J Braga1 , R Rothwell2, M Oliveira 1, D Rodrigues 1, S Fonseca1, R Varandas1 and L Ribeiro1

Background: Choroidopathy is a rare manifestation of systemic lupus erythematosus (SLE). This entity is associated with active phases of severe SLE and it is frequently accompanied by acute kidney failure, central nervous system involvement and coagulopathy. Purpose: To evaluate the choroid thickness of patients with lupus nephritis (LN) without choroidopathy, and to compare this with that of age-matched SLE patients without LN and healthy control subjects. Study design: Cross-sectional case control study. Material and methods: Fifteen women with LN in remission phase (study group), 15 women with SLE in remission without LN (SLE control group), and 15 healthy women (healthy control group), without ocular diseases or significant refractive error, were recruited. Full ophthalmological examination and a macular optical coherence tomography in enhanced depth imaging mode were per- formed. The choroid thickness was measured at nine macular points and six lines of mean choroidal thickness were determined. A comparative analysis between the three groups was performed using the one-way ANOVA test and the paired t-test. The choroid thickness of patients under corticotherapy was also compared to that of patients without corticotherapy.

Additionally, the correlation between choroid thickness and disease duration was evaluated using the Pearson analysis. Results:The mean macular choroidal thickness was 295.73 干67.62 μm in the study group,233.34 干41.01 mm in the SLE control group, and 240.98 干37.93 μm in the control group (p =0.00006 and p =0.0003,respectively). Additionally, the choroid thickness was significantly thicker than in the SLE and healthy control groups at the foveal (p =0.004 and p <0.000), nasal (p <0.000 and p =0.001), superior (p =0.002 and p <0.000) and inferior (p <0.000 and p =0.001) mean lines. The choroidal thickness in this group was not associated with the duration of the disease. The subgroup of patients with LN under corticotherapy did not reveal a significantly different choroidal thickness. Conclusion: This study suggests a relationship between LN and choroidal changes, which may represent an increased risk for choroidopathy in these patients. Choroid thickening was not related with the duration of the disease. This thickening may be correlated with histopathological changes similar to those occurring in kidney glomeruli.Lupus (2019) 0, 1–8.

Keywords: Choroidopathy; lupus nephritis; choroid thickness; systemic lupus erythematosus; EDI-OCT

Introduction
Systemic lupus erythematosus (SLE) is a chronic autoimmune disease characterized by periods of greater disease activity alternating with periods of remission that can virtually affect any organ system. The visual system may also be affected by this disease and any structure of the eye may be involved. The ocular manifestations of SLE vary in severity and visual impact, as these range from keratoconjunctivitis sicca to retinitis and optic neuropathy. Ocular manifestations may occur in up to one-third of SLE patients and seem to be more frequent in patients with active disease.1,2 One particular entity, the choroidopathy, which is con- sidered an indicator of severe systemic disease, is frequently associated with renal and central nervous system coagulopathy.2,3Clinically, choroidopathy manifests with serous detachment of the retina, pigment epithelium or both, causing vision loss due to the involvement of the macular area of the retina.4 Another poten- tially devastating choroidopathy manifestation is uveal effusion with dislocation of the iris, ciliary body and lens anteriorly in the anterior chamber, causing acute angle closure glaucoma.5,6 The pathogenesis of choroidopathy remains uncertain; however, it is likely to involve three Medial patellofemoral ligament (MPFL) different mech- anisms: (a) immune complex deposition in the choroid and choriocapillaris; (b) systemic arterial hypertension, frequently induced by nephropathy; and (c) thrombosis and microangiopathy leading to capillary vaso-occlusions.1–3

Immunofluorescence studies have shown depos- ition of immunoglobulins and immune aggregates in the walls of choroidal capillaries, choroidal blood vessels and basement membrane of the chor- oidal epithelium. Another study demonstrated the presence of autoantibodies directed against retinal pigment epithelium (RPE).7,8 Additionally, the effi- ciency of plasmapheresis in diminishing the serous detachment in choroidopathy is in favor of an immune complex–mediated mechanism for choroidopathy.3Hypertension may play a role in lupus choroido- pathy as it is frequently present in patients with nephropathy and promotes choroidal vascular occlusion, contributing to choroidal ischemia and neurosensory retinal detachments. These manifest- ations must be carefully differentiated from hyper- tensive retinopathy and choroidopathy. In fact, choroidopathy may improve with control of disease activity, even in the presence of sustained arterial hypertension.3Thrombosis and microangiopathy also contrib- ute to the manifestations of choroidopathy. There may be occlusion caused by microthrombus forma- tion in the choroidal vasculature, as seen in disse- minated intravascular coagulation.3

Apparently, lupus choroidopathy develops sim- ultaneously with severe systemic manifestations of SLE.2,3,6 However, other authors report the devel- opment of lupus choroidopathy months before a systemic flare. Therefore, importantly, it could be a signal for the presence of subclinical, reversible nephropathy and neuropathy.The presence of acute choroidopathy may represent an increased risk of morbidity and mortality and maybe an indication for additional immunosuppression. 1–4 Hence, it is rational to study the choroid and chor- iocapillaris of patients with SLE, especially patients with lupus nephritis (LN). The identification of patients with increased risk of choroidopathy would, theoretically, also identify those at greater risk of severe systemic manifestations of SLE.The aim of this study was to evaluate the choroidal thickness in patients with SLE and LN, without choroidopathy, analyzed by spectral domain optical coherence tomography(OCT), and compare with two age-matched control groups(healthy subjects and patients with SLE without LN).

This cross-sectional case control study was per- formed at the Centro Hospitalar Vila Nova de Gaia/Espinho between January and October 2018, after approval by the hospital’s Ethics Committee. Each patient was fully informed of the purpose and procedures of this study and provided oral informed consent in agreement with the Declaration of Helsinki.The patients were divided into three groups: women with LN in remission phase (study group), women with SLE in remission phase without LN (SLE control group), and age-matched healthy woman (healthy control group). The patients in the study group were referred from the nephrology department. The patients in the SLE control group were selected, according to age, from the institu- tional database of patients in screening for hydro- xychloroquine maculopathy, without evidence of maculopathy. All patients were diagnosed with SLE according to the American College of Rheumatology criteria. In the study group, LN was confirmed with kidney biopsy. The women in the healthy control group were healthy volunteers recruited at the ophthalmology department.

Exclusion criteria in the study and SLE control groups included patients with history of significant and/or active ocular diseases, evidence of hydroxy- chloroquine maculopathy,a spheric equivalent >3.5 D, dense media opacities and pregnancy. In the control group, exclusion criteria included previous or active systemic disease, use of medica- tion, ophthalmologic disease and refractive error with spherical equivalent >3.5 D.A full ophthalmological examination was per- formed including Snellen’s best corrected visual acuity, refraction, slit-lamp biomicroscopy, tono- metry and fundus examination. Data regarding complete medical history and medication use were collected.Choroidal imaging was obtained using the enhanced depth imaging(EDI) mode of the Spectralis spectral domain OCT device

Figure 1 Illustration of choroid thickness at the foveal hori- zontal line. (a) 48-year-old woman from the study group. (b) 48-year-old woman from the healthy control group. (c) 48-year-old woman from the systemic YK-4-279 order lupus erythematosus control group.(Heidelberg Engineering, Heidelberg, Germany). A high-resolution scan of 19 horizontal lines, spaced 240 μm apart on a 200 收200 mm2 area centered on the fovea, was used. To improve image quality, the eye tracking software of the device was used, and 100 B scans were averaged together. All the exam- inations were performed by one single experienced technician.For the definition of the choroidal limits, the boundaries of the retina, automatically plotted by the software, were manually readjusted. The line of the internal limiting membrane was moved to the RPE and the line on the RPE was replaced on the posterior limit of the choroid. Figure 1 illustrates choroidal demarcation and thickness measurement at the foveal horizontal line of one subject with LN and two age-matched controls of the two control groups. The boundaries were then confirmed by one co-author, blinded for the patient’s group.The choroidal thickness was measured at nine macular points defined by the intersection points of three horizontal lines (superior line, foveal line and inferior line of the macular scan) with three vertical lines (one on the temporal limit of the scan, one on the fovea and one on the nasal limit of the scan). The mean macularchoroidal thickness was calculated as the mean of the enzyme-based biosensor nine values. Additionally, choroidal thickness was calculated at the four borders of the macular scan (temporal, nasal, superior and inferior) and at the fovea by the mean of the three corresponding macular points.

Figure 2 Schematic representation of the choroidal measure- ments. F:foveal;I:inferior;IN:inferonasal;IT: inferotem- poral; N:nasal; S: superior;SN: superonasal; ST: superotemporal; T: temporal.Statistical analysis was performed using IBM SPSS Statistics software, version 25.0. The Kolmogorov– Smirnov test was used to verify the normal distri- bution of choroidal thickness. One-way ANOVA data analysis was employed to test for differences in variables between the two groups. A p-value <0.05 was considered to be significant. The paired t-test was used to evaluate regional differences of the choroidal thickness between the study and con- trol groups. The Pearson analysis was also applied to determine the presence of a correlation between choroidal thickness and duration of disease in the study group.

Results
This study included 15 women with LN (study group), 15 age-matched women with SLE without LN (SLE control group)and 15 age-matched healthy women (healthy control group). The mean age in the study group was 45.13 干9.0 years (mean 干standard deviation (SD)), 45.2 干9.5 years in the SLE control group, and 45.6 干9.3 years in the control group.There was no statistically significant difference in the mean age of the three groups (p =0.98).In the study group, the mean duration of SLE was 14.0 千6.0 years (mean 千SD), and the mean duration of LN was 13.0 千6.0 years (mean 千SD). All the LN patients had associated articular involvement; five had dermatological manifest- ations; one had chronic hematological alterations; and another had cardiac and pulmonary involve- ment. Every patient of the study group underwent kidney biopsy and LN was diagnosed according to the International Society of Nephrology/Renal Pathology Society 2003 classification of LN. LN class IV was diagnosed in 12 patients, class III in

Figure 3 Boxplot of macular choroidal thickness in each group.
one patient and class V in another. One patient’s classification of LN was unavailable on the medical records.
In the SLE control group, the mean duration of the disease was 8.9 千6.3 years (mean 千SD). In this group, articular manifestations were present in 14 patients, cutaneous in 6, hematological alterations in 5, cardiovascular involvement in 2, pulmonary involvement in one and CNS involvement in another patient.The one-way ANOVA test was used to compare the mean choroidal thickness values between the three groups. Figure 3 illustrates the mean macular choroidal thickness of each group.The macular choroidal thickness of the SLE control group was not significantly different from that of the healthy control group (p =0.457), neither were any of the other mean thickness measurements. The macular choroidal thickness of the study group was thicker than that of the healthy control group (p =0.0003) and the SLE control group(p =0.00006). Additionally, all the mean values were significantly thicker in the study group than in the SLE control group. The foveal, superior, inferior and nasal choroidal thicknesses were also significantly thicker in the study group. The temporal choroidal thick- ness in the study group did not show a statistically significant difference compared to the healthy con- trol group (p =0.08). Table 1 summarizes the chor- oidal thicknesses in the three groups.

The comparison between choroidal thickness at temporal and nasal lines revealed a significant dif- ference in the eyes of both the study and the healthy control groups (p =0.003 and p =0.0000000002, respectively). The comparison between choroidal thickness at the superior and inferior lines also revealed a significant difference in the control group; however, this difference did not occur in the eyes of the study group(p =0.003 and p =0.483, respectively) (Table 2).Of the 15 LN patients,12 were treated with hydroxychloroquine, four of whom in association with mycophenolatemofetil, and one in association with azathioprine. Nine LN patients were under low-dose corticotherapy with prednisolone (5 mg/day or less). The comparison of the choroidalthickness (Pearson correlation 0.137, p =0.47). Additionally, the duration of LN was not asso- ciated with the macularchoroidal thickness (Pearson correlation 0.154, p =0.415).

Discussion
Lupus choroidopathy is commonly reported in cases of severe active disease and is frequently accompanied by CNS involvement and nephropa- thy.1–4 In one review of cases of choroidopathy, nephropathy was present in 64% of patients, and CNS involvement occurred in 36%.3 Hypertension occurred in 54% of cases. Systemic vasculitis and hematologic diseases, such as disseminated intra- vascular coagulation and thrombotic thrombocyto- penic purpura, were also frequently reported in patients with choroidopathy.3
The connection between LN and choroidopathy has been studied by a series of authors.10–12 It is thought that the Bruch-RPE choriocapillaris membrane may morphologically resemble the glomeru- lus. Choriocapillaris may be seen as the capillary tuft of the glomeruli; Bruch’s membrane as the basement membrane; and the RPE as the glomeru- lar epithelium.At the molecular level, Bruch’s membrane and glomerular basement membrane are both selectively and directionally permeable, have a negative electron charge and a high content of heparin sulphate.Additionally, choriocapil- laris and the capillary tuft of the glomeruli are two of the most perfused capillary networks of the human body and may be predisposed to immune complex deposition.

Other studies have demonstrated the association between glomeruli diseases (membranoproliferative glomerulonephritis) and changes in the retina and choroid, including the occurrence of serous detach- ments of the retina.13–15 This association was also corroborated by the evidence of subtle lesions in the indocyanine green angiography of patients with lupus nephropathy that were absent in patients with SLE without nephropathy.10As an indirect measure of activity and (dys)func- tion, the choroidal thickness, which can be deter- mined by the EDI-OCT, is a simple measurement to obtain in an ophthalmological evaluation. The OCT is a non-invasive imaging modality that allows for an in vivo three-dimensional view of the chorioretinal architecture. The EDI mode pro- vides a better visualization of the morphology of the choroidal layers, and has been used in many recent studies concerning chorioretinal diseases.16–19 In this study, the choroidal thickness measurements were made by manual segmentation, by redefining the EDI Spectralis OCT built-in auto- mated retinal segmentation to the choroidal limits. This method has been previously described and applied in other studies and has proven to be highly reproducible, with a small range of variability.

Recent studies have shown variation of the chor- oidal thickness in several systemic and ocular dis- eases. The choroid has been found to be thicker in the eyes of patients with microangiopathy asso- ciated with HIV,22 Cushing syndrome,23 central serous chorioretinopathy, 18 and idiopathic subfo- veal choroidal neovascularization.19In contrast, the choroid seems to be thinner in the eyes of patient swith fibromyalgia,24 apnea-hypopnea syndrome,25 acute hypovolemia.obstructive migraine26 sleep andThe choroidal thickness also varies with age, gender and psychological factors in healthy individ- uals.This study included only female patients that were matched by age with healthy controls. Several authors have also reported diurnal vari- ations of the choroid measurements;29–31 however, the results are conflicting. To minimize possible interferences, all the EDI-OCTs were acquired between 10 am and 3 pm.

To date, few studies have evaluated the choroidal changes associated with SLE and the results are contradictory. Altinkaynak et al. have reported that the choroid is thinner in patients with SLE than in healthy age- and gender-matched controls, and that this thinning correlates with the duration of the disease.Ferreira et al. demonstrated that patients with SLE had thicker choroids than healthy controls.16 However, in both of these studies, the manifestations of SLE were not described and the prevalence of LN in the study populations was not reported. These different findings maybe attributed to the heterogeneity of the disease manifestations and variance of severity. Given the rareness of the choroidopathy manifestations, their association with severe disease, and the apparent physiopatho- logic link with nephropathy, the study of choroidal changes in this subpopulation of SLE patients may give us a better understanding of this entity.
We report significantly thicker macular choroids in the eyes of patients with LN compared to eyes of patients with SLE without LN and healthy con- trols. When compared individually, the five mean lines of thickness were significantly thicker than in the control group, except the temporal line. We observe that the healthy control choroids were thicker at the temporal and superior portions, and thinner at the nasal portion, which is consistent with the normal choroid thickness ously described in other studies.32 seems that the normally thicker choroid at the tem- poral portion suffers less thickening in the LN patients. We found a more significant thickening at the inferior and nasal portions of the macula that, interestingly, were also the thinner areas of the control choroids.

Corticosteroids can affect the choroid circulation by altering the production of nitric oxide, free rad- icals and prostaglandins. These effects may lead to hyperpermeability of the choriocapillaris, venous dilation, hypoperfusion and increased propensity for blood coagulation.Systemic corticotherapy may play a role in the etiology of central serous chorioretinopathy, that is associated with thicker choroids than in control subjects.34 Nevertheless, the effect of systemic corticotherapy in choroid thickness is controversial.35,36 Nine LN patients were medicated with low-dose maintenance sys- temic corticotherapy. However, there was no significant difference in the choroid thickness of these patients. In this study, low-dose systemic corticosteroids did not influence choroid thickness of the study group.The durations of SLE and LN were not correlated with choroid thickness changes in this popu- lation. To our knowledge only one study has evaluated the correlation between SLE duration and choroid thickness. Altinkaynak et al. reported that SLE duration was associated with choroid thinning; 17 however, the presence of LN was not taken into account, nor was the use of corticother- apy. In our study, we did not find a statistically significant difference in choroid thickness of SLE patients without LN when compared to healthy controls. However,in this group, the patients under corticotherapy showed thicker choroids than those without corticotherapy.Interestingly, we also observed that this subgroup of patients had a significantly shorter duration of disease. Hence, it is not possible to determine if this differ- ence is due to a possible thickening of the choroid associated with the corticotherapy, or duetoathin- ning of the choroid with the longer duration of the disease. Severity and activity of the disease and class of LN may have a stronger correlation with choroidal changes than SLE duration. In the study group, all except for two of the patients had LN class IV; therefore, the association of the class of LN with the choroidal thickness could not be evaluated.

All LN patients were in remission phase of dis- ease.Our results suggest that even during remission, there are structural changes of the chor- oid. These changes probably result from multiple co-existing factors such as immune complex deposition, hypertension, and possibly sustained hypervolemia, and microangiopathy. The augmented choroidal thickness might represent a per- manent and latent state of inflammation with vasodilatation and increased permeability. In active phases of SLE, these alterations might cause chor- oidal dysfunction and culminate in serous retinal detachments and uveal effusion.This study does not allow us to establish a chronological relationship between the changes in the glomeruli histology and choroid thickness. The choroidal changes may precede, accompany or follow the changes in the glomeruli. Additional studies may clarify this pathological association. Nonetheless, the EDI-OCT may be a useful tool in identifying SLE patients with higher risk for choroidopathy and nephropathy. In the future, EDI-OCT and choroid measurements may assist in the management of SLE patients with suspected renal involvement and influence the decision to per- form a kidney biopsy.

Conclusion
This study suggests a relationship between LN and choroidal changes, which may represent an increased risk for choroidopathy in these patients.These choroidal changes do not seem to be associated with the use of low-dose cortico- steroids.The choroid thickening was not related to the duration of the disease. This thickening may be correlated with histopathological changes similar to those occurring in the kidney glomeruli.