Study design and ethical considerations

The nationwide registry is named “Japan-acute retinal necrosis (J-ARN) registry”. The present research is a combined prospective and retrospective, multicenter, observational study of ARN in Japan from 1 April 2014–31 March 2029. The use of a combined retrospective and prospective design is due to the limited scope of the research that focuses on a rare and intractable disease with diverse treatment approaches. Recruiting a large number of participants is anticipated to be challenging.

During the prospective recruitment phase, participating sites will enroll consecutive patients as they visit for routine follow-up, to reduce the risk of selection bias, and Fig 1 outlines timepoints of enrolment, interventions, and outcomes. Recruitment commenced on 1 April 2025. For the retrospective component, medical records of the prospectively recruited patients will be reviewed retrospectively back to 1 April 2014, to accurately determine the index date of the initial diagnosis of ARN. This hybrid approach aims to maximize the use of available data while ensuring robust and unbiased participant inclusion, despite the challenges associated with studying a rare disease.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. SPIRIT flow diagram for the schedule of enrolement, interventions, and assessments in our prospective study.

Abbreviations: visual acuity (VA), herpes simplex virus (HSV), varicella zoster virus (VZV), proliferative vitreoretinopathy (PVR).

https://doi.org/10.1371/journal.pone.0323822.g001

The Oita University Hospital Clinical Research Supervision Center, Biomedical Research Ethics Review Board approved the study protocol (No. 2810-D28) on 23 May 2024, and the study is registered in the University Hospital Medical Information Network Clinical Trials Registry (UMIN000056246). The Tokyo Medical University Hospital Research Ethics Committee approved the study protocol (No. E2024-0071) on 1 November 2024. The study will adhere to the principles of the Declaration of Helsinki and the Ethical Guidelines for Life Science and Medical Research Involving Human Patients.

For the prospective component of the study, written informed consent will be obtained from all recruited participants, including consent for the PCR assay. However, written informed consent from patients will not be required for retrospective review of clinical charts, as the study design does not involve randomization. Instead, an opt-out approach will be employed in accordance with the Ethical Guidelines for Medical and Biological Research Involving Human Subjects.

Sample size, inclusion and exclusion criteria

The goal of the study is to register a total of 500 eyes treated for ARN. The sample size was determined based on an estimation of number of eligible patients from a national uveitis survey in Japan [6]. Patients who meet the international diagnostic criteria for ARN and who have given written consent to participate in the prospective study, including intraocular fluid examination, will be eligible for registration [2,8]. Exclusion criteria are: (1) patient with retinal necrosis caused by pathogens other than HSV-1, HSV-2 or VZV, such as cytomegalovirus (CMV), (2) patient deemed unsuitable for the study by the investigator, and (3) patient in whom related diseases cannot be excluded despite meeting the inclusion criteria. In the case of exclusion from the study, clinical data until the date of consent acquisition will be registered, after which the study will be terminated.

Diagnosis

Acute retinal necrosis is diagnosed based on the standard diagnostic criteria [2,8,9], including typical clinical presentation with peripheral necrotizing retinitis, and positive virologic tests using PCR analysis of intraocular fluids.

Registered data

The following data will be collected (Table 1): height, weight, age, sex, laterality of ARN, medication history, history of shingles vaccination and herpetic diseases, drinking and smoking history, systemic disease, ocular disease, days from onset of ocular symptoms to first visit, subjective symptoms, best-corrected visual acuity, spherical equivalent, intraocular pressure, axial length, lens status, slit lamp finding (anterior chamber cells and flares, mutton-fat keratic precipitates), presence of vitreous opacity, fundus appearance (extent of yellowish white lesions), presence of retinal holes/tears, status of optic disc (swelling, redness, atrophy), status of the macula (edema, epiretinal membrane, necrosis), optical coherence tomography (OCT) data, treatment (antiviral agents, systemic corticosteroid, photocoagulation), date of surgery, surgical procedure, and complication.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. List of information collected for J-ARN registry.

https://doi.org/10.1371/journal.pone.0323822.t001

Decimal best-corrected visual acuity will be evaluated at the initial visit and at follow-up after 6, 12, 24, and 36 months. Decimal best-corrected visual acuity will be converted to logMAR units for statistical analysis. The collected data will be anonymized and registered online. The data center will be located within the hosting service of G-INGS Inc. (Fukuoka, Japan), ensuring maximum security for data storage and management. All data will be reviewed at least twice by the committee. The Committee will contact the participating institution for verification if any inconsistency or outlier is identified. If confirmation is not possible, the Committee will make the final decision.

A peripheral blood will be collected at the initial visit and thereafter as needed. Nonspecific inflammation tests such as erythrocyte sedimentation rate and blood count are sometimes included in the routine laboratory evaluation. The results of these tests are also essential for initiating therapy with anti-viral drugs and systemic corticosteroids.

Safety

Because this is an observational study, participation does not involve any risk beyond those associated with the standard care for ARN in Japan. The registry does not contain any information that could identify patients. Each patient is assigned a unique number, and ophthalmologists are responsible for maintaining separate records of patient identities.

Publication plan

The research findings derived from the registry will primarily be disseminated through peer-reviewed publications. These publications are anticipated to guide clinical practice and support the development of educational materials tailored for patients. Furthermore, the results may be shared at international uveitis conferences, aiming to foster global collaboration and knowledge exchange. Authorship of these publications will be attributed to the research group, which will include participating ophthalmologists and other research staff.

Statistical analysis

The clinical profile as well as medical and surgical treatment options will be analyzed using descriptive statistics. Biostatisticians and information technology experts (M.A and M.S) will be involved in the planning, execution, and evaluation of the analyses. Multivariate analysis will be performed on changes in visual function over time and complications to identify influencing factors. Best-corrected visual acuity (BCVA) is converted to logarithm of the minimum angle of resolution (logMAR), and visual acuity worse than 0.01 is converted according to previous report [10] as follows: counting fingers, 2.6 logMAR; hand motions, 2.9 logMAR; light perception, 3.1 logMAR; no light perception, 3.4 logMAR. Data of continuous variables are expressed as mean and standard deviation or median and range, depending on normality of data distribution. Non-parametric data will be analyzed by Mann-Whitney U test or Kruskal-Wallis test, while parametric data will be analyzed by Student’s t-test or one-way ANOVA. Categorical variables will be expressed as total number and percentage, and the data will be analyzed by chi-squared test or Fisher’s exact test. The impact of clinical presentations, viral load, diagnostic methods, treatments, and complications will be estimated by odds ratios (ORs) with 95% confidence intervals (CIs). Survival analyses, especially Kaplan-Meier estimates, will be performed to evaluate visual prognosis, progression-free interval, and complications. Time-to-event data for 6, 12, 24, 36 month from the initial visit, the date of diagnosis, and/or vitrectomy were analyzed. A P value less than 0.05 will be considered statistically significant. Multivariate analyses (such as logistic regression analysis, cox proportional hazards regression, generalized estimating equations, and covariance structure analysis) will consider potential confounding variables related to treatment selection and outcomes.

There are primarily four types of machine learning algorithms: random forests, support vector machines, decision trees, and simple Bayesian classifiers. These algorithms are employed to integrate clinical data and image information for analysis purposes. Support vector machines utilize distinct kernels, and each method option and parameter is selected through cross-validation and bootstrap methods. This process aims to identify the combination that optimizes generalization. In contrast to deep learning, which necessitates a substantial number of cases, these machine learning methods ensure a certain level of accuracy even with a limited number of cases. The machine learning algorithms are designed to predict the outcomes of disease onset, diagnosis, treatment, visual function, and clinical decision-making.

Discussion

Acute retinal necrosis (ARN) is a severe, rapidly progressive viral uveitis that leads to significant visual impairment or blindness if not treated promptly [11]. Despite advances in antiviral therapies and minimally invasive surgery, ARN remains a challenging disease to manage due to its aggressive nature and the complexity of its diagnosis and treatment. This article reports on the development of Japan’s national clinical registry, which will provide real-world data on the clinical course, management, and visual outcome of ARN. Using a multicenter approach, we aim to overcome the challenge posed by the low prevalence of ARN, which has been a major obstacle to advancing research and improving suboptimal outcomes. In addition to tracking visual outcomes associated with different treatment modalities, the registry will serve as a valuable resource for addressing key controversies surrounding the management of ARN. The mainstay for ARN treatment is systemic antiviral therapy, although which specific antiviral agent (e.g., acyclovir, ganciclovir or foscarnet) is superior to the others remains to be identified. In addition to systemic antiviral therapy, numerous studies have evaluated the therapeutic effects of adjunctive treatments such as intravitreal antiviral injections, systemic steroids, systemic immunosuppressants, prophylactic laser therapy, and prophylactic vitrectomy for the management of ARN and the prevention of retinal detachment [12–15]. The role of early and/or prophylactic vitrectomy in eyes without retinal detachment has been the subject of considerable debate. Some studies, such as that of Ishida et al. [16], have reported that prophylactic vitrectomy may be effective in preventing retinal detachment. However, other studies from Japan have shown that prophylactic vitrectomy does not improve visual outcome, because the outcome is largely determined by the extent of retinal and optic nerve damage at presentation, irrespective of the presence or absence of retinal detachment [17]. According to a meta-analysis of 265 eyes with ARN conducted by Fan et al. [18], prophylactic vitrectomy reduces the risk of retinal detachment. These findings suggest that although prophylactic vitrectomy may be beneficial in certain cases, it may not universally prevent retinal detachment in ARN patients due to the variability in efficacy. Therefore, the decision to perform prophylactic vitrectomy should be individualized, taking into account the extent of necrotic lesions and potential postoperative complications. The registry may provide insight into the effectiveness of different treatment regimens and their impact on visual outcome.

This registry has the potential to resolve the controversies by documenting patterns of clinical features, standardizing test procedures such as intraocular fluid sampling, and reviewing cases of misdiagnosis. Another important question that remains unanswered is the optimal approach to vitrectomy in patients with ARN—whether silicone oil tamponade, gas tamponade, scleral buckling, and photocoagulation should be used. The registry can address this issue by tracking visual outcomes after vitrectomy. The underlying challenge is the lack of robust evidence to determine the most effective treatment strategies for ARN patients, such as the type of treatment, whether to perform a vitrectomy alone or a combined procedure, and the timing of adjustment to the treatment regimen.

The second rationale for the proposed study is related to the fact that the number of viral copies in the intraocular fluids of patients with ARN significantly affects the visual prognosis of ARN.^19-21 Calvo et al. [19] identified the presence of more than 5.0 × 10⁶ copies of VZV DNA in the aqueous humor of ARN patients as a poor prognostic factor. Another study [20] and our previous study [4] have reported that VZV-ARN is associated with a higher risk of retinal detachment. In general, VZV-ARN may be associated with worse visual prognosis compared to HSV-ARN [20], but a recent meta-analysis found no evidence to support a difference in visual outcome [21]. Therefore, a better understanding of virus types, viral copy number and clinical features may be important to improve visual prognosis of ARN and guide treatment options. In this regard, the proposed study provides an opportunity to expand current understanding of ARN.

Despite aggressive treatment, ARN carries a poor prognosis with high rates of retinal detachment and severe visual impairment [4]. A recent study reported that initial retinal detachment was detected in 2% of ARN patients, increasing to 47% during the follow-up period [13]. However, even after successful retinal reattachment, visual outcome often remains poor due to complications such as maculopathy and optic neuropathy [22]. The registry will track long-term outcomes including visual acuity, retinal detachment rate, complications, and recurrence, providing valuable data to guide future management strategies. Several studies have reported a significant association between the involved zone and the development of retinal detachment [13,23–26]. Furthermore, patients aged 50 years or younger diagnosed with ARN are more likely to develop retinal detachment and more severe ocular inflammation [27]. Several studies have investigated the prognostic factors that influence the clinical outcome of ARN. Choi et al. [28] identified poor initial visual acuity, necrotic retinitis in the central zone, and delayed retinal detachment onset as factors that significantly increase the risk of severe vision loss. Kim et al. [24] found that early diagnosis correlated strongly with better visual outcomes. Butler et al. [25] reported that zone 1 involvement and widespread retinal lesions were associated with worse visual prognosis. The proposed registry will evaluate the association between the development of retinal detachment and various clinical characteristics, and provide valuable data to guide future management strategies.

This registry study has several limitations. Because ARN is a rare disease, a combined retrospective and prospective study design is adopted to maximize data collection. For example, data extracted retrospectively from medical records may be incomplete due to the inherent limitations of retrospective data collection or insufficient data sources. Consequently, retrospective data may be more susceptible to selection bias than data collected prospectively. Furthermore, in this observational study, treatment methods are determined according to the clinical judgment of the ophthalmologists, making it impossible to control for variability. The number of patients included in the study will be small, and the results may not be generalizable to all regions of Japan. Clinical registries provide valuable opportunities to collect data in real-world clinical settings; however, there are inherent limitations to this research approach. As highlighted by Tan et al. [29] in their review of clinical registries in ophthalmology, limitations of clinical registries include potential selection bias due to the voluntary nature of participation by ophthalmologists. Additionally, missing information and typographical errors in data entry are noteworthy concerns. In future review of the registry, the scope of treatment descriptions could be broadened to include specifics such as the antiviral drugs used (e.g., acyclovir and foscarnet), dosages, and administration methods. This protocol will ensure that the field continues to explore innovative approaches. Accordingly, we expect the scope of data collected in the registry to expand as the project gains momentum.