Clinical manifestations

Presentation and course

Cat-scratch disease is a disorder characterized by chronic localized lymphadenopathy and mild systemic symptoms that follow contact with kittens or cats (12; 08; 37; 37; 74). The disorder is mild and self-limited in more than 75% of patients. Patients are most often young, with most surveys indicating a median age of about 15 years. In more than 90% of cases, contact with cats and scratches or bites from cats can be documented historically and a primary inoculation site can be demonstrated in up to two thirds of affected persons.

After inoculation, the incubation period before clinical signs and symptoms appear is 3 to 7 days. At the inoculation site, erythematous papules develop in 60% to 90% of patients. Lymphadenopathy develops between one and 4 weeks after inoculation and is usually limited to nearby regional nodes; therefore, the most commonly involved nodes are in the axilla, neck, and inguinal regions. Occasionally, more widespread lymphadenopathy occurs, but only rarely does lymphadenopathy become generalized. The affected lymph nodes enlarge, become tender, and are associated with overlying cutaneous erythema and induration. As the disease progresses, between 2% and 13% of nodes suppurate, some observers suggest that this percentage may be as high as 40%. Lymphadenopathy resolves spontaneously within 2 to 4 months, but, on occasion, complete resolution may take 12 months or longer.

Despite the frequent presence of impressive lymphadenopathy, generalized, systemic symptoms are usually mild, affecting 50% to 80% of patients. A low-grade fever develops in one third to one half, with temperatures rarely greater than 39°C and, generally, lasting no more than a few days, although the fever may last longer with presentation as fever of unknown origin (34; 75). Fatigue, weakness, generalized aching, malaise, nausea, and vomiting occurs in about 25% of patients; headache and odynophagia occur in about 10%.

Neurologic involvement in cat-scratch disease is unusual; it has been reported in fewer than 5% of patients with this disease (61; 12; 31; 37). In one of the largest series available, only three patients out of 1200 with cat-scratch disease developed neurologic symptoms (12). The most common neurologic manifestation of cat-scratch disease is generalized encephalopathy and encephalitis that begins abruptly one to 6 weeks after the initial clinical presentation of the disease. Cerebral epidural empyema due to B henselae has been reported (41). Generalized and focal seizures herald neurologic involvement and are followed during the next 1 to 4 days by onset of severe headache, agitated delirium, and, in a few patients, coma. Although seizures usually do not recur or persist, patients occasionally develop chronic, generalized, or complex partial epilepsy as a complication of cat-scratch disease. Status epilepticus may complicate the encephalopathy of cat-scratch disease in affected children (06; 07). Epilepsia partialis continua has also been reported (45; 55).

In a series of 76 patients with neurologic complications of cat-scratch disease, encephalopathy was observed in 61, whereas 15 had either cranial or peripheral nerve involvement (13). Altered level of consciousness with variable focal neurologic features characterized the encephalopathy (13). Convulsions were observed in 46% and combative behavior in 40% (13). Ischemic stroke may occur in some patients, presumably on the basis of cerebral vasculitis, although emboli from B henselae endocarditis have also been observed (66). Cerebral venous thrombosis has been observed, albeit rarely, and may be associated with skull based osteomyelitis (50). Direct evidence confirming the presence of cerebral vasculitis, however, is sparse. Patients may abruptly develop hemiparesis, visual field defects, and dysphasia consequent to ischemic infarction. Choreoathetosis may result from lesions in the basal ganglia (05). When ischemic infarction occurs in the spinal cord, corticospinal tract and dorsal column signs and bowel and bladder dysfunction ensue. Both Brown-Sequard syndrome and transverse myelitis have been observed in cat-scratch disease (32; 52; 59; 68). An inflammatory infiltration of the cauda equina accompanying transverse myelitis has been reported and may be mistaken for Guillain-Barré syndrome on the basis of flaccid paraparesis and hypoactive reflexes (14). Vertebral osteomyelitis with epidural abscess may complicate cat-scratch disease and result in myelopathy (04). Among the other uncommon neurologic complications of the cat-scratch disease are polyradiculitis, peripheral polyneuropathy, mononeuropathies, and cranial neuropathies, the most common being facial nerve paresis. Meningomyeloradiculopathy resulting in back pain, sphincter disturbances, and lower limb weakness, wasting, and numbness that responds to antibiotic therapy may complicate cat-scratch disease (27). An inflammatory transverse myelitis may also occur (62). Among the 15 patients without encephalopathy in the Carithers and Margileth study, two had facial nerve paresis, 10 had neuroretinitis, and three had peripheral neuritis (13). Serous labyrinthitis with rotational vertigo accompanying cervicofacial lymphadenopathy has also been reported (28). Chronic fatigue syndrome has also been reported in association with cat-scratch disease, and at least in one instance accompanied persistent bacteremia with Bartonella (Rochalimaea) henselae (72). The neurologic manifestations of cat-scratch disease have been regarded as one of those infections that can be mistaken for multiple sclerosis (09).

Postinfectious neurologic syndromes that have been observed following cat-scratch disease include optic neuritis, neuroretinitis (56), encephalomyelitis, hydrocephalus (35), parkinsonism (43), and Guillain Barré syndrome (40; 35). Persistent focal seizures following cat-scratch encephalopathy have been reported (22). Fortunately, most neurologic manifestations of cat-scratch disease are reversible, and patients typically recover without sequelae within 12 months of onset.

A wide variety of ocular abnormalities have been reported with cat-scratch disease (49). Parinaud oculoglandular syndrome refers to a follicular conjunctivitis with preauricular lymphadenopathy that occurs when the inoculation site is the conjunctiva. Spontaneous resolution occurs in 2 to 4 months. B henselae is the most common cause of this syndrome (49). Leber neuroretinitis (macular star) is commonly a manifestation of cat-scratch disease and optic neuritis has been observed as well (49). The vitreoretinal manifestations include anterior uveitis, vitreitis, pars planitis, focal retinal vasculitis, a characteristic retinal white spot syndrome, Bartonella retinitis, branch retinal arteriolar or venular occlusions, focal choroiditis, serous retinal detachments, and peripapillary angiomatous lesions (49). A review of 53 patients with cat scratch optic neuropathy demonstrated that bilateral involvement was present in 17%, visual acuity returned to 20/40 or better in 68%, and a macular star was observed in 45% (15). In AIDS patients, ocular manifestations include conjunctival and retinal bacillary angiomatosis (49). A subretinal mass associated with an abnormal vascular network is uniquely characteristic of cat-scratch disease in the setting of HIV infection (17).

Unusual systemic manifestations of cat-scratch disease include thrombocytopenic purpura, erythema nodosum, osteomyelitis, hepatitis, hepatic and splenic abscesses, and hemophagocytic lymphohistiocytosis (33; 34; 03; 53). The latter may occur in the absence of lymphadenopathy. Metagenomic analysis of one patient with hemophagocytic lymphohistiocytosis complicating bartonella infection revealed two gene variants (STXPB2 and IRF5) believed to be contributory (73). Endocarditis, myositis, and soft tissue infection may also result from infection with B henselae (03). Vertebral osteomyelitis and epidural abscess may occur with B henselae (02). These clinical manifestations are more commonly observed in patients with bacillary angiomatosis, an illness resulting from B henselae infection and characterized by neovascular proliferation in skin or viscera. Bartonella-induced neovascular proliferation occurs principally in immunosuppressed individuals and is most often observed in patients with AIDS, although it may occur with other immunosuppressive disorders, including recipients of organ transplantation. Bacillary angiomatosis is characterized by a few widespread, erythematous angiomatous skin lesions, though they may number more than 1000 or may be solitary. The lesions result from angioproliferation. These lesions may be clinically indistinguishable from Kaposi sarcoma. Other organs, including brain, bone, lymph nodes, gastrointestinal tract, respiratory tract, spleen, and liver may be involved by bacillary angiomatosis (26; 63).

Prognosis and complications

The vast majority of patients with cat-scratch disease have mild clinical disease that resolves spontaneously in 2 to 4 months. Lymph nodes return to normal size in several months, but occasionally this may take longer than a year. Prognosis is excellent and sequelae are rare. Death is exceedingly rare. For patients with encephalopathy or other CNS involvement, full recovery is the rule with the exceptional patient suffering from persistent seizures, hemiparesis, visual loss, or myelopathy. Bacillary angiomatosis carries a significant risk of substantial morbidity and even mortality if left untreated (16; 69). Perhaps not surprisingly, disseminated bartonellosis appears to be more common in immunosuppressed individuals, not only those with HIV/AIDS but also organ transplant recipients (54). Death from laryngeal obstruction and disseminated intravascular coagulation has been reported (16). Prompt treatment of this disorder is warranted (03).

Patients with cat scratch optic neuropathy typically have a promising visual outcome. Good visual acuity at presentation is associated with favorable visual outcome (15).

Biological basis

Etiology and pathogenesis

During the past 40 to 50 years, numerous viruses, bacteria, mycobacteria, and fungi have been implicated in causing cat-scratch disease. Wear's discovery of A felis in infected tissue in 1983, however, represented a major step toward understanding the etiology of cat-scratch disease (71). For several years, A felis was considered the cause of cat-scratch disease, but then Bartonella henselae, formerly referred to as Rochalimaea henselae, was identified in infected tissue. B henselae is a small Gram-negative rod, a member of the order Rickettsiales within the class Proteobacteriacae (58; 18). Studies of infected tissue coupled with epidemiologic studies using serologic tests for B henselae, indicate that B henselae, and not A felis, is the principal etiologic organism (20; 74). Additionally, the comparison of polymerase chain reaction detection for B henselae and Afipia felis DNA with serology and skin tests suggests that the latter plays little, if any, role in the pathogenesis of cat-scratch disease (56). B henselae is the etiology of bacillary angiomatosis in patients with AIDS (29; 67).

The precise mode of transmission of cat-scratch disease to humans is not fully known. Inoculation occurs when cats harboring the infective organism scratch or bite a person. Some data suggest that arthropods, such as fleas and ticks, also carry the bacteria and may transmit cat-scratch disease to humans (74; 42). The infection has been reported after blood transfusion (70). Once inoculated into the human, the bacteria infect the involved skin with the resultant formation of papules at the inoculation site. These lesions contain lymphocytes and epithelioid granulomas. Bacteria enter nearby lymph nodes and cause avascular necrosis, lymphocytic infiltration, granuloma formation, and abscess formation. Necrosis of granulomas leads to formation of large pus-filled cavities that may rupture and result in spread of inflammation or in draining cutaneous fistulae. Routine histopathology of lymph nodes is not specific for cat-scratch disease, and the findings may be mimicked by other infectious, inflammatory, and neoplastic diseases. However, Warthin-Starry silver stains will demonstrate intracellular, pleomorphic bacilli within the infected tissue. In bacillary angiomatosis, light microscopy of skin biopsies is most commonly used to confirm the diagnosis (03). The characteristic lesion is a lobular vascular proliferation of plump “epithelioid” endothelial cells (03).

The mechanisms by which neurologic disease develops are not known. Cerebral vasculitis has been demonstrated angiographically in patients with biopsy-confirmed cat-scratch disease (61); however, evidence for vasculitis is lacking in most patients with encephalopathy. Pathological material is scant; however, perivascular lymphocytic infiltrates and microglial nodules in the brain have been described in an otherwise healthy 6-year old who died with systemic B henselae infection (23). B henselae could not be detected in the brain by either special stains or PCR, although it was detected in affected lymph nodes (23). Other studies have also failed to detect direct infection of the CNS by B henselae. The possibility that these organisms elaborate neurotoxins or induce immune-mediated CNS disease has been suggested, but neither of these pathogenetic considerations has been tested experimentally.

Epidemiology

Cat-scratch disease is a worldwide, seasonal, nonepidemic disease. In the United States, estimates of the annual incidence range from 1.8 to 9 per 100,000 people. Approximately 22,000 cases are reported annually in the United States, and of these 2000 are hospitalized. All age groups are affected, but the median age is about 15 years and 80% of persons diagnosed with cat-scratch disease are under 21 years of age (36). The occurrence of cat-scratch disease peaks in the late fall and winter months. Studies from Connecticut indicate that the disease is most likely to be acquired from kittens, especially those infested with fleas (74). More than 90% of infected patients have had exposure to cats within 2 weeks of onset, and some investigators have identified the site of cat scratch or bite in 60% to 90% of patients (12). This contrasts with bacillary angiomatosis in which less than 20% of affected patients recall a cat scratch or bite (63). The results of seroepidemiological studies for B henselae-IgG in healthy populations vary greatly throughout the world, with rates as low as 3.6% in New Zealand to as high as 50% in Switzerland’s healthy adults (51). In one study of the seroprevalence of Bartonella in ocular inflammation, 33.3% of patients with uveitis had a specific IgG class antibody to Bartonella spp.; however, IgM-class antibodies were not detected nor was Bartonella DNA detected by PCR of surgery-extracted intraocular fluid (10).

The ratio of symptomatic to asymptomatic cases is not known. Serologic surveys have shown that many individuals in households with infected cats develop antibodies to B henselae but never develop clinical disease. Veterinarians are probably at increased risk of exposure to infected cats as an occupational hazard. It is also important to note that infected cats display no manifestations of illness of cat-scratch disease; therefore, the only way to identify infected cats is by blood culture or serologic testing.

Atypical cat-scratch disease (retinitis/neuroretinitis, conjunctivitis, neuritis, encephalitis, hepatosplenic disease, osteomyelitis, erythema nodosum, and endocarditis) was estimated to account for 1.5% of all cases with an average annual incidence in the United States of 07 cases/100,000 population (44). Neurologic manifestations accounted for 13.8% and ocular disease for 48.7% of atypical cases (44).

Prevention

No means of preventing infection of cats or humans are available. Vaccines are not available, and prophylactic antibiotic therapy is of no proven value. There are approximately 50 million pet cats in the United States. The infected cat is typically well. Up to 4% of family members scratched by the same cat develop cat-scratch disease. Because there is no practical way to identify these infected cats, there is no effective means of avoiding possible exposure. Generally, cats clear the infection (01), although there is ample justification for the development and employment of a vaccine (47). Person-to-person transmission of cat-scratch disease does not occur, so isolation or quarantine of infected patients is unnecessary.

Special precautions might be considered for patients with AIDS. As a result of their immunocompromised status, AIDS patients infected with B henselae are at risk of developing bacillary angiomatosis or a severe, systemic form of cat-scratch disease with potentially life-threatening hepatitis. AIDS patients with a pet cat might consider having the animal tested for cat-scratch disease and treated for fleas.

Differential diagnosis

A careful history and search for inoculation site in a patient with localized lymphadenopathy usually suggests the correct diagnosis. Other disorders that produce significant lymphadenopathy must be excluded, however. Pyogenic adenitis may be caused by a variety of streptococci and staphylococci as well as anaerobic bacteria, mycobacteria, and fungi. Other infections such as cytomegalovirus, Epstein-Barr virus, human immunodeficiency virus, toxoplasmosis, tularemia, syphilis, and chlamydia may also be associated with prominent lymphadenopathy. Lymphoid, lymphoreticular, and other types of neoplasms must be considered and excluded. Sarcoidosis, Kawasaki disease, and collagen vascular disorders produce symptoms and lymphadenopathy that may initially resemble those produced by cat-scratch disease.

Associated or underlying disorders

The presence of open wounds on the skin may increase the risk of developing cat scratch disease when licked by an infected cat.

Diagnostic workup

The diagnosis of cat-scratch disease rests on finding localized lymphadenopathy in combination with at least three of the following: (1) history of exposure to cats or of cat scratch; (2) exclusion of other infections and neoplasms; (3) positive reaction to cat-scratch skin test antigen; and (4) characteristic histopathologic changes in biopsy material from of skin, lymph nodes, conjunctiva, or liver. General laboratory test results are frequently normal or only mildly and nonspecifically abnormal. The peripheral white blood cell count may be elevated to 15,000 to 20,000 cells/mm with a left shift. Erythrocyte sedimentation rate and C-reactive protein concentration are elevated. Results of blood chemistries are normal. Hepatocellular enzyme concentrations in serum are usually normal but may be elevated if the patient develops cat-scratch disease-induced liver abscesses or granulomatous hepatitis. Suppurative lymph nodes can be drained by needle aspiration and the fluid examined microscopically and cultured. Lymph node excision is indicated in some cases for the definitive exclusion of neoplasm.

Skin tests are available on an investigational basis but are often unnecessary for cat-scratch disease diagnosis. Skin test antigen is injected intradermally on the flexor surface of the forearm, and the amount of induration determined 2 to 4 days later. Positive test results may occur in patients without clinical disease, and negative results may be obtained if the test is performed during the first 2 weeks of illness (12; 37).

Serologic testing may confirm the presence of antibodies against B henselae. In a literature review published in 2022, sensitivities of serological assays for B henselae varied between 10% and 100%, with more than 50% of these studies showing sensitivities below 70% (30). The specificities of serological assays ranged from 15% to 100% with nearly half the studies reporting specificities less than 90% (30). PCR is more likely to prove helpful in establishing the diagnosis. Diagnostic accuracy is improved when PCR is coupled with epidemiological, serological, and histopathological criteria (25). Cultures of blood and lymph node aspirates for B henselae can be grown in some specialty laboratories but are neither generally available nor necessary. Patients with hepatic disease without lymphadenopathy may require hepatic ultrasonography followed by liver biopsy for the correct diagnosis of atypical cat-scratch disease.

For patients with neurologic manifestations, CSF examination, EEG, and neuroimaging procedures may be indicated; however, these tests often yield normal results. In a few patients, CSF protein concentrations are mildly increased and a mild CSF lymphocytic pleocytosis is present. Polymerase chain reaction (PCR) of CSF can detect Bartonella DNA in CSF, and this technology has been incorporated into TaqMan array cards to permit an analysis not only of Bartonella but also other pathogens simultaneously (48). EEGs demonstrate diffuse generalized slowing, but occasionally focal slowing or focal epileptiform activity is seen. Neuroimaging studies are usually normal. Documented neuroimaging abnormalities were seen in only 18.8% in a study (60). In another study of 33 patients with cat-scratch disease encephalopathy (13), CT scans were normal or showed only transient abnormalities, and MRI scans were normal in five of six patients in whom they were obtained. Neuroimaging studies may disclose small focal lesions indicative of ischemic infarctions. Abnormalities are observed in the cerebral cortex, cerebral white matter, thalamus, and, rarely, basal ganglia (05; 60). In a child with status epilepticus complicating cat-scratch disease encephalopathy, focal cortical swelling and hyperintensity was noted on FLAIR and T2 weighted magnetic resonance imaging (46). A few patients with stroke-like presentations have had cerebral angiograms that demonstrate vascular changes consistent with vessel occlusion and localized segmental vasculitis (61).

Management

In the absence of controlled trials in the treatment of cat-scratch disease, published recommendations remain contradictory. Whereas most patients recover in the absence of antibiotic treatment and respond satisfactorily to symptomatic therapy with antipyretics and rest, erythromycin administered as 500 mg four times daily has been regarded as the drug of choice (03). A variety of other antibiotics including doxycycline, gentamicin, trimethoprim-sulfamethoxazole, and others have been also suggested with case reports or small series, demonstrating their utility (03). For suppurative nodes, aspiration is suggested. Lymph node excision or incision and drainage have no place in therapy for cat-scratch disease, as these procedures prolong healing and promote development of chronic draining fistulae. For adults or children over the age of eight years old with neurologic or ocular disease, azithromycin or trimethoprim-sulfamethoxazole coupled with rifampin is recommended (64). Therapy should continue for four to 6 weeks (64). Neurologic symptoms are treated supportively. Anecdotal reports suggest that high-dose parenteral methylprednisolone may be valuable (24). Seizures are treated acutely with anticonvulsants. Postencephalopathy epilepsy and the need for long-term antiepileptic medications are rare.

Special considerations

Pregnancy

There is no specific information that would suggest any special problems for pregnant women who acquire cat-scratch disease. The author is unaware of any instance of maternofetal transmission or any case of congenital cat-scratch disease.

Other conditions. Individuals on tumor necrosis factor-α blockers, eg, adalimumab, abatacept, and tocilizumab, may be at increased risk of cat-scratch disease. Through 2017, at least 10 cases have been reported in the literature, with most presenting with a regional lymphadenopathy and exhibiting a favorable therapeutic response (39).