Epidemiology of Bartonella infection in domestic cats in France

Veterinary Microbiology 80 (2001) 185±198

Epidemiology of Bartonella infection in domestic cats in France A.N. Gur®elda,b,1, H.-J. Boulouisa, B.B. Chomelb,*, R.W. Kastenb, R. Hellerc, C. Bouillina, C. Gandoina, D. Thibaulta, C.-C. Changb, F. Barrata, Y. Piemontc a

UMR 956 INRA/AFSSA/ENVA, Ecole Nationale VeÂteÂrinaire d'Alfort, 94704 Maisons-Alfort, France b Department of Population Health and Reproduction, School of Veterinary Medicine, University of California, Davis, CA 95616, USA c Institut de BacteÂriologie, Faculte de MeÂdecine, Universite Louis Pasteur, 67000 Strasbourg, France Received 15 August 2000; received in revised form 9 January 2001; accepted 9 January 2001

Abstract Blood samples were collected between February and June 1996 from a convenience sample of 436 domestic French cats living in Paris and its environs and were tested for Bartonella bacteremia and seropositivity. Seventy-two cats (16.5%) were Bartonella bacteremic, of which 36 cats (50%) were infected with Bartonella henselae type II (B.h. II) only, 15 cats (21%) were infected with Bartonella clarridgeiae (B.c.) only, and 11 cats (15%) were infected with B. henselae type I (B.h. I) only. Eight cats (11%) were co-infected with B. henselae and B. clarridgeiae (B.h. II/B.c.: ®ve cats; B.h. I/B.c.: three cats). Two cats (2.8%) were concurrently bacteremic with B. henselae types I and II. Risk factors associated with bacteremia included ownership for <6 months (prevalence ratio …PR† ˆ 1:80; 95% con®dence interval …CI† ˆ 1:13ÿ2:85), adoption from the pound or found as a stray (PR ˆ 1:67, 95% CI ˆ 1:05ÿ2:65), and cohabitation with one or more cats (PR ˆ 1:60, 95% CI ˆ 1:01ÿ2:53). Bartonella antibodies to either B. henselae or B. clarridgeiae were detected in 179 cats (41.1%). Risk factors associated with seroposivity paralleled those for bacteremia, except for lack of association with time of ownership. Prevalence ratios of bacteremic or seropositive cats increased with the number of cats per household …p ˆ 0:02†. The lack of antibodies to B. henselae or B. clarridgeiae was highly predictive of the absence of bacteremia (predictive value of a negative test ˆ 97:3%). Multiple logistic regression analysis indicated that bacteremia, after adjustment for age and ¯ea infestation, and positive serology, after adjustment for age, were associated with origin

*

Corresponding author. Tel.: ‡1-530-752-8112; fax: ‡1-530-752-2377. E-mail address: [email protected] (B.B. Chomel). 1 Present address: Department of Agriculture, Weights and Measures, County of San Diego, Of®ce of the County Veterinarian, San Diego, CA 92123, USA. 0378-1135/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 3 5 ( 0 1 ) 0 0 3 0 4 - 2

186

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

of adoption and number of cats in the household. Flea infestation was associated with positive serology. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Bartonella clarridgeiae; Bartonella henselae; Cat; France

1. Introduction Cat scratch disease is caused by the curved, gram-negative bacillus Bartonella henselae (Regnery et al., 1992a, 1992c; Dolan et al., 1993). A new species, Bartonella clarridgeiae (Clarridge et al., 1995; Lawson and Collins, 1996) has also been linked to cases of cat scratch disease (Kordick et al., 1997a; Margileth and Baehren, 1998). The reservoir of B. henselae and B. clarridgiae is the domestic cat (Regnery et al., 1992b; Koehler et al., 1994; Lawson and Collins, 1996; Bergmans et al., 1997; Heller et al., 1997), which can remain bacteremic for months to years (Abbott et al., 1997; Guptill et al., 1997; Kordick et al., 1995). In contrast to cat-to-man transmission, which occurs predominantly via a scratch or bite, the cat ¯ea, Ctenocephalides felis, is the primary vector of Bartonella between cats (Chomel et al., 1996; Higgins et al., 1996; Foil et al., 1998). Unlike human infections that can manifest as a wide spectrum of clinical signs, the majority of naturally infected cats appear to be asymptomatic, although clinical signs have been observed in some experimentally infected cats (Kordick and Breitschwerdt, 1997; Guptill et al., 1998; O'Reilly et al., 1999). Recently, a third feline Bartonella species, Bartonella koehlerae, has been discovered in the blood of cats, but its pathogenic potential for humans remains to be demonstrated (Droz et al., 1999). The spectrum of clinical manifestations of cat scratch disease has been broadened by the identi®cation of these Bartonella species and the development of diagnostic tests. Vascular proliferative diseases, such as bacillary angiomatosis and peliosis hepatis, and in¯ammatory conditions such as endocarditis, encephalitis, abscessation and leukocytoclastic vasculitis represent some of the protean manifestations of B. henselae infection in humans (Regnery et al., 1992a; Welch et al., 1992; Slater et al., 1992, 1994; Hashkes et al., 1996; Baorto et al., 1998; George et al., 1998; Margileth and Baehren, 1998). Additionally, B. henselae isolates differing in their 16S rRNA gene sequences, reported here as type I and type II (previously identi®ed as Houston-1 and BA-TF, respectively, Heller et al., 1997), have been found infecting cats and humans (Bergmans et al., 1996, 1997; Sander et al., 1997, 1998). It has been suggested that these types may differ in their pathogenicity (Bergmans et al., 1996, 1997; Relman, 1998; Sander et al., 1998). Previous epidemiological studies of Bartonella infection in North American domestic cats have demonstrated serological prevalences ranging from 4 to 81%, and bacteremia prevalence of up to 40% (Koehler et al., 1994; Childs et al., 1995; Chomel et al., 1995a; Jameson et al., 1995). In Asia, a seroprevalence of 15% was reported in cats from Japan (Ueno et al., 1995) and 7.2% of 690 Japanese cats were Bartonella bacteremic (Maruyama et al., 2000). The highest bacteremia rate (20%) was found in southern Japan (Okinawa Island). B. clarridgeiae accounted for 10.0% of the Bartonella isolates. One cat was co-infected with B. henselae and B. clarridgeiae. In Indonesia, Marston et al. (1999) reported a seroprevalence of 54% (40/74) and isolated B. henselae from six of 14 feral

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

187

cats (43%) and B. clarridgeiae from two feral cats and one pet cat. In the Philippines, two-thirds of 107 cats tested had Bartonella antibodies and 61% of the 31 cats for which blood was cultured were bacteremic (Chomel et al., 1999). One-third of these 31 bacteremic cats were infected with B. clarridgeiae and two-thirds with B. henselae type I. In Australia, Branley et al. (1996) reported a 35% prevalence of B. henselae bacteremia, with feral cats having a higher prevalence than pet cats, 40 and 16%, respectively. In the Middle East, a seroprevalence of 40% was reported from cats from Israel (Baneth et al., 1996) and only a 12% seroprevalence from cats from Egypt (Childs et al., 1995). In Europe, the highest reported Bartonella seroprevalences were in Dutch cats, with 56% of 50 pet cats and 50% of 113 sheltered cats being seropositive (Bergmans et al., 1997). Twenty-®ve (22%) of these 113 sheltered cats were Bartonella bacteremic, including 10 cats (40%) infected with B. henselae type II, six cats (24%) infected with B. henselae type I, and four cats (16%) infected with B. clarridgeiae. Five of these 25 cats were co-infected with B. clarridgeiae and B. henselae. In Austria, Allerberger et al. (1995) reported that 33.3% of 96 cats were seropositive for B. henselae, whereas in Switzerland, Glaus et al. (1997) reported an 8.3% seroprevalence out of 728 cats tested. In Germany, Haimerl et al. (1999) recently reported a 15% seroprevalence in domestic cats. In this same country, Sander et al. (1997) had previously reported a 13% prevalence of B. henselae bacteremia in domestic pet cats, involving two different B. henselae variants. B. clarridgeiae was not isolated from any of these German cats. In France, Chomel et al. (1995b) reported a B. henselae seroprevalence of 36% in 64 pet cats from the Paris area, with 11% of these 64 cats being B. henselae bacteremic. In contrast, Heller et al. (1997) found 53% of 94 stray cats from Nancy, France, bacteremic with B. henselae or B. clarridgeiae. Among these isolates, 36% were determined to be B. henselae type II, 34% were B. henselae type I and 30% were identi®ed as B. clarridgeiae. However, coinfection was not reported in that study. Using a convenience sample of French domestic cats living in Paris and its environs, we determined and report here the respective prevalences of Bartonella bacteremia and exposure (i.e. presence of antibodies), identi®ed the endemic Bartonella species, and determined risk factors associated with infection. As part of this overall study, coinfection of seven cats with either two Bartonella species or B. henselae types was identi®ed and has been previously reported (Gur®eld et al., 1997). We include these seven cats in the present report. 2. Material and methods 2.1. Bartonella isolation from cat blood A convenience sample of domestic pet cats presented to the Ecole Nationale VeÂteÂrinaire d'Alfort, Maisons-Alfort, France, and four private veterinary hospitals within the Paris area, between February and mid-June 1996, were enrolled in the study. After consent was obtained, owners ®lled out a questionnaire describing the history, lifestyle, and degree of ¯ea infestation of their cats. Three militers of blood were aseptically drawn from peripheral veins. Half was collected into pediatric lysis isolator tubes (Isostat,

188

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

Wampole Laboratories, Cranbury, NJ) and the other half into serum separating tubes. The isolator tubes were spun at 1800g for 75 min at room temperature and the supernatant decanted. The pellet was resuspended in 125 ml of inoculation medium (Koehler et al., 1992) and the volume recorded. Two hundred and ®fty microliters of the suspension were inoculated onto heart infusion agar (Difco Laboratories, Detroit, MI) supplemented with 5% de®brinized rabbit blood. The remaining suspension was inoculated onto a second plate. The plates were incubated for 1 month at 35.58C with 5% CO2 in a humid incubator and checked regularly for bacterial growth. Colonies were Gram stained, sub-cultured, harvested and frozen at ÿ708C in 100% fetal calf serum. Serum separating tubes were centrifuged at 500g, the serum was removed and frozen at ÿ208C. 2.2. Identi®cation of isolates Cat blood isolates were identi®ed using polymerase chain reaction±restriction fragment length polymorphism analysis (PCR±RFLP) of a fragment of the citrate synthase gene (Regnery et al., 1992a; Koehler et al., 1994), and 16S rRNA gene sequence analysis (Relman et al., 1990). DNA was extracted from frozen bacterial suspensions via a Chelex extraction technique (De Lamballerie et al., 1992). An approximately 1500 basepair (bp) fragment of the 16S rRNA gene was ampli®ed from the extracted DNA using two eubacterial universal primers speci®c for the 16S rRNA gene: P8 (AGAGTTTGATCCTGGCTCAG-30 ) and Pc1544 (50 -AAGGAGGTGATCCAGCCGCA-30 ). Rigorous precautions were taken not to contaminate the samples prior to gene ampli®cation. Presence of the ampli®ed products were veri®ed by electrophoresis on an 1% agarose gel followed by staining with ethidium bromide and visualization on a UV transilluminator. The ampli®ed fragment was puri®ed via phenol extraction and isopropanol precipitation (Brow, 1990). The sequencing reaction was performed by using the 50 ¯uorescein isothiocyanate labeled internal primer Pc535 (50 -GTATTACCGCGGCTGCTGGCAC-30 ), Thermo Sequenase DNA polymerase (Amersham, England) and a cyclic sequencing protocol. Electrophoresis and reading of the sequence were performed with an ALF DNA Sequencer (Pharmacia, France) according to the manufacturer's instructions. DNA sequences were compared to other Bartonella sequences contained in the EMBL GenBank database using the Clustal method (Higgins and Sharp, 1988) on DNAStar software (DNAStar, London). An approximately 400 bp fragment of the citrate synthase gene was also ampli®ed using previously described primers and methods (Regnery et al., 1992a). The ampli®ed products were veri®ed by gel electrophoresis and then enzymatically digested using TaqI and HhaI restriction endonucleases (Chomel et al., 1995a). Banding patterns were compared with type strains of B. henselae Houston-1 (ATCC 49882) and B. clarridgeiae (ATCC 51734). 2.3. Serological analysis Antibody titers against B. henselae and B. clarridgeiae were determined using an immuno¯uorescence assay (IFA), as previously described (Regnery et al., 1992c; Chomel et al., 1995a). Brie¯y, serum was serially diluted in PBS and incubated on slides

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

189

containing Felis catus whole fetus (FCWF) cells infected with 16S rRNA sequence con®rmed strains of B. henselae type II (strain U4, UC Davis) and B. clarridgeiae (ATCC 51734). The slides were washed and probed with FITC goat anti-cat IgG …H ‡ L† conjugate (Cappel, Organon Teknika, Durham, NC), and the ¯uorescence graded on a scale from 1 to 4, independently by the same two readers. The endpoint titer was de®ned as the lowest dilution for which an agreement of the two readers was obtained. Any serum with a grade 2 ¯uorescence at a titer of 64 was considered positive (Chomel et al., 1995a). 2.4. Statistical analysis For univariate analyses, non-parametric tests (chi-square, Fisher's exact test) performed with Epi Info 6.04b, were used to test for associations between culture positivity and putative explanatory factors, such as age, sex, origin, time of ownership and ¯ea infestation. Similar tests were performed to test for associations between seropositivity and the same explanatory factors. The association between bacteremia and seropositivity was evaluated by paired analysis using the McNemar chi-square test. Using SAS 6.12, a multiple logistic regression model was constructed to assess the effect of multiple putative risk factors for Bartonella bacteremia and seropositivity, respectively. 3. Results 3.1. Description of the cat population Four hundred and thirty-six cats from Paris and its environs were sampled. They ranged in age from 3 months to 17 years old (mode 0.7 year, mean 2.2 years). Seventy-three percent of the cats were less than 1.5 years old. Forty-six percent of the cats were female, 43% were male, and the sex was not recorded in 11% of the cats. The predominant breed was the European cat (78.8%), equivalent to the domestic short-hair cat in the United States. Other breeds represented included Persian (4.6%), Siamese (2.1%), Burmese (0.7%), Norwegian (0.7%), Chartreuse (0.5%), and other breeds (1.1%). Breed identi®cation was not obtained for 11.5% of cats. Of the cats with a known lifestyle (343/436; 79%), almost two-thirds (224/343; 65%) of the cats lived predominantly indoors and 31% (105/343) lived both outdoors and indoors. Only 3% (14/343) were reported as living predominantly outdoors. Of the 376 cats for which the neutered/spayed status was known, 86% (322/376) were not neutered at the time of testing. Origin of the cats prior to adoption was known for 343 cats (78.7%). Of these 343 cats, 110 (32%) were previously strays (100 cats) or pound cats (10 cats) and 233 cats (68%) were given by relatives or friends (179 cats), bought in stores or from cat breeders (35 cats) or born at home (19 cats). 3.2. Bacteriological and serological data Seventy-two cats (16.5%) were Bartonella bacteremic and 179 cats (41%) were seropositive to B. henselae and/or B. clarridgeiae. Of the 72 bacteremic cats, 62 cats were

190

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

Table 1 Prevalence of Bartonella bacteremia in 436 domestic French cats, Paris and environs, February±June 1996 Bartonella species or type

Number of bacteremic cats and percentage (%)

B. B. B. B. B. B.

11 36 15 2 3 5

henselae type henselae type clarridgeiae henselae type henselae type henselae type

I II I and type IIa I and B. clarridgeiaeb II and B. clarridgeiaec

Total

(2.5) (8.3) (3.4) (0.5) (0.7) (1.1)

72 (16.5)

Isolate percentages among bacteremic cats 15.3 50.0 20.8 2.8 4.2 6.9 100.0

a

Cats previously reported in Gurfield et al. (1997). Two of three cats previously reported in Gurfield et al. (1997). c Three of five cats previously reported in Gurfield et al. (1997). b

only infected with B. henselae type I (B.h. I), B. henselae type II (B.h. II) or B. clarridgeiae (B.c.) (Table 1). Eight cats (11%) were co-infected with B. henselae and B. clarridgeiae. Two cats (2.8%) were concurrently bacteremic with two different B. henselae types (B.h. I/B.h. II). Bacteremia concentrations ranged from 100 to more than 4000 colony forming units (CFUs)/ml. Titers against B. henselae ranged from 0 to greater than 1024 and titers against B. clarridgeiae ranged from 0 to 1024. The titer end points of the seven cats with titers to B. henselae greater than 1024 could not be determined due to exhaustion of the serum banked from these cats. Eighty-four cats (19.3%) had positive antibody titers to both B. henselae and B. clarridgeiae. Fifty-four cats (12.4%) had positive titers to B. henselae only, and 26 cats (5.9%) had positive titers to B. clarridgeiae only. Two hundred and twenty-eight cats (52.3%) were seronegative to either B. henselae or B. clarridgeiae and in 44 cats, titers to B. henselae but not to B. clarridgeiae were obtained due to lack of suf®cient serum to perform both tests. Titers to B. clarridgeiae were on average higher than titers to B. henselae (reciprocal B. henselae positive titer geometric mean 128, reciprocal B. clarridgeiae positive titer geometric mean 256). Seven (9.7%) of the 72 bacteremic cats were seronegative to either B. henselae or B. clarridgeiae. Of these seven cats, three were less than 1 year old. Interestingly, 65 (36.3%) of the 179 seropositive cats were Bartonella bacteremic versus seven (2.7%) of the 257 seronegative cats (predictive value of a negative test: 97.3%). No correlation was detected between seropositivity to B. henselae versus B. clarridgeiae and the species of Bartonella that was isolated. 3.3. Univariate analysis Cats owned for <6 months were more likely to be bacteremic (25.8%; 23 of 89 cats) than cats owned for 6 months (14.4%; 37 of 257 cats) (prevalence ratio …PR† ˆ 1:80, 95% con®dence interval …CI† ˆ 1:13, 2.85). The prevalence of bacteremia decreased with the length of pet ownership (Fig. 1). Cats owned for less than a month also had the highest seroprevalence (54.5%, six of 11 cats). The duration of cat ownership (6 months or more versus <6 months) was not dependent on the adoption origin of the cat …p ˆ 0:186†.

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

191

Fig. 1. Percentage of Bartonella bacteremia and seropositivity as a function of time of cat ownership.

Cats adopted from the pound or found as a stray were more likely to be Bartonella bacteremic (23.6%, 26 of 110 cats) than cats given by friends or family members, obtained from a store or a breeder or born at home (14.2%, 33 of 233 cats) (PR ˆ 1:67, 95% CI ˆ 1:05, 2.65). Cats adopted from the pound or previously stray cats were also more likely to be seropositive (49.5%, 54 out of 109 cats) than cats given by a friend or family member, cats obtained from a store or a breeder, and cats born at home (89 out of 230 cats) (PR ˆ 1:28, 95% CI ˆ 1:00, 1.64) (Fig. 2).

Fig. 2. Percentages of cats from different adoption origins, bacteremic and seropositive to Bartonella organisms.

192

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

Table 2 Prevalence ratios of Bartonella bacteremic and seropositive cats by number of cats per household No. of cats in household

1 cat 2 cats 3 cats 4 cats a b

Bacteremic

Seropositive

PRa

(95% CIb)

PR

(95% CI)

1.0 1.60 1.95 2.73

(1.01, 2.53) (1.15, 3.29) (1.57, 4.75)

1.0 1.45 1.33 1.79

(1.14, 1.85) (0.98, 1.82) (1.32, 2.42)

Prevalence ratio. Confidence interval.

Cats living in households with one or more cats were more likely to be bacteremic (PR ˆ 1:60, 95% CI ˆ 1:01, 2.53) and seropositive (PR ˆ 1:45, 95% CI ˆ 1:14, 1.85) than cats living without other cats. The prevalence ratio of bacteremia and seropositivity increased with increasing numbers of cats owned (p ˆ 0:02 for bacteremia, and p ˆ 0:016 for seropositivity by chi-square test for trend) (Table 2). Non-neutered cats were twice as likely to be bacteremic than neutered cats (18.6% versus 9.2%). However, the difference was not statistically signi®cant (PR ˆ 2:01, 95% CI ˆ 0:85, 4.78). Similarly, cats reported to have been infested with ¯eas during the preceding 6 months were also more likely to be seropositive than cats without ¯eas during the same period (48.7% versus 39%, PR ˆ 1:25, 95% CI ˆ 0:98, 1.59). The breed of cat, sex, environment, housing type, number of dogs living in household, type of ¯ea control employed, and reason for veterinary visit were not risk factors associated with bacteremia nor seropositivity. Additionally, there was no signi®cant difference in the adoption origin of cats living in homes or apartments (i.e. originating from pound or stray versus other sources, p ˆ 0:58). 3.4. Multiple logistic regression analysis After adjustment for age and ¯ea infestation, multiple logistic regression analysis indicated that bacteremia was associated with origin of adoption (stray) and number of cats in the household. Similarly, after adjustment for age, multiple logistic regression analysis indicated that seropositivity was associated with ¯ea infestation, origin of adoption (pound) and number of cats in the household (Tables 3 and 4). 4. Discussion Domestic cats are the main reservoir for two Bartonella species that cause cat scratch disease in humans: B. henselae and B. clarridgeiae (Clarridge et al., 1995; Koehler et al., 1994; Chomel et al., 1995a; Bergmans et al., 1997; Heller et al., 1997). They also can carry the newly identi®ed B. koehlerae, whose pathogenicity in humans remains to be determined (Droz et al., 1999). We studied a population of French domestic cats to determine the prevalences and associated risk factors for Bartonella infection, and to

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

193

Table 3 Multivariate logistic regression odd ratios (OR) and 95% con®dence intervals (CI) for variables associated with Bartonella bacteremia in French catsa Variables

OR

(95% CI)

Age >12 months old 6±12 months 0±6 months

1 2.00 1.83

(0.84, 4.77) (0.17, 19.96)

Flea infestation No Yes

1 1.15

(0.63, 2.12)

Origin of adoption Born at home Pound Stray Family/friend Store/breeder

1 1.92 19.19 1.33 1.33

(0.63, (3.23, (0.46, (0.27,

5.81) 114.16)b 3.88) 6.61)

Length of ownership 12 months 6±11 months 1±5 months <1 month

1 0.66 1.73 1.6

(0.24, 1.85) (0.60, 4.98) (0.33, 7.68)

Number of cats in household (continuous variable)

1.3

(1.03, 1.64)b

a b

Data missing for 109 cats. p < 0:05.

identify endemic Bartonella species. From a convenience sample of 436 domestic cats living in Paris and its environs, 16.5% were bacteremic with either B. henselae or B. clarridgeiae. These results are within the range of previous studies on Bartonella bacteremia in European pet cat populations. Chomel et al. (1995b) previously reported an 11% prevalence in French pet cats and Sander et al. (1997) reported a 13% prevalence in German pet cats. In The Netherlands, Bergmans et al. (1997) reported a slightly higher prevalence of 22%, but all bacteremic cats were from animal shelters. The Bartonella seroprevalence in these shelter cats was 50%. Similarly, in our study, cats adopted from a shelter or found as a stray cat, had an increased prevalence of bacteremia (23.6%) and seropositivity (49.5%). It is not surprising that the highest prevalence came from an animal shelter population of cats. Previous studies have demonstrated that cats from shelters are more likely to be bacteremic (Koehler et al., 1994; Chomel et al., 1995a). Our ®ndings are also supported by the very high bacteremia prevalence reported in stray cats in France (53%) (Heller et al., 1997), in the Philippines (61%) (Chomel et al., 1999), in Indonesia (43%) (Marston et al., 1999) and in Australia (40%) (Branley et al., 1996). B. henselae type II was the most frequent Bartonella species isolated in this study, accounting for 50% of isolates. It also appears to be the most common B. henselae type in cats from Western Europe (Bergmans et al., 1997; Heller et al., 1997; Sander et al., 1997). In contrast, B. henselae type II was not found in any of the 19 bacteremic cats from the

194

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

Table 4 Multivariate logistic regression odd ratios (OR) and 95% con®dence intervals (CI) for variables associated with Bartonella seropositivity in French catsa Variables

OR

(95% CI)

Age >12 months old 6±12 months 0±6 months

1 1.05 1.79

(0.65, 1.68) (0.24, 13.40)

Flea infestation No Yes

1 1.60

(1.01, 2.54)b

Origin of adoption Born at home Pound Stray Family/friend Store/breeder

1 2.24 3.14 1.71 0.79

(0.99, (0.72, (0.78, (0.23,

Number of cats in household (continuous variable)

1.27

(1.01, 1.60)b

a b

5.06)b 13.66) 3.71) 2.65)

Data missing for 113 cats. p  0:05.

Philippines (Chomel et al., 1999), and was present in only one (2%) of the 50 bacteremic cats from Japan (Maruyama et al., 2000). We found a prevalence of 21% for B. clarridgeiae, which is higher than reported in shelter cats from The Netherlands (16%, four of 25 cats) (Bergmans et al., 1997), but lower than in French stray cats (30%) (Heller et al., 1997). Other geographical differences in Bartonella species distribution have been reported for B. clarridgeiae, which only accounted for 10% of isolates in the USA (Kordick et al., 1997b) and 13.6% in Japan (Maruyama et al., 2000). Co-infected cats with B. henselae and B. clarridgeiae accounted for 11% of our bacteremic cats, a percentage quite lower than the 20% reported by Bergmans et al. (1997). It will be important to determine if human cases of CSD in France are more frequently associated with B. henselae type I, as in The Netherlands and Germany, in spite of B. henselae type II being more prevalent than B. henselae type I in Western European cats. This would further support the hypothesis of pathogenicity differences among Bartonella species and types (Bergmans et al., 1996, 1997; Relman, 1998; Sander et al., 1998). Several major ®ndings in this study have important consequences for safe cat ownership and adoption, especially for prospective cat owners who are immunocompromised and are thus more susceptible to developing severe Bartonella infections. This is the ®rst report indicating that cat ownership for <6 months is associated with a higher risk of having a bacteremic cat. Secondly, we also demonstrated for the ®rst time an increased risk of bacteremia or seropositivity in cats from multicat households. Cats living in multicat households have an increased risk for repeated infections by different Bartonella species and types. Similarly, the probability of housing cats that are long-term bacteremic cats or co-infected cats is likely to increase. Strict ¯ea control is also more dif®cult to accomplish in a multicat environment, especially if the cats are allowed to roam outdoors.

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

195

A third important risk factor was adopting a cat from the pound or from the street. Prudence dictates, therefore, that people at risk for developing severe Bartonella infections, i.e. immunocompromised individuals, should refrain from adopting pound or stray cats. Furthermore, it is recommended that immunocompromised individuals do not own several cats, as multiple cat ownership proportionally increases the risk of owning Bartonella bacteremic cats. Con®rming previous studies demonstrating that the cat ¯ea, C. felis, is the vector of B. henselae between cats (Zangwill et al., 1993; Chomel et al., 1996; Higgins et al., 1996; Foil et al., 1998), was the positive association of seropositivity with the history of ¯ea infestation. However, we did not identify an association between bacteremia and ¯ea infestation, in contrast to a previous report (Chomel et al., 1995a). The lack of association between ¯ea infestation and bacteremia may be due to several factors. Our study was performed mainly during cold months, when ¯ea infestation is reduced. Owners of bacteremic cats may also have underestimated the level of infestation of their pets. Therefore, Bartonella antibodies, which persist much longer than bacteremia, may have more accurately re¯ected prior ¯ea infestation and the origin of the cat infection. Some reports have indicated an association between the sex of the cat and bacteremia or association with cat scratch disease cases. These associations have been both with female cats (Sander et al., 1997) and male cats (Zangwill et al., 1993). However, several other studies, including ours, have not determined any sex differences (Glaus et al., 1997; Heller et al., 1997; Foley et al., 1998; Haimerl et al., 1999). Young cats have also been reported to be more likely infected with Bartonella than older cats (Koehler et al., 1994; Chomel et al., 1995a; Sander et al., 1997). We did not observe such a difference in this study. However, most of our cats were less than 1.5 years old, which may explain this lack of association. Seven of the bacteremic cats (9.7%) in our study were seronegative. These cats may have been recently infected and not had time to develop detectable IgG titers or possibly, were cats transiently non-seroreactive (Kordick et al., 1997b). Although the predictive value of seropositive cats being bacteremic was only 36.3%, 97.3% of seronegative cats were abacteremic by our culture methods. The negative predictive value of a cat being seronegative appears to be a more reliable predictor of absence of bacteremia than seropositivity is for bacteremia, as previously reported (Chomel et al., 1995a). Furthermore, the presence of Bartonella antibodies may not be protective against infection (Chomel et al., 1995a; Kordick and Breitschwerdt, 1997). In conclusion, an ideal cat for adoption, especially by an immunocompromised individual, would be a cat, not from the pound nor from the street, but from a donor who has owned the cat for at least 6 months, in a single-cat household, free of ¯eas. Acknowledgements We gratefully thank Daniele Couillard from the Bacteriology and Immunology Laboratory at Ecole Nationale VeÂteÂrinaire d'Alfort, Maisons-Alfort, France, for her technical expertise. We also thank Drs. Fradin and Fayolle, the Alfort Surgery Department, the Alfort veterinary students, and the pet owners for their assistance with

196

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

this work. Nikos Gur®eld was a Chateaubriand grant recipient (French Ministry of Foreign Affairs). This project was also supported by a grant from MeÂrial, Lyon, France. References Abbott, R.C., Chomel, B.B., Kasten, R.W., Floyd-Hawkins, K.A., Kikuchi, Y., Koehler, J.E., Pedersen, N.C., 1997. Experimental and natural infection with Bartonella henselae in domestic cats. Comp. Immunol. Microbiol. Infect. Dis. 20, 41±51. Allerberger, F., Schonbauer, M., Zangerle, R., Dierich, M., 1995. Prevalence of antibody to Rochalimaea henselae among Austrian cats. Eur. J. Ped. 154, 165. Baneth, G., Kordick, D.L., Hegarty, B.C., Breitschwerdt, E.B., 1996. Comparative seroreactivity to Bartonella henselae and Bartonella quintana among cats from Israel and North Carolina. Vet. Microbiol. 50, 95±103. Baorto, E., Payne, R.M., Slater, L.N., Lopez, F., Relman, D.A., Min, K.W., St. Geme III, J.W., 1998. Culturenegative endocarditis caused by Bartonella henselae. J. Ped. 132, 1051±1054. Bergmans, A.M.C., Schellenkens, J.F.P., van Embden, J.D.A., Schouls, L.M., 1996. Predominance of two Bartonella henselae variants among cat-scratch disease patients in The Netherlands. J. Clin. Microbiol. 34, 254±260. Bergmans, A.M.C., de Jong, C.M.A., van Amerongen, G., Schot, C.S., Schouls, L.M., 1997. Prevalence of Bartonella species in domestic cats in The Netherlands. J. Clin. Microbiol. 35, 2256±2261. Branley, J., Wolfson, C., Waters, P., Gottlieb, T., Bradbury, R., 1996. Prevalence of Bartonella henselae bacteremia, the causative agent of cat scratch disease, in an Australian cat population. Pathology 28, 262± 265. Brow, M.A.D., 1990. Sequencing with Taq DNA polymerase. In: Brow, M.A.D., Innis, M.A., Gelfand, D.H., Sninsky, J.J., White, T.J. (Eds.), PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, CA, pp. 189±196. Childs, J.E., Olson, J.G., Wolf, A., Cohen, N., Fakile, Y., Rooney, J.A., Bacellar, F., Regnery, R.L., 1995. Prevalence of antibodies to Rochalimaea species (cat-scratch disease agent) in cats. Vet. Rec. 20, 519±520. Chomel, B.B., Abbott, R.C., Kasten, R.W., Floyd-Hawkins, K.A., Kass, P.H., Glaser, C.A., Pedersen, N.C., Koehler, J.E., 1995a. Bartonella henselae prevalence in domestic cats in California: risk factors and association between bacteremia and antibody titers. J. Clin. Microbiol. 33, 2445±2450. Chomel, B.B., Gur®eld, A.N., Boulouis, H.J., Kasten, R.W., Piemont, Y., 1995b. ReÂservoir feÂlin de l'agent de la maladie des griffes du chat, Bartonella henselae, en reÂgion parisienne: reÂsultats preÂliminaires. Rec. MeÂd. VeÂt. 171, 841±845. Chomel, B.B., Kasten, R.W., Floyd-Hawkins, K., Chi, B., Yamamoto, K., Roberts-Wilson, J., Gur®eld, A.N., Abbott, R.C., Pedersen, N.C., Koehler, J.E., 1996. Experimental transmisson of Bartonella henselae by the cat ¯ea. J. Clin. Microbiol. 34, 1952±1956. Chomel, B.B., Carlos, E.T., Kasten, R.W., Yamamoto, K., Chang, C.C., Carlos, R.S., Abenes, M.V., Pajares, C.M., 1999. Bartonella henselae and Bartonella clarridgeiae infection in domestic cats from The Philippines. Am. J. Trop. Med. Hyg. 60, 593±597. Clarridge III, J.E., Raich, T.J., Pirwani, D., Simon, B., Tsai, L., Rodriguez-Barradas, M.C., Regnery, R., Zollo, A., Jones, D.C., Rambo, C., 1995. Strategy to detect and identify Bartonella species in routine clinical laboratory yields Bartonella henselae from human immunode®ciency virus-positive patient and unique Bartonella strain from his cat. J. Clin. Microbiol. 33, 2107±2113. De Lamballerie, X., Zandotti, C., Vignoli, C., Bollet, C., de Micco, P., 1992. A one-step microbial DNA extraction method using ``Chelex 100'' suitable for gene ampli®cation. Res. Microbiol. 143, 785±790. Dolan, J.M., Wong, M.T., Regnery, R.L., Jorgensen, J.H., Garcia, M., Peters, J., Drehner, D., 1993. Syndrome of Rochalimaea henselae adenitis suggesting cat scratch disease. Ann. Intern. Med. 118, 331±336. Droz, S., Banghee, C., Horn, E., Steigerwalt, A.G., Whitney, A.M., Brenner, D.J., 1999. Bartonella koehlerae sp. nov., isolated from cats. J. Clin. Microbiol. 37, 1117±1122. Foil, L., Andress, E., Freeland, R.L., Roy, A.F., Rutledge, R., Triche, P.C., O'Reilly, K.L., 1998. Experimental infection of domestic cats with Bartonella henselae by inoculation of Ctenocephalides felis (Siphonaptera: Pulicidae) feces. J. Med. Entomol. 35, 625±628.

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

197

Foley, J.E., Chomel, B.B., Kikuchi, Y., Yamamoto, K., Pedersen, N.C., 1998. Seroprevalence of Bartonella henselae in cattery cats: association with cattery hygiene and ¯ea infestation. Vet. Quart. 20, 1±5. George, T.I., Manley, G., Koehler, J.E., Hung, V.S., McDermott, M., Bollen, A., 1998. Detection of Bartonella henselae by polymerase chain reaction in brain tissue of an immunocompromised patient with multiple enhancing lesions. J. Neurosurg. 89, 640±644. Glaus, T., Hofmann-Lehmann, R., Greene, C., Glaus, B., Wolfensberger, C., Lutz, H., 1997. Seroprevalence of Bartonella henselae infection and correlation with disease status in cats in Switzerland. J. Clin. Microbiol. 35, 2883±2885. Guptill, L., Slater, L., Wu, C.C., Lin, T.L., Glickman, L.T., Welch, D.F., HogenEsch, H., 1997. Experimental infection of young speci®c pathogen-free cats with Bartonella henselae. J. Infect. Dis. 176, 206±216. Guptill, L., Slater, L.N., Wu, C.C., Lin, T.L., Glickman, L.T., Welch, D.F., Tobolski, J., HogenEsch, H., 1998. Evidence of reproductive failure and lack of perinatal transmission of Bartonella henselae in experimentally infected cats. Vet. Immunol. Immunopathol. 65, 177±189. Gur®eld, A.N., Boulouis, H.J., Chomel, B.B., Heller, R., Kasten, R.W., Yamamoto, K., Piemont, Y., 1997. Coinfection with Bartonella clarridgeiae and Bartonella henselae and with different Bartonella henselae strains in domestic cats. J. Clin. Microbiol. 35, 2120±2123. Haimerl, M., Tenter, A.M., Simon, K., Rommel, M., Hilger, J., Autenrieth, I.B., 1999. Seroprevalence of Bartonella henselae in cats in Germany. J. Med. Microbiol. 48, 849±856. Hashkes, P.J., Trabulsi, A., Passo, M.H., 1996. Systemic cat-scratch disease presenting as leukocytoclastic vasculitis. Ped. Infect. Dis. J. 15, 93±95. Heller, R., Artois, M., Xemar, V., de Briel, D., Gehin, H., Jaulhac, B., Monteil, H., Piemont, Y., 1997. Prevalence of Bartonella henselae and Bartonella clarridgeiae in stray cats. J. Clin. Microbiol. 35, 1327±1331. Higgins, D.G., Sharp, P.M., 1988. Clustal: a package for performing multiple sequence alignment on a microcomputer. Gene 73, 237±244. Higgins, J.A., Radulovic, S., Jaworski, D.C., Azad, A.F., 1996. Aquisition of the cat scratch disease agent Bartonella henselae by cat ¯eas (Siphonaptera: Pulicidae). J. Med. Entomol. 33, 490±495. Jameson, P., Greene, C., Regnery, R., Dryden, M., Marks, A., Brown, J., Cooper, J., Glaus, B., Greene, R., 1995. Prevalence of Bartonella henselae antibodies in pet cats throughout regions of North America. J. Infect. Dis. 172, 1145±1149. Koehler, J.E., Frederick, D.Q., Berger, T.G., LeBoit, P.E., Tappero, J.W., 1992. Isolation of Rochalimaea species from cutaneous and osseous lesions of bacillary angiomatosis. New England J. Med. 327, 1625±1631. Koehler, J.E., Glaser, C.A., Tappero, J.W., 1994. Rochalimaea henselae infection. A new zoonosis with the domestic cat as reservoir. J. Am. Med. Assoc. 271, 531±535. Kordick, D.L., Breitschwerdt, E.B., 1997. Relapsing bacteremia after blood transmission of Bartonella henselae to cats. Am. J. Vet. Res. 58, 492±497. Kordick, D.L., Wilson, K.H., Sexton, D.J., Had®eld, T.L., Berkhoff, H.A., Breitschwerdt, E.B., 1995. Prolonged Bartonella bacteremia in cats associated with cat-scratch disease patients. J. Clin. Microbiol. 33, 3245± 3251. Kordick, D.L., Hilyard, E.J., Had®eld, T.L., Wilson, K.H., Steigerwalt, A.G., Brenner, D.J., Breitschwerdt, E.B., 1997a. Bartonella clarridgeiae, a newly recognized zoonotic pathogen causing inoculation papules, fever and lymphadenopathy (cat scratch disease). J. Clin. Microbiol. 35, 1813±1818. Kordick, D.L., Papich, M.G., Breitschwerdt, E.B., 1997b. Ef®cacy of enro¯oxacin or doxycycline for treatment of Bartonella henselae or Bartonella clarridgeiae infection in cats. Antimicrob. Agents Chemoth. 41, 2448± 2455. Lawson, P.A., Collins, M.D., 1996. Description of Bartonella clarridgeiae sp. nov. isolated from the cat of a patient with Bartonella henselae septicemia. Med. Microbiol. Lett. 5, 64±73. Margileth, A.M., Baehren, D.F., 1998. Chest-wall abscess due to cat-scratch disease (CSD) in an adult with antibodies to Bartonella clarridgeiae: case report and review of the thoracopulmonary manifestations of CSD. Clin. Infect. Dis. 27, 353±357. Marston, E.L., Finkel, B., Regnery, R.L., Winoto, I.L., Graham, R.R., Wignal, S., Simanjuntak, G., Olson, J.G., 1999. Prevalence of Bartonella henselae and Bartonella clarridgeiae in an urban Indonesian cat population. Clin. Diag. Lab. Immunol. 6, 41±44. Maruyama, S., Nakamura, Y., Kabeya, H., Tanaka, S., Sakai, T., 2000. Prevalence of Bartonella henselae,

198

A.N. Gurfield et al. / Veterinary Microbiology 80 (2001) 185±198

Bartonella clarridgeiae and the 16S rRNA gene types of Bartonella henselae among pet cats in Japan. J. Vet. Med. Sci. 62, 273±279. O'Reilly, K.L., Bauer, R.W., Freeland, R.L., Foil, L.D., Hughes, K.J., Rohde, K.R., Roy, A.F., Stout, R.W., Triche, P.C., 1999. Acute clinical disease in cats following infection with a pathogenic strain of Bartonella henselae (LSU16). Infect. Immun. 67, 3066±3072. Regnery, R.L., Anderson, B.E., Clarridge III, J.E., Rodriguez-Barradas, M.C., Jones, D.C., Carr, J.H., 1992a. Characterization of a novel Rochalimaea species, R. henselae sp. nov., isolated from blood of a febrile human immunode®ciency virus-positive patient. J. Clin. Microbiol. 30, 265±274. Regnery, R.L., Martin, M., Olson, J.G., 1992b. Naturally occurring Rochalimaea henselae infection in domestic cat. Lancet 340, 557±558. Regnery, R.L., Olson, J.G., Perkins, B.A., Bibb, W., 1992c. Serological response to ``Rochalimaea henselae'' antigen in suspected cat-scratch disease. Lancet 339, 1443±1445. Relman, D.A., 1998. Editorial response: are all Bartonella henselae strains created equal? Clin. Infect. Dis. 26, 1300±1301. Relman, D.A., Loutit, J.S., Schmidt, T.M., Falkow, S., Tompkins, L.S., 1990. The agent of bacillary angiomatosis: an approach to the identi®cation of uncultured pathogens. New England J. Med. 323, 1573± 1580. Sander, A., BuÈhler, C., Pelz, K., von Cramm, E., Bredt, W., 1997. Detection and identi®cation of two Bartonella henselae variants in domestic cats in Germany. J. Clin. Microbiol. 35, 584±587. Sander, A., Ruess, M., Bereswill, S., Schuppler, M., Steinbrueckner, B., 1998. Comparison of different DNA ®ngerprinting techniques for molecular typing of Bartonella henselae isolates. J. Clin. Microbiol. 36, 2973± 2981. Slater, L.N., Welch, D.F., Min, K.W., 1992. Rochalimaea henselae causes bacillary angiomatosis and peliosis hepatis. Arch. Intern. Med. 152, 602±606. Slater, L.N., Pitha, J.V., Herrera, L., Hughson, M.D., Min, K.W., Reed, J.A., 1994. Rochalimaea henselae infection in acquired immunode®ciency syndrome causing in¯ammatory disease without angiomatosis or peliosis. Demonstration by immunocytochemistry and corroboration by DNA ampli®cation. Arch. Pathol. Lab. Med. 118, 33±38. Ueno, H., Muramatsu, Y., Chomel, B.B., Hohdatsu, T., Koyama, H., Morita, C., 1995. Seroepidemiological survey of Bartonella (Rochalimaea) henselae in domestic cats in Japan. Microbiol. Immunol. 39, 339±341. Welch, D.F., Pickett, D.A., Slater, L.N., Steigerwalt, A.G., Brenner, D.J., 1992. Rochalimaea henselae sp. nov., a cause of septicemia, bacillary angiomatosis, and parenchymal bacillary peliosis. J. Clin. Microbiol. 30, 275± 280. Zangwill, K.M., Hamilton, D.H., Perkins, B.A., Regnery, R.L., Plikaytis, B.D., Hadler, J.L., Cartter, M.L., Wenger, J.D., 1993. Cat scratch disease in Connecticut. Epidemiology, risk factors, and evaluation of a new diagnostic test. New England J. Med. 329, 8±13.