Treatment Regimen and Outcomes in Acute Retinal Necrosis – an IRIS^® Registry Study

Abstract

IRIS (Intelligent Research in Sight) Registry Study showing that acute retinal necrosis cases treated with systemic antivirals alone vs combined with intravitreal antivirals or early vitrectomy had statistically similar outcomes at 6 and 12 months.

Acute retinal necrosis (ARN) is a rare condition, often complicated by retinal detachments (RD) and poor visual outcomes.¹ Systemic antiviral medications are the most common treatment for ARN.² Clinicians often add adjunctive intravitreal antiviral agents ³ or early pars plana vitrectomy (PPV),¹ but the literature remains controversial on their benefits. ^1,4,5,6 This study aimed to determine the association between initial treatment regimen and RD rates and visual outcomes in patients with ARN in the American Academy of Ophthalmology IRIS^Ⓡ Registry (Intelligent Research in Sight). The study was exempt from Institutional Review Board approval and adhered to the tenets of the Declaration of Helsinki.

We identified patients with a new diagnosis of ARN between 2016 to 2020 by the International Classification of Disease (ICD)-10 codes in the IRIS Registry (Supplementary Methods, https://www.aaojournal.org). Figure 1 presents our cohort selection process. Only patients with documented systemic antiviral medications initiated within ± 14 days of the first instance of an ARN ICD-10 code were included; and we excluded cases (1) without known laterality; (2) who did not remain on systemic antivirals 4–8 weeks following diagnosis or did not receive an antiviral intravitreal injection within the same period; and (3) with a diagnosis of RD (ICD-10 codes in Supplementary Methods, https://www.aaojournal.org) within 1 year preceding or up to 31 days after the initial diagnosis of ARN. The last criterion was established because we were interested in cases of RD after ARN.

Figure 1.

Schematic representation of study design and selection criteria. ARN – acute retinal necrosis, RD – retinal detachment, NDC – National Drug Code, CPT - Current Procedural Terminology; PPV – pars plana vitrectomy: IV – intravenous.

*Cases that were initiated on treatment before having an ARN ICD-10 diagnosis code were included because in clinical practice they may initially present with intraocular inflammation of unclear etiology and may be started on the appropriate treatment, but ARN may not be coded until the clinical picture evolves.

Using these criteria, a total of 410 eyes of 358 patients were included. Based on their initial treatment regimen, the following groups were defined:

1. systemic antiviral medications alone (group A): treated with valacyclovir, oral famciclovir or intravenous acyclovir (National Drug Codes (NDC) in Supplementary Methods, https://www.aaojournal.org);

2. intravitreal injections with systemic medications (group B): received at least one intravitreal injection of foscarnet or ganciclovir with matching laterality for ARN diagnosis (Supplementary Methods, https://www.aaojournal.org), in addition to systemic medications;

3. systemic antivirals with PPV +- intravitreal injections (group C): received any of the systemic treatments mentioned ± intravitreal injections and had early PPV (CPT codes 67036, 67039, 67040). As we were interested in “prophylactic” PPV, only procedures within 31 days of initial diagnosis with matching laterality and without a concomitant RD code were included.

Supplementary Table 1 (https://www.aaojournal.org) presents the final study population characteristics. RD occurrence at 6 or 12 months after ARN diagnosis was considered as the primary outcome. RD was defined by having an ICD-10 code (Supplementary Methods) more than 31 days after the diagnosis of ARN and with matching laterality. An RD was documented in 20% (n=82) of the cases at 6 months: 18.67% (n= 56) in group A, 22.58% (n=21) in group B, and 29.41% (n=5) in group C. No statistically significant differences were seen among the groups (p>0.244) in the final backward selection model, retaining age at ARN onset (p=0.013), race (p>0.093), ARN laterality (p<0.01) and best-recorded VA at onset (p=0.172) as covariates. Similar results were seen at 12 months, with rates of RD of 22.2% (n=91) for the entire population; 21.0% (n=63) for group A, 24.73% (n=23) for group B, and 29.41% (n=5) for group C (p>0.391). No significant associations were observed between RD rates and number of intravitreal injections within first month of diagnosis (p>0.921). Mixed-effects Cox proportional hazards regression models were then fitted to evaluate the risk of having an RD over time and log-rank tests were performed. Compared to systemic antivirals alone, treatment with adjunctive intravitreal injections or early PPV did not affect time to occurrence of RD (p>0.248) accounting for the covariates retained in the final backward selection model (age at ARN onset, race and ARN laterality) – Supplementary Figure 1. Additionally, a sensitivity analysis was done on a subset of the eyes from immunocompromised patients (n=12) and immune status did not influence the described outcomes (p>0.723).

For eyes with data available on best-recorded VA at 6 months and at 12 months (Supplementary Table), treatment with early PPV or adjunctive intravitreal antivirals was not associated with significantly different VA outcomes at 6 (p>0.360) and 12 months (p>0.232), accounting for baseline VA, age at ARN onset and ARN laterality. All analyses were conducted using R (version 4.0.3) and a significance level of P < 0.05 was used both for analyses and selection procedures.

This work is not without limitations. Namely, (i) there is currently no single ICD-10 identifier for ARN. We followed a rigorous procedure for case selection, but our criteria might have been too strict and eliminated some cases, especially of milder disease. Even though others have reported similar 1-year rates, ⁷ the observed rates of RD were lower than what has been described in some series, which may be related to excluding RD cases documented within 1 month of initial diagnosis; (ii) the IRIS Registry is limited to US participating practices, and coding for systemic medications is likely suboptimal; (iii) there is no information on initial treatment strategy rationale and it is possible that adjunctive treatments (i.e intravitreal injections or PPV) were used for more severe cases, thus this study is prone to indication bias; (iv) the number of eyes that had a PPV as initial treatment strategy is small which may affect our statistical calculations; (iv) there is no information available on initial extent of disease or causal virus and these are important factors; (v) we were unable to assess how the different treatment strategies controlled the degree of intraocular inflammation and the progression of the disease clinically. Despite these limitations, to our knowledge, this is the largest dataset of patients with ARN analyzed to date.

In conclusion, our data suggests that, compared to systemic treatment alone, initial treatment with adjunctive intravitreal antiviral medications or early pars plana vitrectomy did not significantly change rates of RD or VA outcomes at 6 and 12 months in patients with ARN. These findings are important because they support that, especially in areas of the world where adjunctive treatments are not readily available, initial treatment with systemic antiviral medications alone may not significantly compromise patient outcomes.

IRIS Registry Analytic Center Consortium Authors:

Suzann Pershing, MD², Leslie Hyman, PhD³, Julia A. Haller, MD³, Aaron Y. Lee, MD MSCI^4,5, Cecilia S. Lee, MD MS⁵, Flora Lum, MD⁶, ⁺Joan W. Miller, MD¹, ⁺Alice Lorch, MD MPH¹

Affiliations:

1 Massachusetts Eye and Ear, Harvard Medical School, Boston, MA

2 Stanford University, Palo Alto, CA, USA

3 Wills Eye Hospital, Philadelphia, PA, USA

4 eScience Institute, University of Washington, Seattle, WA, USA

5 Department of Ophthalmology, University of Washington, Seattle, WA, USA

6 American Academy of Ophthalmology, San Francisco, CA, USA