FormalPara Key Summary Points

1. Central serous chorioretinopathy (CSCR) is not characterized by any specific values of axial length of the eyeball or refraction error.

2. CSCR-affected eyes and their fellow eyes demonstrate a significantly greater choroidal thickness compared with healthy individuals.

3. Cases with longer disease duration present with a loss of retinal thickness and volume.

Introduction

Central serous chorioretinopathy (CSCR) is a common disorder affecting young and active individuals, predominantly men [1]. CSCR can lead to significant visual impairment, especially in its chronic form; thus, the potential risk factors for its development have attracted the attention of clinicians around the world [2,3,4,5]. Most often, researchers concentrate on the systemic and exogenous factors predisposing individuals to the onset of CSCR, and local risk factors are reported significantly less often [6,7,8,9,10]. It is only recently that CSCR has been classified as a disorder belonging to the spectrum of pachychoroid diseases, which are characterized by increased choroidal thickness measured by spectral domain or swept-source optical coherence tomography (SD-OCT, SS-OCT) [11, 12]. Other local conditions most frequently analyzed in association with the occurrence of CSCR are the objective refraction and axial length of the eyeball (AL); nevertheless, presented results are scarce [13,14,15,16,17].

The main goal of our study was to determine the biometric parameters characteristic for CSCR. The secondary goals included defining such variables for the subgroup of patients with acute and chronic forms of CSCR, and determining their relationship with the duration of the disease.

Methods

The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by a local bioethics committee board (Komisja Bioetyczna at OIL in Gdansk, approval no. KB-46/21 dated 7.12.2021.). Informed consent was obtained from all the subjects involved in the study. The study group included 66 eyes of 60 consecutive patients who were diagnosed with CSCR in the Dobry Wzrok Ophthalmological Clinic between January 01 2021 and June 30 2021. Diagnosis of CSCR was made on the basis of modern criteria formulated by the Central Serous Chorioretinopathy International Group [18]. Six patients had active binocular disease with the presence of subretinal fluid in both eyes. The control group consisted of 75 eyes of 38 consecutive healthy individuals who underwent a routine ophthalmological check up as part of an occupational medicine program. Amblyopic eyes or eyes with any ophthalmic pathology were excluded from the control group. The study group and the control group were matched for age. Fellow eyes were included in the analysis only if without any retinal disorder, amblyopia, or previously resolved CSCR episode noted in the medical history or detected upon ophthalmological examination. Finally, 39 fellow eyes were included in the analysis. All remaining 21 eyes showed signs of previous resolved CSCR episode detected on fluorescein angiography (FA), fundus autofluorescence (FAF), and SD-OCT exams.

The following ophthalmological procedures were performed on the affected and collateral eyes of all study patients during the initial examination: measurement of distance best-corrected and uncorrected visual acuity (BCVA, UCVA), optical biometry performed with SD-OCT (REVO NX Optopol Technology, Poland 2018), automated refraction (Huvitz, Huvitz Co. Ltd., Korea 2018), assessment of anterior segment and fundus examination at the slit lamp, fundus autofluorescence (Visucam 524, Carl Zeiss Meditec AG, Germany 2019), and SD-OCT and OCT angiography (OCTA; REVO NX Optopol Technology, Poland 2018). Fluorescein angiography was either performed at the initial visit or only analyzed if available in cases previously diagnosed in other centers. The diagnosis of CSCR was made on the basis of medical history and the picture emerging from multimodal imaging: the presence of subretinal fluid on SD-OCT, leakage points recognized on FA, and anomalies of the retinal pigment epithelium (RPE) on FAF. The presence of subretinal neovascularization was ruled out by the OCTA. BCVA values were tested on the Snellen chart and converted to logMAR values. Manifest refraction values were calculated to the spherical equivalent (SE). Subfoveal choroidal thickness (SFCT) was measured manually under the foveal center with the use of the software provided in the SD-OCT device. During the examination, all the patients from the control group underwent the same procedures as the study group, with the exception of FA due to its invasive character.

The analysis included the following parameters: age, BCVA, SE, AL, central subfoveal thickness (CST), minimal central retinal thickness (CRTmin), average central retinal thickness (CRTA), macular volume (MV), and SFCT. The CST parameter refers to the average retinal thickness within the central circle of 1 mm diameter, CRTmin corresponds to the minimal retinal thickness noted within the central circle of 1 mm diameter, the CRTA parameter reflects the average retinal thickness measured within the central circle of 6 mm diameter, MV refers to the retinal volume above the central circle of 6 mm diameter, and SFCT refers to the choroidal thickness measured under the foveola. The areas of measurements refer to Early Treatment Diabetic Retinopathy Study (ETDRS) grid. The parameters were compared among the affected eye group, fellow eye group, and control group. Additionally, the same comparison was performed between the affected eyes with an acute or short-lasting form of CSCR and the chronic or long-lasting form of CSCR. For the purpose of the study, we adopted the threshold value of disease duration for classification as acute versus chronic at 4 months (most of the published reports set this value between 3 and 6 months) [1]. We realize that this division does not reflect the complexity of the disease; however, it enables the retinal damage to be associated with disease duration. Finally, the correlation between the duration of CSCR and the above-mentioned parameters was examined. Disease duration was assessed on the basis of documented medical history or subjective patient’s report of the duration of symptoms. In recurrent cases, the duration of the disorder was considered as the time from the first episode of CSCR.

Statistical Analysis

Categorical variables were presented using integers and percentages. Numerical traits were depicted via their mean, median, standard deviation, and minimum–maximum values. The Anderson–Darling test was used to assess the probability distribution. The homogeneity of variances was appraised using Levene’s test. The statistical significance of between-group differences in the numerical variables was tested using a multifactor analysis of variance (ANOVA) without replication (for normally distributed variables) or the Kruskal–Wallis test (for non-normally distributed ones). The correction for multiple comparisons was applied when building and performing the multivariate models. Generalized linear models with robust standard errors were fitted when implementing a multifactor model incorporating non-normally distributed traits. Spearman’s correlation coefficients were computed to expose relationships between the investigated numerical variables. The chi-square test was applied to assess the independence of variables in the analysis of the distribution of refraction error.

A p-value < 0.05 was considered statistically significant. The study sample size was determined for statistical significance at p < 0.05, a statistical power of at least 80%, and a minimal relevant difference in parameters of 0.5 mm (for AL) and equaled 64 eyes at the minimum. All the statistical procedures were performed using Statistica, release 14 (TIBCO Software Inc., Palo Alto, CA, USA).

Results

The study group included 46 male and 14 female subjects with a mean age of 48.8 ± 10.0 years. Healthy individuals in the control group were of a similar age, i.e., 48.8 ± 11.10 years. The mean duration of CSCR was assessed on the basis of available medical history and patients’ reported symptoms and equaled 44.1 ± 65.40 months (range 1–240 months), with a median of 24 months.

The analysis confirmed a lack of significant differences between the groups according to age, AL, refraction error, and MV. The analysis of simple effects of variance revealed that BCVA was significantly lower in the affected eyes compared with the control eyes and collateral eyes (p < 0.001). Similarly, CST was significantly higher in the affected eyes compared with the control (p < 0.001) and collateral eyes (p = 0.001); however, such a difference was not noted between the collateral and control eyes (p = 0.086). Pairwise comparisons for CRTA and MV revealed substantially higher values in the CSCR eyes compared with the collateral and control eyes (p < 0.01). Significant differences were also noted in the variance of SFCT. Both affected eyes and collateral eyes had significantly higher SFCT values compared with the control eyes (p < 0.001), and the study eyes presented with higher SFCT than the fellow eyes (p < 0.001). A summary of the findings in the three analyzed groups is presented in Table 1. Statistical results for pairwise comparisons among the groups are presented in Table 2.

Table 1 Descriptive statistics for the selected ophthalmological measures in the study participants’ affected eyes versus fellow eyes versus the control group (n = 180 eyes)
Full size table
Table 2 p-Values for pairwise comparisons of CSCR eyes, fellow eyes, and controls
Full size table

The distribution of refraction errors in the study groups is presented in Table 3. Emmetropia was considered when automated refraction values were between –0.25 D and +0.25 D (0.25 D refers to the precision of measurements of the device). The differences in the chi-square test for the whole model were not significant, with p = 0.101. Pairwise comparisons revealed a statistically significant difference for CSCR-affected eyes versus controls, with p = 0.030; however, it should be noted that the presence of subretinal fluid in CSCR produces transient hypermetropia independent of the actual refractive error. Hence, comparison of the fellow eyes with the controls is more reliable for such an analysis as CSCR fellow eyes are devoid of SRF and usually present with a similar refraction error to the affected eyes (no significant anisometropia was reported in the patients’ medical history). This comparison did not reveal any significant difference in the chi-square test, with p = 0.418.

Table 3 Distribution of refraction errors in the study groups
Full size table

The analysis of differences between the affected eyes according to the type of CSCR revealed the following: patients with acute CSCR were on average 8 years younger than patients with chronic CSCR; however, this difference only bordered on statistical significance (p = 0.064). No difference in the AL or refraction error was noted between the control group and CSCR subtypes. However, statistically significant variations were observed for retinal and choroidal measurements. All the results are presented in Table 4.

Table 4 Descriptive statistics for selected ophthalmological measures in the study participants’ affected eyes by phase of disease versus control group (n = 180 eyes)
Full size table

The statistical results of the pairwise comparisons between these groups are presented in Table 5. All the retinal and choroidal measurements were significantly higher in the acute and chronic groups compared with the controls. Nevertheless, when acute and chronic cases were compared, only the choroidal thickness, but not the retinal parameters, was significantly greater in chronic cases.

Table 5 Results of pairwise comparisons of retinal and choroidal parameters among the acute, chronic, and control groups
Full size table

Analysis of the correlation between the duration of the disease and the SD-OCT parameters of the eyeball revealed the occurrence of smaller MV and CRTA values with longer duration. Such a tendency was also noted for CST, CRTmin, and AL; however, in our results, they only bordered on statistical significance. BCVA values in this study were independent of disease duration. The results of the analysis are presented in Table 6.

Table 6 Relationship between ocular parameters and disease duration
Full size table

Discussion

The results of our study do not prove a relationship between axial length or refraction error and the occurrence of CSCR. Nevertheless, patients with CSCR presented with significantly greater choroidal thickness in both affected eyes and healthy fellow eyes. Moreover, chronic cases of CSCR demonstrated higher SFCT values compared with acute cases. The active phase of CSCR was associated with increased retinal thickness and volume. These increased SD-OCT parameters tended to decrease with the duration of the disease, which was clearly visible for the CRTA and MV parameters.

The link among axial length, refraction error, and the onset of CSCR was analyzed previously in a few studies. The presented results were not always consistent and sometimes contradictory. A recent study by Terao et al. on a large number of patients reported statistically shorter AL in CSCR patients versus controls (23.19 mm for bilateral cases and 23.75 mm for unilateral cases versus 24.85 mm in the control group, p < 0.001) [14]. Refraction error (SE) was also significantly greater in the bilateral CSCR subgroup compared with the unilateral and control groups (0.25 D in the bilateral group versus –0.81 D in the unilateral group and –1.28 D in the controls, p < 0.001). Moon et al. reported shorter AL in CSCR patients versus controls (23.52 ± 0.84 mm versus 23.94 ± 1.05 mm) with marginal significance (p = 0.044) [16]. Such a difference was not revealed for the fellow eyes. No differences among affected, fellow, and control eyes were noted regarding SE. On the contrary, Oh et al. reported significantly shorter AL in CSCR-affected eyes compared with fellow eyes (23.76 ± 1.343 mm versus 24.0 ± 1.349 mm, p < 0.001) and similarly, no significant differences in SE between both groups (p = 0.443) [17]. On the other hand, Arora et al. reported results consistent with our findings, i.e., no significant differences in AL and SE among CSCR eyes, CSCR fellow eyes, and controls. Substantial differences between the groups were noted in SFCT measurements [19].

It cannot be excluded that the results obtained in some studies might depend on the statistical methods used for the analysis. AL is an objective parameter; however, the comparison of the results of the study group with those of the fellow eyes and controls is a matter of the precision of the statistical test and the representativeness of the control group. In our research, the sample size was calculated for the study, all the patients were consecutive, and the control group was matched for age and thus representative of the studied population (discussed further below).

The results of our study do not prove a relationship between AL and the incidence of CSCR; nevertheless, it cannot be excluded that a shorter AL affects the course of the disease. A shorter AL has been associated with the presence of pachyvessels, and hence, might add to CSCR severity [20]. In a study by Maltsev et al., a relationship between the occurrence of multiple and scattered leakage points on FA and a shorter AL was revealed [13].

The question of the correlation between the SE value and the occurrence of CSCR requires an understanding of the mechanisms of subjective and objective measurements of refraction. Retinal elevation due to the presence of subretinal fluid and retinal edema produces a hyperopic shift recorded by refraction measurements. The refraction error decreases after the resolution of the fluid [15]. Taking into account such a scenario, it should be stated that the results of refraction obtained in the active phase of CSCR do not reflect the patient’s actual refraction error, and as such, should not be used in the analysis of risk factors of CSCR. For example, the largest study so far (published by Ersoz et al.) analyzed risk factors for CSCR in the medical records of 811 patients versus 816 healthy controls [21]. Comparison of the refraction error between CSCR patients and controls revealed a significant difference in SE: –0.17 ± 1.74 versus –1.01 ± 2.15 (p < 0.001). This finding has to be interpreted as the presence of a transient hypermetropic shift in the active phase of CSCR but does not indicate hyperopia as a risk factor for CSCR. Healthy fellow eyes could be used for such considerations, provided that amblyopia and/or anisometropia are not noted or reported in the medical history before the onset of this condition. In our study, we performed such an analysis, which did not reveal a link among refraction error, hyperopia in particular, and the incidence of CSCR.

Some of the studies directly named myopia as a protective factor against the development of CSCR [22]. Maynath et al. compared the refraction error of 262 patients diagnosed with acute CSCR with that of 237 matched healthy controls. Only 1.1% of patients in the CSCR group had myopia, defined as refraction of –0.5 D or less compared with 29.5% in the control cohort. However, again, the authors did not take into account the hypermetropic shift resulting from retinal elevation due to edema and the presence of SRF; thus, these findings have to be treated with caution. Moreover, a recent analysis of diagnostic imaging of myopic patients with CSCR showed similar imaging and clinical characteristics as in other CSCR patients, apart from the choroidal thickness, which was lower in myopia [23].

The question of bias in the selection of the control group has to be raised as errors might occur in every environmental study. The distribution of refraction error in our control group was practically identical to that reported in a large environmental study by Nowak et al., who reported myopia in 24.1% and hyperopia in 37.5% of cases in a Polish population study including 998 participants [24]. Biometric parameters in a Central European Caucasian population were also assessed in a large study of 2340 patients by Popov et al. [25]. The mean AL was calculated as 23.33 ± 1.01 mm, a value practically identical to the value of 23.33 ± 1.19 noted in our control group. Thus, we believe that our control group is representative according to the distribution of refraction and AL, and the results of the comparative analysis presented in our study are devoid of patient selection bias.

As mentioned in the introduction, CSCR is nowadays considered a pachychoroid entity characterized by increased choroidal thickness. In this respect, our study’s results are consistent with the findings reported by other authors [26,27,28]. We found higher SFCT values in CSCR patients compared with controls and fellow eyes, a finding also reported frequently in other studies [19, 27]. Our study showed that fellow eyes had increased SFCT in reference to the control group. This result puts fellow eyes at risk for CSCR occurrence, which remains in consent with results from other studies. Analysis of patients’ medical history, especially angiographic findings of CSCR, proves bilateral involvement during the course of the disease in a large percentage of cases, sometimes as high as 40% [29,30,31,32]. On the other hand, bilateral symptomatic cases with subretinal fluid detected on the SD-OCT exam in both eyes are not that frequent, and usually are reported in between 10% and 15% of the CSCR patients [21, 33]. In our study, bilateral symptomatic cases occurred in 10% of patients; however, signs of resolved CSCR episode in the fellow eye was detected in 21 patients, thus altogether, bilateral involvement was noted in 35%. This group of patients remains a therapeutic and diagnostic challenge, which becomes clear with application of a new classification of CSCR [18]. According to the recent research by Singh et al., bilateral cases present more commonly as complex CSCR, with numerous retinal pigment epithelium (RPE) alterations and leakage points, contrary to unilateral cases more often presenting with simple phenotype (minor RPE alterations and leakage points) [32]. As such, they more often require more invasive treatment, such as photodynamic therapy, rather than observation or subthreshold micropulse laser [34].

In our results, chronic cases had a statistically significantly thicker choroid than acute cases. Similar findings were reported by Hamzah et al. [35]. Nevertheless, it should be noted that in some research, the opposite phenomenon was observed, i.e., SFCT was greater in acute than in chronic cases [36,37,38]; alternatively, no such difference was observed [21, 26, 39]. Overall, these findings suggest variability in the amount of SFCT increase in CSCR eyes, which is probably not strictly dependent on disease duration. In any case, the difference between choroidal thickness in acute and chronic CSCR patients was relatively small compared with the difference in choroidal thickness between CSCR-affected and healthy eyes.

Analysis of correlation between retinal thickness and disease duration showed that a decrease in CRTA and MV values is associated with a longer duration of this clinical entity. This relationship was observed, despite retinal edema or sponginess, which typically accompanies active phase of CSCR and produces increase in CST and MV, clearly visible in acute cases. It has to be noted that precise assessment of variation of retinal parameters occurring with the persistence of CSCR is always difficult. The first problem is the objective retinal thickness assessment in the active phase. The second is defining disease duration, which in chronic forms has a variable course with periods of resolutions and recurrences. In spite of the technical obstacles, some studies showed loss of retinal layers, observed in SD-OCT measurements occurring with the duration of the disease. [2, 4, 40, 41]. We believe that the present study confirms the existence of significant tendency for retinal loss with the course of CSCR; however, the exact values of correlation coefficients have to be treated with caution.

Strengths and Limitations of the Study

The idea of performing of this study was to provide reliable data on biometric characteristics of CSCR patients. The subject of correlation between the diagnosis of CSCR and biometric parameters is not new, nevertheless, published up-to-date data are limited, and applied research methods not always relevant and consistent. That is why we decided to perform a precisely designed study and compare its results with reports available in the literature so far. We believe that the major strength of this study is its design, which included separate analyses for acute and chronic cases, fellow eyes, and the control group. Additionally, our analysis of refraction error in CSCR patients excludes the bias of retinal edema producing hypermetropic shift in the active phase of the disease. The results for the control group were consistent with measurements reported in other large environmental studies, which confirms its representativeness. Another advantage of this study is its multifactorial approach to the local biometric risk factors of CSCR and its inclusion of AL, SE, SFCT, and retinal parameters.

The limitation of this study is its moderately high number of cases. With a larger sample, some correlations analyzed in our study could prove statistically significant, such as the differences in biometric parameters between acute and chronic cases.

Conclusions

CSCR is a clinical entity that can occur in patients with every type of refraction error. A shorter AL of the eyeball is not associated with a higher prevalence of CSCR. Both acute and chronic CSCR eyes, as well as their fellow eyes, demonstrate significantly greater choroidal thickness compared with healthy eyes. Finally, cases with longer disease duration present with substantial loss of retinal thickness and volume.