lg mL - 1 in 0.05 M carbonate bicarbonate, pH 9.6),and incubated overnight at 4�C. One hundred microliters perwell of deer sera, diluted 1:100 in PBS + 0.05% Tween 20(PBST) + 1% dried skimmed milk (PBST-M), was incubatedfor 1 h at 37�C. After three washes with 0.2 mL PBST and onewash with 0.2 mL PBS, 100 lL per well of rabbit anti-deer IgGconjugated to horseradish peroxidase and diluted 1:1000 inPBST-M was added and incubated for 1 h at 37�C. The wa-shes were then repeated. The substrate solution, orthophe-nylenediamine dichloride (0.5 mg mL - 1; Pierce, Rockford,IL), was added at 100 lL per well and developed for 20 minat room temperature. The reaction was stopped with 50 lL of2 M H2SO4. Absorbance values were read at 490 nm in anautomatic ELISA reader (Biotek EL808 9 Channel absor-bance reader; BioTek Instruments Inc., Winooski, VT).B. burgdorferi-positive (n = 10) and B. burgdorferi-negative seracontrols (n = 40) from wild white-tailed deer as determined byimmunofluorescence antibody assay were obtained from theConnecticut Agricultural Experiment Station, New Haven.All reagents were tested with the positive and negative con-trols to optimize working conditions and ensure standardi-zation. The diagnostic accuracy of the developed ELISA wasevaluated using receiver operating characteristic curve (ROC)analysis. This allows assessment of the discrimination per-formance of a test, or the capacity to correctly identify positiveand negative samples (Gardner and Greiner 2006; Tripepiet al. 2009). ROC curves were also used to identify the cut-offfor the positive value of the ELISA test (the value that de-termines a positive result) by ‘‘identifying the optical density(OD) value that maximizes the Sensitivity (Se) and Specificity(Sp) of the test’’ (Gardner and Greiner 2006; Tripepi et al.2009). ROC curves, including Se and Sp of the selected cut-offfor positive value, were calculated using SPSS statisticalsoftware (PAWS statistics 18; SPSS Inc., IBM, Chicago, IL).Other measures of ELISA performance, such as within-assayand inter-assay precisions, were calculated with coefficientsof variation (CV) expressed as a percentage (mean/SD;Stockham and Scott 2008). The within-assay CV was calcu-lated by three replicates of eight positive and four negativecontrols within the same ELISA plate. The inter-assay CV wascalculated from the values observed for four positive controlson eight different days. The ELISA was considered of goodprecision if the intra- and inter-assay CV were below 10% and20%, respectively. The same positive and negative controlswere always included in each plate. To minimize variability,all conditions were kept as constant as possible on the various

Borrelia burgdorferi SEROSURVEY IN WILD DEER 449

FIG. 1. Map showing sampling locations (England and Wales) and total numbers of samples collected.

450 ALONSO ET AL.

days of testing. ELISA plates were prepared and stored at 4�Cand used within 5 days of preparation. Commercial antigenfrom the same batch was used in all experiments.

Data analysis

STATA software (STATA 9.0; StataCorp LP, CollegeStation, TX) was used to analyze the data. The spatial andtemporal relations between seropositivity and the level oftick infestation were described. The proportion of seroposi-tive animals (seroprevalence) was calculated for each of theeight locations in the study and compared to the level oftick infestation in the area, estimated as the proportion ofanimals recorded as carrying ticks. For those sampling siteswhich provided more than 100 samples (3 locations), therelation between seroprevalence and tick infestation wasexplored over time. For this purpose, time was split into fourcategories (winter: December to February; spring: March toMay; summer: June to August; autumn: September to No-vember). Moreover, in each of those 3 locations, the influenceof season on seropositivity was further assessed (Pearson’schi-square test). At the individual animal level, the asso-ciation between specific animal characteristics (includingtick infestation, time of sampling, species, age, and sex), andthe serological findings was also assessed. This was doneusing conditional logistic regression in order to account forthe degree of clustering of the samples in the eight specificlocations.

Results

Deer sampling

A total of 684 blood samples from deer were collected over12 months. The sampling rate was not homogenous acrosslocations, and ranged between a minimum of 20 to a maxi-mum of 139 samples (Fig. 1). Twelve samples were receivedfrom stalkers outside the FCs, and were combined with the FCsamples for the purposes of analysis. Thirty-two samples didnot report the origin and were therefore excluded from theanalysis. The characteristics of deer sampling are listed inTable 1. Red deer were sampled in only three locations, due tothe more restricted distribution of this species in the country,and represented no more than 30% of the total samples foreach area. Fifty-seven percent of the deer were infested withticks, and the majority of those were reported as having a lowinfestation level. All ticks (n = 2207) were identified as Ixodesricinus, with female ticks being three times more frequentlycollected than males. Co-infestation with insects of the orderDiptera (n = 16) was observed in 10 deer.

Antibody testing

The results of the ROC analysis demonstrate the robustaccuracy of the developed test (area under the curve = 0.975;85% CI 0.932,1.00). Sera with an OD on the ELISA higher than0.55 were considered positive. This cut-off for the positivevalue corresponded to 90% sensitivity (95% CI 0.54,0.99) and93% specificity (95% CI 0.79,0.98). The reproducibility of thetest was good, with within-assay and inter-assay coefficientsof variation of 8% and 13.7%, respectively.

Overall, 156 blood samples (23%) were found to be positivefor B. burgdorferi s.l. Seroprevalence and tick presence by lo-cation are summarized in Figure 2. The proportion of sero-

positive animals varied among locations in the study, as didthe proportion of animals presenting with ticks. East Angliashowed the highest seroprevalence, followed by the rest of thesouthern locations. In the northern areas, seroprevalencediffered greatly across locations. Thirty-seven percent (37%)of animals in the North West District had a positive serology,while the Kielder District produced very few positive animals(1.5%). The observed serological pattern followed, to a greatextent, the degree of tick infestation, with higher ser-oprevalence observed in areas with higher tick infestations.This relationship was very clear in Kielder District, wherevery few animals presented with ticks (21%), and only twodeer had a positive serology. It was also very conspicuous inEast Anglia, where high infestation rates were coupled withhigh proportions of seropositive animals. The Lake (Wales)and Kielder Districts were the areas with lower numbers ofinfested deer. Indeed, none of the deer sampled in Wales(n = 30) were infested with ticks, and none had antibodiesagainst B. burgdorferi.

The seroprevalence and tick infestation over time are de-scribed for three locations in Figure 3. In general, the pro-portion of seropositive animals varied over the year, withmore positive animals observed in spring (March to May).When exploring this observation further, we found a strongassociation between the season and the outcome of serology( p < 0.01) in two locations (East Anglia and Peninsula), whileno association was found in the New Forest District ( p = 0.12).The proportion of deer infected with ticks increased fromApril onwards, which reveals a long-lasting presence of ticksin forest areas, with decreased activity only in the wintermonths.

At an individual level, animals infested with ticks weremore likely to be seroreactive to B. burgdorferi ( p < 0.0001).However, when accounting for the effect of time and location,the presence of ticks on an animal was not associated with apositive serology ( p = 0.105, OR = 1.47, 95% CI 0.92,2.33). Apositive serology was found to be more frequent in males( p < 0.001), with 32% of males being seropositive compared to

Table 1. Characteristics of the Sampled Deer

Variable Number of deer (%)

SpeciesRoe deer 80 (13.2)Red deer 524 (86.8)Totala 604

SexMales 296 (49)Females 307 (51)Total 603

Age< 1 year 99 (32.5)1–2 years 133 (43.5)‡ 3 years 73 (24)Total 305

TicksNone 276 (42.9)Low 344 (51.9)High 33 (5.2)Total 643

aNumber of samples with information on that variable (out of atotal of 652 samples).

Borrelia burgdorferi SEROSURVEY IN WILD DEER 451

18% of females (OR = 2.2, 95%CI 1.4,3.5, p = 0.001), even whenthe effect of time was taken into account in the model. Roedeer were more likely to be seropositive compared to red deer(OR = 2.95, 95% CI 1.4,3.2; p = 0.004). Age was not found to beassociated with positive serology.

Discussion

This study presents the first available data on the dis-tribution of antibodies against B. burgdorferi s.l. in wild deerin the U.K. A relatively low proportion (23%) was foundwhen compared to reports from other European countries.

In Slovakia, seroprevalence rates of up to 45% were foundin various studies (Travnicek et al. 2003; Bhide et al. 2004),a finding in agreement with the theory that Borrelia spe-cies are more predominant in the central areas of Europe(Smith and Takkinen 2006). Relatively lower deer ser-oprevalence rates, ranging from 28% to 39%, were alsofound in different parks in Ireland (Gray et al. 1996).Significant differences in seroprevalence were found inDenmark across the country, ranging from 27% to 47%(Skarphedinsson et al. 2005). Our results also showedgreat diversity in seroprevalence across the various loca-tions studied. The proportion of animals presenting with

FIG. 3. Seroprevalence and tick infestation over time.Shown is the seroprevalence in wild deer over time in threedifferent locations. The dashed line indicates the proportionof tick-infested animals. (A) East Anglia. (B) Peninsula. (C)New Forest (Winter: January–March; Spring: April–June;Summer: July–September; Autumn: October–December).

FIG. 2. Seroprevalence and tick infestation by location. Chart showing seroprevalence in wild deer (95% CI) by location,and the proportion of animals infested with ticks.

452 ALONSO ET AL.

I. ricinus ticks was also variable, probably reflecting a het-erogeneous distribution of ticks across the country (Pietzschet al. 2005). Consistently higher seroprevalence rates werefound in the southern regions, which are the areas wherethe majority of reported human Lyme disease cases origi-nate (Department for Environment Food and Rural Affairs,2008).

There is still no agreement on the actual effect of deerdensity on the spread of Borrelia infection. Though theiramplifying effect on transmission has been described in en-demic countries (Hinrichsen et al. 2001), some authors sug-gest that as dead-end hosts for Borrelia, deer could reduce theburden of infection on vectors, and therefore reduce thetransmission rate (LoGiudice et al. 2003; Ostfeld et al. 2006;Pugliese and Rosa 2008). It would be valuable to evaluatehow our findings compare to the disease incidence in hu-mans in each of the studied locations. Our results show thatthe degree of pathogen exposure in wild deer decreases to-wards the north of the country. In accord with this, the ser-oprevalence in the Kielder District was very low, and onlyvery few animals presented with ticks at the time of culling.This confirms the findings of extensive field studies con-ducted in that location, where ticks positive for Borrelia werenot found (Richard Birtles, personal communication).Nevertheless, in our study North West England produced arelatively large number of positive deer. This suggests thatlatitude may only explain part of the variability seen in thespatial distribution of Borrelia in the U.K. This was supportedby Gilbert (Gilbert 2010), who highlighted that specific mi-croclimates and landscape characteristics are factors that cansupport the presence of I. ricinus in unexpected latitudes.However, our findings in northwest England should be in-terpreted with caution, due to the limited number of samplesobtained in that location. Further testing should be con-ducted to confirm whether North West England may be anorthern focus for borreliosis.

The higher Borrelia seroprevalence rates found fromMarch to May in the present study corresponds with themonths of increased tick activity (Randolph et al. 2000;Pietzsch et al. 2005). Similar findings were reported instudies in the United States (Magnarelli et al. 1995), withseroprevalence rates decreasing over the winter until April.Nevertheless, a small number of samples was collectedfrom June to August, and this limits our capacity to makean accurate interpretation of the seroprevalence trend overthe summer months. We observed a direct relationshipbetween tick infestation and deer seropositivity over thesummer months, but the strength of that relationship de-creased over the following months (i.e., the tick load washigh, but fewer deer had positive serology). This is notsurprising, given that spring and summer are known to bethe seasons of highest transmissibility of infection by ticks(Department for Environment Food and Rural Affairs,2008). It is also possible that antibodies in deer decay overtime, and deer infected during the spring and summer maynot show antibodies by autumn. There is limited informa-tion in the scientific literature on the duration of the im-mune response to Borrelia in deer. Luttrell and associates(Luttrell et al. 1994) detected antibodies by ELISA in fouranimals 2–3 weeks after experimental inoculation with B.burgdorferi. Antibodies were still present 10 weeks post-infection, although two animals started to show antibody

decay by that time. These results would be compatible withincreased pathogen transmission from ticks in the springand summer months, with a subsequent decay of anti-bodies in infected deer. However, seroreversion (frompositive to negative status in two consecutive tests) wasstudied by Magnarelli and colleagues in recapture studies,and they found that only a few animals (4/31) had theirserological status reversed in a period of 10 months(Magnarelli et al. 1995). The seroprevalence trend observedover time in our study is compatible with a decay of anti-bodies gradually after exposure; if the antibody response indeer lasted longer, we would have observed constant ser-oprevalence levels throughout the year and across areas inour study.

Unsurprisingly, the majority of collected ticks were I. rici-nus (Pietzsch et al. 2005; Jameson and Medlock 2011). Itshould be emphasized that our study did not aim to under-take a survey of tick infestation in deer. Our identification ofticks is likely to be biased towards the most prevalent tickspecies in the country (Pietzsch et al. 2005; Jameson andMedlock 2011), and we cannot exclude the presence of othertick species in the animals. A more complete survey wouldrequire systematic examination of animals and collection of alltick species present on the animal, but this was outside thescope of the present study.

Surprisingly, gender and species were found to be factorsassociated with positive serology, with males and roe deershowing highest seroprevalence rates. However, due to thestructure of the culling program in the U.K., females wereonly culled in the winter months (from November to March),and as a result only samples from males were available for themonths when highest seroprevalence would be expected. Thestructure of the sampling over time may have further biasedour results, since in those months only roe deer were sampled.Gender was not found to be a risk factor in other studies fromEurope and the U.S. (Travnicek et al. 2003; Skarphedinssonet al. 2005; Murdock et al. 2009).

This study aimed to provide the spatial distribution ofwild deer seropositivity for B. burgdorferi s.l., which couldserve as an indication of the presence of the pathogen in theareas where serological responses were detected. There isongoing debate about the reliability of deer serology as anindicator of the presence of Borrelia, due in part to the factthat antibodies do not persist in these animals for life, andalso due to doubtful sensitivity and specificity of the avail-able diagnostic tests. Reported human cases, and in somecases active tick collection and testing, can be used as mea-sures of the presence of the pathogen. The sensitivity andspecificity of these methods is also questionable. Studies toevaluate the association between seropositivity in wild deerand infection status of ticks in the area could help elucidatethe extent to which both estimates are in agreement, andwould clarify the extent to which systematic testing of wilddeer represents a viable means of monitoring the pathogen’sspread.

In conclusion, evidence of antibodies against B. burgdorferiindicates that infection is prevalent in the deer population inEngland and Wales, with high variability seen across dif-ferent regions. This study identified areas that are potentialhot spots for human Lyme borreliosis, including southernEngland, where further studies should be carried out. Spatialanalysis of these results could provide information on the

Borrelia burgdorferi SEROSURVEY IN WILD DEER 453

factors that may determine the spread of the pathogen inthe U.K.

Acknowledgments

This project was funded by an internal grant from theRoyal Veterinary College. Sample collection was funded bythe British Deer Society. The authors would like to thankthe British Deer Society for their financial support for con-ducting the sampling, and the Deer Initiative for the coor-dination of the sampling. We thank also the ConnecticutAgricultural Experiment Station for providing control sera,and the University of Veterinary Medicine (Kosice, Slova-kia) for their initial support. We also thank LaurianneTavernier and Daphne Boulicault who helped with theserological testing.

Author Disclosure Statement

No competing financial interests exist.

References

Barbour AG. Fall and rise of Lyme disease and other Ixodes tick-borne infections in North America and Europe. Br Med Bull1998; 54:647–658.

Bhide MR, Curlik J, Travnicek M, et al. Protein A/G dependentELISA a promising diagnostic tool in Lyme disease ser-oprevalence in game animals and hunting dogs. Comp Im-munol Microbiol Infect Dis 2004; 27:191–199.

Bohm M, White PC, Chambers J, et al. Wild deer as a source ofinfection for livestock and humans in the UK. Vet J 2007;174:260–276.

De Meneghi D. Wildlife, environment and (re)-emergingzoonoses, with special reference to sylvatic tick-borne zoo-noses in North-Western Italy. Ann Ist Super Sanita 2006;42:405–409.

Department for Environment Food and Rural Affairs. 2006.Zoonoses Report.

Department for Environment Food and Rural Affairs. 2008.Zoonoses Report.

European Concerted Action on Lyme Borreliosis (EUCALB;www.eucalb.com).

Gardner IA, Greiner M. Receiver-operating characteristic curvesand likelihood ratios: improvements over traditional methodsfor the evaluation and application of veterinary clinical pa-thology tests. Vet Clin Pathol 2006; 35:8–17.

Gilbert L. Altitudinal patterns of tick and host abundance: apotential role for climate change in regulating tick-borne dis-eases? Oecologia 2010; 162:217–225.

Gray JS, Hayden TJ, Casey S, et al. A Lyme disease serosurveyof deer in Irish National Parks. Biology and Environment-Proceedings of the Royal Irish Academy 1996; 96B:27–32.

Health Protection Scotland. Lyme disease, Annual Totals,www.hps.scot.nhs.uk; Last accessed 2012.

Hillyard P. Ticks of north-west Europe : keys and notes foridentification of the species. Published for The Linnean Societyof London and The Estuarine and Coastal Sciences Associa-tion by Field Studies Council Shrewsbury, 1996.

Hinrichsen VL, Whitworth UG, Breitschwerdt EB, et al. Asses-sing the association between the geographic distribution ofdeer ticks and seropositivity rates to various tick-transmitteddisease organisms in dogs. J Am Vet Med Assoc 2001; 218:1092–1097.

Jaenson TG, Eisen L, Comstedt P, et al. Risk indicators for thetick Ixodes ricinus and Borrelia burgdorferi sensu lato in Sweden.Med Vet Entomol 2009; 23:226–237.

Jameson LJ, Medlock JM. Tick surveillance in Great Britain.Vector Borne Zoonotic Dis 2011; 11:403–412.

Linard C, Lamarque P, Heyman P, et al. Determinants ofthe geographic distribution of Puumala virus and Lymeborreliosis infections in Belgium. Int J Health Geogr 2007;6:15.

LoGiudice K, Ostfeld RS, Schmidt KA, et al. The ecology ofinfectious disease: effects of host diversity and communitycomposition on Lyme disease risk. Proc Natl Acad Sci USA2003; 100:567–571.

Lovett JK, Evans PH, O’Connell S, et al. Neuroborreliosis in theSouth West of England. Epidemiol Infect 2008; 136:1707–1711.

Luttrell MP, Nakagaki K, Howerth EW, et al. Experimental in-fection of Borrelia burgdorferi in white-tailed deer. J Wildl Dis1994; 30:146–154.

Magnarelli LA, Anderson JF, Cartter ML. Geographic distribu-tion of white-tailed deer with ticks and antibodies to Borreliaburgdorferi in Connecticut. Yale J Biol Med 1993; 66:19–26.

Magnarelli LA, Denicola A, Stafford KC 3rd, et al. Borreliaburgdorferi in an urban environment: white-tailed deer withinfected ticks and antibodies. J Clin Microbiol 1995; 33:541–544.

Martinez A, Salinas A, Martinez F, et al. Serosurvey for selecteddisease agents in white-tailed deer from Mexico. J Wildl Dis1999; 35:799–803.

Murdock JH, Yabsley MJ, Little SE, et al. Distribution of anti-bodies reactive to Borrelia lonestari and Borrelia burgdorferi inwhite-tailed deer (Odocoileus virginianus) populations in theeastern United States. Vector Borne Zoonotic Dis 2009; 9:729–736.

Ostfeld RS, Canham CD, Oggenfuss K, et al. Climate, deer, ro-dents, and acorns as determinants of variation in Lyme-disease risk. PLoS Biol 2006; 4:e145.

Pietzsch ME, Medlock JM, Jones L, et al. Distribution of Ixodesricinus in the British Isles: investigation of historical records.Med Vet Entomol 2005; 19:306–314.

Pugliese A, Rosa R. Effect of host populations on the intensity ofticks and the prevalence of tick-borne pathogens: how to in-terpret the results of deer exclosure experiments. Parasitology2008; 135:1531–1544.

Randolph SE, Green RM, Peacey MF, et al. Seasonal synchrony:the key to tick-borne encephalitis foci identified by satellitedata. Parasitology 2000; 121(Pt 1):15–23.

Riera C, Valladares JE, Gallego M, et al. Serological and para-sitological follow-up in dogs experimentally infected withLeishmania infantum and treated with meglumine antimoniate.Vet Parasitol 1999; 84:33–47.

Scharlemann JP, Johnson PJ, Smith AA, et al. Trends in ixodidtick abundance and distribution in Great Britain. Med VetEntomol 2008; 22:238–247.

Simpson VR. Wild animals as reservoirs of infectious diseases inthe UK. Vet J 2002; 163:128–146.

Skarphedinsson S, Jensen PM, Kristiansen K. Survey of tick-borne infections in Denmark. Emerg Infect Dis 2005; 11:1055–1061.

Smith R, O’Connell S, Palmer S. Lyme disease surveillance inEngland and Wales, 1986–1998. Emerg Infect Dis 2000; 6:404–407.

Smith R, Takkinen J. Lyme borreliosis: Europe-wide coordinatedsurveillance and action needed? Euro Surveill 2006;11:E060621–E060622.

454 ALONSO ET AL.

Stockham SL, Scott MA. Introductory concepts. In: Fundamentals ofVeterinary Clinical Pathology. London: Wiley-Blackwell Oxford.

Travnicek M, Stefancikova A, Nadzamova D, et al. Im-munoglobulin G antibodies to Borrelia burgdorferi in gameanimals and small mammals in eastern Slovakia. Rev Sci Tech2003; 22:1035–1041.

Tripepi G, Jager KJ, Dekker FW, et al. Diagnostic methods 2:receiver operating characteristic (ROC) curves. Kidney Int2009; 76:252–256.

Williams LN, Sonenshine DE, Lane RS, et al. Ticks (Ixodida). NewYork: Academic Press, 2002.

Address correspondence to:Silvia Alonso

Veterinary Epidemiology and Public Health GroupRoyal Veterinary College

Hawkshead LaneHatfield

Hertfordshire, AL9 7TAUnited Kingdom

E-mail: salonso@rvc.ac.uk

Borrelia burgdorferi SEROSURVEY IN WILD DEER 455