Introduction

The eye is often involved in disseminated cat-scratch disease (CSD) caused by Bartonella spp. [1]. CSD is defined as a benign, self-limiting systemic condition that typically presents with fever and lymphadenopathy. In most cases, the microorganism is transmitted to human after a cat scratch or bite as cats consist its natural reservoir. Approximately 95% of patients mention a history of cat contact, and about 73% of patients have had a cat scratch [2]. The cat flea (Ctenocephalides felis) has been recently recognized as an arthropod vector of the disease [3, 4]. Various Bartonella species have been associated with human diseases, but it appears that Bartonella henselae is most commonly implicated in intraocular inflammations. In this article, which comes in continuity with our previous work on Bartonella species [5], we similarly focus on intraocular inflammations caused by Bartonella henselae and Bartonella quintana.

The aim of this retrospective study is to evaluate the wide spectrum of clinical features associated with ocular involvement of CSD and analyze the management and follow-up of this clinical entity and its potential complications.

Clinical presentation

Lymphoid CSD has been described as the most common clinical manifestation. The infection is a result of a cat scratch or bite and is usually followed by the development of non-tender erythematous pustules or papules at the site of primary cutaneous inoculation. Within the next one to two weeks, patients develop flu-like systemic symptoms (e.g., fever and fatigue), as well as regional lymphadenopathy (LAP). This particular stage of CSD is self-limited and usually resolves within a few weeks. LAP is typically unilateral affecting a single lymph node (50%), multiple lymph nodes (20%) or even multiple lymph node regions (30%). In some cases, LAP can be painful and suppurative. Symptoms of headache, nausea, vomiting, anorexia and sore throat have also been described. Additionally, some individuals may present non-specific maculopapular rash or erythema nodosum [6, 7].

In some rare cases, approximately 5–14% [8], CSD can manifest as a disseminated disease. The eye is considered as the most commonly involved organ in the disseminated course of CSD. Ocular bartonellosis presents in approximately 5–10% of patients with CSD [9]. The clinical features of eye involvement include Parinaud oculoglandular syndrome, neuroretinitis, intermediate uveitis, anterior uveitis, choroiditis, retinal infiltrate, branch retinal vessel occlusion, choroidal mass, serous retinal detachment and acute endophthalmitis [10, 11]. Apart from the ocular involvement, hepatosplenic disease (i.e., granulomatous hepatitis, splenomegaly or splenic abscess), pneumonia, endocarditis, encephalitis, osteomyelitis and paronychia have also been described [6, 7, 12]. The ocular manifestations have been thoroughly described in our previous study [5] and therefore are summarized in Table 1.

Table 1 Ocular and systemic symptoms recorded in patients with Bartonella spp. Infection
Full size table

Materials and methods

The medical records of 13 patients that were diagnosed, treated and followed up from 2011 to 2018 at the Department of Ocular Inflammations and Infections of the University Eye Clinic of Ioannina (Greece) for CSD with ocular involvement were analyzed retrospectively. The study was conducted in accordance with the Declaration of Helsinki (2008). The information recorded includes demographic data, past medical and ocular history, history of cat contact especially scratch/bite, visual acuity, intraocular pressure (IOP), anterior chamber inflammatory activity, anterior and posterior segment findings, laboratory findings, the treatment methods and outcome. The evaluation of anterior chamber and vitreous cells was based on the Standardization of Uveitis Nomenclature criteria [13].

The imaging techniques included fundus photography, fundus fluorescein and optionally indocyanine green angiography, optical coherence tomography (OCT) and echography of the eye. Visual fields examination was performed when needed. Magnetic resonance imaging of the brain and orbit was also performed in patients with intermediate uveitis and optic nerve involvement. Laboratory examination of sera for Bartonella henselae and quintana IgG and IgM was carried out at the Pasteur Institute of Athens. According to the Pasteur Institute instructions, IgG titers > 1:64 were considered positive and titers = 1:64 concerned suspected disease and were interpreted according to positive history of cat’s scratching or even licking, presence of human body lice, preceding illness compatible with CSD and intraocular inflammation not attributed to other causes according to the results of clinical examination of systems and laboratory investigation.

In addition, a follow-up of at least 6 months was considered along with the short- and long-term complications description and management.

Results

Thirteen patients (7 males and 6 females) with unilateral intraocular inflammation (except one with bilateral affection) attributed to Bartonella (either henselae or quintana) were included in this retrospective study. The mean age at presentation was 39.2 (10–77) years. In 12 patients (92.3%), there was a positive history of traumatic cat contact (scratching) and the causative organism was B. henselae. Only one case was positive for B. quintana; this patient was a worker in the fur industry, indicating a possible tick or flea bite. Table 2 shows the profile of patients with intraocular inflammation considered to be of Bartonella origin. The ocular clinical findings included neuroretinitis (Figs. 1 and 2) in 5 eyes (38.5%), vasculitis (Figs. 3 and 4) in 3 eyes (23.1%), iridocyclitis in 2 eyes (15.4%), intermediate uveitis (Fig. 4) in 2 eyes (15.4%), posterior uveitis in 1 eye (7.7%), panuveitis in 1 eye (7.7%), retinochoroiditis in 2 eyes (15.4%), vitritis in 1 eye (7.7%) and peripheral choroidal granuloma (Fig. 4) in 1 eye (7.7%). Therefore, neuroretinitis was the most common clinical feature and in two of these cases was accompanied by either posterior uveitis (Fig. 1) or panuveitis (Fig. 2). On the other hand, retinal vasculitis (occlusive or non-occlusive) was not an uncommon manifestation of intraocular inflammation induced by Bartonella.

Table 2 Profile of patients with intraocular inflammation associated with Bartonella infection
Full size table
Fig. 1
figure 1

a–d Patient (no 4), woman, 77 years old with posterior uveitis (apparent vitritis and optic disk edema in B-mode echography) with vitreous haze 2 + (photograph of the fundus) and neuroretinitis along with cystoid macular edema (fundus autofluorescence and OCT). (IgG titers for Bartonella henselae, 1: 1024)

Full size image
Fig. 2
figure 2

a–e Panuveitis (including findings of neuroretinitis) in a 10-year-old girl (case no 3) with Bartonella henselae infection (IgG titers, 1: 1024) along with a juxtapapillary chorioretinal granuloma (a, b), complicated by a full thickness macular hole (c) closed after prolonged therapy (d, e) [note the presence of epiretinal membrane and shrinking at the nasal macular area (c) and the absence of this feature after treatment (e)]. The optic disk pallor is a late consequence of neuroretinitis

Full size image
Fig. 3
figure 3

a, b Focal retinochoroiditis and vasculitis with an ipsilateral retinal ischemia in a young man (no 5), 18 years old (titers IgG for Bartonella henselae, 1: 256). In fluorescein angiography, an apparent branch retinal artery occlusion is noted due to the inflammatory process. c Improvement of the inflammatory lesion with absorption of hemorrhage and restoration of the retinal circulation (absence of ischemic edema). The bright white spot inferiorly to the optic disk consists an artifact due to the fundus reflection while acquiring the fundus photograph

Full size image
Fig. 4
figure 4

a–e Optic nerve head involvement in a 22-year-old woman (case no 2) (IgG titer for bartonella henselae, 1: 64 (a) along with peripheral choroidal granuloma (b) due to intermediate uveitis (c). There is a leakage from peripheral neovascularization (at the area of granuloma) (d and e). The main reasons of neovascularization in patients with intermediate uveitis are the inflammation itself and a precedent occlusive retinal vasculitis often associated with the intermediate uveitis

Full size image

Eight out of 13 cases (61.5%) did not present a systemic or another ocular illness prior to the intraocular inflammatory manifestations (Table 2). Treatment with antibiotics was efficient in all cases. The main antibiotics administered in our cases were rifampicin, doxycycline and azithromycin. Ciprofloxacin and ceftriaxone were also used in cases with treatment failure, allergy and early side effects. However, per os methylprednisolone (initial dose depended on age and uveitis severity) with tapering was administered in 6 patients in an effort to control the inflammation and its sequences (Table 3, cases number 1, 2,3,4,6 and 12). On the other hand, peripheral retinal neovascularization was treated using laser photocoagulation (Tables 2 and 3, second case). Table 3 summarizes the visual acuities before and after treatment, the antibiotics used and the complications recorded during the period of active inflammation and follow-up.

Table 3 Treatment and final visual outcome in 13 patients with intraocular inflammation attributed to Bartonella spp.
Full size table

In one patient (Table 2, second patient), very high rates of serum IgM and IgG titers for Bartonella henselae were noted 3 years after the initial diagnosis, suggesting a current disease (reinfection or relapse of a chronic indolent bartonellosis). Investigation for Q fever, chlamydia and rickettsiae (as a possibility of cross reaction) was negative.

Finally, the majority of patients, 8 out of 13 (61.5%) experienced complications, seven of them ocular (2 cases with an irreversible complication—those with macular scar) and one systemic (Table 3).

Discussion

CSD is a zoonotic disease with no gender or race predilection. Although it may insult individuals of any age, the vast majority of recorded cases involve children and adolescents [1]. In our study, 12 out of the 13 patients were adults, whereas only one child (7.7%) was diagnosed with intraocular infection attributed to Bartonella.

Interestingly, it has been demonstrated that 90–95% of patients with CSD report a history of cat contact. However, it has been also reported that ocular CSD may present in patients without a history of cat contact [14]. In the current study, 12 out of 13 patients had a cat contact, but none of them had been previously diagnosed with ocular or systemic CSD.

Neuroretinitis, defined by sudden and painless vision loss, consists the most typical and common sign of ocular CSD but cannot be considered as pathognomonic. Despite the fact that Bartonella henselae is the etiologic factor in approximately two-thirds of neuroretinitis cases, it is always important to exclude other causes such as toxoplasmosis, Adamantiades-Behçet’s disease, tuberculous uveitis or even spirochetal diseases [15,16,17]. Although neuroretinitis is mainly unilateral, bilateral cases have also been described [18]. Visual acuity in the affected eye at presentation can vary from perception of light to 1.0, while vision may impair suddenly within a few days. Many patients are found to have relative afferent pupillary defect, distortion in color vision, and central, cecocentral or arcuate visual field defects. Chi et al. [19] reported 53 cases of cat-scratch optic neuropathy. To our knowledge, Bartonella-related optic neuropathy with or without other ocular findings (i.e., macular star, panuveitis, posterior uveitis, anterior and intermediate uveitis, and retinal vasculitis) is the 30, 43% of our cases (including those reported in our previous published work) [5]. In our center, we recently revealed B. henselae as a cause of papillitis without other ocular findings, but this case was not incorporated in the current case series. Older studies have described neuroretinitis accompanied by macular star [20] and/or isolated foci of retinitis or choroiditis [21] the most common clinical findings. Macular star may be detected a few days after the onset of vision loss and becomes more distinct within the next 2–3 weeks [15, 16, 18]. In this cohort of patients, 3 patients (23.1%) were diagnosed with isolated unilateral neuroretinitis and 2 patients (15.4%) with unilateral neuroretinitis and uveitis (1 with posterior uveitis and 1 with panuveitis). Isolated optic neuritis has also been described in a previous study [22], underlying the importance of excluding infectious agents like Bartonella henselae before commencing treatment with pulse methylprednisolone, especially in children. Interestingly, in our previous study intermediate uveitis was the most prominent clinical entity (64.2%) [5]. Ocular CSD can also present with retinal artery occlusion, retinal infiltrates resembling cotton-wool exudates, or endophthalmitis [1]. The pathogenetic mechanism of retinal infiltrates is probably associated with ischemia secondary to retinal arteriole occlusion [19]. It is critical for the ophthalmologist to differentiate the superficial retinal infiltrates in ocular CSD from retinitis or retinal infiltrates observed in patients with sarcoidosis, Adamantiades-Behçet’s disease, and toxoplasmosis or rickettsia infection [22]. Ophthalmic vascular occlusions related to CSD have been reported in the literature in several case reports and series [5, 21, 23,24,25,26,27,28]. Interestingly, the alterations and distortion in visual acuity are correlated with the location of the affected artery or vein [22].

In some cases of secondary epiretinal membrane (ERM) (including those of uveitic origin), a spontaneous release of the ERM is observed. It has been suggested that this phenomenon is facilitated by the treatment [29,30,31]. The release of ERM leads to reducing or releasing the tractional forces on the macula and therefore permits the closing of the macular hole. On the other hand, in cases with permanent tractional epiretinal membrane (as a complication of uveitis) with or without macular hole development, the appropriate management is pars plana vitrectomy and ERM peeling.

CSD is expected to have a more severe systemic impact in immunocompromised or immunosuppressed patients. Characteristically, it has been reported to cause bacillary angiomatosis in HIV-positive patients [32]. In our current and previous [5] case series, none of the patients were immunosuppressed and none was found with signs of angiomatosis.

The diagnosis of CSD is based on systemic and/or ophthalmologic symptoms (Table 4) [5, 19, 21, 22, 33,34,35] and clinical findings, whereas serologic tests can be used to confirm the final diagnosis. Indirect fluorescent antibody assay (IFA) for anti-B. henselae IgG is considered as the gold standard, due to its high specificity [36]. Elevated B. henselae or quintana IgM titer is indicative of a recent infection, and values become normal again within a 3-month period. However, the sensitivity of IFA for anti-B. henselae IgM (IgM-IFA) is low [37]. B. henselae or quintana IgG titer can increase over the course of time and remain positive for up to 2 years. A positive B. henselae IgM and mainly an elevated B. henselae IgG titer can establish a diagnosis for CSD [38]. In the current study, all patients were found to be positive for B. henselae or B. quintana IgG (Table 2).

Table 4 Most common posterior segment findings in patients with ocular infection attributed to Bartonella spp.
Full size table

Doxycycline and rifampicin and alternatively one of them combined with azithromycin are the antibiotics of choice in individuals without systemic diseases, such as diabetes, or immune deficiency [5]. Trimethoprim–sulfamethoxazole, quinolones or intravenous aminoglycosides have also been reported as effective alternatives [1]. In our case series, the most commonly administered antibiotic was doxycycline, and in the majority of patients, it was administered in combination with rifampicin, macrolide and/or quinolone. There is a wide spectrum of opinions regarding the duration of treatment with antibiotics. An overall treatment between 10 days to 3 weeks has been shown to be adequate for cases with ocular involvement. However, it has been suggested that immunosuppressed patients (e.g., HIV-positive) should continue their treatment for an overall of 2–4 months [6]. The efficacy of systemic steroids remains debatable, but in cases with severe inflammatory findings administration of systemic steroids could be beneficial.

Taking into account that CSD is a zoonotic infection, it is crucial to establish and maintain preventive measures [39]. It has been recommended that strict ectoparasite control can limit the risk of Bartonella infection from arthropod vectors to domestic animals and pets and therefore prevent the transmission of pathogen from animals to humans. Hygiene education after contact with cats, especially of children and immunosuppressed individuals, could also contribute to reducing the risk of infection [1].

According to our results, Bartonella uveitis is more commonly induced by Bartonella henselae and rarely by quintana (the later observed in a worker if furs elaboration industry) and this intraocular inflammation mimics that caused by other diseases.

Conclusion

Intraocular inflammation caused by B. henselae and B. quintana is being diagnosed with increasing frequency worldwide. Although neuroretinitis and optic neuritis remain the most common manifestations, ocular CSD can present with a wide spectrum of clinical findings suggesting that bartonellosis is a new emerging mimicker. The history of cat contact can be the key for setting a clinical diagnosis, and Bartonella serologic investigations are substantial in the differential diagnosis between other diseases with similar clinical features. Preceding systemic symptoms must also be taken into consideration. Detailed recording and observation of the epidemiologic features can be helpful in establishing preventive measures especially in high-risk regions and populations. Raising awareness among clinicians of both ocular and systemic findings can contribute setting a prompt and early diagnosis, leading to a more efficient management of the disease. Well-designed randomized control trial studies are essential to define the most favorable treatment especially for sight-threatening cases.