Article 234 0.1 CEC

Jeffrey J. Windsor, AbenJstwyth Public Hellltlr LaboflltOMJ, Brollg/nis Hospital,

Abel1jstwy l'h, Wales

Cat-scratch Disease: Epidemiology, Etiology, and Treabnent

Jeffrey J. Windsor

Cat-scratch disease (CSD) is a clinical syndrome that usually presents as a self-limiting lym­phadenopathy associated with a cat scratch or bite. Commonly affecting children and young adults, it has a worldwide distribution. In tem­perate climates, higher rates are reported in the autumn and winter, which can be attributed to the seasonal breeding of the domestic cat. The ~rganism responsible was identified in 1983, having eluded detection for 50 years. Initially, Afipia felis was thought to be the cause, how­ever, subsequent study failed to confirm a link. During the 1990s, it was demonstrated conclu­sively that Rochalimaea hellselae,later reclassi­fied as Bartonella henselae, was the cause of CSD. B. henselae has been isolated from bacteremic cats, with transmission among cats thought to be via the cat flea. Although other Bartonella species are transmitted by arthropod vectors, it is unlikely that the cat flea is involved directly in human infection, but plays a role in amplifying the reservoir. B. henselae is difficult to culture, and either serology or the polymerase chain reaction are considered to be the best methods of detection. Genetic variation occurs among B. henselae strains, perhaps explaining the inconsistency of some diagnostic techniques. A separate serogroup (Marseilles) has been re­ported in a seronegative patient with CSD, and B. clarridgeiae has the potential to cause the disease. Atypical presentation is seen in up to 25% of cases, and manifests itself as ocular involvement, encephalopathy, granulomatous hepatitis, hepatosplenic infection, endocardi­tis, and osteomyelitis. The majority of CSD cases resolve spontaneously and do not require antibiotic treatment. In complicated CSD, treat­ment with trimethoprim-sulp.hamethoxazole, ciprofloxacin, or azithromycin is recommended, wi th gentamicin being reserved for the severely ill patient.

Editor's Note: Reprinted from. the British )o!1Il1n1 of Bio­medico I Science 200];58:101-110. Biomedical EssiJY com­missioned by the Editor, British JO l/ rna l of Biomedicnl Sci­enc~, and based on a structured reading essay submitted as part of the ISMS Continuing Professional Develop­ment Program. Accepted 16 February 2001.

Introduction Cat-scratch disease (CSD) is characterized

usually as a self-limiting regional lymphaden­opathy, associated with a cat scratch or bite. l

Although CSD is normally benign, the causative organism(s) may cause severe systemic or recur­ren t disease. l Originally considered rare, it is now recognized to affect children and young adults conunonly, and occurs worldwide. Indeed, in some areas itis the most common cause of chronic lymphadenopathy in this age group.2 CSD oc­curs more frequently in certain geographical ar­eas and shows a marked seasonality.3 In the USA, it is estimated that over 24,000 cases occur annu­ally, with 2,000 patients requiring hospitaliza­tion, and a resulting health-care expenditure of $12 million.4

Despite many clinical descriptions (750 by the early 1980s)/ the etiological agent remained elusive. Various organisms have been implicated since the early 1900s, including viruses, chlamy­dia, acid-fast bacilli, and gram-positive bacte­ria;'I.6 and it was not proven to be a bacterial infection until 1983.5 However, the isolation and characterization of the causative organism(s) would prove to be more difficult. Indeed, it is only in the past decade, using powerful molecu­lar techniques, that the true etiological picture has emerged.

Historical overview CSD was described initially by Debre and

coworkers7 in 1950, although Parinaud is cred­ited with the first report of symptoms consistent with CSD some 60 years earlier, when he de­scribed Pal'maud's oculoglandular syndrome in 1889.8 Debre examined his firs t patient, a 10"year­old boy who presented with suppura tive adeni­tis and severe cat scratches, in 1931.

Although physicians in the USA were aware of patients with CSD as early as 1932, the first published American case was not reported until 1951.9 Foshay, a professor of microbiology at the University ofCincllu1ati, produced a crude intra­dermal skin test in 1947.3•9 TI1e antigen used was derived from pus obtained from the lymph nodes of individuals suspected of having the disease.

Professor Mollaret, at the Pasteur Institute, stud­ied CSD independently, and also prepared a skin antigen test.3

Many extensive clinical reviews followed, including Warwick'slO description of 160 cases in 1960, and Carithers'll review of 1,200 cases, pub­lished in 1985. Surprisingly, the clinical descrip­tion of CSD changed litUe over the decades. To date, over 900 publications have appeared in the scientific literature, reflecting the high level of interest shown in this intriguing syndrome.3,12

Etiology The etiological agent of CSD eluded detec­

tion for 50 years. Although man y infectious agents were suspected,3 it was not until 1983 that Wear and coworkerss demonstrated that it was a bacte­rial infection. Using the War thin-Starry silver stain, they detected small, delicate pleomorphic bacilli in 34 out of 39 lymph nodes taken from patients with suspected CSD. Using an irnmunoperoxidase staining technique, the or­ganisms in sections of lymph node stained in­tensely when exposed to convalescent sera from CSD patients, whereas they did not react with control sera.

These results were supported by the work ofMargileth and coworkers,13 who used the same staining methods and detected bacilli in both biopsies of inoculation papules and the lymph nodes draining the site of inoculation in CSD patients.

In 1988, English and coworkersl isolated a bacterial organism from the lymph nodes of ten patients with CSD, after extended incubation on specialized media. After inoculation into a nine­banded armadillo, it caused dermal lesions iden­tical to those seen in humans. The group was able to reisolate the organism from these lesions, thereby fulfilling Koch's postulates, and consid­ered it to be the causative agent of CSD.

Initially, it was known simply as the "cat­scratch disease bacillus./I14 However, in 1991, when Brenner and coworkers 15 described the genus Afipia, it was given the specific name Afipia felis. Afipia was taken from the Armed Forces Institute of Pathology, where the original organ­ism was isolated; and feIis referred to the cat as the probable vector.3 Subsequent reports failed to confirm a strong link between A. felis and CSD, however.16 Indeed, further doubt was cast when patients with CSD failed to mount either a hu­moral or cellular response to A. felis antigen.3 In addition, other investigators were unable to iso­late A. felis from CSD patients; and, despite most patients reporting exposure to cats, no clear link could be demonstrated between cats andA.felis.

m 1990, ReIman and coworkers17 using a

novel method for identifying uncultured patho­gens, identified the causative agent of bacillary angiomatosis (BA). This is an infectious disease that causes cutaneous lesions, and primarily af­fects human immunodeficiency virus (HIV)-posi­tive patients. IS The organism was visualized pre­viously in infected tissue stained using the Warthin-Starry method, and looked similar to that associated with CSD.

Using the polymerase chain reaction (PCR), this group17 amplified prokaryotic 16S ribosomal RNA (rRNA) gene products directly from tissue samples. The primers used were common to all bacteria, and the amplified DNA was sequenced and found to be related to Rochalimaea quintana.

At the same time, in Oklahoma, Slater and coworkers 19 isolated Rochalimaea -like organisms from the blood of two HIV -positive patients pre­senting with persistent fever. Similar organisms were isolated subsequently from blood and lymph;2D.2l and, after genetic analysis, were named as a new species, R. 11ellselae. 19,2D,22 Then, Koehler and coworkersls succeeded in isolating R. he/1seJae from the cutaneous lesions of patients with SA.

Recently, the genus Rochalimaeawas united with Bartonella, and Rocltalimaea henselae is now reclassified as Bartonella he/1selae.23

An indirect fluorescence antibody (lFA) technique was developed to detect B. henselae antibodies. 24 Used initially on serum samples from both HIV-infected BA patients and HIV­infected controls, several BA patients showed high-titre antibodies, but only one HIV-infected control gave a strong reaction, Perhaps it was fortuitous that this patient had CSD.

Subsequently, Regnery and coworkers25

used this IFA test on 41 patients suspected of having CSD, and 36 (88%) had serum titres to B. henseiae antigen of >1 in 64, Furthermore, there was only a low prevalence (3%) of significant tit res in healthy controls. Interestingly, when the sera was tested against A. felis, very few samples gave significant titres.

Anderson and coworkers26 confirmed the association between B, 11enseiae and CSD by PCR amplification of B.lzenselae DNA from the antigen used for skin testing. Bergmans and coworkers27

compared PCR detection of B. henselae andA.Jelis DNA with serology and skin tests. They could not detect A. felis DNA in clinical samples, and concluded that CSD is caused by bartonellae.

Given the body of evidence, it would ap­pear that B. llenselae is the primary cause of CSD, rather thanA.felis. However, Alkan and cowork­ers amplified th~ 16S rRNA gene from lymph node sections in 12 patients with histological features of CSD, and detected B, llenselae in five, A.felis in three, and both organisms in two. Giladi

and coworkers29 c™orrunented on the rare role of A. felis in CSD, after isolating the organism from a lymph node of a patient with CSD. Although they were unable to detect A.felis DNA from 32 B. henselae DNA-positive patients with CSD, they felt that A. felis might playa role in its pathogen­esis. La Scola and Raoult3'l isolated A. felis from a hospital water supply, and suggested thatprevi­ous isolation of A. felis from lymph nodes may have been due to contamination.

In 1997, a newly described organism, B. clarridgeiae, was recognized as causing CSD in a veterinarian.6 A Bartonella-like organism was isolated from the patient's cat, and the patient's sera did not react against B. henselae, B. quintana, or B. elizabetlwe, but did react (titre: 1 in 24) against the cat isolate. The following year, Margileth and Baehren31 described another adult patient with a chest-wall abscess due to CSD, and antibodies to B. clarridgeiae.

Animal reservoir Although the association of CSD with cats

is an obvious one, given the name of the syn­drome, it was only proven conclusively in the 19905. In 1992, Regnery and coworker?32 isolated B. hel1selae from an asymptomatic cat on two occasions, and suggested that the domestic cat (Felis domesticus) may be a reservoir for human Bartonella disease.

Zangwill and coworkers33 compared CSD patients with age-matched, cat-owning subjects, and found that CSD patients were more likely to own a kitten, been scratched or bitten by a kitten, and to have at least one kitten infested with fleas. Furthermore, serological tests demonstrated that 39 of 48 (81%) patients' cats had antibodies to B. henselae, compared with 11 of 29 (38%) control cats.

Sander and coworkers31 investigated blood cultures from 100 household cats, and isolated B. henselae from 13%. They found that positive cul­tures were more likely to be found in young female cats «24 months) than in male and older cats.

Koehler and coworkers35 studied cats in the San Francisco area and confirmed the domestic cat to be a large asymptomatic reservoir for hu­man B. henselae infection. The organism was cul­tured from 25 of 61 (41%) pet and impounded cats, and was found in seven cats belonging to four patients with BA. In addition, B. henselae DNA was detected by PCR in a single flea, and was isolated from Ctenocephalides felis infesting a culture-positive cat. Interestingly, B. henselaecould not be isolated from the fleas of a culture-negative cat.

It has been suggested that contamination of

cat nails with B. henselae may occur following excoriation of skin during scratching; or peri­odontal disease may result in saliva becoming contaminated.35 As arthropod vectors are respon­sible for transmitting other members of the genus (i.e., B. qttintana by the human body louse, and 8. bacilli/armis by the sandfly), it would seem plau­sible that they may be involved in the transmis­sion of B. he11selae .36

Chomel and coworkers37 provided experi­mental evidence that cat fleas can transmit B. henselae from one animal to another. Fleas re­moved from cats that were bacteremic with B. henselae transmitted the organism to five specific­pathogen-free kittens in two separate experi­ments, and all developed B. 17enselae bacteremia. As the cat flea consumes an average amount of 13.6 ilL blood a day, which may contain millions of B. henselae organisms per mL, they considered the potential for ingesting large numbers of or: ganisms to be substantial.

Some fleas can infect mammalian hosts with Bartonella spp. via infected feces. As large amounts of potentially infective flea feces are present in the fur of infested cats, direct inocula­tion during scratching may lead to feline infec­tion, and be passed directly to humans via claws contaminated with flea feces. 37 They concluded that the cat flea plays a substantial indirect role in human disease by amplifying the size of the cat reservoir. Although epidemiological data do not support transmission from cat to human via the cat flea, the potential exists.

Whether or not other animals can act as a reservoir for B. henselae is unknown. However, two cases of CSD associated with a dog scratch have been described; one in a nine-year-old boy presenting with osteomyelitis,38 the other in a ten­year-old boy with bilateral cervicallymphaden­opathy.39 In both cases, the only animal contact was with a dog; however, in the second case, Bartonella DNA was found in the gingiva and buccal membrane of the puppy. Tsukahara and coworkers,39 using IFA, detected significant B. henselae antibody titres in four out of 52 dogs (three titres were 1 in 64, and one was 1 in 128).

Epidemiology CSD has a worldwide distribution, having

been reported in the United States,2.Z.~·211·3J . .j(}-~2 Eu­rope (Germany,43-16 The Netherlands,27.47-so Swit­zerland,51.52 France,53,54 Spain,55 ItalyS6 and the UK57,5S), Japan,39 Israel29.3s.s9 and Australia.60- 62 In the UK, very few cases were reported after the late 1970s, and this can be attributed directly to the withdrawal of the skin antigen test because of concerns about its safety. However, Harrison and Doshis8measured Bartonella antibodies, and

found that infection is common in the UK, with serological evidence present in 20% of their prob­able CSD group.

In temperate climates, a higher rate of CSD cases has been reported in the autumn and winter months.3,16,36,63 Some authors suggest that this may be explained by the seasonal breeding of the domestic cat, which governs when kittens are available to new owners.3,16

Bass and coworkers9 considered many vari­ables that might affect the prevalence of CSD, including geographical location, climate and sea­son, and cat-associated variables such as density, age, exposure, and degree of flea infestation. Serological studies in the USA found an increase in B. henselae antibodies in both humans and cats in warm, humid areas, compared with those in the rest of the country.9

Jackson and coworkers4 analyzed three national databases in the USA, and found CSD to be an important cause of morbidity, especially in children, where it accounted for more than 15% of lymph-node biopsies performed. CSD was the diagnosis in 0.77--0.86 per 100,000 hospital dis­charges, and the incidence in ambulatory pa­tients was higher, at 9.3 per 100,000 population. There are an estimated 57 million pet cats in the USA; therefore, the B. henselae reservoir is large and the potential for human infection immense.33

Clinical presentation Typically, a non-tender papule 0.5-1 cm in

diameter develops in the scratch line, three to ten days after exposure;3.9.60 usually healing without scarring in 2 to 3 weeks. Regionallymphaden­opathy follows in more than 80% of cases, but resolves usually within 2 to 3 months. In about 10% of cases, the lymph nodes may suppurate and become fluctuant. 60

,64 In uncomplicated CSD, patients usually remain afebrile and do not ap­pear to be ill.

Nowadays, atypical CSD symptoms are seen in up to 25% of cases, which is considerably more than reported previously; however, this increase probably is associated with increased clinical awareness/ and it is suggested that these atypical presentations are manifestations of Bar­tonella infection rather than CSD.9 Nevertheless, many authors continue to refer to such infection as atypical CSD, presumably because of an asso­ciation with a scratch or bite.3,16.64

One of the most common atypical manifes­tations of CSD is ocular involvement, either via direct inoculation or after rubbing the eye follow­ing contact with a cat. Parinaud's oculoglandular syndrome is a type of conjunctivitis described in up to 6% of patients with CSD.62 Other oph­thalmic manifestations include retinitis,

panuveitis, subacute orbital abscess, chorioditis, optic nerve masses, branch retinal artery occlu­sion, and peripapillary angioma.1W,50.6S-<i7 Most ocular symptoms resolve with or without treat­mentwithin 4 months, but severe and permanent vision loss has been reported.67

Neurological involvement occurs in a small proportion (1-7%) of cases, usually characterized by encephalopathy.3,6B It presents often as the sudden onset of a seizure or coma that may last for several days, but usually resolves with rapid recovery. However, Noyola and coworkers69 de­scribed a CSD patient with recurrent encephal­opathy who had two distinct episodes of sei­zures. Carithers and Margileth70 reported a clini­cal account of 76 patients with neurological com­plications (encephalopathy: 61; convulsions: 35). Armengol and Hendley68 described six children who presented with sudden onset of status epilepticus (seizures >30 minutes duration) that were diagnosed with CSD, based on clinical cri­teria and elevated B. henselae titres. As CSD en­cephalopathy is usually self-limiting without neurological sequelae, they recommended sup­portive care rather than specific antimicrobial treatment.

Other reported atypical manifestations in­clude granulomatous hepatitis,71 hepatosplenic infection/s,n endocarditis,73.74 paronychia,75 and osteomyelitis.38•49.60 CSD may be difficult to diag­nose and has been mistaken clinically for lym­phogranuloma venereum,s7lymphopharyngeal cancer with lymph-node metasteses,4< and tuber­culous lymphaderutis.61 Systemic CSD has been reported in both immunocompetent patients,S1.76 and those infected with HIV.77 The clinical presenta­tion may differ greatly in immunocompromised patients and may become life-threatening.

Clinical diagnosis Until fairly recently, the diagnosis of CSD

relied on meeting at least three of four criteria: 1) animal contact (usually a cat or kitten), with a scratch or eye lesion; 2) a positive CSD skin test; 3) regional lymphadenopathy that develops 2 weeks after contact; and 4) typical histopatho­logical changes in a lymph node biopsy.3.27 How­ever, the skin test antigen is not available com­mercially and is prepared from a pool of tissue from CSD patients. Consequently, it carries an inherent safety risk and may be capable of trans-

. mitting pathogens.78 In addition, it has been re­ported to be less sensitive and less specific than other diagnostic methods, and, as such, is no longer recommended by some workers.9.27

Laboratory diagnosis B. l1enselae can be seen in stained sections of

lymph node (i.e., using Warthin-Starry silver stain or Brown Hopp's Grain stain5,33). As biopsy of the lymph node is an invasive procedure, it should not be used in routine cases but reserved for atypical CSD.3 The organisms are best ob­served early in the illness, and may be very difficult to see once the lymph node suppurates.

Although B. henselae is fastidious and diffi­cult to grow routinely, it has been isolated suc­cessfully from human lymph node tissue,21 cuta­neous BA lesions,18 and from blood. 2O,22,53,79-S2 This is unlikely to be undertaken in a routine diagnos­tic laboratory, and has only been performed in a handful of laboratories worldwide. Growth is slow, especially on blood agar, and isolates re­quire 5 to 15 days incubation in CO/2 Colonies of B. lzenselae are firm, adherent, often embedded in the agar, and have been described as "cauli­flower-like."20 B. henselae are small, curved gram­negative bacilli, with morphology similar to Camptjlobacter Spp.22

Isolation from human blood may be diffi­cult because B. henselae does not produce detect­able CO

2, in some blood culture systems.83 Tierno

and coworkers81 isolated B. henselae from the blood of five febrile immunocompromised pa­tients, using the automated BacT / Alert™ system (Organon Teknika, Durham, NC);however, they found it more difficult to isolate from CSD pa­tients.

Gram staining of blood culture specimens often fails to detect Bartonella spp; therefore, it is recommended that acridine orange be used to stain blood cultures before discarding them as negative.83 Brenner and coworkers79 examined the effect of different methods of blood collection and handling on the isolation of B. henselae. Using blood specimens from B. henselae-infected cats, they found that freezing and thawing the sample yielded the largest number of colonies, and con­cluded that cell lysis, disruption of cell mem­branes, and dispersal of bacterial aggregates im­proved the sensitivity of B. henselae culture from blood.

B. henselae are catalase- and oxidase-nega­tive, and display twitching motility; they are non­reactive in routine biochemical tests, and are neither hemolytic nor acid-fast.82 Cellular fatty acid analysis reveals the major fatty acids to be octadecanoic acid eCI!!;). 56%); octadecanoic acid (C i8,O, 25%), and hexadecanoic acid (C;6:O' 13%).80 Such specialized methodology is beyond th.escope of most clinicalmicl'obiology laboratories; how­ever, Welch and coworkergEO compared the.abil­ity of six comme.rcial identification systems to distinguish B. henselae and B. quintana, and found that the MicroScan Rapid Anaerobe Pan:el'l'M (Baxter Diagnostics, Deerfield, IL) could separate

the two, and identify B. lzenselae successfully. The most common laboratory method used

to diagnose CSD is serological detection of B. henselae antibodies. In 1992, Regnery and co­workers25 developed the IF A testfor the detection of B. l1enselae antibodies in humans. The antigen was prepared from B.l1enselae co-cultivated with vero cells to reduce the autoadherent nature of the organism. Other workers used an enzyme immunoassay (EIA) to detect IgG, IgM and IgA antibodies in patients with suspected CSD.s4 Barka and coworkers84 found the EIA method to be highly specific and more sensitive (95%) than the IFAmethod.

Bergrnans and coworkers48 evaluated both methods of detecting B. henselae antibodies. Set­ting the specificity of the assays at ~95% by ana­lyzing serum from blood donors, they found the sensitivities of the IgG assays to be very low. However, by increasing the significant titre to ensure that 95% of healthy donors were negative, they may have introduced some bias. For ex­ample, Zbinden and coworkersS5 studied 120 blood donors who were positive for B. henselne, and found 11 (9.2%) to have antibody titres > 1 in 256.

The presence of B. lzenselae antibodies in healthy individuals may depend on many vari­ables, but it appears to be higher in Europe than in the USA.46 Bergmans and coworkers46 found the IgMEIA tohavethehighestsensitivity (71.4%) in patients fulfilling two or more criteria for CSD, and 80.6% in patients with a positive Bartonella PCRresult. Furthermore, they recommended that if only IgG IFA is performed, two-point serology is necessary to detect a developing B. henselae infection.

Drancourt and coworkers86 described ge­notypic variation among B. henselae strains and described a separate serogroup (Marseille). Given the-inconsistency of serological results in CSD, they suggest that this antigenic variability is partly to blame. Mainardi and coworkersS-l reported CSD due to B. henselae serotype Marseille in a seronegative patient, emphasizing the need to incorporate this serotype into future immunofluo­rescence assays.

Anderson and coworkers87 characterized a 17-kDa B.ltenselae antigen that provoked a strong antibodyresponsein CSD patients. Using recom­binant DNA techniques, they isolated clones that expressed tI1is antigen, and this may prove useful in the development of a simpler serological test. In addition, if the antigen elicits a protective inunlUle response, it may be useful in the devel­opment of a suitable vaccine.

PCR techniques have been used to detect B. l1enselae. Bartonella spp. are very similar, both

genotypically and phenotypically, and molecu­lar techniques are powerful tools for definitive species identificationY In addition, these tech­niques have been used to good effect in elucidat­ing the etiology of CSD. For example, Bergmans and coworkers27 used PCR to detect B. l1e11selae DNA in pus and lymph node specimens from CSD patients, but failed to detect A. felis DNA.

Many different molecular techniques have been employed successfully, including amplifi­cation ofBartonella DNA and dot-blothybridiza­tion with a specific B. henselae probe,4O and ampli­fication of B. henselae genes encoding for citrate synthase or a 60 kDaheat -shock protein. 59,68 Sander and coworkersB9 reconunended at least two dif­ferent primer pairs for higher sensitivity, as false­negative reactions are likely when performing PCR for B. henselae on fixed and paraffin-wax­embedded tissue.

Molecular methods proved invaluable when typing strains of B. henselae, with many au­thors reporting at least two genotypes.34,43,63,89-91 It would appear that some B. herzseiae genotypes are more prevalent in different geographical loca­tions. Furthermore, different genotypes were found in cats and humans, which may be related to pathogenicity.B9 Sander and coworkers43 com­pared different DNA fingerprinting methods for molecular typing of B. henselae and found that genetic heterogeneity was high. This may prove useful in future epidemiological studies.

Treatment The majority of CSD cases resolve sponta­

neously within a month or two and do not require antibiotic treatment. Zangwill36 conunented on the lack of controlled studies performed to date and regarded much of the information on CSD treatment to be anecdotal. He reconunended therapy only in patients with unresolved lym­phadenopathy, where lymphadenopathy is as­sociated with significant morbidity, severe sys­temic disease, or an underlying medical disorder complicated by severe CSD.

Margileth3 retrospectively reviewed anti­biotic therapy for CSD and described 202 pa­tients, in whom 18 different antimicrobial agents were used, 'only four of which (rifampicin [87% effective], ciprofloxacin [84%], gentamicin [73%], and trimethoprim-sulphamethoxazole [58%]) proved effective. Ciprofloxacin has been used successfully to treatfive adult patients with CSD, but is contraindicated in both children and preg­nant women.92

B. henselae is sensitive to many antibiotics when tested on agar, with minimum inhibitory concentration (MIC) <0.125 mg/L for gentami­cin, tetracycline, doxycycline, minocycline,

azithromycin, clarithromycin, amoxycillin, cefotaxime, ceftriaxone and rifampicin.93 How­ever, these antibiotics are less effective when used clinically, as treatment failure and relapse are conunon, B, henselae is intracellular and there­fore MIC testing should be performed in a tissue culture system, where antibiotics have to pen­etrate the cells.

Musso and coworkers94 used a cell-line as­say to test B. hense/ae against various antimicro­bial agents, and found that only aminoglycosides were bactericidal. Ives and coworkers95 also used a cell-line assay (vero) to evaluate in vitro suscep­tibilities of B. henselae against macrolide antibiot­ics. They measured the ability of the antibiotics to inhibit rather than eliminate B. henselae, and con­cluded that clarithromycin, roxithromycin and aZithromycin may be useful in the treatment of CSD.

Margileth and Baehren31 considered the treatment of a healthy patient with typical CSD to be supportive. In patients with thoracic or pulmo­nary disease associated with prolonged symptoms and/ or severe multisystem disease, they recom­mended oral trimethoprim-sulphamethoxazole, ciprofioxacin, orazithromycin. Rifampicin was sug­gested as an alternative, with gentamicin reserved for the severely ill patient.

In 1998, Bass and coworkers41 published a prospective, randomized, double-blind placebo­controlled study of the use of azithromycin in CSD. They concluded that patients with typical CSD showed significant clinical benefit after a 5-day course, indicated by a decrease in lymph­node volume within the first month of treatment. However, Zangwill96 was critical of this study because 10% patients, who had not undergone serological testing, were excluded from the final analysis; and considered a combination oftherapeu­tic restraint and reassurance to be the key with patients presenting with uncomplicated CSD.

Antibiotic treatment is reconunended in immunocompromised patients with CSD. Pa­tients with BA, bacillary peliosis hepatis, or re­lapsing bacteremia should be treated with macrolides or tetracycline, preferably for 6 weeks.9.36 Acquired immune deficiency syndrome (AIDS) patients with BA respond dramatically to various antibiotics,2.3 and HIV-infected patients require at least 6 to 8 weeks treatment. In immunocompromised patients who relapse, treatment should be continued for 4 to 6 months, and repeated relapses should be treated indefi­nitely.9

Conclusions Cat-scratchdiseaseisafairlyconunonclini­

cal syndrome that has puzzled microbiologists

for many years. There is now little doubt that B. henselae causes CSD, and is transmitted to man via infected cats; however, in some CSD patients the organism cannot be detected, even by sensi­tive molecular techniques. This raises the possi­bility that more than one causative agent is in­volved.

A. fe/is has been isolated from the lymph nodes of CSD patients, and could be responsible for a small number of cases; however, the lack of a serological response to this organism in hu­mans, and its absence from cats, would seem to argue against this hypothesis.

Genotypic variation among B. henselae iso­lates may result in negative diagnostic tests in CSD patients, especially as the Marseilles serogroup is not screened for routinely. In addi­tion, other Bartonella spp, such as B. clarridgeiae, also may be associated with CSD.

At the beginning of the 21 ,I century we have a much better understanding of the etiology and epidemiology of CSD; however, there is much still to discover, and the vexing question of how the organism is transmitted from the blood of an asymptomatic cat to man largely remains unan­swered.

Acknowledgments The authorwouldlike to thank Dr. T. C. Harrison, Respiratory and Systemic Infection Laboratory, PHLS Central Public Health Laboratory, UK, for help and guidance.

References 1. EngUsh CK, Wear DJ, M<!rgileth AM, et al: Cat scratch

disease: Isolationand culture ofthe bacterial agent. JAMA 19B8;259:1347-1352.

2. MargilelhAM: An tibiot ic therapy fOl'cal-scratchdisease: Clinical stud.y of therapeutic outcome in 268 patients and a review of the Iiteratme. Pedinlr infect Dis J 1992;11:474-478.

3. Midani S, Ayoub EM, Anderson B: Cat-scratch disease. Adll Paedintr 1996;43:397-423.

4. Jackson LA, Peikins BA, Wenger JD: Cat scratch disease in the United States. Am JPublic HeaItlI1993;83:1707-171l.

5. Wear DJ, Margileth AM, Hadfield TL, et al: Cat scratch disease: A bacterial infection. Science 1983:221 :1403-1405.

6. Kordick DL, Hiliard EJ, Hadfield TL, et al: Bartonella clarridgeiae, a newly recognized zoonotic pathogen caus­ing inoculation papules, fever, and lymphadenopathy (cat scratch disease). J CIIIl MicrobioI1997;35:1813-1818.

7. Debre R, Lary M, Jemmet ML, et al: La maladie des griffes du chat. Sem Hop Paris 1950;26:1B95-1904.

8. Parinaud H: ConJunctivite infectieuse transmise par les animaux. A lin d'Oculistiql.lf! 1B89;101:252-253.

9.13ass]W, Vincent 1M, Person DA; The expanding spec­trum of Bartonella jJ,fection. II. Cat-scratch disease. Piled/lllr Infect Dis J 1997;16:163-179.

10. Warwick W]: The cat scratch syndromes, many diseases or one disease? Prog Med ViroI1967:9:256--301.

11. Carithers HA: Cat scratch disease. An overview based on a study of 1,200 patients. Am J Dis Child 19B5;139:1124-1131.

12. Regnery R, Tappero J; Unr,weling mysteries associated with cat-scratch disease, bacillary angiomatosis, and re­lated syndromes. Ewerg bifecl D;sl995;1:16-21.

13. Margileth AM, Wear OJ, Hadfield TL, et al: Cat-scratch disease: Bacteria in skin at the primary inoculation site. JAMA 1984:252:938-942.

14. Solley WA, Martin DF, Newman NJ, at ai. Cat scratch disease: Posterior segment manifestations. Opiltllnlmol­ogy 1999;106:1546--1553.

15. Brenner OJ, Hollis DG, Moss CW, et al. Propos,11 or Aftl.lin gen. nov" WiUl AJ/pia felis sp. nov. (formerly the cat scratch disease bacillus), lifipin c/l!!letflJ1riclIs is sp. nov. (fonnerly the Cleveland Clinic Foundatloll sh'ain), Afipin braomellc sp. nov., il nd three unnamed genospecies.J Clill MicrobioI1991;29:24SQ-02460.

16. Anderson BE, Neuman MA: Bflrtonella spp. as emerging human pathogens. Clill Microbiol Rell1997;10:203-219.

17. Reiman DA, Loutit JS, Schmidt TM, et al: The ngent of bacillary angiomatosis: An approach to the identification of uncultured pathogens. N Engl J Med 1990;323:1573-1580.

18. Koehler jE, Quinn FD, Berger TG, et ai: Isolation of Roc!ralimaen species from cutaneous and osseous lesions of bacillary angioimatosis. N Engl J Met! 1992;327:1625-1631.

19. Slater l.N, Welch DF, Hensel D, et al: A newly recognized pa thogen as a cause of fever and bacteremia. N Engl J Med 1990;323: l.587-1593.

20 . .Regnery RE, AndersQIl BE, Clarridge fE III, et al: Charac­terization of a novel Rodmli/llflea species, R. iIense/ne sp. nov" isolated from blood of a febrile, human immunode­ficiencyvirus-positive patient. J Clill Microbio/1992;30:265-274.

21. Dolan Mj, Wong MT, Regnery RL, et ili: Syndrome of Roc/rnlillrnea Ilen.clne adenitis suggesting cat-scratch dis­ease. AmI Intel'll Mcd 1993;1l8:331-~36.

22. WeJchDF,PickettDA,SlaterLN,etal:Roc/w/i/llneaiIcnselne sp. nov., a cause of septicaemia, bacillary angiomatosis, and parenchymal bacillary peliosis. J Clin Microbiol 1992;30:275-2BO.

23. Brenner DJ, O'COImor SP, Winkler HH, et al: Proposals to unify the genera Bartonella and Rochalimaea, with de­scriptions of Bartonella quintana comb. nov., Bartonelln vinsonii comb. nov., Bartonella he/lselae comb. nov., and Bar1oncJlneliznbetlwecomb.nov., and to remove the family Bartonellaceae from the order Rickettsiaies. ll'lt J Syst BncterioI1993;43:777-7B6.

24. Tappero J, Regnery R, Koehler j, et al: Detection of sero­logic response to RoclJIIlilllaea /tellseine i1\ patients with bacilliny angiomatosis (BA) by immunofiuorescental1ti­body (rEA) testing. 32nd lutcrscimceCDllferellC(!OII Alltillli­c]'cbinl Agel1ls Ilmi C!te/llQllzempy, Anah eim. CA. Octnber 10-14, 1992. Abstract no. 674.

25. Regnery RL, Olson JG, Perkins BA, et al: Serological response to 'RodlllllllllJen Ilel1se1ne' antigen insuspected cat scratch d ise<tse. ulIlcet 1992:339:1443-1445.

26. Anderson B, Kelly R, Threlkei R, et al: Detection of Rocllalimaea henselne in cat-scratch disease skin test anti­gens. J Infec Dis 1993;168:1034-1036.

27. Bergmans AMC, Groothedde J-W, Schellekens JFP, et ai: Aetiology o( cal scratch disease: comparison of poly­merase chain reaction detection of Bartonella (formerly Rochalimaea) and Afipin fells DNA with serology and skin tests. J lIz/ect D,s 1995;171:916-923.

28. AlkanS, MorganMB,SandinRL, et al: Duni role for Afipia felis and Rod1ll1imaea lIenselne in cat-scratch disease. La"cct 1995;345:385.

29. Giiadi M, Avidor B, Kletter Y, et al: Cat scratch disease: The rare role of Afipiafe1is. J Clin MicrobioI1998;36:2499-2502.

30. La Scola B, Raoult D: Afipiafelis in hospital water suppiy in association with free-living amoeba. Lancet 1999;353:1330.

31. Margileth A, Baehren DF: Chest-wall abscess due to cat scratch disease (CSD) in an adult with antibodies to Bartollella dnn'idgeiae: Case report and review of the thoracopulmonary manifestations of CSD. Clin Infect Dis 1998;27:353-357.

32.Regnery R, Martin M, Olson J: Naturally occurring Rodwlimnen lIenselae i.nfection in domestic cats. Lancet

1992;340:557-558. 33. Zangwill KM, Hamilton DEL, Perkins BA, et al: Cat

scratch disease in Connecticut. Epidemiology, risk fac­tors, and evaluation of a new diagnostic test. N Eng/ J Med 1993;329:8-13.

34. Sander A, Buhler C, Pelz K, et al: Detection and identifi­cation of two Bartonella hense/ae variants in domestic cats in Germany. J Clin Microbio/1997:35:584-587.

35. Koehler JE, Glaser CA, Tappero JW: Rocha/imaea hense/ne infection. A new zoonosis with the domestic cat as reser­voir. JAMA 1994;271:531-535.

36. Zangwill KM: Bartonella infections. Sem Paediatr Infect Dis 1997;8:57-{i3

37. Chomel BB, Kasten RW, Floyd-Hawkins K, et al: Experi­mental transmission of Bartonella hense/ae by the cat flea. J Clin Microbio/1996;34:1952-1956.

38. Keret D, Giladi M, Kletter Y, et al: Cat-scratch disease osteomyelitis from a dog scratch. J Bone Joint SHrg 1998;80:766-767.

39. Tsukahara M, Tsuneoka H, Lino H, et al: Bartonella henselae infection from a dog. Lancet 1998;352:1682.

40. Anderson B, Sims K, Regnery R, et al: Detection of Roclmlimaea hense/ae DNA in specimens from cat scratch disease patients by PCR. J ClinMicrobio/1994;32:942-948.

41. Bass ]W, Freitas BC, Freitas AD, et al: Prospective ran­domized double-blind placebo-controlled evaluation of azithromycin for treatment of ca t-scratchdisease. Paediatr Infect Dis 1998;17:447-452.

42. Golnik KC, Marotto ME, Fanous MM, et al: Ophthalmic manifestations of Rocllnlimaea species. Am J Ophtllll/mol 1994;118:145-151.

43. Sander A, Ruess M, Bereswill S, et al: Comparison of different DNA fingerprinting techniques for molecular typing of Bartonella henseille isolates. J Clin Microbiol 1998;36:2973-2981.

44. Ridder GJ, Richter B, Laszig R, et al: A farmer witha lump in his throat. Lancet 1998;351:954.

45. Arvand M, Wendt C, Regna th T, et al: Characterization of Bartonella henselae isolated from bacillary angiomatosis lesions in a human immunodeficiency virus infected patient in Germany. Clin Infect Dis 1998;26:1296-1299.

46. Sander A, Posselt M, Oberle K, et al: Seroprevalence of antibodies to Bllttfmella helT.elnein patients with cat scratch disease and in hea lthy controls: Evaluation and compari­son of two commercial serological tests. Clin Diagn Lnb Inl1lHtI1oI1998;5:486-490.

47. Rothova A, Kerkhofff, Hooft HI, et al: Bartonella serology for patients with intraocular inflammatory disease. Retina 1998;18:348-355.

48. Bergmans AMC, Peeters MF, Schellekens JFP, et al: Pit­falls and fallacies of cat scratch disease serology: Evalua­tion of Bartonella !tenselne-based indirect fluorescence as­say and enzyme-linked immunoassay. J c/in Microbiol 1997;35:1931-1937.

49. Hulzebos CV, Koetse HA, Kimpen JLL, et al: Vertebral osteomyelitis associated with cat-scratch disease. Clin Infect Dis 1999;28:1310-1312.

50. Kerkhoff FT, Ossewaarde JM, de Loos WS, et al: Pre­sumed ocular bartonellosis. Br J Ophtha/mo/1999;83:270-275.

51. Waldvogel K, Regnery RL, Anderson BE, et al: Dissemi­nated cat-scratch disease: Detectionof RoclJalil1laea !renselae in affected tissue. Eur J Pediatr 1994;153:23-27.

52. ZbindenR: Bartonella hense/ae-based indirect fluorescence assays are useful for diagnosis of cat scratch disease. J Clill MicrobioI1998;36:3741-3742.

53. La Scola BL, Raoult 0 : Culture of Bar/oncllll quintana and Bartonella hellse!ae from human samples: A 5-year experi­ence (1993 to 1998). J Clill M1crobioI1999;37:1899-1905.

54. MainardiJ-L, Figliolini C, GoldsteinFW, et al. Cat scratch disease due to Bartonella henselaeserotype Marseille (Swiss cat) in a seronegative patient. J Clin Microbio/1998;36:2800.

55. Fortuno a, Uriel JA, Lomba E, et al: Hepatosplenic cat scratch disease diagnosed by serology. Eur J Pediatr 1999;158:432

56. Not T, Canciani M, Buratti E, et al: Serologic response to Bartonella henselae in patients with cat scratch disease and

in sick and healthy children. Acta Pediatr 1999;88:284-289. 57. Hagley M, Came CA, Gorgees N: A case of cat scratch

disease masquerading as lymphogranuloma venereum. Int J STD AIDS 1999;10:334-335.

58. Harrison TG, Doshi N: Serological evidence of Bartonella spp. infection in the UK. Epidemiol Infect 1999;123:233-240.

59. Avidor B, Kletter Y, Abulatia S, et al: Molecular diagnosis of cat scratch disease: A two-step approach., CUn Microbio/ 1997;35:1924-1930.

60. Robson JMB, Harte GJ, Osborne DRS, et al: Cat scratch disease with paravertebral mass and osteomyelitis. Ciill Infect Dis 1999;28:274-278

61. Gottlieb T, Atkins BL, Robson JMB: Cat scratch disease diagnosed by polymerase chain reaction in a pa tien t with suspected tuberculous lymphadenitis. Med J Allst 1999;170:168-170.

62. Grando O,Sullivan LJ, FlexmanjP, et al: Bar/onella !rense/ae associated with Parinaud's oculoglandular syndrome. CUn Infect Dis 1999;28:1156-1158

63. Sander A, Ruess M, Deichrnalill K, et al: Two different serotypes of Bnrtonella !rense/ae in children with ca t-scratch disease and their pet cats. Scand J Infect Dis 1998:30:387-391.

64. Seals JE, Oken HA: Cat scratch encephalopathy. Md Med J 1999;48:176-178.

65. Gaebler jW, Burgett RA, Caldermeyer KS: Subacute or­bital abscess in a four-year-old girl with a new kitten. Paediatr Infect Dis J 1998:17:844-846.

66. Kerrison JB, Bennett MD, Newman NJ, et al: Atypical mass lesions associated with cat-scratch disease. Clill Infect Dis 1999;29:221-223.

67. Gray AV, Reed jB, Wendel RT, et al: Bartonella hellselne infection associated with peripapillary angioma, branch retinal artery occlusion, and severe vision loss. Am J OplJt!lIlllloI1999;127:223-224.

68. Armengol CE, Hendley 10: Cat-scratch disease encepha­lopathy: A cause of status epileplicus in school-aged children. J Pediatr 1999;134:635--{;38.

69. Noyola DE, Holder or, Fishman MA, et al: Recurrent encephalopathy in cat-scratch disease. Pediatr Infect Dis J 199;8:567-568.

70. Carithers HA, Margileth AM: Cat-scratch disease: Acute encephalopathy and other neurologic manifestations. Am J Dis Gild 1991;145:98-101.

71. Lenoir AA, Storch GA, Deschryver-Kecskemeti K, et al: Granulomatous hepatitis associated with cat scratch dis­ease. Lnncet 1988;1:1132-1136.

72. Tan TQ, Wagner .ML. Kaplan SL: Bartonella (Roc1l1lli",nen) lreJlselaehepatospleniciniectionoccurringsimultaneously in two Siblings. Clin Infect Dis 1996;22:721-722.

73. Baorto E, Payne M, Slater LN, et al: Culture-negative endocarditis caused by Bartonella !rense/ae. J Paediatr 1998;132:1051-1054.

74. Holmes AH, Greenough TC, Balady G], et al: Bartonella Irenselaeendocarditis in an immunocompetent adult. Clil1 Infect Dis 1995:21:1004-1007.

75. Sander A, Frank B: Paronychia caused by Bartonella hense/ae. Lnncet 1997;350:1078.

76. Rosenburg A: Systemic cat scratch disease: A brief case report and review of the literature. Conn Med 1998;62:205-208

77. Wong MT, Dolan MJ, Lattuada CP, et al: Neuroretinitis, aseptic meningitis, and lymphadenitis associated with Bnrtonella (Rochlllimaea) hense/ae infection in immuno­competent patients and patients infected with human immunodeficiency virus type 1. Ciin Infect Dis 1995;21:352-360.

78. Arvand M, Mielke MEA, Sterry K, et al: Detection of specific cellular immtUle response to Bartonella henselae in a patient with cal' scratch disease. C/il1 Illfect Dis 1998;27:1533-1534.

79. BrennerSA, Rooney JA, Manzewitsch P, et al: Isolation of Bartonella (Roc1Ullimaea) hense/ae: Effects of methods of blood collection and handling. J CIiI1Microbio/1997;35:544-547.

80. Welch OF, Hensel DM,Pickett DA, etal: Bacteraemia clue

Article 234 0.1 CEC

to Roc/mUll/ilea Ilellselae in a child: practical identification of isolates In the clinical laboratory. J Clin Microbiol 1993;31:2381-2386.

81. Tierno PM, Inglima K, Parisi MT: Detection of Bartonella (Roc/llllimaea) henselae bacteremia using BacT/Alert blood culture system. Am J CIin PatllOl 1995;104: 530-536

82. Lucey D, Dolan MJ, Moss CW, et al:.Relapsing i1lnes-5 due to Roclmlilllllea IlIlI/ selae jn immunocompetent hosts: Im­plications for therapy and new epidemiological associa­tion. CUI/II/feet Dis 1992;14:683-638.

83. Larson AM, Dougherty MJ, Nowowiej'ski DI, eta!. Detec­tion of Bnrtollelln (Rochalillmea) qlli/l/ll1m by routine acri­dine staining of broth blood cultures. J eli" Microbial 1994;32:1492-1496.

84. Barka NE, Hadfield T, Patnail M, et al: EIA for detection of Rochnlill/llen !tell seine-reactive IgG, IgM, and IgA anti­bodies in patients with suspected cat-scratch disease. I Infect Dis 1993;167:1503-1504.

85. Zbinden R, Michael N, Sekulovki M, et al: Evaluation of cOmmercial slide for detection of immunoglobulin G against Bllrlonella lumselne by indirect immunofluores­cence. ElIr JClin Micr.obioJ II/fect Dis 1997;16:648-651

86. Drancourt M, Birlles R, Chammentin G, et al: New sero­type of Bartonella henselae in endocarditis and cat-scratch disease. lAncet 1996;347:441-443.

87. Anderson l3, Lu E, Jones D, et al: Characterization of a 17-kilodalton antigen of Bartonella henselae reactive with the sera from patients with catscratch disease. I Clin Microbial 1995;33:2358-2365.

88. Norman AF, Regnery R,Jameson P, eta.l: Differentiation of Bartonella-llke isolates at the species level by PCR­restri tion fragment length polymorphism in the citrate

synthase gene. I c/in Microbio/1995;33:1797-1803. 89. Sander A, Posselt M, B6hm N, etal: Detection of Bartollella

henselae DNA by MO different PCR assays and determi­nation of the genotypes of strains involved in histologi­cally deffned cat scrat.ch disease. I Clill Microbiol 1999;37:993-997.

90. Rodriguel.-B~rradas MC, Hamill RJ, Houston ED, et al: Genomicfingerprinting ofBnrlondla species by repetitive element PCR for distinguishing species and isolates. J Clill Microbio/1995;33:1089-1093.

91.l3ergmans AMC,Schellekens JFP, vanEmbden JDA, et al: Predominance of MO Bartonella henselae variants in The Netherlands. J Clin Microbio/1996;34:254-260.

92. Holley HPJr: Successful treatment of cat-scratch disease with ciprofloxacin. lAMA 1991;265:1563-1565.

93. Maurin M, Raoull D: Antimicrobial susceptibility of ROc/lIlliI1ll7enqllilllmla, Roc/lalilllaea vinson ii, and the newly recognized Roclllllillll1en ilel1selae. J An limic/'obinl CI,e/1/olller 1993;32:587-594.

94. Musso D, Drancourt M, Raoult D: Lack of bactericidal effect of antibiotics except aminoglycosides on Barlollelln (Rociwlimaen) Ilenselae. J Antimicrobiol Cllemollrer 1995;36:101-108.

95. IvesIT, Manzewitsch P. j{egnery.RL. et ~l : In vi lro suscep­tibilities of Bnr/ollella llenselae, B. quilltalla. B. eliznbetlme, Rit:k~tlsla ric.kellsli. R. conorii, R. IIknti, and R. pI'DwlIZfkii to macrolJde antibiotics as determined by immunotluores­cent-antibody analysis of infecled vl!ro cell m0I101a.yers. An/illlicrob Agellis Chemot/rer 1997:41:578-582.

96. Zangwill K: Therapeutic options for cat-scratch disease. Pediatr Infect Dis J 1998;17:1059-1061.

Questions for STEP Participants \

Answer questions only on the official STEP answer sheet. !fyou do nothav~ the official STEP answer sheet, a year's supply can be obtained (a t no cost) simply by writing to: STEP Program Answer Sheets, American Medical Technologists, 710 Higgins Road, Park Ridge, IL 60068-5765.

In addition to marking your answers, be sure to include all the required information on the answer sheet and a processing fee of $2.00 per article. In the following, choose the one best answer for each question.

1 Cat scratch disease (CSD) was first described 6 The most common laboratory method used to in 1864. (True or false.) diagnose CSD is serological detection of B.

2 henselae antibodies. (True or false.)

The etiological agent for CSD eluded detection until 1983. (True or false.) 7 One of the criteria for diagnosis of CSD is

3 regional lymphadenopathy, which develops 3

The disorder is so named because it is days after contact. (True or false.) associated with a cat bite or scratch. (True or

8 false.) The majority of CSD cases:

4 A. resolve spontaneously

Nearly 25% of the 24,000 cases annually B. require intensive treatment require hospitalization. (True or false.) C. require antibiotic therapy

5 D. are in temperate climates

Typically a papule develops in the scratch line in about: 9 Because there are over 57 million cats in the

A. 3 to 5 days United States, there is an immense reservoir B. 3 to 7 days for potential infections. (True or false.) C. 3 to 10 days

10 D. 3 to 15 days CSD in immunocompromised patients may be life-threatening. (True or false.)