1958 Scientific Reports: Retrospective Study JAVMA, Vol 224, No. 12, June 15, 2004

SMALLANIMALS

Following the introduction of a bivalent vaccine againstLeptospira interrogans serovars canicola and ictero-

haemorrhagiae in the early 1970s, leptospirosis in dogs inthe United States decreased.1-3 An increasing prevalence ofleptospirosis in dogs during the 1980s and 1990s hasbeen identified.4 Although this increasing prevalence hasbeen documented in specific regions; for example, themidwest region of the United States during the 1990s5;there have also been numerous reports1,3,5-13 of casesthroughout North America during this same period.

Reasons for this increase remain speculative; how-ever, infection of dogs by serovars other than ictero-haemorrhagiae and canicola are likely to be important.Beginning in approximately 1990, reports1,2,6,7,14,15 of lep-tospirosis caused by Leptospira interrogans serovarsbratislava and pomona and Leptospira kirschneri serovargrippotyphosa were published. Livestock (pigs andhorses) and wildlife (skunks and opossums) are thelikely reservoir for L interrogans serovars bratislava andpomona, and wildlife (voles, raccoons, skunks, andopossums) are the likely reservoir host for L kirschneriserovar grippotyphosa.1,2,8,14,16,17 Thus, increased oppor-tunity for contact between dogs and wildlife, livestock,or livestock waste may partially explain the increase inleptospirosis prevalence in dogs in the UnitedStates.1,2,6,14 Male dogs; dogs between 4 and 10 years ofage; and herding dogs, hounds, working dogs, andmixed-breed dogs have been identified as at-risk groupsfor clinical leptospirosis.4

A vaccine containing L interrogans serovarscanicola, icterohaemorrhagiae, and pomona and L kirschneri serovar grippotyphosa antigens firstbecame available to veterinarians during 2000. Forthis vaccine to be used effectively, relative serovar-specific infection of dogs at risk of leptospirosis andrisk factors need to be identified. Additionally,knowledge of the prevalence of different infectingserovars would permit better understanding of thereasons for the increase in leptospirosis prevalencesince the early 1980s.4 The purposes of the studyreported here were to estimate serovar-specificprevalence of leptospirosis by use of veterinaryteaching hospital and laboratory submission data;describe annual and seasonal patterns of leptospiro-sis; and identify risk factors for age, sex, and breed.

Criteria for Selection of CasesThe Veterinary Medical Database (VMDB) was

used to identify records of dogs examined at the PurdueUniversity Veterinary Teaching Hospital from January 1,1997, through December 31, 2002, with a diagnosis ofleptospirosis (VMDB diagnosis code 010017200) asdetermined by a clinician on the basis of history andclinical signs, results of CBC, serum biochemical analy-ses, and radiography, and response to treatment, sup-ported by microscopic agglutination test (MAT) or his-tologic examination (Leptospira-like organisms seen viasilver stain of liver or kidney tissue) results. Detailsobtained from the record for each dog included monthand year of examination, age at examination (< 2 weeks,2 to 8 weeks, 9 to 24 weeks, 25 weeks to 1 year, 1.1 to1.9 years, 2 to 3.9 years, 4 to 6.9 years, 7 to 9.9 years, 10

Serovar-specific prevalence and risk factors for leptospirosis among dogs:

90 cases (1997–2002)

Michael P. Ward, BVSc, MSc, MPVM, PhD; Lynn F. Guptill, DVM, PhD, DACVIM; Annalisa Prahl, DVM; Ching Ching Wu, DVM, PhD

Objective—To estimate serovar-specific prevalenceof leptospirosis by use of veterinary teaching hospitaland laboratory submission data; describe annual andseasonal patterns of leptospirosis; and identify riskfactors for age, sex, and breed.Design—Retrospective study.Animals—90 dogs with leptospirosis.Procedures—Hospital records of dogs examined atPurdue University Veterinary Teaching Hospital with adiagnosis of leptospirosis and laboratory records ofdogs from which sera were tested for antibodiesagainst Leptospira spp at Purdue University AnimalDisease Diagnostic Laboratory from 1997 through 2002were reviewed. The likely infecting Leptospira serovarwas identified. Seasonal and annual prevalences werecalculated by use of hospital population at risk (hospitalcases) or serologic testing submissions (diagnostic lab-oratory cases). Age-, sex-, and breed-specific risk fac-tors for hospital cases were estimated by odds ratios.Results—Of the 39 hospitalized dogs identified, 34had been serologically tested, and 22 of those wereinfected with Leptospira kirschneri serovar grippoty-phosa. Of the 51 diagnostic laboratory cases, 59%had a reciprocal titer ≥ 800 against serovar grippoty-phosa. Diagnostic laboratory cases were more com-mon in summer, whereas hospital cases of lep-tospirosis were more common in fall. Male dogs wereat significantly greater risk of leptospirosis thanfemale dogs; and dogs 4 to 6.9 years old were at sig-nificantly greater risk than dogs < 1 year old.Conclusions and Clinical Relevance—L kirschneriserovar grippotyphosa infection was associated withmost cases of leptospirosis in dogs. Use of an effec-tive vaccine that includes this serovar is advisable fordogs at risk of leptospirosis. (J Am Vet Med Assoc2004;224:1958–1963)

From the Departments of Veterinary Pathobiology (Ward, Wu) andVeterinary Clinical Sciences (Guptill, Prahl), School of VeterinaryMedicine, Purdue University, West Lafayette, IN 47907-2027.

Address correspondence to Dr. Ward.

03-07-1097.qxd 5/21/2004 10:20 AM Page 1958

JAVMA, Vol 224, No. 12, June 15, 2004 Scientific Reports: Retrospective Study 1959

SMALLANIMALS

to 15 years, > 15 years, or unknown), sex (female ormale) and reproductive status (sexually intact, neutered,or unknown), breed (purebred [name of breed wasrecorded], mixed breed, or unknown), and results oftesting (MAT titer). The MAT was performed by mixingdilutions of sera with live serovar bratislava, canicola,icterohaemorrhagiae, pomona, grippotyphosa andLeptospira borgpetersenii serovar hardjo leptospires. Thetest result was read as the antibody titer representing thehighest dilution of serum that agglutinated ≥ 50% of liv-ing leptospires.1 Sera were screened at a dilution of 100,and if results were positive, titrated to a maximum dilu-tion of 51,200.

Records of submissions to the Purdue UniversityAnimal Disease Diagnostic Laboratory from January 1,1997, through December 31, 2002, were searched forsubmissions requesting Leptospira serovar testing ofsera from dogs. For each record identified, results ofMAT testing and date of sampling were extracted.Information on age, sex, and breed was not reliablyrecorded; therefore, this information was not extracted.To avoid dogs with leptospirosis from being countedtwice, the 34 dogs admitted to the Veterinary TeachingHospital from which sera were submitted for testingwere excluded from the search of laboratory records.For dogs examined at the hospital and diagnostic labo-ratory submissions, information on relevant vaccina-tion status was recorded in too few records to permituseful analysis.

ProceduresFor each hospitalized dog, the medical record was

reviewed and information on the Leptospira serovarcausing the disease was determined. Diagnosis wasmade on the basis of serologic test results (4-foldincrease in MAT titers in paired serum samples or anMAT titer ≥ 800 in a single serum sample1; referencelimit, < 100) when available. The maximum serovar-specific titer in sera collected within 4 weeks of date ofhospital admission was used to identify the likelyserovar responsible for disease.1 For diagnostic laborato-ry cases, each submission was classified with respect tomaximum titer and serovar. A case was defined as thepresence of a reciprocal titer of ≥ 800 against 1 or moreserovars. In general, MAT titers against Leptospira expo-sure (natural or vaccination) are short lived (1 to 3months), and titers in response to vaccination rarely are> 100 to 400.1 Therefore, we assumed that a reciprocaltiter of ≥ 800 in a serum sample that was submitted by aveterinarian because of clinical signs consistent withleptospirosis represented clinical leptospirosis.1 For dogsin which an equal maximum titer was recorded in serumtested, all serovars with that titer were recorded as thepossible infecting serovar for serovar-specific analyses,but only once for overall analyses, for hospital and diag-nostic laboratory estimates for prevalence of leptospiro-sis. Hospital prevalence of leptospirosis and its 95% con-fidence interval (CI)18 were calculated for each year,1997 through 2002, as number of hospital cases dividedby the total number of dogs examined at the VeterinaryTeaching Hospital each year. The same procedures wererepeated for each infecting serovar identified during thestudy period. A simple weighted linear regression model

was used to test for trends in annual hospital prevalanceof leptospirosis; year (1997 through 2002) was the inde-pendent variable, prevalence was the dependent vari-able, and the annual number of dogs examined at theVeterinary Teaching Hospital was the weighting variable.For seasonal analysis, each hospital case of leptospirosiswas allocated, on the basis of date of hospital admission,to winter (December 22 to March 19), spring (March 20to June 20), summer (June 21 to September 22), or fall(September 23 to December 21). Similarly, dogs exam-ined at the hospital from 1997 through 2002 were allo-cated (on the basis of date of examination) to each sea-son, and seasonal-specific prevalences of hospital casesof leptospirosis were calculated. The same procedureswere used to estimate serovar-specific seroprevalence fordiagnostic laboratory cases. Those dogs from which aserum sample with an MAT titer ≥ 800 had been sub-mitted were considered as diagnostic laboratory cases.The denominator was the total number of submissionsrequesting serologic testing for Leptospira during thecorresponding time periods. Correlations between MATserovar-specific titers were estimated by use ofSpearman rank correlation (rsp), and distributions oftiters were described by skewness.a

To identify risk factors, odds ratios (ORs) wereestimated for the risk of being a hospital case of lep-tospirosis by use of age < 1 year old, sexually intactfemale dogs, and companion dogs as reference cate-gories, respectively. These reference categories werechosen to permit comparison of results with a previousstudy4 of risk factors for leptospirosis in the UnitedStates and Canada. Confidence intervals were calculat-ed for OR estimates, assuming a χ2 approximation.b

Age categories used were < 1, 1 to 1.9, 2 to 3.9, 4 to6.9, 7 to 9.9, and ≥ 10 years. Sex categories were sexu-ally intact female, spayed female, sexually intact male,and castrated male. Breed categories were companiondogs, gun dogs, herding dogs, hounds, terriers, work-ing dogs, and mixed-breed dogs; a general reference19

on dog breeds was used to allocate individual breeds toeach of the breed categories.

ResultsBetween 1997 and 2002, 15,680 dogs were exam-

ined at the Veterinary Teaching Hospital. A total of 39hospitalized dogs with leptospirosis were identified (249cases/100,000 dogs examined; 95% CI, 179 to 343). In34 of the 39 dogs, serologic testing had been performed.In the other 5 dogs, clinical diagnosis was confirmed onnecropsy via silver stain of liver or kidney tissue. In 5and 2 of the 34 dogs in which serology was performed,equal maximum titers were observed against 2 and 3serovars, respectively. In all except 2 of these dogs,serovar grippotyphosa was 1 of the serovars againstwhich a maximum titer was observed. During the sametime period, sera were submitted from 470 dogs exam-ined by practitioners in Indiana (94.5%), Illinois (3.9%),and Michigan (1.6%) for serologic testing for Leptospiraantibodies. Overall, sera from 51 of these 470 (10.9%;95% CI, 8.3 to 14.1) dogs reacted to 1 or more serovarsat a dilution of ≥ 800 and were defined as diagnostic lab-oratory cases of leptospirosis. Two of these dogs hadequal reciprocal titers (1,600 and 25,600) against 3

03-07-1097.qxd 5/21/2004 10:20 AM Page 1959

1960 Scientific Reports: Retrospective Study JAVMA, Vol 224, No. 12, June 15, 2004

SMALLANIMALS

serovars (grippotyphosa, bratislava, and icterohaemor-rhagiae or pomona, respectively), and 6 of these dogshad equal reciprocal titers (range, 800 to 25,600) against2 serovars (including grippotyphosa, bratislava, ictero-haemorrhagiae, or pomona).

Temporal distribution of leptospirosis—Thehighest seasonal prevalence of hospital cases (518cases/100,000 dogs examined; 95% CI, 332 to 1,099)was during fall, whereas the highest prevalence of diag-nostic laboratory cases (16.3%; 95% CI, 10.8% to23.7%) was during summer (Fig 1). The highest annu-al prevalence of hospital cases (612 cases/100,000 dogsexamined; 95% CI, 332 to 1,099) was during 2000, andthe highest prevalence of diagnostic laboratory cases(25.6%; 95% CI, 15.5% to 34.67%) was during 1998(Fig 2; Tables 1 and 2). No significant linear trend inannual prevalence of hospital (P = 0.28) or diagnosticlaboratory (P = 0.18) cases was detected.

Serovar-specific distribution of leptospirosis—The highest serovar-specific prevalences for hospitaland diagnostic laboratory cases were estimated forserovars grippotyphosa (134 cases/100,000 dogs exam-ined; 95% CI, 85 to 209 cases/100,000 dogs examinedand 6.4%; 95% CI, 4.4% to 9.1%, respectively) andbratislava (51 cases/100,000 dogs examined; 95% CI,24 to 105 cases/100,000 dogs examined and 2.8%; 95%CI, 1.5% to 4.8%, respectively).

The highest annual prevalence of hospital cases(219 cases/100,000 dogs examined; 95% CI, 65 to 560)was estimated for serovar grippotyphosa during 2001(Table 1). The highest annual prevalence of diagnosticlaboratory cases (14.6%; 95% CI, 8.1% to 24.6%) wasestimated for serovar grippotyphosa in 1998 (Table 2).The correlation between annual hospital and diagnosticlaboratory prevalence was 22.3% (P = 0.67; Fig 2).Hospital prevalence of leptospirosis associated withserovar grippotyphosa infection was the highest orequal highest in every year of the study, ranging from 34(1999) to 219 (2001) cases/100,000 dogs examined.Serovar bratislava was detected in 4 of 6 years of thestudy, and serovar pomona was detected in 3 years.Serovar icterohaemorrhagiae was detected in 2 years(2000 and 2001) of the study, and serovar canicola andL borgpetersenii serovar hardjo were detected in only 1year (2000). Diagnostic laboratory prevalence of lep-tospirosis associated with serovar grippotyphosa infec-tion was highest in every year of the study except dur-

ing 2000, when prevalence associated with serovaricterohaemorrhagiae infection was the highest. Serovarbratislava was detected in every year of the study;serovars pomona and icterohaemorrhagiae were detect-ed in 5 and 3 years, respectively. A reciprocal titer ≥ 800against L borgpetersenii serovar hardjo was detectedonly once, and no submitted sera had a reciprocal titeragainst serovar canicola ≥ 800. The median titer (ofthose ≥ 100) against serovars canicola and icterohaem-orrhagiae was 100; against L borgpetersenii serovar hard-jo, and serovars pomona and bratisava was 200;

Figure 2—Annual prevalence of leptospirosis among dogs exam-ined at Purdue University Veterinary Teaching Hospital (No. of hos-pital cases/100,000 dogs examined, bar) and annual prevalence ofsera with MAT titers ≥ 800 against Leptospira spp submitted toPurdue University Animal Disease Diagnostic Laboratory (percent-age of submissions, line) from 1997 through 2002.

Figure 1—Seasonal prevalence of leptospirosis among dogs exam-ined at Purdue University Veterinary Teaching Hospital (No. of hos-pital cases/100,000 dogs examined, bar) and seasonal prevalenceof sera with microscopic agglutination test (MAT) titers ≥ 800against Leptospira spp submitted to Purdue University AnimalDisease Diagnostic Laboratory (percentage of submissions, line)from 1997 through 2002.

Year

Serovara 1997 1998 1999 2000 2001 2002 1997–2002

Leptospira kirschneri serovar grippotyphosa 121 148 34 204 219 99 134Leptospira interrogans serovar bratislava 0 37 0 204 73 49 51L interrogans serovar icterohaemorrhagiae 0 0 0 102 110 0 32L interrogans serovar pomona 30 0 0 102 73 0 32L interrogans serovar canicola 0 0 0 51 0 0 6Leptospira borgpetersenii serovar hardjo 0 0 0 51 0 0 6

All serovars 151 259 34 612 365 197 249

aMaximum microscopic agglutination test (MAT) titer; 7 dogs had equal maximum titers against � 2 serovars.

Table 1—Serovar-specific prevalence (No. of hospital cases/100,000 dogs examined) of leptospirosisamong dogs examined at Purdue University Veterinary Teaching Hospital from 1997 through 2002

03-07-1097.qxd 5/21/2004 10:20 AM Page 1960

JAVMA, Vol 224, No. 12, June 15, 2004 Scientific Reports: Retrospective Study 1961

SMALLANIMALS

and against serovar grippotyphosa was 400 (Table 3).All distributions were positively (right) skewed. Thedistribution of titers was least skewed for serovars grip-potyphosa and bratislava and most skewed for serovaricterohaemorrhagiae and L borgpetersenii serovar hard-jo. Titers against serovar grippotyphosa were signifi-cantly (P ≤ 0.01) correlated with those in the same sam-ple against serovars pomona (rsp = 0.66), canicola (rsp =0.73), and bratislava (rsp = 0.69).

Risk factors for leptospirosis—Age was recordedfor 15,648 (99.8%) of the 15,680 dogs examined at theVeterinary Teaching Hospital and for all hospital cases(Table 4). A diagnosis of leptospirosis was 5.7 times

more likely in dogs 4 to 6.9 years old (95% CI, 1.54 to21.3) than dogs < 1 year old. Sex was recorded for15,547 (99.2%) of the 15,680 dogs examined and forall hospital cases. A diagnosis of leptospirosis was 4.2times more likely in sexually intact male dogs (95% CI,1.04 to 16.60) than sexually intact female dogs. Breedwas recorded for 15,679 (99.99%) of the 15,680 dogsexamined and for all hospital cases. No significant (P >0.05) associations were found between diagnosis ofleptospirosis and breed categories.

DiscussionLeptospira kirschneri serovar grippotyphosa was

associated with most of the cases of leptospirosis

Table 2—Number (prevalence [%]) of sera submitted to Purdue University Animal Disease Diagnostic Laboratory from 1997 through2002 with MAT titers ≥ 800 against 6 Leptospira serovars

Year

Serovara 1997 1998 1999 2000 2001 2002 1997–2002

L kirschneri serovar grippotyphosa 3 (7.5) 12 (14.6) 2 (2.7) 2 (3.6) 5 (4.2) 6 (5.9) 30 (6.4)L interrogans serovar bratislava 1 (2.5) 5 (6.1) 1 (1.4) 1 (1.8) 4 (3.4) 1 (1.0) 13 (2.8)L interrogans serovar icterohaemorrhagiae 0 0 1 (1.4) 4 (7.1) 0 2 (2.0) 7 (1.5)L interrogans serovar pomona 1 (2.5) 4 (4.9) 1 (1.4) 2 (3.6) 2 (1.7) 0 10 (2.1)L interrogans serovar canicola 0 0 0 0 0 0 0L borgpetersenii serovar hardjo 0 0 1 (1.4) 0 0 0 1 (0.2)

All serovars tested 40 82 73 56 118 101 470MAT � 800 5 (12.5) 21 (25.6) 6 (8.3) 9 (16.1) 11 (9.3) 9 (8.9) 51 (10.9)

aMaximum MAT titer; 8 dogs had equal maximum titers against � 2 serovars.

Table 3—Distribution of MAT titers in sera from dogs submitted to Purdue University Animal Disease Diagnostic Laboratory from 1997through 2002 and tested against 6 Leptospira serovars

Mediana

Serovar �� 100 100 200 400 800 1,600 3,200 6,400 12,800 �� 25,600 titer Skewness

L kirschneri serovar grippotyphosa 301 25 13 15 7 4 10 0 2 6b 400 3.09L interrogans serovar bratislava 310 31 13 11 3 3 5 5 1 2c 200 3.47L interrogans serovar icterohaemorrhagiae 353 20 2 4 2 3 0 0 0 0 100 5.43L interrogans serovar pomona 339 15 9 9 5 1 1 4 0 1 200 4.41L interrogans serovar canicola 349 22 9 3 1 0 0 0 0 0 100 4.30L borgpetersenii serovar hardjo 339 15 9 9 5 1 1 4 0 1 200 5.04

aSera with MAT titers � 100. bMAT titers � 25,600, 51,200, and 102,400 were measured in 4, 1, and 1 serum samples, respectively. cMAT titer =25,600.

MAT titer

Table 4—Risk factors for leptospirosis among dogs examined at Purdue University Veterinary TeachingHospital from 1997 through 2002

No. of No. Cases/100,000 Odds Variable Category cases at-risk dogs ratio 95% CI

Age (y) � 1 2 2,918 69 1.00 NA1–1.9 3 1,534 196 2.85 0.52, 15.82–3.9 6 2,496 240 3.51 0.78, 15.74–6.9 13 3,316 392 5.72 1.54, 21.37–9.9 8 2,848 281 4.10 0.98, 17.1� 10 7 2,536 276 4.03 0.94, 17.2

Sex Sexually intact female 2 2,403 83 1.00 NASpayed female 15 5,631 266 3.20 0.79, 12.9Sexually intact male 11 3,181 346 4.15 1.04, 16.6Castrated male 11 4,332 254 3.05 0.73, 12.8

Breed Companion dog 6 2,283 263 1.00 NAGun dog 10 3,821 262 1.00 0.36, 2.74Herding dog 1 1,474 68 0.26 0.04, 1.84Hound 2 1,298 154 0.59 0.12, 2.85Terrier 5 1,256 398 1.51 0.47, 4.91Working dog 6 2,137 281 1.07 0.35, 3.31Mixed-breed dog 9 3,410 264 1.00 0.36, 2.82

CI = Confidence interval. NA = Not applicable.

03-07-1097.qxd 5/21/2004 10:20 AM Page 1961

1962 Scientific Reports: Retrospective Study JAVMA, Vol 224, No. 12, June 15, 2004

SMALLANIMALS

detected from 1997 through 2002. Although the pop-ulation at risk was predominantly dogs in northernand central Indiana and some geographic variationwithin the United States with respect to the relativeimportance of various Leptsopira serovars is expected,results suggested that the increase in prevalence ofleptospirosis since the 1980s4 could partly be causedby an increase in serovar grippotyphosa infections.Several wildlife species are reservoir hosts for L kirschneri serovar grippotyphosa.1,2,8,14,16,17 In particu-lar, raccoon populations appear to be endemicallyinfected in parts of the midwestern and southernUnited States.20,21 Exposure of dogs to sites contami-nated by this serovar is likely to be increased by a dog’sroaming habits and wildlife density (especially rac-coons). Detailed mechanisms of how dogs are exposedneed to be determined. If results from the presentstudy can be extrapolated to dog populations through-out the United States, practitioners who recommendto their clients vaccination of dogs at risk against lep-tospirosis should consider using a vaccine that incor-porates serovar grippotyphosa.

Several seroprevalence surveys10,22-25 of dogs forLeptospira exposure have been performed, but to theauthors’ knowledge, none have been performed in theUnited States. Estimates of the prevalence of lep-tospirosis from surveys of dog populations are difficultto compare because of different MAT cutoff titers,serovars that are included in the MAT, and studydesigns used. In the literature, results of studies10,22-26

with cutoff titers ranging from ≥ 50 to ≥ 800 have beenreported. The MAT remains the standard reference testfor the serodiagnosis of leptospirosis.3 Interpretation ofMAT results is influenced by cross-reactivity of anti-bodies, antibody titers induced by vaccination, andlack of consensus regarding correlation between activeinfection and serologic titers.1 In general, the infectingserovar is assumed to be that to which the highest titerdevelops. Dogs develop low (usually < 400) MAT titersthat persist for 1 to 3 months in response to vaccina-tion.1 Clinical signs consistent with leptospirosis plusan MAT titer ≥ 800 in a single serum sample providegood presumptive evidence of leptospiral infection.1,11

A reciprocal titer of ≥ 800 was used as the cutoff titerin the study reported here in an attempt to reduce thenumber of dogs falsely classified as leptospirosis cases(conservative bias). This decision rule was definedprior to data analysis, and we assumed that if a dog hada titer of < 100 prior to the disease event and serologictesting, a titer of ≥ 800 would correspond to the stan-dard case definition of a 4-fold rise in paired sera. Asindicated by serovar-specific distributions of MATtiters in this study, estimates of prevalence can changegreatly depending on the cutoff titer used. Informationon the prevalence of leptospirosis may be derived fromsera submitted to diagnostic laboratories for Leptospiratesting (generally because of a clinical suspicion of lep-tospirosis) or from purpose-designed surveys in whichsera may or may not be systematically collected. Theserovars tested in such studies are generally deter-mined by those serovars believed to be prevalent in thepopulation of interest.1 For example, Leptospira interro-gans serovar autumnalis is not included in the panel of

MAT serovars used at the diagnostic laboratory inwhich sera were tested in the present study becauseclinical disease is not believed to develop in the popu-lation at risk, despite the reported ability of this serovarto cause disease in dogs in other regions.27 Therefore,overall prevalence of leptospirosis in this study mayhave been underestimated. However, because of cross-reactivity between Leptospira serovars,1,12 such underes-timation of overall prevalence may be neglible. In addi-tion, dogs with clinical signs of leptospirosis can havenegative MAT results during the first 1 to 2 weeks ofinfection,3,12,13 also resulting in underestimation ofprevalence determined by diagnostic laboratory data.The choice of serovars tested for has an importantimpact on determination of serovar-specific diagnosis.Because of cross-reactivity, a truly definitive diagnosisof the infecting serovar through serologic testing israrely possible without incorporating many serovars.Longitudinal studies, performed by use of serologicand bacteriologic techniques, are required to clarify theassociation between exposure, infection, and disease.To the authors’ knowledge, the study reported here isthe first that has focused on identifying cases of lep-tospirosis by use of 2 different sources (hospital anddiagnostic laboratory cases) from the same populationat risk during the same period. Results of this studysuggested that both sources are useful, and perhapsnecessary, for understanding the epidemiology of lep-tospirosis in dogs.

Age and sex were identified as risk factors for lep-tospirosis. The risk of leptospirosis was greatest inmiddle-aged (4 to 6.9 years old) dogs, consistent withprevious case reports and results of retrospective stud-ies.1,8,9,14 Middle-aged dogs may be more active outsidetheir normal home environment than young or olddogs, increasing potential exposure to Leptospiraserovars. The sex category at greatest risk (OR, 4.15) ofhaving a diagnosis of leptospirosis was sexually intactmale dogs; this finding also was consistent with resultsof other studies.1,7,8,28 Male dogs may be more likely tospend time outside their homes and be exposed to lep-tospires in the environment.4,8 Breed was not associat-ed with a significant increased risk of clinical diagnosisof leptospirosis. This finding contradicts results of aprevious study4 of dogs in the United States andCanada, in which herding, hound, working, andmixed-breed dogs were at greater risk of leptospirosisthan companion dog breeds. Inconsistencies mayreflect regional differences in breed risk factors; alldogs in mostly rural areas may be at risk of leptospiro-sis, whereas in more urban areas only dogs used forsporting or working purposes, or that are more likelyto travel with their owners to recreational areas, maybe at increased risk. Further studies of risk factorsshould examine breed and the environment becausethis information can be used to guide the vaccinationrecommendations veterinarians make to their clients.

A combination of history and clinical signs andbiochemical, hematologic, and serologic testing is usedto diagnose leptospirosis, and this diagnostic approachappears to have remained constant among literaturepublished in the 1980s and a more recently publishedreport8 and was used in the present study. The VMDB,

03-07-1097.qxd 5/21/2004 10:20 AM Page 1962

JAVMA, Vol 224, No. 12, June 15, 2004 Scientific Reports: Retrospective Study 1963

SMALLANIMALS

a secondary data source, was used to provide popula-tion information. Study findings may be subject to biasin recorded information on risk factors and diagnosisof leptospirosis. Referral bias may also have existed inthe hospital data. Although Purdue UniversityVeterinary Teaching Hospital has a general practicecomponent, dogs examined may not necessarily havebeen representative of the general dog population.Diagnostic laboratory data probably represent moreclosely the population of interest.

Information on prevalent Leptospira serovars andrisk factors for leptospirosis in dogs is needed to devel-op effective disease control strategies. This knowledgewill assist veterinarians in identifying dogs with lep-tospirosis and will help prevent leptospirosis by pro-tecting dogs against the most common infectingserovars and permitting dog owners to modify expo-sure of their dogs to leptospires. Middle-aged and maledogs are at greater risk of leptospirosis, and serovargrippotyphosa appears to be an important infectingLeptospira serovar. Further research is needed to iden-tify specific environmental characteristics that increasethe risk of leptospirosis and the role of breed as a riskfactor. This information will permit sound vaccinationstrategies and disease prevention programs to be madeavailable to dog owners.

aStatistix, version 7.0, Analytical Software, Tallahassee, Fla.bWin Episcope, version 2.0, Department of Veterinary Clinical

Studies, Royal (Dick) School of Veterinary Studies, University ofEdinburgh, UK. Available at: www.clive.ed.ac.uk/winepiscope.Accessed Aug 30, 2000.

References1. Bolin CA. Diagnosis of leptospirosis: a reemerging disease of

companion animals. Semin Vet Med Surg Small Anim 1996;11:166–171.2. Hanson LE. Leptospirosis in domestic animals: the public

health perspective. J Am Vet Med Assoc 1982;181:1505–1509.3. Sykes JE. Canine leptospirosis—a re-emerging disease, in

Proceedings. 19th Annu Meet Am Coll Vet Intern Med 2001;763–765.4. Ward MP, Glickman LT, Guptill LF. Prevalence of and risk

factors for leptospirosis among dogs in the United States and Canada:677 cases (1970–1998). J Am Vet Med Assoc 2002;220:53–58.

5. Ward MP. Clustering of leptospirosis among dogs in theUnited States and Canada. Prev Vet Med 2002;56:215–226.

6. Brown CA, Roberts AW, Miller MA, et al. Leptospira interro-gans serovar grippotyphosa infection in dogs. J Am Vet Med Assoc1996;209:1265–1267.

7. Rentko VT, Clark N, Ross LA, et al. Canine leptospirosis. Aretrospective study of 17 cases. J Vet Intern Med 1992;6:235–244.

8. Adin CA, Cowgill LD. Treatment and outcome of dogs with

leptospirosis: 36 cases (1990–1998). J Am Vet Med Assoc 2000;216:371–375.

9. Forrest LJ, O’Brien RT, Tremelling MS, et al. Sonographicrenal findings in 20 dogs with leptospirosis. Vet Radiol Ultrasound1998;39:337–340.

10. Prescott JF, McEwen B, Taylor J, et al. Resurgence of lep-tospirosis in dogs in Ontario: recent findings. Can Vet J 2002;43:955–961.

11. Ribotta M, Fortin M, Higgins R. Canine leptospirosis: serol-ogy. Can Vet J 2000;41:494–495.

12. Harkin KR, Gartrell CL. Canine leptospirosis in New Jerseyand Michigan: 17 cases (1990–1995). J Am Anim Hosp Assoc 1996;32:495–501.

13. Kalin M, Devaux C, DiFruscia R, et al. Three cases of canineleptospirosis in Quebec. Can Vet J 1999;40:187–191.

14. Birnbaum N, Center SA, Schermerhorn T, et al. Naturallyacquired leptospirosis in 36 dogs: serological and clinicopathologicalfeatures. J Small Anim Pract 1998;39:231–236.

15. American Public Health Association. Leptospirosis. In: ChinJ, ed. Control of communicable diseases manual. 17th ed. Washington,DC: American Public Health Association, 2000;293–296.

16. Pilgrim S, Threlfall WR. A serologic study of leptospirosisin mares. Equine Pract 1999;21:20–23.

17. Ellis WA, O’Brien JJ, Cassells JA, et al. Leptospiral infectionin horses in Northern Ireland: serological and microbiological find-ings. Equine Vet J 1983;15:317–320.

18. Fleiss JL. Statistical methods for rates and proportions. 2nded. New York: John Wiley & Sons, 1981.

19. Wilcox B, Walkowicz C. Atlas of dog breeds of the world. 3rded. Neptune City, NJ: TFH Publications, 1989.

20. Mitchell MA, Hungerford LL, Nixon C, et al. Serologic sur-vey for selected infectious disease agents in raccoons from Illinois. J Wildl Dis 1999;35:347–355.

21. Shotts EB Jr, Andrews CL, Harvey TW. Leptospirosis inselected wild mammals of the Florida panhandle and southwesternGeorgia. J Am Vet Med Assoc 1975;167:587–589.

22. Van den Broek AHM, Thrusfield MV, Dobbie GR, et al. Aserological and bacteriological survey of leptospiral infection in dogsin Edinburgh and Scotland. J Small Anim Pract 1991;32:118–124.

23. Dickeson D, Love DN. A serological survey of dogs, cats andhorses in south-eastern Australia for leptospiral antibodies. Aust Vet J1993;70:389–390.

24. Hilbink F, Penrose M, McSporran K. Antibodies in dogsagainst Leptospira interrogans serovars copenhageni, ballum and cani-cola. N Z Vet J 1992;40:123–125.

25. O’Keefe JS, Jenner JA, Sandifer NC, et al. A serosurvey forantibodies to Leptospira in dogs in the lower North Island of NewZealand. N Z Vet J 2002;50:23–25.

26. Scanziani E, Origgi F, Giusti AM, et al. Serological survey ofleptospiral infection in kennelled dogs in Italy. J Small Anim Pract2002;43:154–157.

27. Wohl JS. Canine leptospirosis. Compend Contin Educ Pract Vet1996;18:1215–1224.

28. Baldwin CJ, Atkins CE. Leptospirosis in dogs. Compend ContinEduc Pract Vet 1987;9:499–507.

03-07-1097.qxd 5/21/2004 10:20 AM Page 1963