Association of Subretinal Fluid Duration and Baseline Chorioretinal Structure With Optical Coherence Tomography in Central Serous Chorioretinopathy

Abstract

Purpose: This study investigated the association between subretinal fluid (SRF) duration and baseline chorioretinal structure in patients with naïve (first-episode and untreated) central serous chorioretinopathy (CSC).

Methods: We divided 59 patients into non-prolonged (<3 months) and prolonged (>3 months) groups based on SRF fluid persistence from the initial visit. The follow-up period varied as the shorter duration was chosen between SRF disappearance time and 3 months from the initial visit. We measured the central retinal thickness (CRT), central choroidal thickness (CCT), SRF height (SRFH), and outer nuclear layer thickness (ONL) using spectral-domain optical coherence tomography (SD-OCT) at the initial visit and recorded SRF duration. We compared these parameters between the groups, conducted multivariate analysis for SRF duration of more than 3 months, and investigated the correlation among CCT and CRT, SRFH, or ONL, and among SRF duration and CRT, CCT, SRFH, or ONL.

Results: CCT was significantly thicker in the prolonged than in the non-prolonged group at the initial visit (P = 0.044) and significantly correlated with CRT and SRFH (P = 0.007, r = 0.35 and P = 0.002, r = 0.39). SRF duration significantly correlated with CRT and SRFH (P = 0.009, r = 0.40 and P = 0.003, r = 0.41). The optimal model for SRF duration more than 3 months included age (P = 0.054) and CCT (P = 0.008).

Conclusions: Thicker baseline CCT can lead to a longer SRF duration in naïve CSC.

Translational Relevance: This association aids in early detection of cases at a higher risk of prolonged SRF.

Introduction

Central serous chorioretinopathy (CSC) usually occurs in middle-aged men and accumulates subretinal fluid (SRF) in the macula with various subjective symptoms, such as distorted vision and central scotoma.^1,2 The pathogenesis of this disease is believed to involve increased permeability and congestion of the choroidal vessels.^3–6 

The natural course of CSC is often self-limiting, and spontaneous resolution and complete fluid re-absorption often occur.⁷ However, chronic SRF accumulation commonly leads to permanent structural damage to the neuroretina and retinal pigment epithelium (RPE), with irreversible long-term vision loss.^8–11 This is because the outer retina is separated from the RPE by the SRF, and the outer retina is damaged owing to insufficient oxygen supply from the RPE.¹² Therefore, it is important to prevent photoreceptor and RPE damage by early detection of SRF, follow-up cases with a high likelihood of persistent SRF, and treat such cases when SRF does not resolve spontaneously. 

To date, there is no consensus regarding the duration threshold or therapeutic protocols for non-resolving CSC cases. Spaide et al. reported that central choroidal thickness (CCT) is increased in eyes with CSC compared to normal eyes, using enhanced depth imaging-optical coherence tomography (EDI-OCT).^13–15 Few studies have examined the difference between the initial OCT findings in cases with spontaneous resolution of SRF and the initial OCT findings in cases with prolonged SRF when naïve (first-episode and untreated) CSC were followed up without treatment.^16–18 They did not investigate choroidal structure. Another study examined the relationship between baseline choroidal thickness and the need for treatment in patients with untreated CSC.¹⁹ However, this study included not only first-episode cases but also recurrent cases. 

To the best of our knowledge, no study has examined the relationship between choroidal thickness in the acute phase and SRF duration in more than 50 patients with naïve CSC. The main objective of our study was to investigate the relationship between baseline OCT findings in the acute phase, including choroidal thickness and SRF duration, in patients with naïve CSC. This study has the potential to predict cases with a higher risk of prolonged SRF duration based on initial OCT findings and to treat such cases at the appropriate time. 

Materials and Methods

This study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Shinseikai Toyama Hospital (approval number: 230313-3). Our study was retrospective, and we used an opt-out consent process; informed consent was waived by the Institutional Review Board of Shinseikai Toyama Hospital. 

The patient database at Shinseikai Toyama Hospital was searched and patient records from December 2010 to August 2022 were reviewed. 

The clinical diagnosis of CSC was based on symptoms, reduced visual acuity with or without vision impairment, metamorphopsia, micropsia, dyschromatopsia, or central scotoma, and the presentation of SRF detected by both fundus examination and EDI-OCT (Spectralis OCT, Heidelberg Engineering, Heidelberg, Germany or Avanti, Optovue Inc., Fremont, CA, USA). All subjects underwent a thorough ocular examination of both eyes at the initial visit, including best-corrected visual acuity (BCVA) using a Landolt decimal acuity chart, which was converted into the logarithm of the minimal angle of resolution; intraocular pressure; slit-lamp biomicroscopy; OCT examination using a horizontal scan including the fovea; and recording of data on SRF duration. The onset was inferred from the patient's complaints in all cases. Moreover, the presence of chronic SRF accumulation leading to RPE damage was supplementally checked in some cases with fundus autofluorescence (FAF) using a confocal scanning laser ophthalmoscope (cSLO, Heidelberg Retina Angiograph 2 system; Heidelberg Engineering, Heidelberg, Germany). Some patients were administered oral kallidinogenase 150 mg/day for up to 3 months from the initial visit at the discretion of the examining doctor. 

Exclusion criteria were as follows: initial presentation later than 1 month after symptom onset; FAF indicating chronic CSC; either clinical signs or a history of any other intraocular disease in either eye; topical treatment such as photocoagulation, photodynamic therapy (PDT), or intravitreous injection of anti-VEGF agents before 3 months had passed since the initial visit; a former episode of CSC in either eye; steroid use; or inability to define symptom onset. The follow-up scheme for this observational study included monthly clinical evaluations for >3 months or until the SRF was resolved. 

The central retinal thickness (CRT) was measured by an experienced investigator (author C.O.) who was blinded to the information on the patients' demographic data using EDI-OCT scans as the axial distance from the inner aspect of the internal limited membrane at the fovea to the inner aspect of the RPE. A retinal specialist (author T.S.) measured the CRT and acted as an arbiter when CO had difficulty or a lack of confidence in the measurement. The CCT was measured in a similar manner as the axial distance from the outer aspect of the RPE to the outer choroid/scleral interface. Subretinal fluid height (SRFH) was measured similarly as the axial distance between the foveal ellipsoid zone and the inner aspect of the RPE. The outer nuclear layer (ONL) thickness was measured similarly at the fovea. 

We divided all 59 eyes of the 59 patients in our study into 2 groups by defining chronicity as the persistence of SRF for more than 3 months from the initial visit, according to previous reports.^20–22 The non-prolonged group consisted of patients whose SRF resolution time was ≤3 months and the prolonged group consisted of patients whose SRF resolution time was >3 months from the initial visit. Age, time from onset to initial visit, oral kallidinogenase use, spherical equivalent (SE), BCVA, CRT, CCT, SRFH, and ONL at the initial visit were compared between the non-prolonged and prolonged groups. Moreover, the BCVA at 3 months was compared between the non-prolonged and prolonged groups, and was compared with that at the initial visit in each group. Multivariate analysis was also conducted in order to determine the factors including age, sex, SE, CRT, CCT, SRFH, and ONL associated with SRF resolution time longer than 3 months. In addition, the correlation between CCT and CRT, SRFH, or ONL in all cases was calculated, and the correlation between SRF duration and CRT, CCT, SRFH, or ONL in 50 out of the 59 eyes was also calculated. This is because the remaining 9 eyes received topical treatments, such as photocoagulation, PDT, or intravitreous injection of anti-VEGF beyond 3 months after the initial visit. These treatments were considered to affect SRF duration beyond 3 months after the initial visit. 

Statistical Analysis

Continuous variables are presented as the mean ± standard deviation. 

Fisher's exact test was used to compare sex and oral kallidinogenase use. The Wilcoxon rank-sum test was used to compare the sex, age, time from onset to initial visit, SE, BCVA, CRT, CCT, SRFH, and ONL between the non-prolonged and prolonged groups. 

Wilcoxon signed-rank test was used to compare BCVA at 3 months with that at the initial visit. 

To determine the best explanatory variables for SRF duration longer than 3 months, multivariate logistic regression analysis with model selection using Akaike information criterion index (AIC) was also performed. Among the parameters of age, SE, CRT, CCT, SRFH, and ONL, the optimal model for SRF duration longer than 3 months was identified from 2⁶ patterns. 

Spearman's correlation was used to determine the relationship between CCT and CRT, SRFH, or ONL, and the relationship between SRF duration and CRT, CCT, SRFH, or ONL. 

Statistical significance was set at P < 0.05. Statistical analyses were performed using R version 4.0.2 (The R Foundation for Statistical Computing, Vienna, Austria). 

Results

Patient Characteristics

The patient characteristics at the initial visit are shown in Table 1. There were 34 eyes of 34 male patients and 8 eyes of 8 female patients in the non-prolonged group, and 14 eyes of 14 male patients and 3 eyes of 3 female patients in the prolonged group. Sex, age, time from onset to initial visit, and oral kallidinogenase use were not significantly different between the groups (P = 1, P = 0.492, P = 1, P = 0.179, and P = 0.569, respectively). BCVA at 3 months was recorded in all cases in the prolonged group (n = 17), and in only 27 cases in the non-prolonged group. This is because 15 patients with early SRF resolution in the non-prolonged group stopped visiting our hospital after less than 3 months. In 26 out of the 59 eyes examined, FAF measurements were taken within 1 month from the initial visit. Notably, in 25 of these cases, no observable changes or uniform changes in decreased autofluorescence were recorded in areas where SRF existed. This phenomenon, termed “blocked FAF,”²³ is believed to be present in the early stages of the disease, considering the short duration of symptoms observed in the previous study.²³ The remaining eye exhibited a distinctive “mottled FAF pattern”²³ characterized by diffusely speckled hyper FAF. This pattern is likely related to a prolonged duration of SRF, contrasting with the blocked FAF pattern. However, it is pertinent to note that the average duration of symptoms is reported to be 28.3 days²³ in previous research, and our study's inclusion criteria was eyes where the period from onset to the initial visit was limited to within a month. Consequently, the objective assessment of onset using FAF was considered to be consistent with onset based on the patient's complaint. 

Table 1.

 

Summary of Results for the Current Study

 

Comparison Between the Non-Prolonged and Prolonged Groups

Representative cases from the non-prolonged and prolonged groups are shown in Figure 1. SE, BCVA, CRT, SRFH, and ONL were not significantly different between the non-prolonged and prolonged groups at the initial visit (P = 1, P = 0.217, P = 0.245, P = 0.128, and P = 0.183). BCVA at 3 months was not significantly different between the non-prolonged and prolonged groups (P = 0.208). BCVA at 3 months was not significantly different from that at the initial visit in both the non-prolonged group and prolonged group (P = 0.344 and P = 0.906, respectively). The CCT in the non-prolonged group was 365.8 ± 94.9 µm and that in the prolonged group was 432.7 ± 123.2 µm. The CCT in the prolonged group was thicker than that in the non-prolonged group (P = 0.044). SRF duration in the non-prolonged group was 53.8 ± 25.2 days and that among the 8 cases without any treatment in the prolonged group was 201.9 ± 112.6 days (see Table 1). 

Figure 1.

 

Representative cases in the current study. OCT image at the initial visit in the non-prolonged group (A) shows that CRT was 466 µm, CCT was 382 µm, SRFH was 348 µm, and ONL was 67 µm. OCT image at the initial visit in the prolonged group (B) shows that CRT was 536 µm, CCT was 486 µm, SRFH was 409 µm, and ONL was 88 µm. The SRF improved in the non-prolonged group after 3 months (C), while it persisted in the prolonged group for 3 months (D). OCT, optical coherence tomography; CRT, central retinal thickness; CCT, central choroidal thickness; SRFH, subretinal fluid height; ONL, outer nuclear layer.

The Optimal Model for SRF Duration More Than 3 Months

As a result of model selection, the optimal model for SRF duration more than 3 months included age (P = 0.054) and CCT (P = 0.008, AIC = 67.014; Table 2). 

Table 2.

 

The Optimal Model for SRF Duration More Than 3 Months

 

Correlation Among CCT and CRT, SRFH or ONL, and Among SRF Duration and CRT, CCT, SRFH, or ONL

CCT was significantly correlated with CRT and SRFH, but not with ONL (P = 0.007, r = 0.35; P = 0.002, r = 0.39, and P = 0.294, respectively). SRF duration was significantly correlated with CRT and SRFH, but not with CCT or ONL (P = 0.009, r = 0.40; P = 0.003, r = 0.41; P = 0.236; and P = 0.690, respectively; Table 3, Figures 2–5). 

Table 3.

 

Correlation Among CCT and CRT, SRFH, or ONL, and Among SRF Duration and CRT, CCT, SRFH, or ONL

 

Figure 2.

 

Correlation between CCT and CRT. CCT was significantly correlated with CRT (P = 0.007, r = 0.35).

Figure 3.

 

Correlation between CCT and SRFH. CCT was significantly correlated with SRFH (P = 0.002, r = 0.39).

Figure 4.

 

Correlation between SRF duration and CRT. SRF duration was significantly correlated with CRT (P = 0.009, r = 0.40). SRF, subretinal fluid.

Figure 5.

 

Correlation between SRF duration and SRFH. SRF duration was significantly correlated with SRFH (P = 0.003, r = 0.41).

Discussion

In the current study, baseline CCT in the prolonged group was thicker than that in the non-prolonged group. Age and baseline CCT were closely correlated with SRF duration longer than 3 months. SRF duration was significantly correlated with baseline CRT and SRFH, and baseline CCT was significantly correlated with baseline CRT and SRFH. Patients with CSC have increased permeability and congestion of choroidal vessels.^3–6 Strong choroidal vascular hyperpermeability or congestion was considered to cause water leakage from the choroidal vessels, resulting in a thicker CCT, more SRF, consequently higher CRT, and longer SRF duration. In multivariate analysis, CCT was selected as the optimal model for SRF duration longer than 3 months, but CCT was not significantly correlated with SRF duration in Spearman's correlation. With regard to Spearman's correlation in this study, the SRF duration was investigated only in cases of the non-prolonged group and 8 cases in the prolonged group, because the other 9 patients in the prolonged group received topical treatments that are considered to affect SRF duration such as photocoagulation, PDT, or intravitreous injection of anti-VEGF 3 months after the initial visit. If we could have examined the SRF duration in all cases or included more cases considered to have strong choroidal vascular permeability and congestion, CCT might have also correlated significantly with SRF duration in Spearman's correlation. Thus, we believed that the insufficient number of cases was one of the possible reasons for the lack of the significant correlation. 

Age was also selected as an optimal model in our study. A previous study¹⁶ reported that baseline subfoveal choroidal thickness more than 500 µm, RPE elevation at leakage site more than 50 µm, and age over 40 years were the explanatory factors for SRF duration in naïve CSC. These findings were consistent with ours. RPE repair capacity was considered to decrease with aging related metabolic stress.¹⁶ Thus, we believed that older age was associated with SRF duration. 

ONL was not significantly different between the two groups or selected as the optimal model in our study. Referring to a previous report, ONL becomes thinner in longer duration CSC.²⁴ Therefore, our result indicates that the cell body of the photoreceptors was yet to be damaged because both groups included naïve CSC cases in the acute phase. 

The BCVA at the initial visit and at 3 months did not significantly differ between the non-prolonged and prolonged groups. During the early phases of the acute CSC, BCVA was relatively good despite the presence of SRF. Rather than the mere presence of SRF, the damage to photoreceptors or the ellipsoid zone, as observed through OCT imaging, is the aspect primarily associated with BCVA in cases of chronic CSC.^25–28 In addition, chronic SRF accumulation commonly leads to permanent structural damage to the neuroretina and RPE, with irreversible long-term vision loss.^8–11 The cases in our study were naive CSC, and they were considered to be with relatively limited damage in the neuroretina, including photoreceptors and RPE within the 3 month timeframe. As a result, regardless of the presence of SRF, all cases included in our study demonstrated relatively good visual acuity both at the initial visit and at 3 months. Therefore, we believe that there was no significant change in BCVA from the initial visit to 3 months, and there was no significant difference in BCVA between the 2 groups at each respective time point. 

A previous report investigated the relationship between the baseline choroidal thickness and the need for treatment during a 12-month follow-up period in patients with CSC.¹⁹ That retrospective study included 38 eyes of 33 patients who were untreated but included recurrent cases, unlike our study. The 38 eyes are divided into two groups (group A = CCT <356 µm and group B = CCT >356 µm). They reported that patients with a lower baseline CCT required more treatment during the follow-up period. This report mentioned that the eyes with thicker baseline CCT were in the acute phase of CSC, whereas eyes with thinner baseline CCT were in the chronic phase and required treatment, such as laser photocoagulation or PDT, to the leak point. In contrast, our study included only naïve CSC cases within 1 month from the onset but did not include cases with atrophic choroidal thinning in the chronic phase. Thus, we believe that the difference in baseline CCT results between our study and the previous report was due to the difference in patient demographics. In actual clinical practice, we should first assess whether a patient is in the acute or chronic phase, and then predict SRF duration and treatment resistance for each case by referring to the results of our study and the previous report, as appropriate. 

A previous study on naïve CSC cases focused on SRF duration, although the report did not focus on choroidal structure, which is the main body of CSC pathology.¹⁷ They retrospectively examined 93 eyes diagnosed with naïve CSC that presented within 30 days of symptom onset. They were divided into 2 groups (group 1 and group 2) based on SRF duration from the initial visit: group 1 ≤ 3 months and group 2 > 3 months. This study concluded that female gender and the presence of hyper-reflective dots (HRDs), which are believed to be macrophages and disintegrated outer photoreceptor remnants that become deposited in the subretinal or the intraretinal space secondary to the failed pumping action of the RPE,^20,29,30 and RPE bumps in the baseline OCT are associated with longer SRF duration. This study differs from ours in that the choroidal thickness was not examined. In addition, this study used topical non-steroidal anti-inflammatory drugs (NSAIDs) in all patients. NSAIDs may have affected the results, such as SRF duration.³¹ In contrast, our study did not use eye drops but used oral kallidinogenase for up to 3 months. Kallidinogenase may affect choroidal blood flow.³² However, there was no difference in the time from onset to initial visit between the non-prolonged and prolonged groups in our study, unlike a previous report,¹⁷ and we could compare the two groups without significant differences in patient demographic data. The values of HRDs and RPE bumps depend on the direction of the OCT image. However, CCT can obtain almost the same values as the OCT image in any direction. Therefore, the CCT is considered a better and tractable biomarker in actual clinical practice. 

The limitations of this study are as follows. This study was retrospective and the definition of the first episode was based on the patient's first subjective symptoms, such as distorted vision and central dark spots, which may not be accurate. We did not consider the results of fluorescein angiography and indocyanine green angiography (FA/IA) because FA/IA is rarely performed in patients with CSC at the first visit. Oral kallidinogenase treatment might have affected these results. We might have missed a later recurrence of SRF in the non-prolonged group due to the short follow-up period. Therefore, the non-prolonged group did not always reflect the absence of chronicity. 

In conclusion, a thicker baseline CCT can lead to a longer SRF duration in patients with naïve CSC. Thus, it can allow us to predict cases at a higher risk of prolonged SRF duration and to initiate treatment in such cases at the appropriate time. In the future, we would like to prospectively examine the factors affecting the clinical prognosis of CSC, including FA/IA results, in a larger sample population. 

Acknowledgments

The author would like to thank Editage (www.editage.com) for English language editing. 

Author Contributions: T.S. was involved in the study design, data collection, analysis of results, and manuscript drafting. H.S. participated in the study design and manuscript review and editing. C.O. and H.T. participated in data collection. Y.U. reviewed and edited the manuscript. All the authors have read and approved the final manuscript. 

Disclosure: T. Suzuki, None; H. Sasajima, None; C. Otaki, None; Y. Ueta, None; H. Tate, None