Seventeen patients (19 eyes) were included (Tables 1, 2). Eleven patients (65%) were female. The mean age at presentation was 28.3 ± 12.3 years (range, 13-60 years). Eight patients (47%) identified as European, eight patients (47%) identified as Māori or Pasifika, and one patient (6%) identified as Asian. The mean presenting visual acuity (Early Treatment Diabetic Retinopathy Study [ETDRS] testing) was 50 ± 23 (range, 4-90). The mean final visual acuity was 78 ± 12 (range, 45-90). Simultaneous bilateral involvement occurred in two patients (12%). The mean duration from presentation to diagnosis was 11.8 days (range, 0-42 days). The mean duration of follow-up was 207 days (range, 7-1,100 days). Eleven patients (65%) had cats or kittens at home, with only one (6%) recounting a history of being scratched. Six patients (35%) presented with general malaise concurrent with ocular symptoms, consisting of nausea, vomiting, diarrhea, headaches, night sweats, fever, chills, reduced appetite, and a new rash.

All presented with optic disc swelling. Eight patients (47%) demonstrated the characteristic macular star (proteinaceous leakage from disc vasculature arranged in a stellate configuration) in conjunction with optic nerve head swelling at initial presentation. All patients eventually developed the macular star within a mean duration of 6 days from initial presentation. Four patients (24%) had macular oedema, three (18%) had vasculitis, two (12%) had uveitis, two (12%) had disc granuloma, and one (6%) had multifocal chorioretinitis. Five patients (29%) had a relative afferent pupillary defect documented on presentation.

All patients in our cohort had serological workup for infectious and inflammatory etiologies including sarcoid, leptospira, toxoplasma, and syphilis. Nine patients (53%) had neuroimaging to investigate for other causes of optic nerve swelling such as demyelination or compressive lesion. Additionally, 12 (71%) had chest imaging as part of their infective screen. Five (29%) had fundus fluorescein angiogram to assess causes for vasculitis and vascular occlusions. Seven patients (41%) demonstrated both immunoglobulin G (IgG) and IgM seropositivity for B. henselae, whilst 10 (59%) were seropositive for IgG and seronegative for IgM. Six patients (35%) were treated with rifampicin (300 mg twice daily for 1 month) and doxycycline (100 mg twice daily for 1 month), one patient (6%) with rifampicin and azithromycin (250 mg once daily for 1 month), two patients (12%) with doxycycline only, two (12%) with co-trimoxazole monotherapy (960 mg twice daily for 6 weeks), and one patient (6%), who was breast-feeding, was treated with erythromycin. Four patients (24%) did not have receive any treatment (Table 2).

A mean visual acuity of 74 letters was achieved by the untreated cohort (4 patients), compared to 80 achieved by the treated cohort (12 patients). There was no documented final visual acuity for three in the treated cohort due to nonattendance. The mean presenting visual acuity was 45 letters in those that did not receive treatment, compared to 51 in those that were treated.

A statistically significant difference was observed in posttreatment visual acuity with a final visual acuity of 80 letters (95% confidence interval, 15.8-44.4; p = 0.001). Statistical analysis was not possible for eyes which did not receive treatment due to very small numbers.

Multivariable analysis was performed to ascertain association of signs and their impact on final visual acuity of eyes that were given treatment. No clinical features showed significant association to posttreatment visual acuity except preoperative visual acuity improvement (Table 3).

This is the largest study of the clinical characteristics, ocular complications, treatment, and visual outcomes of patients with neuroretinitis in New Zealand to date, comprising 19 eyes in 17 patients. The diagnosis of neuroretinitis was made based on clinical presentation, serological results, and clinical progress. All patients presented with optic nerve head swelling and all either presented with or went on to develop the macular star. Despite the absence of cat scratch in all but one patient, the diagnosis of Bartonella neuroretinitis was made as it has been increasingly demonstrated that disease transmission can occur through other arthropod vectors such as ticks, biting flies, and other organisms yet to be identified [1,5,6]. All patients had infectious and inflammatory serological workup for other etiologies of neuroretinitis such as toxoplasma, toxocara [7], leptospirosis, syphilis, and sarcoid.

There is a subset of patients who present with optic disc oedema with macula star where the underlying etiology is a mimicker of neuroretinitis [8]. Conditions that fall into this category include hypertensive retinopathy, papilledema, anterior ischemic optic neuropathy, and toxic causes [3,9-12]. It is unusual for hypertensive retinopathy to present unilaterally [10]. In the 10 patients that did not demonstrate seropositivity to both Bartonella IgG and IgM antibodies, 4 patients had neuroimaging which would have excluded causes of papilledema such as space-occupying lesion or idiopathic intracranial hypertension. Arteritic anterior ischemic optic neuropathy primarily affects those above the age of 50 years [13], and only one patient fit this criteria of the remaining patients demonstrated improvement in visual function by 15 ETDRS letters. Toxic causes of optic disc oedema with macula star include agent such as bis-chloethyl-nitrosurea, procarbazine, and immune checkpoint inhibitors. None of the patients in this cohort reported taking any of these medications.

Four patients (24%) did not receive any systemic treatment, and on average achieved a worse final visual acuity of 74 letters compared to 80 letters in those who were treated. Those that were not treated also had a lower mean visual acuity at presentation of 45 compared to 51 in those who received treatment. Six patients (35%) were treated with both rifampicin and doxycycline, and has been reported in the literature to accelerate visual recovery [14]. The four patients who did not receive treatment were noted to have made clinical progress without treatment prior to Bartonella titers returning. The role of treatment in neuroretinitis remains a disputed area. Solley et al. [15] conducted a retrospective case series on 24 patients with choroidal, retinal, or optic disc manifestations of Bartonella disease and did not observe a difference in final visual acuity between 11 patients who received treatment and 13 patients who were not treated. Chi et al. [16] found no difference in the final visual acuity between treatment with systemic antibiotics including azithromycin, ciprofloxacin, tetracycline, co-trimoxazole, systemic corticosteroids, or combination treatment of both. Conversely, a noncontrolled small study by Reed et al. [14] did suggest efficacy of early antimicrobial treatment with doxycyline and rifampicin. Habot-Wilner et al. [17] found better final visual acuity in those who received combination treatment of antimicrobials (including doxycycline and rifampicin in combination as well as monotherapy, fluoroquinolones or macrolides monotherapy, or in combination with rifampicin) and corticosteroids, although the authors were unable to evaluate whether drug treatment was superior to no treatment. In their literature review of neuroretinitis from CSD, Purvin et al. [3] found treatment to be very heterogenous in the literature where 14 patients received no treatment, 37 patients were treated with antibiotics alone, 10 patients were treated with steroids and antibiotics, and 3 patients treated with steroids alone. The authors noted that based on small case series, combination treatment of doxycycline and rifampicin [14] and ciprofloxacin [18]. Based on the above, no definitive conclusions can be drawn about the efficacy of treatment and remains an area of controversy in the management of neuroretinitis.

Systemic symptoms have been reported to be associated with 50% to 70% of CSD cases [3,16,17,19]. In our study, six patients (35%) reported preceding systemic symptoms consisting of fevers, chills, reduced appetite, headaches, nausea, vomiting, diarrhea, and new rash, which is lower than that reported in the literature. Only one patient reported a history of being scratched by a cat, whilst 11 patients (65%) had cats or kittens at home. It is known that owning a cat or kitten or being scratched, bitten, or licked by an infected feline is a significant risk factor for CSD; however, up to 25% of people with CSD do not report close contact with cats, though transmission from another infected person has been reported [20].

All patients in our cohort demonstrated IgG seropositivity to B. henselae, whilst seven were also seropositive for IgM. The two main methods of detecting anti-B. henselae antibodies are through indirect immunofluorescence assay (IFA) and enzyme linked immunosorbent assay (ELISA). The former technique was used in this study for both IgG and IgM. IgG remains positive long after infection for up to 2 or more years, hence seropositivity may reflect past as opposed to current infection [21]. This contrasts with IgM, which are reported to be present in the acute phases of disease and decreases with time [22]. By 3 months, only 4% of patients with CSD are seropositive for IgM [23]. Both IFA and ELISA are reported to have low sensitivity in diagnosing CSD [21]. Using IFA, IgG sensitivity is reported to be 36% to 67% and specificity of 82%, and IgM sensitivity of 46% to 53% and specificity of 93% [21,24,25]. Based on these numbers, the probability of a false-negative IgG result in our cohort is up to 64%, and a false-negative IgM result of 54%. Furthermore, serological positivity also depends upon the genotype of B. henselae and also the cross-reactivity to other pathogens such as chlamydia. The genotype I of B. henselae has high rates of serological positivity (38%-75%) as compared to genotype II (7%-67%) [21]. It is therefore important to consider the entire clinical picture when interpreting serological results. In the setting where specialist laboratories are accessible, patients with inconclusive Bartonella serology can undergo polymerase chain reaction testing on a lymph node aspirate [26]. This technique is limited by its high cost as well as potential failure to detect atypical presentations [27]. Furthermore, metagenomics next generation sequencing is a novel technique that is highly sensitive and able to detect multiple pathogenic organisms using a single tissue sample, and has been reported to be able to diagnose B. henselae when standard techniques fail [26,28]. Metagenomics next generation sequencing is limited by its high cost, lack of universal reference standards, and interpretation difficulties.

A hallmark of neuroretinitis, the macular star, the result of accumulation of lipid-rich fluid in Henle nerve fiber layer was noted on initial presentation in eight patients (47%). Habot-Wilner et al. [17] in their large study of ocular manifestations of CSD comprising of 107 eyes from 86 patients reported neuroretinitis at presentation in 64% of their patients, and Chi et al. [16] documented 43% of their cohort of 62 eyes from 53 patients developed a macular star, which is comparable to our study.

This case series is limited by its small sample size and retrospective nature, with incomplete information documented in some patients’ records. A larger study population is required for validation purposes, ideally in the form of a double-blind randomized controlled trial to determine the efficacy of treatment in CSD as such a study is yet to be found in the literature.

The characteristic macular star is not always manifest at initial presentation for patients with neuroretinitis secondary to B. henselae, nor is there always a preceding history of cat exposure. Common systemic manifestations of neuroretinitis include fevers, chills, reduced appetite, headaches, nausea, vomiting, diarrhea, and new rash. Common ocular complications of neuroretinitis include macular oedema, vasculitis, uveitis, disc granuloma, and multifocal chorioretinitis. We found that B. henselae does not always yield positive IgM and IgG titers, even on convalescent samples. Not all patients require treatment to have a good visual outcome, and the immune status of the patient is important. Optic disc oedema in the absence of a macular star should not immediately exclude B. henselae infection, because some patients with neuroretinitis may not develop a macular star or develop it later in the disease process.