Corresponding Author: Se Joon Woo, MD, PhD. Department of Ophthalmology, Seoul National University Bundang Hospital, Seoul National University College of Medicine, 82 Gumi-ro 173beon-gil, Bundang-gu, Seongnam 13620, Korea. Coats disease, initially reported in 1908, is characterized by idiopathic retinal vasculopathy with telangiectatic vessels and exudative retinopathy [1]. This disease primarily affects young male patients unilaterally and is known for its faster disease progression and greater severity in the childhood-onset form than the adult-onset form [2,3]. Typical retinal findings of Coats disease include intraretinal and/or subretinal exudates, exudative retinal detachment, retinal telangiectasia, peripheral nonperfusion, capillary dropout, and light bulb-like microaneurysms, which are particularly well observed on fluorescein angiography [2,4].
The introduction of ultra-widefield fluorescein angiography (UWF-FA) has improved the visualization of peripheral vascular abnormalities in Coats disease [5,6]. Although Coats disease is predominantly noted in the temporal peripheral region, it can also be found in the nasal, inferior, and superior quadrants [7,8]. Previous literature has reported vascular abnormalities on UWF-FA images even in the fellow of Coats disease patients with normal conventional color fundus and f luorescein angiographic appearance [8,9]. However, to our knowledge, there have been no studies investigating the association between UWF-FA features and the final visual outcome in Korean patients with presumed unilateral Coats disease.
Therefore, this study aimed to investigate the retinal vascular abnormalities in both affected and fellow eyes of presumed unilateral Coats disease patients using UWF-FA and their association with visual prognosis in the affected eye.
Materials and Methods
Ethics statement
This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (No. B-2406-906-103), with a waiver of informed consent due to its minimal risk to subjects. The study was conducted in accordance with the Declarations of Helsinki.
Study design
We retrospectively reviewed the medical records of patients diagnosed with Coats disease at a tertiary referral hospital (Seoul National University Bundang Hospital, Seongnam, Korea) from March 2003 to May 2024. Coats disease was clinically diagnosed based on idiopathic retinal vascular telangiectasias, aneurysmal dilatations, and/or associated exudates [1,10]. A total of 60 patients were diagnosed with Coats disease during the study period. Among them, 33 patients underwent UWF-FA at least once during the follow-up period. 30 patients (30 affected eyes, 29 fellow eyes) were selected as the final study population, excluding three patients with concurrent wet age-related macular degeneration, diabetic macular edema, or panuveitis in the affected eye, which could be other causes of exudation. The patients were grouped based on the age of first symptom onset or diagnosis. Adult-onset Coats disease was defined as disease onset at 18 years of age or older [7]. If a patient exhibited visual symptoms related to Coats disease, such as severe decreased visual acuity, or had been diagnosed with Coats disease at another hospital before the age of 18 years, they were classified into the childhood-onset group. Therefore, the age at presentation may differ from the actual age of disease onset.
Included patients were evaluated for the following clinical characteristics and multimodal retinal findings: (1) demographic and clinical data at presentation, including the classification of Coats disease as described by Shields and Shields [2] in 2001, which includes stage 1 (retinal telangiectasia only), stage 2 (telangiectasia and exudation), further subdivided into stage 2A (extrafoveal exudation) and stage 2B (foveal exudation), stage 3 (exudative retinal detachment), divided into stage 3A (subtotal detachment with 3A1 being extrafoveal and 3A2 being foveal), and stage 3B (total retinal detachment). Stage 4 includes total retinal detachment with glaucoma, and stage 5 describes advanced end-stage disease; (2) multimodal retinal imaging findings of the affected eye and fellow eye, including ultra-widefield fundus photography (UWF-FP), optical coherence tomography (OCT), and UWF-FA; and (3) long-term visual outcomes.
We used the Optos California equipment (Optos, Nikon) for UWF-FP and UWF-FA and adjusted the image contrast initially using V2 Vantage Pro software (Optos) during the imaging process. At the time of the retrospective study, we used INFINITII PACS G3 (Infinitt Healthcare) to further refine the contrast and perform quantitative and qualitative analysis of vascular abnormalities. The initial UWF-FA image was evaluated by dividing the retina into four quadrants (superior, temporal, inferior, and nasal), with the optic disc at the center including the macular region. Evaluation measures included retinal telangiectasia, peripheral nonperfusion area, capillary dropout area, late-phase leakage, peripheral granular hyperfluorescence, microaneurysms, and exudate. The peripheral nonperfusion area was defined as extending 2 or more disc diameters from the vascular termini to the ora serrata. When the ora serrata was not visible, measurements were taken only from the vessel termini to the edge of the image, potentially underestimating the actual extent [11-13]. Capillary dropout was defined as the obliteration of capillaries inward from the vascular termini. Retinal exudate, macular edema, and retinal hemorrhage were evaluated using UWF-FP and OCT. The extent of retinal abnormalities such as telangiectasia, peripheral nonperfusion, capillary dropout, and exudate was quantified in clock hours, ranging from 1 to 12, with the optic disc as the center [14]. Specifically, 1 clock hour represents a 30° segment. The mean extent of exudate is the average extent of total exudate across all quadrants, ranging from 0 to 12 clock hours. The mean extent of macular exudate is the average extent of exudate specifically within the macular region, quantified in disc diameters.
We have included all 30 patients in the demographic and initial UWF-FA findings analysis to provide a comprehensive overview. However, the mean follow-up period and final best-corrected visual acuity (BCVA) are reported only for the 28 patients who had follow-up for at least 5 months, which was also the criterion applied for the regression analyses evaluating final visual outcomes.
Severe visual loss was defined as final Snellen BCVA less than 20/200, and moderate visual loss as 20/200 or more but less than 20/66 [15]. In the multivariable logistic regression analysis, we used “moderate to severe visual loss” as the dependent variable instead of “severe visual loss” due to the small sample size of severe cases (n = 5).
Main outcome
The primary outcome was to identify baseline factors associated with final visual outcomes. Secondary outcomes include clinical characteristics comparing baseline demographic information, multimodal imaging focusing on UWF-FA, treatment methods, and final visual outcomes by disease onset types, and UWF-FA features in the fellow eye.
Statistical analysis
We performed statistical analyses using IBM SPSS ver. 19.0 (IBM Corp). For normality testing, the Shapiro-Wilk test was employed. To compare the data, we used the Mann-Whitney-Wilcoxon nonpaired test, chi-square test, or Fisher exact test. Linear regression analysis was conducted to investigate factors associated with the final visual acuity. Additionally, logistic regression analysis was used to investigate prognostic factors associated with final moderate to severe visual loss. Variables with correlation coefficients below 0.9, tolerance above 0.3, and variance inflation factor below 4, identified through collinearity analysis, were included in the multivariable regression analysis. Statistical significance was set at p < 0.05.
Results
All 30 patients were diagnosed with presumed unilateral Coats disease at presentation, comprising 11 childhood-onset (36.7%) and 19 adult-onset patients (63.3%). The mean age at presentation was 16.6 ± 15.0 years (median, 15 years; range, 2 to 57 years) in the childhood-onset group and 39.5 ± 15.8 years (median, 39 years; range, 19 to 73 years) in the adult-onset group (p < 0.001). Patient demographics and characteristics, including age, sex, laterality, follow-up period, initial and final visual acuity, and disease classification at presentation are summarized in Table 1. Representative images of childhood- and adult-onset Coats disease are shown in Figs. 1A-1G and 2A-2F, respectively. The mean initial logarithm of the minimum angle of resolution (logMAR) BCVA in the affected eyes was significantly worse in the childhood-onset group compared to the adult-onset group (0.91 ± 0.96 vs. 0.19 ± 0.51, p = 0.014). The proportion of final moderate to severe visual loss was 60.0% (6 of 10) in the childhood-onset group and 27.8% (5 of 18) in the adult-onset group. Stage 2 was predominant (90.0%), with childhood-onset patients significantly more likely to present with stage 2B or higher compared to adult-onset (72.7% vs. 26.3%, p = 0.023). Decreased vision was the most common presenting symptom/sign (overall, 36.7%), followed by incidental discovery during routine examination (overall, 33.3%). The adult-onset group had significantly higher rates of focal/scatter laser treatment (94.7 vs 54.5%, p = 0.016) and the total number of laser treatments (2.7 vs 2.0, p = 0.011) compared to childhood-onset. Among the patients diagnosed with Coats disease in the fellow eye through UWF-FA, two underwent treatment for the fellow eye. In one patient, a single prophylactic scatter laser treatment was performed for extensive peripheral telangiectasia and nonperfusion areas. In another patient, a single intravitreal anti-vascular endothelial growth factor injection was administered for macular edema and subfoveal exudate (Fig. 2). We observed no progression in all patients, including those who received treatment in the fellow eye during the follow-up period.
Retinal vascular abnormalities of the affected eye
The retinal vascular abnormalities of the affected eye at presentation are summarized in Supplementary Table 1 for frequency analysis and Table 2 for quantitative analysis. Fig. 3A-3C presents a comparative graph based on the data from Table 2. Mean extent of telangiectasia was 7.1 clock hours, predominantly located in the temporal quadrant (100%), followed by nasal (83.3%), inferior (56.7%), and superior quadrants (53.3%), with macular involvement seen in 40% of patients. The extent of telangiectasia was greatest in the temporal quadrant (2.8 clock hours), followed by nasal (2.0 clock hours), superior (1.2 clock hours), and inferior quadrants (1.1 clock hours). There was no significant difference in the extent of telangiectasia between the childhood- and adult-onset groups (8.0 clock hours vs. 6.6 clock hours, p = 0.185). The mean extent of exudate was 3.6 clock hours, predominantly located in the temporal (90.0%), followed by superior (40.0%), inferior (26.7%), and nasal quadrants (23.3%). Compared to the adult-onset group, the childhood-onset group had a significantly larger mean extent of exudate (5.2 clock hours vs 2.7 clock hours, p = 0.033) and mean extent of macular exudate (3.2 disc diameters vs. 1.3 disc diameters, p = 0.033). The mean extent of exudate, which covers all quadrants, ranges from 0 to 12 clock hours. In contrast, the mean extent of macular exudate is quantified in disc diameters. Additionally, the childhood-onset group showed a significantly higher frequency (63.6% vs. 5.3%, p = 0.001) and larger extent of exudate in the inferior quadrant (1.4 clock hours vs. 0.1 clock hours, p = 0.006) compared to the adult-onset group. The extent of exudate exhibited a positive correlation with the extent of telangiectasia in the affected eye, which was statistically significant (r = 0.43, p = 0.02). The peripheral nonperfusion was observed in 76.7% of affected eyes, predominantly in the temporal quadrant (73.3%), followed by the superior (23.3%), inferior (13.3%), and nasal quadrants (13.3%). There was no significant difference in the extent of peripheral nonperfusion between the childhood- and adult-onset groups (3.4 clock hours vs. 2.3 clock hours, p = 0.328). Capillary dropout was observed in 96.7% of affected eyes, most commonly in the temporal (86.7%), followed by the superior (46.7%), nasal (33.3%), and inferior quadrants (23.3%), with a significantly larger extent of capillary dropout in the childhood-onset group compared to the adult-onset group (5.0 clock hours vs. 2.8 clock hours, p = 0.023). Retinal hemorrhage was observed in 56.7% of cases, predominantly in the temporal (40.0%), followed by the superior (10.0%), nasal (10.0%), and inferior quadrants (6.7%), with no significant differences between the childhood- and adult-onset groups (p > 0.999). Macular edema was present in 63.6% and 31.6% of the childhood- and adult-onset groups, respectively, without statistical significance (p = 0.132).
Retinal vascular abnormalities of the fellow eye
The retinal vascular abnormalities of the fellow eye at presentation are summarized in Supplementary Table 1 for frequency analysis and Table 2 for quantitative analysis. Fig. 3 presents a comparative graph based on the data from Table 2. On UWF-FA images, telangiectasia was observed in 51.7% of the fellow eye, predominantly in the temporal (48.3%), nasal (24.1%), inferior (13.8%), and superior quadrants (10.3%), with macular involvement seen in 6.9%. The extent of telangiectasia was greatest in the temporal quadrant (1.2 clock hours), followed by the nasal (0.5 clock hours), superior (0.1 clock hours), and inferior quadrants (0.1 clock hours). There was no significant difference in the extent of telangiectasia between the childhood- and adult-onset groups (2.0 clock hours vs. 1.8 clock hours, p = 0.550). Late-phase leakage was present in 51.7%, primarily in the temporal quadrant (44.8%). Peripheral granular hyperfluorescence was seen in 69.0%, predominantly in the temporal quadrant. The adult-onset group showed a significantly higher proportion compared to the childhood-onset group (83.3% vs. 45.5%, p = 0.048). Microaneurysms were observed in 65.5% of patients, predominantly in the temporal quadrant (62.1%). However, there were no significant differences between subgroups (p > 0.999). Coats disease was diagnosed in 15 of 29 fellow eyes (51.7%), using UWF-FA images; stage 1, 13 patients (44.8%); stage 2A, one patient (3.4%); and stage 2B, one patient (3.4%) (Table 3). Through UWF-FA, patients finally diagnosed with bilateral Coats disease were 7 of 11 patients (63.6%) in the childhood-onset group and 8 of 18 patients (44.4%) in the adult-onset group, with no statistically significant difference (p = 0.160). Notably, the extent of telangiectasia and exudate (clock hours) in the affected eye did not show a significant correlation with the extent of telangiectasia (clock hours) in the fellow eye (r = 0.12, p = 0.54; r = −0.13, p = 0.49, respectively).
Factors associated with final visual loss outcomes in the affected eye
For the analysis of final visual outcomes, we included 28 patients who had a minimum follow-up period of 5 months. The mean follow-up period for these patients was 64.8 ± 46.5 months (median, 53.9 months; range, 5.1 to 187.1 months). Factors associated with final visual outcomes are summarized in Tables 4 and 5. On univariable linear regression analysis, final logMAR BCVA was significantly associated with initial logMAR BCVA (coefficient β, 0.665; 95% confidence interval [CI], 0.288 to 1.043; p = 0.001), and macular abnormalities such as telangiectasia, exudate, or edema (coefficient β, 0.546; 95% CI, 0.009 to 1.083; p = 0.046). Other factors, including age, sex, childhood-onset, extent of telangiectasia and exudate in the affected eye, and extent of telangiectasia in the fellow eye, did not show a statistically significant association (Table 4). In multivariable linear regression analysis, initial logMAR BCVA was significantly associated with final logMAR BCVA (coefficient β, 0.593; 95% CI, 0.158 to 1.029; p = 0.010), whereas macular abnormalities did not show a statistically significant association in the affected eye (coefficient β, 0.188; 95% CI, −0.359 to 0.734; p = 0.486) (Table 4). The extent of telangiectasia in the fellow eyes showed no significant correlation with the final logMAR BCVA in the affected eyes (r = −0.21, p = 0.27). The mean extent of telangiectasia in the temporal and nasal quadrants (odds ratio [OR], 12.759; p = 0.043) and the initial logMAR BCVA of the more affected eye (OR, 11.841; p = 0.024) were identified as prognostic factors for final moderate to severe visual loss (Snellen BCVA <20/66). No significant associations were found with the extent of telangiectasia, defined as the total extent observed across all four quadrants (OR, 0.402; p = 0.092) nor with its extent in any quadrant of the fellow eyes.
We conducted an additional regression analysis with treatment status and the number of treatments as independent variables. The univariable linear regression analysis showed a significant association between intravitreal bevacizumab injection (vs. without) and final logMAR BCVA (coefficient β, 0.735; 95% CI, 0.191 to 1.280; p = 0.010), as well as between the number of intravitreal bevacizumab injections and final logMAR BCVA (coefficient β, 0.189; 95% CI, 0.083 to 0.296; p = 0.001). In the multivariable linear regression analysis, the number of intravitreal bevacizumab injections was significantly associated with final logMAR BCVA (coefficient β, 0.162; 95% CI, 0.004 to 0.320; p = 0.045), whereas intravitreal bevacizumab injection (vs. without) did not show statistical significance (coefficient β, 0.178; 95% CI, −0.570 to 0.925; p = 0.692). In the multivariable logistic regression analysis with final moderate to severe visual loss (Snellen BCVA < 20/66) as the dependent variable, neither intravitreal bevacizumab injection (vs. without; OR, 4.338; p = 0.335) nor the number of intravitreal bevacizumab injections (OR, 1.499; p = 0.437) showed statistical significance.
Discussion
In this study, we conducted a retrospective study on presumed unilateral Coats disease patients to investigate the UWF-FA features of the affected and fellow eyes, and to identify baseline factors related to the final visual prognosis in the affected eyes.
In our study, all 30 patients with Coats disease showed an earlier disease stage, predominantly stage 2 (90.0%), followed by stage 1 (6.7%), and stage 3A2 (3.3%) (Table 1). This differs from findings in other race populations reported previously [16-18]. However, our study showed comparable results to the Korean study by Lee et al. [8], which reported 95.2% at stage 2 and 4.8% at stage 3. This difference might be due to higher medical accessibility in Korea and well-established pediatric ophthalmic screening, enabling early diagnosis at earlier disease stages [8,19]. Like the study by Kang et al. [7], our study showed a higher proportion of adult-onset Coats disease. The visual acuity at presentation in the adult-onset group was significantly better compared to that in the childhood-onset group (p = 0.014), and the disease was less severe. Childhood-onset group was significantly more likely to present with stage 2B or higher compared to the adult-onset group (72.7% vs. 26.3%, p = 0.023).
Shields et al. [4] in 2001 analyzed vascular abnormalities in 158 affected eyes primarily in childhood-onset Coats disease patients (mean age, 11 years; range, 1 month to 63 years) using fluorescein angiography. In their study, telangiectasia was observed in the temporal quadrant (42%), followed by the inferior (25%), superior (5%), and nasal quadrants (4%), with 22% showing diffuse involvement across two or more quadrants. In our study using UWF-FA, telangiectasia was observed in the following order: temporal (100%), nasal (83.3%), inferior (56.7%), and superior quadrants (53.3%) (Supplementary Table 1). There were no significant differences in frequency or extent between subgroups (Table 2 and Supplementary Table 1). Exudate was observed in the temporal quadrant (90.0%), followed by the superior (40.0%), inferior (26.7%), and nasal quadrants (23.3%). The mean extent of exudate across all quadrants and macular exudate were significantly larger in the childhood-onset group compared to the adult-onset group (5.2 clock hours vs. 2.7 clock hours, p = 0.033; 3.2 disc diameters vs. 1.3 disc diameters, p = 0.033, respectively). Unlike previous studies, our study identified higher frequencies of vascular abnormalities in affected eyes. It is believed that UWF-FA allows for more detailed identification of far peripheral pathological changes.
Brockmann et al. [9] analyzed the fellow eyes of 19 unilateral Coats disease patients using UWF-FA, reporting that 100% of them exhibited vascular abnormality predominantly located in the temporal periphery. Lee et al. [8] investigated the fellow eyes of 21 unilateral Coats disease patients using UWF-FA, reporting peripheral vascular abnormalities in 52.4% of cases. Similarly to previous studies, our study found telangiectasia in 51.7%, late-phase leakage in 51.7%, and microaneurysms in 65.5% of the fellow eyes, predominantly located in the temporal quadrant. Coats disease was diagnosed in 15 of 29 fellow eyes (51.7%) of fellow eyes with stage 1 being the most common with 13 patients (44.8%), while stages 2A and 2B were one patient (3.4%) each. The significant prevalence of retinal vascular abnormalities in the fellow eye supports the idea that Coats disease exhibits bilateral disease characteristics, suggesting a potential underlying genetic cause [5,9]. The extent of telangiectasia and exudate in the affected eyes did not exhibit a significant correlation with the extent of telangiectasia in the fellow eye (r = 0.12, p = 0.54; r = −0.13, p = 0.49, respectively). This suggests that there was no association between disease severity in the affected eye and abnormal findings in the fellow eye. Wang et al. [20] reported that 12.65% of eyes showed peripheral granular hyperfluorescence when analyzing the far peripheral retina using UWF-FA in 166 normal eyes of 83 patients. However, in our study, peripheral granular hyperfluorescence was observed in 69.0% of cases, and the adult-onset group exhibited a significantly higher frequency compared to the childhood-onset group (83.3% vs. 45.0%, p = 0.048). Wang et al. [20] and Panda-Jonas et al. [21] suggested that peripheral granular hyperfluorescence could be a normal appearance due to the uneven distribution of physiological retinal pigment epithelium cells in the far periphery and might show a higher frequency with age. However, given the higher frequency observed in our study, further research is needed to determine whether it is indeed a normal appearance or a vascular abnormality.
Previous studies have reported that a greater extent of telangiectasia and exude at presentation, dense macular exudate, subfoveal nodule, and younger age are factors associated with poorer visual outcomes [2,22-24]. Similarly to the previous study, univariable linear regression analysis in our study showed significant associations between macular abnormalities (e.g., telangiectasia, exudate, or edema) and final logMAR BCVA (coefficient β = 0.546, p = 0.046), as well as between initial logMAR BCVA and final logMAR BCVA (coefficient β = 0.665, p = 0.001). However, in multivariable linear regression analysis, only initial logMAR BCVA remained significantly associated (coefficient β = 0.593, p = 0.010). Since approximately 96.7% of our patients were at Coats disease stage 2 or below, the presence of macular abnormalities such as macular telangiectasia, exudate, or edema at presentation may not have had a significant impact on poorer final logMAR BCVA due to effective treatment. Additionally, the variable “macular abnormalities” does not account for quantitative factors such as the extent of telangiectasia, exudate, or edema. Therefore, initial logMAR BCVA might be a more effective indicator of disease severity. This could explain why, in the multivariable linear regression analysis, initial logMAR BCVA was a more significant predictor than macular abnormalities.
Furthermore, the small sample size in our study may have been one of the reasons why macular abnormalities did not show statistical significance in the multivariable analysis. Notably, in the multivariable logistic regression analysis, the mean extent of telangiectasia in the temporal and nasal quadrants (OR, 12.759; p = 0.043) and the initial logMAR BCVA of the more affected eye (OR, 11.841; p = 0.024) were identified as prognostic factors for final moderate to severe visual loss (Snellen BCVA <20/66). No significant prognostic factors were found for the extent of peripheral nonperfusion or capillary dropout in the affected eye. Therefore, conducting a detailed investigation of the extent of telangiectasia, including up to the peripheral regions, particularly in the nasal and temporal quadrants, through earlier UWF-FA, may aid in predicting final moderate to severe visual loss. Given the small sample size of 30 participants, subgroup analysis between childhood and adult-onset groups was not performed, suggesting the need for future large-scale studies.
Additional multivariable linear regression analysis, with treatment status and the number of treatments as independent variables, found that only a higher number of intravitreal bevacizumab injections was significantly associated with worse final logMAR BCVA (coefficient β, 0.162; p = 0.045). This suggests that a greater number of bevacizumab injections may be associated with a more severe condition of Coats disease, which could have required additional treatments and is likely related to poorer visual outcomes.
Our study has several limitations. First, it was conducted in a single tertiary center with a retrospective design, and the population was ethnically homogeneous, making generalization difficult. Second, there was no appropriate control group included. Third, the inclusion of only patients who underwent UWF-FA may have introduced selection bias. Fourth, in our study, we used the concept of clock hours as an indirect method to quantify vascular abnormalities due to the limitations of UWF-FA resolution. We anticipate that future improvements in UWF-FA resolution will enable direct and accurate quantification of abnormalities across the entire field using software such as ImageJ (US National Institutes of Health). Fifth, the Optos equipment used for UWF-FA may show a wider field in the nasal and temporal quadrants compared to the superior and inferior quadrants, which introduces a possibility of overestimation [25]. However, the strengths of our study include the long-term follow-up of a relatively large number of patients with a rare disease, allowing us to identify prognostic factors related to final visual outcomes. Additionally, using UWF-FA, we discovered that a significant proportion of presumed unilateral Coats disease patients have bilateral Coats disease. In the future, we expect that large-scale prospective studies using UWF-FA will provide more detailed insights into factors associated with final visual outcomes, including comparisons between childhood and adult-onset groups.
In conclusion, through UWF-FA, about half of presumed unilateral Coats disease cases could be diagnosed as bilateral Coats disease with asymmetric severity. Visual prognosis is associated with the peripheral retinal telangiectasia in the temporal and nasal quadrants as well as initial visual acuity in the affected eyes. Although retinal vascular abnormalities in the fellow eye were not associated with the visual prognosis of the affected eye, a thorough examination of peripheral telangiectasia particularly in the nasal and temporal quadrants, as well as in the macula of the affected eye is crucial for assessing visual prognosis. Consequently, early use of UWF-FA is essential for accurate diagnosis and prognosis.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
Print
