Rubella virus-associated uveitis: The essentiality of aqueous humor virological analysis

Aims / background Rubella virus-associated uveitis (RVAU) classically presents with the clinical features of Fuchs uveitis syndrome (FUS). We report a series RVAU, and discuss the relevance of available diagnostic strategies, and how vaccination could potentially prevent disease. Methods We retrospectively included patients with RV-positive aqueous humor (AH) with RT-PCR and/or intraocular RV-IgG production, between January 2014 and December 2019. RV-IgG titers from AH and serum were compared with other virus-specific IgG titers (VZV and/or CMV and/or HSV-1), to determine the derived Goldmann-Witmer coefficient (GWC’). Clinical findings at presentation and during follow-up are reported, as well as the anti-RV vaccination status. Results All 13 included patients demonstrated intraocular synthesis of RV-IgG (median GWC’: 9.5; 3.2–100). RV-RNA was detected in one patient while PCR results were negative for other HSV1, VZV and CMV. The mean delay in diagnosis was 13 ± 12.6 years, with an initial presentation of FUS in only 3 patients (23%). Only four patients had been vaccinated, but all after the recommended age. Conclusion As RVAU is a pleiomorphic entity, virological analysis (RV RT-PCR and GWC’) of aqueous humor is essential to improve the diagnosis and management of this entity. Improper vaccination against RV appears to be implicated in RVAU.


Introduction
Rubella virus (RV) is a strictly human pathogen transmitted via aerosols to infect the respiratory tract. 1 RV infections of unimmunized children or adults are usually benign but when it occurs during embryogenesis it can cause congenital rubella syndrome (CRS), 2 which typically includes cataracts, cardiac abnormalities and sensorineural deafness. 3,4 Rubella and CRS are close to being eliminated in many countries due to implemented immunization programmes. However, since the early 2000's, RV has been associated with the occurrence of chronic uveitis. [5][6][7] The clinical picture of RV-associated uveitis (RVAU) is somewhat comparable with the so-called Fuchs uveitis syndrome (FUS) in a significant proportion of cases. This clinical entity is characterized by low grade, typically unilateral inflammation of the anterior segment associated with peculiar iris heterochromia and nodules, diffuse keratic precipitates (KPs), a predisposition to develop cataracts and glaucoma and a mild vitritis. 8 FUS has been linked to infection with RV, Herpes simplex virus type 1 (HSV-1) and cytomegalovirus (CMV). 7,[9][10][11] The latter is common in Asian patients, with an onset of disease between the third to fifth decade, while RV predominates in Europeans and occur usually in the second to third decade of life. 7,10,12 Groen-Hakan recently reported a large series of RVAU with none of the patients being vaccinated against RV, which was compelling for the still-debated role of RV in causing uveitis. 7 Visual prognosis associated with FUS and more generally with RVAU, is considered good. 7,13 However, diagnosis can be very challenging in various situations. In children, the disease may present initially with an incomplete clinical picture. 14 Whereas clinical FUS landmarks (heterochromia) can be missed in adult, especially those with dark irides, 15,16 or with bilateral diseases. 17 Furthermore, corticosteroids are ineffective and deleterious, hastening the occurrence of cataract and glaucoma. 8 Thus, FUS and RVAU misdiagnoses lead to inappropriate management and iatrogenic complications.
Virological analysis of aqueous humor (AH) is of major importance, 7 by influencing the prevention of useless corticosteroids or antiviral therapy and ultimate iatrogenic complications. 18 However, virological diagnosis is not yet standardized. 19 Two virological techniques are currently available: reverse transcriptase polymerase chain reaction (RT-PCR) to detect RV RNA in AH, or an immunoassay to detect intraocular anti-RV IgG. 20 In this study, we report a series of proven RVAU and discuss the relevance of available diagnostic strategies, to improve patient management and how vaccination could potentially prevent disease.

Patients and methods
Patients with RT-PCR confirmed RV-positive AH and/or intraocular RV-IgG production, between January 2014 and December 2019 in the Department of Virology, Paul Brousse hospital were considered for inclusion. Patients for whom clinical data were missing were excluded from analysis.
Data collected from patient files included: sex, ethnicity, age at onset of symptoms, age at diagnosis of RVAU (date of positive AH analysis and considered as the date of inclusion) and rubella vaccination history. In France, 2 rubella immunizations are recommended for all children from 12 months, since 1986. Rubella vaccination history was available for all patients born after 1986. Patients born before 1986 or born in a country lacking a national rubella vaccination programme, were considered unimmunized unless proven otherwise. The diagnostic delay defined time between first ophthalmic symptoms and a positive AH rubella test Virological analysis RV RT-PCR and anti-RV IgG testing was performed on the AH of all patients. One eye only was tested in cases of bilateral disease.

Antibody detection and derived goldmann-witmer coefficient calculation
Immune status of the patients was assessed for RV, CMV, HSV-1 and VZV, for which seroprevalences in France are 90%, 50%, 70% and 95% respectively. Patients seronegative for at least two viruses (excluding RV) were not considered for inclusion in the study. Measuring specific IgG in AH was only performed if the patient was seropositive. IgG antibody titers were determined using Enzygnost specific IgG enzyme-linked immunosorbent assay kits (Siemens, Marburg, Germany) for RV, HSV-1, VZV and CMV, according to manufacturer's instructions. AH and blood serum were paired for analysis in the same analytical run. The volume used for measuring IgG in AH was the same for each virus (20 μL). Optical density was evaluated as arbitrary concentration units (AU), by reference to a standard curve. RV-IgG titers from AH and serum were compared with other virus-specific IgG titers (VZV and/or CMV and/or HSV-1), to determine the Goldmann-Witmer coefficient 21,22 (GWC'): GWC' = (AH RV-IgG × serum other virus-IgG) / (serum RV-IgG × AH other virus-IgG). The test was considered positive only if two or three ratios (when three ratios were available) were greater than 3.

Ophthalmological data and clinical presentation
Ophthalmological data included best-corrected visual acuity (BCVA) and slit-lamp findings. Patients were categorized into two groups based on the clinical presentation. Group 1 included typical FUS patients, 7 with all criteria required for diagnosis (including: i) chronic low-grade anterior segment inflammation, ii) diffuse stellate KPs, iii) iris heterochromia, iv) lack of posterior synechia and v) cataract (Figure 1)). Group 2 included patients not fulfilling typical FUS criteria, or with additional clinical presentations such as: predominant posterior segment involvement, or retinal vasculitis. Complications experienced were also collected at virological diagnosis and during the follow-up and included: i) ocular hypertension (OHT) (defined by intraocular pressure > 21mmHg without glaucomatous optic neuropathy), ii) glaucoma (i.e. typical glaucomatous optic nerve head and visual field alterations) and iii) surgical procedures.
The study adhered to the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of the French Society of Ophthalmology (IRB 00008855 Société Française d'Ophtalmologie IRB#1). Informed consent was obtained for all the participants (or from a parent for participants under 18).

Demographics and rubella vaccination status
Between January 2014 and December 2019, 38 patients tested positive for rubella by either RV RT-PCR and/or GWC'. The complete medical files of 13 patients were retrieved (6 women, 7 men), aged 49.6 ± 22.2 (7-86) years at inclusion ( Table 1). The delay in diagnosis was 13 ± 12.6 years (0-42). Most patients (9/13) were not immunized against rubella, while the remaining (4/13) were immunized after the recommended age (between 6 and 28 years) ( Table 1). Most patients were Caucasian and born in France (10/13), while 2 originated from Asia, where they were born (case #10 from India and case #13 from Vietnam) and one was born in Africa (from African descent) (case #11).
Virological data. RV RT-PCR as well as HSV-1, VZV and CMV PCR were performed in 11 cases and RV-RNA was detected in the AH of one patient (case #12). PCR results for other viruses were all negative.
All patients were seropositive for RV, HSV-1 and VZV. Additionally, ten patients were also seropositive for CMV. Therefore, GWC' could be calculated for all patients using two other viruses, and for ten patients, three other viruses. AH anti-RV-IgG was demonstrated in all cases, with a median GWC' of 9.5 (range 3.2-100) ( Table 2). GWC' was negative for all other viruses in all patients.
Ophthalmological data and clinical presentation. Three patients were classified into Group 1, fulfilling all FUS criteria, and 10 were classified in Group 2. Patient's initial complaints were vision loss (n = 11, 85%) and floaters (n = 3, 23%). At presentation, mean BCVA of the affected eyes was 0.6. Ten patients had cataract in the affected eyes (3 in group 1 and 7 in group 2) and 7 patients had either OHT or glaucoma (1 in group 1 and 6 in group 2). Low grade inflammation of the anterior segment was present diffuse KPs were present in 8 patients (3 in group 1 and 5 in group 2 for both). Iris atrophy, heterochromia, or nodules were noted in 6, 3 and 3 patients, respectively (Table 3). Vitreous involvement was present in 10 eyes (9 patients) and retinal vasculitis in 5 eyes (4 patients). One patient (case #10) had bilateral disease.
At follow-up of 96 ± 103 months (2-306), BCVA was similar (0.6) (p = 0.61). Cataract was detected in one additional eye; surgery was required in 10 eyes (9 patients). Glaucoma occurred in 3 additional eyes and surgical procedures were required in 6 eyes of 6 patients. The mean delay between onset of symptoms and virological diagnosis was 157.6 ± 153 months.
Among the 8 eyes (8 patents) with OHT or glaucoma during follow-up, 7/8 initially had a positive history for Possner-Schlossman phenotype, with OHT peaks concomitant with inflammation flare-ups. In these cases, IOP elevation was not caused by steroid eye drops instillation. Six out of this seven cases ended with chronic IOP elevation with associated glaucomatous optic neuropathy. The last case (case #13), who had no obvious clinical signs of FUS (aqueous sampling was performed during glaucoma surgery as part of an exhaustive check-up for severe unilateral glaucoma), had elevated IOP at diagnosis, and no history of IOP peaks.
Among the 5 patients who underwent fluorescein angiography, sectorial peripheral non-occlusive capillaropathy and vasculitis was observed in 4 cases (cases #4, 5, 7, 10) (Figure 2A, 2D). In one patient, a concomitant focal staining on a large vein was noted (Case #4, Figure 2A). Disc hyperfluorescence was observed in 3/5 cases ( Figure 2C). OCT did not show macular edema ( Figure 2B). Indocyanin green angiography was performed in 4 patients  and did not reveal significant abnormalities. We did not observe retinal scars in affected eyes included in our series.

Discussion
The clinical spectrum of RVAU RVAU appears to be a pleiomorphic entity. As in previous studies, 7 our cohort had a minority of patients fulfilling FUS criteria (3/13 patients). In group 2, some patients had completely atypical pictures where retinal vasculitis and posterior segment features were predominant, while other had "incomplete" FUS that could be explained by individual particularities or the moment in the course of the disease where clinical examination was performed. Diagnosis of heterochromia depends on the initial amount of iris pigment and the degree of atrophy. Therefore, it is often absent in dark-colored irides and during the first years of diseases, 12,14 and present in approximately 1/4 of RVAU patients. 23 The common clinical denominator of RVAU in the previous series and in our study, however, was the absence of posterior synechiae and macular edema. 7 These clinical findings are not specific for RVAU, and overlap with CMV anterior uveitis 24 and some HSV-1-and VZV-related uveitis. [9][10][11] In most cases of our series, elevated IOP and glaucoma were part of the spectrum of rubella uveitis, rather that steroid induced. In the series by Groen-Hakan et al. 50% of patients with RVAU had OHT or glaucoma while half of them did not receive steroids. 7 We acknowledge that fluorescein angiography was not performed in all patients, and this could lead to underdiagnosis of posterior segment subclinical features. In RVAU, posterior segment findings could be either secondary to vitreous inflammation or primitive. In a case series, Bouchenaki and Herbort reported disc hyperfluorescence in 98% of FUS patients (43/44) who underwent fluorescein angiography, but it was not clear whether these patients had cataract surgery before FA was performed. 25 These authors also reported peripheral retinal vascular leaking in 14% of patients. 25 Vasculitis involving large vessels and retinitis have been reported in isolated case reports of RVAU. 26,27 Retinal scars, most often of small size (< 1 papillary diameter), that may be found both in central and peripheral retina, have also been reported in RVAU. 7 The essential role of virological Ah analysis Altogether, there is weakness in clinical diagnosis alone, while virological assays contribute significantly to correctly managing patients with uveitis. Accurate diagnosis can improve management, by avoiding unnecessary corticosteroids or antiviral treatments if Herpesviridae is suspected. In contrast, some FUS may be caused by CMV or HSV-1 and thus require specific antiviral treatments. 10,28 Furthermore, some patients may be co-infected with RV and HSV-1. 11 It is difficult to diagnose RVAU with RV-RNA detection alone and combining immunological and molecular assays may improve the diagnosis of infectious uveitis. [29][30][31] RVAU diagnosis relies on RV-IgG-positive AH and/or presence of RV-RNA. 7 While all our patients were GWC' positive, only one was positive by RT-PCR. Thus, RT-PCR diagnosis alone would have missed 90% of cases. Our results are consistent with previous studies, where 10 to 20% of suspected cases were positive by RT-PCR, whereas 97 to 100% of AH samples were RV-IgG-positive. 6,7,11 Low viral loads and the small volumes of AH fluid contribute to the low sensitivity of RV RT-PCR, which also contain inhibitory compounds. 32 In order to improve detection, viral culture could have been attempted in order to amplify viral load before PCR.  However, this technique would have required the entire AH sample. Besides, the delay and uncertainty of results with this process (due to difficulties in standardization of these techniques) is not suitable for clinical management.
In the future, deep sequencing has the potential to detect any foreign genomic sequence in small clinical samples (20-50μL) and could greatly improve the diagnosis of infectious uveitis. 33

Pathophysiology of ocular infections with Rv
Detection of RV-RNA from AH in some patients and anti-RV IgG in all, demonstrates that RVAU can be caused by intraocular infection and possibly persistent viral replication associated with an immune-mediated inflammation triggered by systemic RV infection. Thus, the seric anti-RV IgG indicates a necessity for RV viremia, whether postnatal or in utero. 34 Winchester et al. reported high titre of anti-RV IgG and RV RT-PCR-positive AH in a CRS patient with bilateral FUS in his third decade. 35 None of patients in our series had history or signs of CRS but as we did not conduct in depth investigations, CRS cannot be excluded.
Also consistent with results of previous studies is the absence of Rubella vaccination in RVAU patients. 7,36 Meanwhile, recent epidemiological studies demonstrate a decline in FUS incidence since the introduction of mass rubella vaccination and this trend was not observed in an unimmunized control group. 37 Why Rv still occurs in immunized patients?
In France, vaccination against rubella is recommended since 1986, and the double-dose regime given between 12 and 18 months of age is mandatory since 2018. 38 Stunf et al. reported that 4/11 anti-RV-positive AH RVAU patients were immunized. 39 However, RV infection occurring before vaccination cannot be ruled out. Indeed, in our study, 4/13 RVAU patients reported incomplete vaccination. Three were immunized late during childhood, between ages 6 and 7, leaving them exposed to rubella prior to vaccination, in the early 80's, a period where RV was highly prevalent in France. Our third 'partially immunized' case was born in Vietnam when rubella vaccination were unavailable, and was subsequently immunized at the age of 28, in France, after the birth of a child upon negative RV-IgG in the early 80's. However, a serologic control performed after vaccination was still negative and there were no ocular complaints. In contrast, the patient had seroconverted to RV-IgG-performed with ELISA assay in our study was positive by the time of our study. It is still possible that her initial test was were false negative due to the poor sensitivity of serologic assays (inhibition hemagglutination assay) performed at this time. 40 The rubella immunogen is a live-attenuated virus, and so it cannot be completely ruled out that vaccination could be involved in some rare cases of RVAU, as it is the case for live attenuated VZV vaccination. 41 However, to date this has never been demonstrated. Sequencing and phylogenetic analysis could help to confirming or excluding this possibility. However, RV-RNA sequencing in the AH is extremely challenging, 42 and it failed in our only RV-PCR-positive patient, due to insufficient sample volume.

Treatment strategy in RVAU
To our knowledge, there are no specific therapeutic guidelines for RVAU. However, therapeutic guidelines for FUS may be applicable in most cases. As recently reminded by Sun & Ji, anti-inflammatory treatment is not indicated as these patients are usually resistant to corticosteroid therapy. 8 An occasional and short course of corticosteroids may be indicated if a symptomatic exacerbation occurs, whereas long-term steroids are not indicated in FUS.
In summary, RVAU is a pleiomorphic entity for which clinical diagnosis can be easily undermined. AH analysis combining RV RT-PCR and detection of intraocular specific anti-RV IgG are essential in the management of patients with compatible clinical findings. Detection of other pathogens should also be performed. Overall, it seems that improper vaccination against RV is implicated in RVAU.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.