Bartonella are a genus of intracellular bacteria that infect mammalian reservoir hosts and are typically transmitted through arthropod vectors [1, 2]. Bartonella henselae is the most well described; cats are considered the reservoir host and transmission occurs through the vector Ctenocephalides felis or cat flea [3, 4]. Bartonella henselae may also be transmitted to humans and other incidental hosts through arthropod bites or via direct inoculation through cat bites and scratches contaminated with arthropod feces [1, 2]. For this reason, Bartonella henselae has been referred to as Cat-Scratch disease, and has a spectrum of clinical including fever, regional lymphadenopathy, ocular and neurologic involvement, culture-negative endocarditis, hepatosplenic lesions, and angio-proliferative lesions [1, 2, 10].

However, understanding of Bartonella has been rapidly expanding in the past decades. Multiple Bartonella species have been identified in a wide variety of mammalian hosts including humans, cats, dogs, rabbits, rodents, horses and other wild animals [2,3,4, 11]. A number of additional biting arthropods have also been implicated as vectors including fleas, lice, ticks, bed bugs, flies and mites [2–4, 11]. Bartonella henselae specifically has been identified both in other animals such as dogs, as well as in a variety of vectors including ticks, however their capacity to transmit infection to humans has been theorized but not confirmed [2, 3]. Nonetheless, there are numerous reports of human Bartonella henselae infection occurring in the absence of any cat or cat flea exposure, suggesting that alternative methods of transmission do occur, and dogs have specifically been implicated in a number of cases [3, 4, 11]. Similarly, our case did not identify any cat or flea exposure, and the dog scratch was the only likely transmission event identified in the period proceeding symptom onset. This raises both One Health concerns about the potential for zoonotic transmission between humans, arthropods and pets, as well as the potential for missed or delayed diagnoses if clinicians only consider cats as an epidemiologic risk factor [3].

Diagnosis is another area of uncertainty because Bartonella are fastidious organisms that do not grow in typical bacterial cultures [2, 10]. Diagnosis is often made clinically based on exposure history and compatible presentation supported by histology (granuloma with pyogenic abscess) and serology [10]. While direct detection methods including culture and PCR have high specificity, neither is sensitive nor used in routine diagnosis though they may play a greater role in atypical presentations [10]. Serology is the most commonly used diagnostic test with IgG values > 1:256 considered highly suggestive of recent infection; IgM production is usually brief and not routinely tested. However, sensitivity of serology is also limited and there is significant cross-reactivity between Bartonella species [2]. In our patient, the diagnosis of Bartonella henselae was based on the combination of typical presenting syndrome (painful regional lymphadenopathy), compatible exposure, strongly positive serology and response to appropriate treatment. Nonetheless we cannot rule out the possibility that the infection was caused by another cross-reactive Bartonella species [10].

Treatment of human bartonellosis depends on the specific presenting syndrome. Lymphadenitis can be treated supportively or with a short course of azithromycin [1, 10]. Neurologic disease is treated with 4–6 weeks of combination therapy with rifampin and doxycycline, while endocarditis is treated with six weeks of doxycycline and rifampin or gentamicin followed by a further 6 weeks of doxycycline alone [1]. For disseminated infection other than neurologic involvement or endocarditis, there is limited data to guide management although azithromycin and rifampin for two weeks has been suggested [1, 10].

HLH is a life-threatening inflammatory condition where unregulated activation of macrophages results in cytokine storm, tissue damage and organ failure [12]. A diagnosis of HLH is made using the HLH-2004molecular or clinical criteria [Table 2] [13, 14]. The three components of HLH treatment include supportive care, treatment/removal of trigger and immunosuppression [13, 14].

HLH is classified as genetic (primary) or acquired (secondary) [12, 14]. Acquired causes of HLH are extensive and include infection, autoimmune disease, malignancy, and medications [14]. Early identification of the trigger for secondary HLH is essential for treatment [12]. Infections are a precipitant in up to half of acquired cases, and in some cases treatment of the underlying infection is sufficient without immunosuppression [14]. Viral infections, particularly Epstein-Barr Virus, are the most common triggers followed by bacterial, fungal and parasitic infections [14]. Bacterial infections constitute 9% of acquired cases of HLH, with the most common bacterial triggers being Mycobacterium tuberculosis, Rickettsia species, Staphylococcus species, and Escherichia coli. [14]. The pathophysiology by which such infections trigger HLH is an area of ongoing research.

Bartonella species have been uncommonly reported as a precipitant of HLH; however, the spectrum of clinical presentations of Bartonella is still being elucidated. It is hypothesized that intracellular bacteria may trigger HLH because they reside in host immune cells and may continuously activate immune receptors [13]. There are only five published reports linking HLH with Bartonella infection, all with Bartonella henselae [5,6,7,8,9]. All but one of these cases occurred in immunocompromised hosts and at least three presented initially with lymphadenopathy [5,6,7,8,9]. The majority of the cases had been symptomatic for a prolonged period without diagnosis before presenting in HLH, suggesting a missed opportunity for diagnosis prior to the development of complications [5, 7,8,−9]. Three of the cases had a confirmed history of exposure to cats, but this was not reported for the other cases [5, 7, 8].

In each of the four cases that described management, Bartonella was treated with prolonged courses of either doxycycline or combination therapy regardless of clinical syndrome suggesting that the development of HLH was considered sufficient evidence of disseminated infection by the treating clinicians even in the absence of other findings [5, 7,8,−9]. However the exact treatment regimens differed, particularly duration of therapy which varied from 1 month to 1 year. In all cases, immunosuppression for HLH was also required.

The case we present here is unusual in that the patient did not have profound immunosuppression. Although he had a diagnosis of SLE, he had been off immunosuppression for nearly six months at the time of diagnosis. Only one prior case of Bartonella henselae leading to HLH in an immunocompetent patient has been reported, in which the patient had two genetic variants associated with susceptibility to HLH [8]. However, it is possible that Bartonella has been underdiagnosed given evolving knowledge of risk factors and diagnostic testing discussed here. Given that Bartonella frequently present alongside other co-infections, it may also be missed if other serologies are incidentally positive. This may be especially relevant since Bartonella has effective treatments while many viral triggers do not.

Although management of HLH requires prompt initiation of treatment for infection, it is unclear how the development of HLH affects the treatment when other clinical manifestations would typically require only short courses of antibiotics. Despite lymphadenitis being the main presenting symptom in our patient, we elected to consider the infection disseminated as had been done in other reported cases. There is minimal evidence to guide treatment for disseminated bartonellosis and only case reports were available regarding treatment in the setting of HLH.