eMedicine Specialties > Dermatology > Bacterial Infections

Catscratch Disease

Author: Kerrie J Spoonemore, MD, PharmD, Clinical Instructor, Department of Dermatology, Pacific Medical Centers
Coauthor(s): Jill McKenzie, MD, Resident, Division of Dermatology, University of Washington School of Medicine; Gregory J Raugi, MD, PhD, Professor, Department of Internal Medicine, Division of Dermatology, University of Washington at Seattle; Chief, Dermatology Section, Primary and Specialty Care Service, Veterans Administration Medical Center of Seattle
Contributor Information and Disclosures

Updated: Apr 4, 2008

Introduction

Background

Catscratch disease (CSD), also known as catscratch fever or subacute regional lymphadenitis, is caused by infection with the gram-negative bacillus Bartonella henselae. Typically a benign and self-limited disease in patients who are immunocompetent, only a small percentage of patients have complications involving the skin, lymph nodes, eyes, liver, spleen, or nervous system. Infection of immunocompromised patients with the same organism leads to a very different disease, bacillary angiomatosis-peliosis, which is characterized by angioproliferative lesions resembling those of Kaposi sarcoma in the skin, liver, spleen, bone, and other organs.

Henri Parinaud¹ sometimes is given credit for the first description of CSD in 1889. However, his oculoglandular syndrome of conjunctivitis with an enlarged preauricular lymph node ultimately was shown to be only a small subset of the possible clinical presentations of CSD, the result of inoculation of the CSD agent into the conjunctivae. Parinaud did not make the association with cat exposure; thus, his contribution is of limited scope.

The history of CSD has been reviewed comprehensively by Carithers² in 1970 and by Margileth³ in 1987 and is summarized here.

Debre and a colleague, Semelaigne, observed an unusual case of suppurating epitrochlear adenitis in a 10-year-old boy at the University of Paris and noted a number of cat scratches on the affected side. These were believed to be portals of entry for tubercle bacilli. When the tuberculin skin test results turned out to be negative, the investigators pursued an infectious cause of feline origin. While bacteriologic investigations yielded no clues, the physicians continued to observe similar cases of spontaneously remitting regional lymphadenitis associated with cat scratches in their pediatric population. Debre postulated tularemia, pasteurellosis, infectious mononucleosis, or tuberculosis as possible etiologic agents, but with no convincing proof.

Foshay, a microbiologist at the University of Cincinnati, suspected CSD to be a possible manifestation of tularemia. On meeting Debre in 1947, the 2 investigators compared notes on "catscratch disease" (Debre) and "cat fever" (Foshay). Foshay had produced an antigen from the pus of affected patients and achieved what was believed to be a diagnostic reaction after intradermal injection. Debre and his colleagues subsequently developed a similar antigen and demonstrated reactions in both old and new cases of CSD. These results⁴ were presented and published in 1950. These investigators also recorded failure of transmission of CSD to 15 different species of animals and possible human-to-human transmission in 1 of 4 cases.

In 1951, Greer and Keefer⁵ published the first report of CSD in American literature, in which they described a broader spectrum of CSD manifestations. In the late 1950s, William Warwick of the University of Minnesota collaborated with Robert Good, MD, in an attempt to transmit CSD to "every variety of lab animal from the monkey to the mouse." Their only positive result was the development of cutaneous lymphadenopathy in a monkey given intracerebral injections of ground lymph nodes and pus. The first thorough review of the world's literature, published in 1967, included 567 references and detailed the manifold clinical presentations. The landmark publication⁶ of a series of 1200 cases evaluated by one observer put the varied clinical presentations into perspective and provided the first realistic analysis of the spectrum of disease.

Discovery and classification of the etiologic agent for CSD is one of the triumphs of modern microbiology. The elegance and power of molecular taxonomy applied to the CSD agent revealed unexpected connections with other well-recognized infectious diseases and a deeper understanding of the pathogenesis of CSD.

Both viruses and Chlamydia had been proposed as possible etiologic agents for CSD, until a small gram-negative motile coccobacillus was observed in infected lymphatic tissue using a Warthin-Starry stain and Brown-Hopp tissue Gram stain in 1983 at the Armed Forces Institute of Pathology.⁷ In 1984, Margileth et al,⁸ using the same staining technique, demonstrated identical organisms in biopsy specimens taken from CSD inoculation papules. The first successful isolation and culture of the CSD organism was performed by English et al⁹ in 1988. Their further studies fulfilled Koch's postulates, and the organism was determined to be the cause of CSD.

One of the isolates from the study by English et al was investigated at the US Centers for Disease Control and Prevention, along with additional specimens from Tripler Army Medical Center in Honolulu. From these specimens, the CSD organism was determined to be a new entity and given the name Afipia (from Armed Forces Institute of Pathology) felis.

Reports associating another agent (Rochalimaea henselae) with CSD began appearing in 1992. Although they are not closely related, R henselae and A felis are members of the alpha-2 subclass of Proteobacteria and share a similar microscopic appearance and affinity for the Warthin-Starry stain. R henselae already had been implicated in the pathogenesis of bacillary angiomatosis, an angioproliferative condition observed in patients who are immunocompromised. Reports of R henselae –associated CSD appeared, and new immunological data subsequently supported a major role for R henselae as the etiologic agent in CSD. Although R henselae now is believed to be the principal pathogen in CSD, both organisms have been reported in some patients with CSD.¹⁰

When the sequences of 16S bacterial rRNA from R henselae and Bartonella were compared, these organisms were determined to be so clearly closely related that they belonged in the same genus. Because Bartonella had historical precedence, R henselae was renamed Bartonella henselae.

A Medscape General Medicine article that may be of interest is " Do Bartonella Infections Cause Agitation, Panic Disorder, and Treatment-Resistant Depression? "

Pathophysiology

Feline infection with B henselae is common and asymptomatic. In the United States, 28% of surveyed cats had antibodies against the organism. In California, blood cultures were positive in 56% of domestic cats younger than 1 year and in 34% of cats older than 1 year. More than three fourths of all cats in California had antibodies to B henselae as evidence of prior infection; however, only 21% of pet cats were bacteremic, compared with 61% of stray cats.

A similar survey of cats in the Baltimore area found seropositivity in 12-14% of domestic cats versus 44% of feral animals. Cats can be asymptomatically bacteremic for months, even while antibody titers are developing. The organism has been isolated from fleas residing on infected cats. Studies have shown that flea-vectored transmission of infection among cats occurs with high efficiency and that in the absence of fleas, infected cats do not transmit the infection to uninfected cats. Although flea-vectored transmission to humans has not been documented, it could explain some cases in which the patient has no history of exposure to cats. Treatment of cats with doxycycline is associated with a reduction of bacteremia, but whether such treatment prevents relapse or reinfection is unknown.

Familial and household clustering of cases of CSD have been reported. However, only one member of a family in contact with an infected cat usually is affected. Zangwill et al¹¹ found an 18% prevalence rate of seropositivity to B henselae among family members of patients with CSD. Upon further questioning, 43% of these individuals reported symptoms consistent with CSD during the previous 2 months. In the same study, matched control subjects not exposed to cats exhibited a 3.6% seropositivity rate. Carithers⁶ found similar results in a series of 1200 patients; 18.5% of asymptomatic family members had positive CSD antigen skin test results.

Frequency

United States

CSD is not a reportable infection. Seroprevalence rates vary greatly throughout the world, ranging from 0.6-37%. Approximately 22,000 cases occur annually in the United States. The prevalence is estimated to be 9.3 cases per 100,000 population. Cases of CSD occur throughout the United States. The incidence is greater in regions with higher temperatures and humidity (eg, Hawaii, Pacific Northwest, southeastern states, coastal California), while Alaska, the Rocky Mountains, and midwestern states have a prevalence lower than the median. Incidence peaks in the fall and winter months. These trends parallel the feline flea population density. The number of pediatric hospitalizations in the United States for CSD complications was estimated at 437 cases in 2000 children (0.60 in 100,000 children younger than 18 y and 0.86 in 100,000 children younger than 5 y).

Mortality/Morbidity

CSD generally is a self-limited infection, manifesting as a subacute regional lymphadenitis persisting for 3 weeks or more. Very few deaths (2 reported^12,13 ) from CSD have occurred in immunocompetent patients. However, significant morbidity occurs in 5-10% of cases, usually because of involvement of the central or peripheral nervous system or because of multisystem disseminated disease.

Race

CSD has no documented racial predisposition.

Sex

In some case series, a slightly higher incidence of CSD appears to occur in male patients, while others show equal rates between males and females. One hypothesis to explain a greater incidence among males is the tendency toward rougher play with kittens and cats.

Age

CSD affects persons in all age groups, but most patients are younger than 21 years. The younger age of individuals most likely to acquire CSD reflects their likelihood of exposure to the major risk factor (ie, kittens). A bias may exist in the literature because pediatricians have collected many of the large case series. Adults are more likely to manifest atypical features of CSD.

Clinical

History

Margileth et al studied a series of 1312 patients referred for chronic lymphadenopathy. The diagnosis of CSD was confirmed in 1174 patients via catscratch antigen (CSA) testing, similar to the tuberculosis purified protein derivative (PPD) test. Of this cohort, 88% exhibited a typical course characterized by lymphadenopathy lasting an average of 3 months.  

• Fever and malaise/fatigue were described in approximately one third of patients, none of whom was symptomatic for longer than 1 month.
• An unusual presentation was experienced by 11.6% of the cohort. Of these, Parinaud oculoglandular syndrome (POS) was found in half the patients, followed in decreasing order of frequency by encephalopathy (2.3%), systemic disease (2%), erythema nodosum (0.6%), atypical pneumonia (0.2%), breast tumor (0.2%), and thrombocytopenic purpura (0.1%).
• A subset of patients with severe systemic disease was observed for 7 years.
• Differences between the group with severe disease and those patients with typical CSD included more frequent primary skin or mucous membrane lesions, fewer animal contacts and cat scratches, greater frequency and duration of fevers, and other constitutional symptoms.
• Skin eruptions (eg, urticaria, vesiculopapular lesions, erythema nodosum) occurred 4 times more often in the severe CSD group compared with the typical CSD group. Demographics of the severe disease group included a greater percentage of adult males, who tended to have larger and multiple affected lymph nodes. Fortunately, all patients fully recovered. The greatest period of recuperation was 4.5 years.
• Another interesting finding from this series was evidence of recurrent CSD. Three of the patients in the severe disease category experienced recurrences at 4- to 20-month intervals. One patient presented with recurrent cervical lymphadenitis, whereas the other 2 patients experienced systemic symptoms with periodic fever, malaise, headache, and weight loss. In one patient, the episodes lasted 1-3 months, with a 20-month interval, while the second patient experienced similar episodes with a 10-month interval.

The incubation period ranges from 3-30 days. Patients may remember a self-healing lesion resembling an insect bite on the hand, arm, face, or neck. The most common presenting symptom of patients seeking medical care is tender regional lymphadenopathy, typically cervical, axillary, or epitrochlear nodes. Approximately 50% of patients experience systemic symptoms (eg, fever, headache, malaise, myalgias, arthralgias, exanthemas).

When questioned, patients may recall being scratched, licked, or bitten by a cat in the previous 2-8 weeks. The scratch is not mandatory because transmission can occur by petting alone with subsequent self-inoculation via a mucous membrane, skin break, or conjunctiva. Chronic fluctuant lymphadenitis develops in approximately 1 month and usually occurs in a single node or group of regional nodes draining the inoculation site. More than 65% of cases involve the nodes in the axillae or anterior or posterior triangles of the neck. Lymphadenopathy at multiple sites occurs in 37% of cases. Lymphadenopathy remains regional and regresses over a period of 2-4 months. Rarely, it may persist for a year or more. Approximately 10-15% of nodes suppurate.

Open table in new window

Other problems (ie, atypical CSD) occur in approximately 10% of cases. The following are considered atypical manifestations of CSD:  

• POS is the most common atypical presentation of CSD and is characterized by unilateral conjunctivitis with adjacent preauricular lymphadenopathy. Examination of the palpebral conjunctiva on the involved side reveals either a characteristic granulomatous lesion 2-4 mm in diameter or a frank scratch. While POS has been associated with other infections (eg, tuberculosis, tularemia, lymphogranuloma venereum), it most commonly is associated with CSD. Prognosis is identical to that of typical CSD.
• Central nervous system involvement can include encephalopathy, transverse myelitis, cranial or peripheral nerve involvement, and neuroretinitis.
  • Encephalopathy is a complication of CSD described in up to 5% of patients, occurring 1-6 weeks after typical CSD.^15,16,17,18 Most patients are young, in parallel to the spectrum of patients diagnosed with CSD. Symptoms usually begin with confusion or combative behavior and rapidly progress within hours to seizures and coma. Seizures occur in one half to three fourths of patients and may be self-limited or progress to status epilepticus. Coma complicated by respiratory depression requiring intubation and ventilatory assistance often is reported. Laboratory findings from examination of cerebrospinal fluid (CSF) often are normal. However, when CSF results are abnormal, no consistent patterns emerge. Abnormal findings from EEG, when observed, are nonspecific and transient. Recovery is usually complete in 2-10 days, with no sequelae. CT tomographic or MRI study findings of the brain are usually normal or reveal transient abnormalities.
  • Transverse myelitis presenting as Brown-Sequard syndrome has occurred in a 44-year-old man.¹⁹ The patient experienced rapid resolution of symptoms in a few days and was asymptomatic within 3 months.
  • Cranial or peripheral nerve involvement may occur. Two children with oculoglandular CSD were reported to have developed transient facial nerve paresis. Three adult women reportedly developed peripheral neuritis lasting 1-4 months associated with the onset of lymphadenitis due to CSD.
  • Neuroretinitis (Leber stellate neuroretinitis, Leber idiopathic stellate retinopathy,²⁰ Leber idiopathic stellate maculopathy) was first described in 1916. Patients present with painless, unilateral (rarely bilateral) decreased vision associated with central scotomata. Funduscopic examination reveals optic disc swelling and macular star formation. Permanent loss of vision is not reported, although recovery may take 1-3 months.
• Hepatosplenic CSD is a rare form of atypical CSD found in patients who are immunocompetent. Individuals present with daily fevers up to 104°F (40°C) and no identifiable cause. Physical examination findings are usually normal, with occasional detection of well-healed cutaneous scars secondary to cat scratches. Abdominal discomfort, without focal findings, is commonly reported.  Lymphadenopathy is present in approximately half the cases. Hepatosplenomegaly, jaundice, and elevated transaminase levels are not associated with this condition. The diagnosis is based on characteristic filling defects in the liver, spleen, or both as detected by ultrasonography, CT scanning, and a positive B henselae titer result. Intravenous administration of aminoglycoside antibiotics has been reported to be helpful. Most patients become afebrile within 48 hours of initiating treatment; however, in a few cases, fever has persisted for up to a month, even with antibiotic therapy.
• Blood culture–negative endocarditis has been reported.²¹ Bartonella species account for approximately 3% of all cases of endocarditis, and, now, many cases of blood culture–negative endocarditis are believed to be the result of infection with Bartonella species. Some cases diagnosed as Chlamydia endocarditis also are likely to be caused by Bartonella species because of the high prevalence of cross-reacting antibodies. Of 22 patients proven to have Bartonella endocarditis, 4 were infected with B henselae. Contact with cats was a significant risk factor in this group.
• Osteomyelitis secondary to B henselae, while still rare, is being reported with increased frequency. Hajjaji et al²² reviewed the literature and found 47 cases that met criteria and had sufficient data. They noted the vertebral column and pelvic girdle were the most commonly affected locations. MRI, scintigraphy, and CT scanning can all be used to demonstrate bone involvement. Immunosuppression was not a risk factor. Affected patients responded well to therapy without significant long-term effects. In a case report, a child with osteomyelitis and epidural extension confirmed by polymerase chain reaction responded to intravenous clindamycin and gentamicin.²³

Physical

Careful examination may reveal an inoculation papule in up to 90% of patients. Given the tendency to hold cats against one's chest, the lesion is found most often on the head or upper extremities. Include the scalp, finger web spaces, eyelids, and conjunctiva in a thorough inspection. Multiple sites may be infected.

The initial lesion evolves from a small, 2- to 5-mm reddish-brown macule or vesicle to a papule or pustule over the course of several days. Often, it is mistaken for an insect bite. Lesions typically are nonpruritic and heal in days to months without scarring.

Conjunctival CSD may manifest as nonsuppurative conjunctivitis or ocular granuloma.

Lymphadenitis involving 1 or more nodes in the proximal drainage area of regional lymphatics occurs in all CSD cases approximately 2 weeks (range 5-50 d) after initial inoculation. Nodes are tender and range from 1-5 cm, occasionally exceeding 10 cm. The overlying skin may be erythematous, but only rarely is an associated cellulitis present. The most commonly involved nodes are in the cervical, axillary, inguinal, femoral, preauricular, supraclavicular, and epitrochlear areas. Lymphadenopathy involving multiple sites occurs in approximately one third of patients. Lymphadenitis resolves over 2-4 months, rarely persisting longer than a year.

Other dermatologic manifestations occur in approximately 5% of patients and are more likely to occur in patients with more severe or atypical disease. These manifestations include macular, maculopapular, morbilliform, and petechial rashes. They typically are nonpruritic and resolve in days to weeks. Cases of erythema multiforme, erythema nodosum, erythema annulare, and ecchymoses associated with thrombocytopenia have been reported in patients with CSD, but these are rare.

Open table in new window

Causes

B henselae, the etiologic agent for CSD, is a small, fastidious, slow-growing, gram-negative, aerobic, nonmotile, pleomorphic bacillus. Although domestic cats are the principal reservoir for this bacterium, occasional cases of infection associated with dog and monkey bites have been reported. Another species, Bartonella clarridgeiae, has rarely been associated with cases of CSD.

Risk factors for acquiring CSD include ownership of a cat younger than 12 months, having been bitten or scratched by a kitten, and owning at least 1 kitten with fleas. More than 90% of patients with CSD have a history of exposure to cats, and 75% of patients have a history of a cat scratch or bite, usually from a healthy kitten.

• Search for CME/CE on This Topic »

• Catscratch Disease (Otolaryngology and Facial Plastic Surgery)
• Catscratch Disease (Pediatrics: General Medicine)
• Catscratch Disease (Emergency Medicine)
• Catscratch Disease (Infectious Diseases)