Long-term Observation of Chinese Idiopathic Retinitis, Vasculitis, Aneurysms, and Neuroretinits (IRVAN) Patients and suggestion of a Revised Staging System

Background To investigate the clinical characteristics and intervention effects on Chinese patients with idiopathic retinitis, vasculitis, aneurysms, and neuroretinitis (IRVAN). Methods The consecutive series patients received eye examinations including fundus uorescein angiography (FFA), optical coherence tomography (OCT) and systemic testing. Laser, including pan-retinal photocoagulation (PRP), oral corticosteroid and Pars plana vitrectomy (PPV) were used during the investigation period. Results Forty-two eyes of 21 patients aged 15-58 years old, 19 females and 2 males, initial decimal BCVA NP ~1.5 (0.55±0.38), were included. Eighteen eyes reached Stage 2; 21 eyes with neovascularization of disc or elsewhere and/or vitreous hemorrhage (VH) reached Stage 3; one eye had neovascular glaucoma (NVG) at Stage 5. Two eyes had proliferative vitroretinopathy (PVR). Thirty-four eyes of 20 patients of stage 2 and 3 accepted retinal photocoagulation. 27 eyes completed PRP. PPV was performed for 3 eyes at the rst visit due to VH or PVR. Three eyes developed to Stage 3 from Stage 2. One eye of Stage 3 developed to PDR with retinal detachment. Additional 5 eyes, 3 heavy VH, 1 PVR and 1 macular epiretinal membrane eye, received PPV during the follow-up. Intra-retinal microvascular abnormality (IRMA) was found in 7 eyes. Aneurysms on the optic nerve head and artery bifurcations disappeared in 8 eyes and decreased in number in 2 eyes after one year of photocoagulation. Seven eyes were found with BCVA ≤ 0.1, of which 3 experienced PVR, 2 exudative maculopathy, 1 acute macular neuroretinopathy (AMN) eye and 1 eye due to NVG. The BCVA of the last visit was NLP~1.2 (0.53±0.38), no signicant difference was found with initial BCVA.

eyes and decreased in number in 2 eyes after one year of photocoagulation. Seven eyes were found with BCVA ≤0.1, of which 3 experienced PVR, 2 exudative maculopathy, 1 acute macular neuroretinopathy (AMN) eye and 1 eye due to NVG. The BCVA of the last visit was NLP~1.2 (0.53±0.38), no signi cant difference was found with initial BCVA.
Conclusions Female is more susceptible to IRVAN. IRMA and AMN are rstly described in IRVAN patients. PVR and exudative maculopathy, the major causes for severe visual impairment, are suggested in the revised staging system.

Background
Idiopathic retinitis, vasculitis, aneurysms, and neuroretinits (IRVAN) may lead to blindness due to the development of proliferative vitroretinopathy (PVR) and neovascular glaucoma (NVG) [1,2]. IRVAN is considered to be diagnosed by three major criteria (retinal vasculitis, aneurismal dilatations at arterial bifurcations and neuroretinitis) and three minor criteria (peripheral capillary nonperfusion, retinal neovascularization, and macular exudation) [1,2]. As IRVAN is rare, the condition was often misdiagnosed as peripheral retinal vasculitis or Eales disease until Chang et al [1] described the clinical features of IRVAN in ten patients and Samuel et al [2] published a paper with observations of twenty-two IRVAN patients, including the above-mentioned ten patients. A stage system was also proposed by Samuel et al [2] in 2007. However, the diagnosis and management of IRVAN is still a challenge to ophthalmologists.
We rstly noticed a young girl with typical characteristics of IRVAN over 10 years ago. Her condition was diagnosed according to the criteria suggested by Chang et al [1]. Since then more IRVAN cases have been recognized and diagnosed in our outpatient department. We would like to present our long time observation outcomes of our series of Chinese IRVAN in this paper.

Methods
We conducted a retrospective study of 21 consecutive cases of patients diagnosed as IRVAN at the outpatient department of the Eye & ENT Hospital of Fudan University between April 2003 and March 2018. The diagnoses were made according to IRVAN's three major criteria [2]. The diagnosis was considered for both eyes if one eye of the patients met the three major criteria even the fundus of the other eye could not been seen due to vitreous hemorrhage (VH) and/or PVR.
Optical coherence tomography (OCT) (Heidelberg Engineering, Heidelberg, Germany) was also performed for patients when it was available.
Patients were classi ed to different stages according to the system suggested by Samuel et al [2] at the rst and last visits: Stage 1 includes macroaneurysms, exudation, neuroretinitis and retinal vasculitis; Stage 2 FFA evidenced peripheral capillary nonperfusion; Stage 3 neovascularization of the disc or elsewhere and/or VH; Stage 4 anterior segment neovascularization; Stage 5 NVG (Table 1).
The methods of the initial management and subsequent treatments were reviewed. The nal clinical manifestation features and reasons for poor visual acuity (≤ 0.1) were assessed.
The BCVA at rst and last visit and between eyes of different stages (Stage 2 and Stage 3) or groups was analyzed using paired t test or t test with 95% con dence intervals. P values < .05 were considered statistically signi cant. Statistical analysis was performed using commercially available software (SPSS software for Windows, Version 16.0; SPSS Inc, Chicago, Illinois, USA).

Results
Twenty-one Chinese patients (19 female and 2 male) aged between15-58 (mean 37±11) years old were included in this study. The presenting symptoms were usually oaters, blur vision and even lost vision of one eye for 1 to 6 months. One patient had no symptoms and was found retinal vascular abnormality and hemorrhage during health examination. The initial mean decimal BCVA was 0.55±0.38 (LP ~ 1.5). All the patients were FFA proved IRVAN bilaterally at the rst visit except 4 patients who had one eye with VH (2 eyes) and/or PVR (2 eyes). Two eyes with VH were proved with typical IRVAN by FFA after pars plana vitrectomy (PPV). Two PVR eyes did not have FFA examination, although one of the eyes accepted PPV. Two patients were lost follow-up including the patient with a NVG eye. The follow-up time of the rest 19 patients 38 eyes was 6-120 (70±36.4) months. The demographic information and clinical features are summarized in Table 2.
No eye in this series cases agreed with Stage 1. Eighteen eyes (18/42) with FFA evidenced peripheral capillary non-perfusion ( Fig. 1) that agreed with Stage 2, BCVA 0.04-1.2 (0.59±0.32). Twenty-one eyes (21/42) with neovascularization peripherally at the posterior edge of the non-perfusion, one eye with macular and two eyes with papillary neovascularization and/or VH reached Stage 3, BCVA 0.01-1.5 (0.57±0.33). One eye (1/42) had NVG at Stage 5 (BCVA 0.05) although scattered photocoagulation was performed before the patient came to the authors. The staging of the eyes is illustrated in Table 3.Two eyes (2/42) (BCVA 0.02 and LP respectively) had an obscuring view of the fundus and B-ultrasound scanning revealed a whole RD 2-4mm in the LP eye, with severe PVR in both eyes. These two eyes did not agree with any stages suggested by Samuel et al [2] (see Table 1).
Laser photocoagulation of non-perfusion areas were intended to be performed in eyes only with capillary non-perfusion of stage 2. Pan-retinal photocoagulation (PRP) for stage 3 or more severe eyes was indicated. Direct photocoagulation of the aneurysms was forbidden. For younger patients with obvious vessel leakage elucidated by FFA, oral corticosteroid was prescribed according to the doctor's judgment. With heavy VH and PVR or if the VH developed and did not resolved within at least two weeks at the initial visit or during the follow-up period, PPV was performed according to the clinical manifestation and doctor's consideration.
Retinal photocoagulations were performed by several doctors including doctors of other hospitals. Ten eyes (10/18) received photocoagulation of the non-perfusion areas only ( Figure 2) and pan-retinal photocoagulations (PRP) were ful lled in 8 eyes of stage 2 (8/18) nally ( Figure 3). Nineteen eyes (19/21) of stage 3 accepted PRP. Two VH eyes of stage 3 and 1 PVR eye were indicated to PPV at the rst visit and had PRP during the operation. Patient 4, with whole RD of the left eye, refused to accept any operation. Seven eyes had recurrent VH during or after photocoagulation, 4 eyes recovered spontaneously, 3 eyes with heavy VH received PPV. Among them, case 3 had PRP completed in both eyes but her left eye had fog-like blurred vision now and then due to VH and when the recurrent VH decreased the vison to 0.5 and persisted for 2 weeks, PPV was performed for the eye. The vision recovered to 1.0 and VH then did not appear till the last visit.
Overall, 34 eyes accepted retinal photocoagulation, 27 eyes completed PRP. Three eyes developed to Stage 3 from Stage 2. One eye of Stage 3 developed to PDR with retinal detachment, which had not been classi ed in any of the stage by Samuel et al [2] system. PPV was performed for 3 eyes at the rst visit due to VH or PVR. Additional 5 eyes, 3 heavy VH, 1 PVR and 1 macular epiretinal membrane eye, received PPV during the follow-up ( Table 2).
The BCVA was NLP~1.2 (mean decimal 0.53±0.38) at the last visit. There was no signi cant change in BCVA between the initial and last visit of the patients. The nal mean decimal BCVA of Stage 2 and Stage 3 eyes was 0.58±0.33 and 0.55±0.41 respectively. There was no signi cant difference between the initial and the nal vision either (Table 3).
Severe visual impairment (BCVA≤0.1) was estimated and found in 7 eyes (7/42) ( Table 4). One eye (lost follow-up) was found at the rst visit due to NVG. Three eyes experienced RD or/and PVR (among them 2 eyes had no light perception at the last visit) and 2 eyes due to exudative maculopathy with dense macular lesion and 1 eye had macular atrophy due to acute macular neuroretinopathy (AMN) which was 0.4 at baseline and 0.05 at the last visit.

Special Fundus Findings and Evolution
Aneurysms on optic disc head and artery bifurcations Forty-one eyes of twenty-one patients were found aneurysms on the optic disc head and artery bifurcations at the baseline which were illustrated more clearly with FFA. During the follow-up visits, aneurysms on the optic nerve head and artery bifurcations disappeared in eight eyes of four patients ( Figure 3). The aneurysms were found decreased in number and size in both eyes in one case. The shortest time observed for this phenomenon was one year. All these patients had received laser photocoagulations. However there was no signi cant difference in BCVA between the initial and last visit in these eyes (Table 3).

Intra-Retinal Microvascular Abnormality (IRMA)
IRMA was found in the macular area, or nearby, in 7 eyes (Figure 3). Macular neovascularization was also found in the right eye of case 14. One eye with IRMA developed macular neovascularization in the later visit. IRMA was not found to retreat in any of the eyes during the follow-up time. No signi cant difference was found in BCVA between IRMA group and non-IRMA group, neither between the initial and last visit in eyes with IRMA (Table 3).

Venous abnormality
Thirty eyes of15 patients demonstrated mild to moderate venous uorescein leakage either as segmented on the main branch of veins and/or peripheral venules at the rst visit. Fluorescein leakage usually disappeared after photocoagulation in all the patients at the last visit.
Exudates, edema of maculae and exudative maculopathy Exudates were found in 28 eyes at the rst visit. The exudate varied in amount from several dots or radiant stars to heavy yellow exudates and as well as in area from peri-papillary or involved the macular area. Massive macular and foveal exudates with macular edema were found in 2 eyes of 2 cases and dense ovoid lesions developed in these heavily exudate eyes and eventually led to severe damage to vision due to maculopathy (BCVA≤0.1) (2/42). The exudates of other eyes reduced gradually during the observation period.
Exudates did not found in the posterior area of 11 eyes of 7 patients; unknown in three eyes because VH or PVR meant that fundus could not been seen. One patient had obvious exudates in the right eye, while the left eye had no exudates at all (Figure 3

Discussion
Although the syndrome was rst described by Kincaid and Schatz [3], it was named IRVAN by Chang et al [1] and a stage system was proposed by Samuel et al [2] after observation of 22 patients. Several papers about IRVAN patients, describing one to seven cases (mostly one case) for short-term observation were published subsequently [4][5][6][7][8][9][10]. However the diagnosis of IRVAN remains a challenge to ophthalmologists as the disputed diagnoses of IRVAN cases can even be found in published papers. The etiology of the disease is still not clear. There are no obvious in ammatory cells in the anterior segment and vitreous as well as the corticosteroids does not give a de nite control of the disease [2], the in ammatory and immunopathogesis of the disease is not so convinced. Arterial involvement is also rarely seen in vasculitis [11]. The macroaneurysms on the optic disc and artery, particularly capillary nonperfusion are more suggestive for ischemia origin close to Coates disease and diabetic retinopathy with an arising of vascular endothelial Growth factor (VEGF) [10].
The number of patients in this series (21 cases) are similar to the Samuel et al [2] (22 cases), although ours were Chinese while theirs were including mainly Caucasian (13/22) and other races, but no Chinese. The average age of this group at diagnosis time was 37  year-old which is comparable to that of the patients (average 31.5, 9-60 years old) reported by Samuel et al [2]. Most patients in our group were female 90.5% (19/21) and it was 68.2% (15/22) in the report of Samuel et al [2]. The follow-up time was also comparable; ours was 70 ± 36.4 (6-120) months while theirs was 50 (2-132) months.
There was no systemic nding in our group of patients as Samuel et al [2] reported.
It seems young female are more susceptible to IRVAN as shown both in our group and Samuel et al [2]. Young females have been considered to have higher risk in many vascular retinopathies such as AMN [12] and diabetic retinopathy [13] due to the narrower venule calibers and lower ratio of arterial length to diameter than males.
Patients in our group usually came to the clinic because of oating or even severe visual decrease. We also had patients exhibiting no symptoms but found retinal hemorrhage due to routine health examination. The diagnosis of all the patients was con rmed by FFA evidence of large extant areas of capillary non-perfusion and even peripheral neovascularization in all eyes except 2 eyes with PVR. We had patients who came with VH in one eye and were found to have IRVAN by FFA in both eyes after PPV. As IRVAN is usually considered bilaterally [1][2], therefore we diagnosed bilateral IRVAN in these patients even when the fundus of one eye could be seen as in case 4 who did not accept PPV due to PVR and complete RD.
The natural course of IRVAN is now believed to be severe or even devastating if without proper intervention [1][2]. The eyes with severe visual impairment in our group were 16.7% (7/42) which is less than that of Samuel et al [2] (8/32). The reasons for severe visual impairment are also assessed in this group of IRVAN patients. PVR, maculopathy caused by macular exudates and NVG which agreed with Samuel et al [2] and AMN which has not been described in IRVAN patients before, contributed to severe damage to vision in this group of eyes. The eyes undergoing PPV at the rst visit with vision remained stable even improved. During the follow-up period, many eyes with recurrent VH during or after photocoagulation recovered spontaneously. Three eyes developed heavy VH and received PPV and stabled ever after. PPV is believed to remove the elevated vascular endothelial growing factor (VEGF) and other in ammation factors and early PPV for persisted VH may ease the proliferative or leakage of the retinal vessels of IRVAN.
IRMA which has not been reported in IRVAN patients was found in 7 eyes of our IRVAN patients. One eye was already and the other eye developed with macular neovascularization. IRMA is considered an indication of the severity of retinal vascular disease and the precursor of the neovascularization as described in diabetic retinopathy [14].
Laser photocoagulation, corticosteroid systemically and/or intra-vitreally have been tried in treatment for IRVAN [1][2][3][4][5][6][7][8][9]. Systemic corticosteroid was also used on younger patients of our group of patients for months, but no obvious effect was observed as many previous papers [1,2] reported. The method of laser photocoagulation is also controversial. In our group of patients, those with peripheral capillary nonperfusion and neovascularization received pan-retinal photocoagulation. Patients only with peripheral capillary non-perfusion (stage 2) were suggested to be coagulated on the non-perfusion area and under close follow-up.
Samuel et al [2] proposed a function-staging system including ve stages as described previously and in Table 1. According to this staging system, initial stage of the eyes were classi ed into Stage 2 (18/42 eyes); Stage 3 (21/42 eyes) with neovascularization of the disc or elsewhere and/or VH, and one eye had NVG at Stage 5. However two eyes which had PVR and RD (BCVA20/1,000-NLP) did not match any of the ve stages. In order to solve this dilemma, we suggest increasing a stage of PVR and RD to re ect the natural course of IRVAN (Table 1).
The exudates and edema at the macular area, particularly involving fovea, were found as important factors impairing visual acuity [2]. The staging systems of familial exudative vitreoretinopathy (FEVR) [15] and Coats' disease [16], both retinal vascular diseases, give a detailed presentation of involvement of the exudates and RD and a guide for proper treatment. In diabetic retinopathy, another retinal vascular disease, macular exudates are also considered seriously as a factor for impairment of vision [14], although they were not included in the staging. So we would strongly suggest that macular exudates should also be considered in the revised staging system. Therefore, according to the staging criteria of Samuel et al [2] for IRVAN, Kashani et al [15] for FEVR, and Shields et al [16] for Coats disease, our suggestion for the revised stage system is proposed (Table 1).

Study Limitation
This study included the second largest reported and the largest of Chinese IRVAN patients in number. However, as IRVAN is a rare disease, it is di cult to collect larger cohort of patients for evaluate the patients. Our group of patients takes more than a decade to include and the examination methods varied even FFA experienced lm, data and nally wide eld images in recent cases. OCT was only performed in part of the cases and do not mention about the optical coherence angiography because of only available in recent years. Treatment was performed by the judgment of different doctors and patients' willingness. These may all in uence the estimation and nal outcome of the patients.

Conclusions
In this study, 90.5% patients are female. IRMA and AMN are rstly described in IRVAN patients. PVR and maculopathy are found to be the major reasons for visual impairment, so that a revised staging system with an increase of the PVR and retinal detachment stage and macular exudates are recommended. We suggest that bilateral diagnosis of IRVAN could be made if one eye with VH or PVR blocks the fundus and the contra-lateral eye agrees with the IRVAN diagnosis. Earlier laser photocoagulation and PPV for heavy VH are recommended for better prognosis.