Abstract
Background:
Thyroid eye disease (TED) is a debilitating autoimmune disorder linked to thyroid dysfunction. There is limited knowledge of TED in Asian populations. This multicenter study characterizes the clinical features and treatment response of TED in a large Chinese cohort.
Methods:
A retrospective multicenter study included 4157 patients with TED from nine Chinese hospitals from February 2016 to July 2023. Disease severity and activity were evaluated according to the European Group on Graves’ Orbitopathy standards. We examined associations of variables including sex, age, smoking status, I131 treatment, consultation department, and geographical region with clinical outcomes. Logistic regression and nomogram models were developed to examine associations with sight-threatening TED and, in a subgroup analysis (n = 126), patients’ responsiveness to intravenous glucocorticoid (IVGC) therapy.
Results:
We included 4157 patients with mean age and standard deviation (SD) 45.96 ± 16.44 years. Of these, 63.6% (n = 2644) were females. Over half (55.6%, n = 2310) of participants were in the inactive phase, with a mean clinical activity score of 2.19 ± 1.61 (SD) for all patients. TED severity was categorized as mild (9.3%, n = 385), moderate-to-severe (82.5%, n = 3428), and sight-threatening (8.2%, n = 344). The average degree of exophthalmos was 20.04 ± 5.27 mm, and 48.8% (n = 2029) of patients had diplopia. Patients treated with I131 had higher disease activity (47.5%, n = 468, vs. 43.5%, n = 1379, p < 0.05). Coastal region patients exhibited more severe TED (sight-threatening cases: 10.1%, n = 195, vs. 7.2%, n = 147) and higher diplopia scores (1.00 ± 1.10 vs. 0.86 ± 1.09, p < 0.001) than inland counterparts. TED severity was also greater in patients treated in Ophthalmology departments (mild cases: 6.0%, n = 213; moderate-to-severe cases: 85.6%, n = 3055) compared with Endocrinology departments (mild cases: 29.3%, n = 172; moderate-to-severe cases: 63.5%, n = 373). Nomograms had an area under the receiver operating curve of 0.742 (confidence interval [CI] 0.716–0.768) for sight-threatening TED and 0.759 (CI 0.674–0.843) for IVGC therapy responsiveness.
Conclusions:
We characterized the clinical features and treatment response of TED in a large Chinese cohort. These findings offer valuable insights informing TED risk stratification in Asian patients and forming a foundation for future prospective studies.
Introduction
Thyroid eye disease (TED), also known as thyroid-associated orbitopathy or Graves’ orbitopathy (GO), is an organ-specific autoimmune disease closely related to thyroid disease. 1 TED is the most common extrathyroidal manifestation of Graves’ disease (GD), occurring in 25 − 40% of patients with GD. 2 TED can cause exophthalmos, diplopia, and even blindness in patients, and may adversely affect social interactions and psychological health to varying degrees. 3,4 Thus, patients’ quality of life may be greatly reduced by the presence of TED. 5
This large multicenter study reports the clinical characteristics of Chinese patients with TED and examines associations of clinical variables with disease activity/severity and treatment response.
Methods
Study design and participants
This multicenter, retrospective study was led by Shanghai Ninth People’s Hospital (SH9H), Shanghai Jiao Tong University School of Medicine. Patients who visited nine centers across different regions of China from February 2016 to July 2023 were included. Departments of both Ophthalmology and Endocrinology from general hospitals, as well as ophthalmology specialty hospitals, were included. Only patients with a confirmed diagnosis of TED and complete clinical data were included. Patients with incomplete information on key demographic or clinical data were excluded. The diagnosis of TED was based on the Bartley criteria. 6 This study was in accordance with the Declaration of Helsinki and approved by SH9H, Shanghai Jiao Tong University School of Medicine Ethics Committee (No. SH9H-2023-T11-3). Only de-identified retrospective data were used for analysis, without the active involvement of participants.
Data collection and definitions
We retrieved existing medical records from Electronic Medical Record databases of multiple centers to identify all eligible patients diagnosed with TED. For each patient, the selected time point for clinical data collection was the initial consultation of TED at the center where they were seen. We collected data on demographic characteristics, medical history, clinical risk factors for TED, and clinical presentation. In order to ensure the accuracy and authenticity of the data, the research team adopted a standardized data extraction and verification process, established a unified data entry protocol, implemented double-checking, and used an error correction mechanism. 7 These records were retrospectively reviewed and analyzed for the study. The demographic information included sex, age, and place of residence. Data on the disease duration captured the date of initial consultation, the onset date of TED, and the type and onset date of thyroid disease. The treatment history indicated the treatment patients received prior to enrollment. Risk factors for TED included the smoking habits and the history of I131 treatment (defined as prior I131 therapy documented in the patient’s medical records prior to study enrollment). Clinical characteristics documented for each patient included symptoms of TED, activity and severity of TED, and the degree of exophthalmos and diplopia.
The severity of TED was determined by the European Group on Graves’ Orbitopathy. 3 The activity was determined by the clinical activity score (CAS). 8 A CAS score of ≥3/7 indicates an active disease phase, whereas a CAS score of <3/7 indicates clinical inactivity. 9 The degree of exophthalmos was measured using the Hertel exophthalmometer. Diplopia was recorded based on the modified Bahn–Gorman grading. 10 Data on exophthalmos and diplopia were collected for both eyes of each patient, with the higher value from either eye used for analysis. To explore the potential associations of geographical factors with TED, we divided the patients into a coastal group and an inland group according to their place of residence. According to the classification of the State Planning Commission and the National Bureau of Statistics, 11 the provinces where patients usually reside can be divided into coastal provinces and inland provinces.
Therapy response to intravenous glucocorticoid (IVGC) treatment was assessed within three months after the last administration of IVGC. The definition of “responsive” and “unresponsive” was based on the standard proposed by Bartalena et al. 3 The responsive group included those with an improvement of at least two of the following in one eye after treatment: (1) reduction of lid aperture ≥2 mm, (2) reduction of exophthalmos ≥3 mm, (3) eye motility with an increase of ≥8, and (4) reduction in five-item CAS (not including spontaneous or gaze-evoked pain) of ≥1 point; without concomitant deterioration in the other eye. Deterioration was defined by the occurrence of dysthyroid optic neuropathy (DON) or worsening of at least two of the four components mentioned above. The unresponsive group was comprised of those who did not meet the aforementioned criteria.
Statistical analysis
Categorical variables were presented as counts and percentages, while continuous variables were summarized as mean ± standard deviation (SD) or median (interquartile range) otherwise. Group comparisons employed Wilcoxon rank-sum tests for non-normally distributed data and chi-square/Fisher’s exact tests for categorical variables (p < 0.05). All statistical analyses were performed using R (v4.3.3) by the biostatistics team at Clinical Research Institute, Shanghai Jiao Tong University School of Medicine.
Development of mediation effect model
Independent variables evaluated included sex and age, with mediating variables consisting of smoking, and dependent variables describing TED severity. Mediation analysis was utilized to examine the influence of sex and age through the mediating variables on disease severity, involving calculations of average causal mediation effects, average direct effects, total effects, and the proportion mediated. Stratified analyses based on sex and age were conducted to identify subgroup differences.
Development and validation of nomograms
A multivariable logistic regression model was used to develop a nomogram evaluating the associations with sight-threatening TED, and in a subgroup analysis, IVGC responsiveness. Statistically significant associations were first identified through univariate regression and then incorporated into the multivariable model.
Multivariable logistic regression with backward elimination (AIC-based) was used to evaluate the association with sight-threatening TED, utilizing a nomogram (rms package for R). Performance was assessed via the area under the receiver operating characteristic (ROC) curve (AUC), C-index, and calibration curves. IVGC responsiveness was modeled similarly, incorporating sensitivity analyses and Wald statistics for variable importance. Missing data were imputed using MICE, and model calibration was evaluated via Brier scores and Hosmer–Lemeshow tests.
Each nomogram was developed based on the model’s β coefficients, providing a quantitative tool for risk assessment by assigning weighted scores to predictive factors. The model’s performance was evaluated using ROC curve analysis and calibration assessment, with internal validation conducted via the Concordance Index to ensure robustness. Sensitivity analyses were performed to assess model stability under different parameter adjustments, with calibration curves generated.
Results
Baseline
This study included 4157 patients (Table 1, Fig. 1). There were 1513 males (36.3%) and 2644 females (63.6%). The average age was 45.96 ± 16.44 years, with males being 49.37 ± 12.84 years and females 44.01 ± 17.9 years. A demographic distribution map was developed based on the patients’ place of residence (Fig. 2A). From a coastal and inland perspective, 1939 cases (46.6%) resided in coastal areas, while 2035 cases (53.4%) lived in inland China. Most patients (4015; 96.6%) were hyperthyroid. Only 102 patients (2.4%) were hypothyroid, and 40 patients (1.0%) were euthyroid. There were 766 smokers (18.4%). A total of 985 cases (23.7%) had received I131 treatment. A total of 1229 patients (29.6%) had symptoms for less than 9 months. We found 2310 patients (55.6%) to be in the inactive phase. The mean CAS was 2.19 ± 1.61. There were 385 patients with mild disease (9.3%), 3428 with moderate-to-severe disease (82.5%), and 344 with sight-threatening disease (8.2%). The average degree of exophthalmos was 20.04 ± 5.27 mm, and 48.8% of patients had diplopia.
Flowchart of study participants included and excluded in analyses.
Demographic characteristics and clinical characteristics in diverse age groups.
Baseline Characteristics of Patients with Thyroid Eye Disease Patients
CAS, clinical activity score; SD, standard deviation.
Age
In order to stratify the sample sizes across various age groups, we categorized age into five groups: under 30 years (487 patients), 30–39 years (875 patients), 40–49 years (1100 patients), 50–59 years (602 patients), and over 60 years (1083 patients) (Table 2). Results showed the peak age of onset for males to be 50–59 years, with that of females to be 40–49 years (Fig. 2B). The proportion of patients in the active phase of the disease increased with age (p < 0.001) (Fig. 2C). Specifically, 33.06% of participants under 30 years of age were in the active phase, which rose to 56.31% in those over 60 years. The severity of TED also increased with age (p < 0.001): only 7.19% of the participants under 30 years of age were considered to have sight-threatening TED, compared with 12.62% of those over 60 years (Fig. 2D). The exophthalmos varied statistically significantly across different age groups (p < 0.05) (Fig. 2E). The average diplopia score increased with age (p < 0.001), from 0.52 ± 0.88 in the participants under 30 years of age to 1.35 ± 1.14 in those over 60 years (Fig. 2F).
Clinical Characteristics According to Age Groups
Sex
Overall, males in the study had more severe disease than females. More than half of female patients (1565 patients, 59.2%) were in the inactive phase, whereas 49.2% (745 patients) of male patients were in the inactive phase (p < 0.001) (Table 3). Sight-threatening disease was present in 191 females (7.2%) and 153 males (10.1%) (p < 0.001). Males exhibited greater degrees of exophthalmos (20.7 ± 4.7 mm) compared with females (19.7 ± 5.5 mm) (p < 0.001). The average diplopia score for females was 0.79 ± 1.05, compared with 1.20 ± 1.14 for males (p < 0.001).
Clinical Characteristics According to Clinical, Demographic, and Treatment Characteristics
Smoking habits
The association of smoking habits with TED outcomes was analyzed (Table 3). Although the proportion of smokers in the active phase was slightly higher than the nonsmokers, the difference between the two groups did not reach statistical significance (p = 0.193). The proportion of sight-threatening disease was significantly higher among smokers compared with nonsmokers (p < 0.001), with 7.3% of nonsmokers and 12.8% of smokers with this degree of severity. Smokers exhibited higher exophthalmos, with an average of 20.55 ± 3.21 mm compared with 19.72 ± 2.90 mm in nonsmokers (p < 0.001). Furthermore, smokers had higher mean diplopia scores of 1.30 ± 1.13 compared with 0.86 ± 1.08 in nonsmokers (p < 0.001).
Prior I131 treatment
We explored the association of prior I131 treatment and TED (Table 3). A higher proportion of patients who received prior I131 treatment compared with those who did not were in the active phase (p < 0.05): 47.5% (468 patients) versus 43.5% (1379 patients), respectively. There was no significant difference in disease severity between the two groups (p = 0.079). However, patients with a history of I131 treatment had worse exophthalmos (p < 0.01), averaging 20.10 ± 2.77 mm versus 19.81 ± 3.03 mm, respectively. Patients with a history of I131 treatment had lower diplopia scores (p < 0.05), averaging 0.89 ± 1.11, compared with 0.96 ± 1.10, in those who did not receive I131 treatment.
Region of participant residence
Regarding the activity of TED relative to region (Table 3), 39.2% of patients residing in coastal areas were in the active phase compared with 49.6% in inland areas (p < 0.001). Sight-threatening cases were more prevalent in coastal areas (10.1%) compared with inland areas (7.2%). Mean diplopia scores were significantly higher in coastal patients 1.00 ± 1.10 compared with 0.86 ± 1.09 in inland areas (p < 0.001). However, exophthalmos data were not statistically different (p = 0.465).
Consulting department
In this study, 53.8% of patients who had consults in Ophthalmology departments were in the inactive phase, compared with 66.4% of Endocrinology clinic patients (p < 0.001). There were also significant differences in disease severity between patients in the Ophthalmology and Endocrinology departments (p < 0.001), particularly in mild and moderate-to-severe cases. In the Ophthalmology department, 6.0% (213 patients) had mild disease severity and 85.6% (3055 patients) had moderate-to-severe disease severity, while in the Endocrinology department, 29.3% (172 patients) had mild disease and 63.5% (373 patients) had moderate-to-severe disease. Patients in Ophthalmology department had worse exophthalmos (p < 0.001), mean measurement of 20.00 ± 2.86 mm compared with 19.15 ± 3.52 mm in Endocrinology department. Moreover, the diplopia score was higher in patients from Ophthalmology department (p < 0.001), with an average score of 0.99 ± 1.13 versus 0.66 ± 0.85 in Endocrinology department.
Construction of mediation effect model
This study employed a mediation effect model to rigorously explore how sex and age might influence severity of TED through the intermediary of smoking. Our analysis revealed that 22.82% of the total effect of sex on disease severity was mediated through smoking, demonstrating statistical significance (p < 0.001). Similarly, 16.75% of the impact of age on disease severity was mediated through this behavioral factor (p < 0.001). These findings validated the mediation effect model in elucidating the complex interactions between biopsychosocial factors.
Development of the nomogram for sight-threatening TED
A nomogram was developed examining associations with sight-threatening TED (Fig. 3A). A total of 3974 patients with sight-threatening TED were included. A total of nine variables—sex, age group, smokers, I131 treatment, treatment history, region type, diplopia, eye redness, and duration—were included in the model. The AUC of the nomogram was 0.742 (confidence interval [CI] 0.716–0.768). The Concordance Index was 0.725 (CI 0.680–0.771).
Nomograms for sight-threatening TED and intravenous glucocorticoid (IVGC) therapy responsiveness.
Nomogram examining IVGC therapy responsiveness
A nomogram was also developed to examine the association with TED IVGC therapy responsiveness in a subgroup of patients for whom data on IVGC responsiveness was available (n = 126). A total of 126 patients with data on IVGC responsiveness were included. The nomogram included statistically significant clinical variables including sex, age group, CAS, exophthalmos, lid aperture, and duration (Fig. 3B). Among these variables, lid aperture, CAS, and exophthalmos emerged as the three most influential variables in the model. The AUC was 0.759 (CI 0.674–0.843).
Discussion
This study is among the world’s largest multicenter, retrospective studies of TED. It was conducted across nine centers in China, reporting the clinical characteristics of 4157 patients with TED. To the best of our knowledge, this study is the first to link regional variations and geographic factors with clinical presentations. Through multivariable regression analysis, the study delineates the association of clinical variables with severity and treatment response of TED.
Some clinical manifestations observed in our cohort differed from previous reports. For instance, the proportion of mild cases (9.3%) was lower, 12 –15 and the proportion of sight-threatening cases (8.2%) was higher than those reported in European populations (3 − 4%). 16,17 These discrepancies may stem from several factors. First, our data were primarily sourced from tertiary care centers, which typically manage referrals with more severe cases. Second, this bias was evident in the Ophthalmology department, where moderate-to-severe cases predominated (85.6%), and mild cases constituted only 6.0%. In contrast, the Endocrinology department saw a higher proportion of mild cases (29.3%). Finally, anatomical and ethnic differences may play a role. Asians generally have shallower and narrower orbits, which can increase the risk of blood vessel and optic nerve compression, predisposing them to DON. 18
This study also revealed a lower mean CAS among Asian patients (2.19 ± 1.61) compared with Caucasians (4.34 ± 1.55). 19 This difference may also be attributable to anatomical differences. The confined nature of orbital swelling in Asians, along with tighter orbital septa and unique tendon sheath structures, may obscure external signs of inflammation, leading to lower CAS scores. 18,20,21 Similarly, the mean exophthalmos in our cohort (20.04 ± 5.27 mm) was lower than the reported average for Caucasians (23.7 ± 3.4 mm). 22 Asians tend to have a rounder and shallow orbit, black people have a rectangular orbit, and Caucasians are in between. The anthropological difference may explain the wide variation of exophthalmos and base values among different population groups. 23 Consistent with this, average exophthalmos measurements in Caucasians are 16.55 ± 2.58 mm for males and 15.46 ± 2.33 mm for females, whereas average Hertel values in Chinese individuals are 15.0 ± 2.0 mm for males and 14.9 ± 2.0 mm for females. 22 These findings underscore the importance of contextualizing diagnostic criteria for TED based on population-specific characteristics. Western clinicians should interpret lower CAS scores and exophthalmos in Asian patients with caution, as these may underrepresent disease severity and risks such as DON.
This study provides a detailed analysis of variables associated with TED in a Chinese population. We observed that the severity of TED shows a clear progressive trend across five age groups, which is consistent with previous studies. 24 Male patients tend to present with more severe TED compared to females, with sight-threatening cases being significantly more prevalent among males (10.1% in males vs. 7.2% in females), aligning with existing studies. 25 –27 Additionally, the proportion of sight-threatening cases among smokers is significantly higher than among nonsmokers (12.8% in smokers vs. 7.3% in nonsmokers). The results of the mediation effect model suggest that smoking may mediate the age-related disease severity of TED. Previous research highlighted smoking as the strongest modifiable risk factor for GO. 28
The impact of I131 treatment on the incidence and severity of TED remains inconclusive. Some studies suggest that it may trigger exacerbations of the disease, 17,29,30 while others indicate no significant relationship. 31 We observed that patients who had previously received I131 treatment exhibit more active disease and greater severity of exophthalmos compared with patients who did not receive I131 treatment. This may be due to the use of I131 treatment after antithyroid drugs in clinical settings where hyperthyroidism is poorly controlled, 32 thus leading to more severe TED. I131 treatment may be associated with a higher incidence or exacerbation of TED. 33,34 These findings suggest that clinicians should exercise caution when planning I131 treatment for patients with TED.
This is the first study to describe the regional differences in TED manifestations between inland and coastal areas of China. We found a higher proportion of active-phase TED cases inland, while sight-threatening cases were more common in coastal regions. Given that both populations are predominantly Han Chinese, genetic factors likely play a minor role, and other factors may contribute to these differences. Some factors that may explain these regional differences may include: (a) environmental differences (as inland and coastal regions differ in temperature, humidity, and air pollution, all of which have been implicated in thyroid disorders, autoimmune eye diseases and other eye diseases), (b) differences in iodine content of water (which is significantly higher than in coastal areas compared with inland areas 35 ), (c) socioeconomic factors and related and work-life stress (which may vary in regions of China 36 ), and (d) differences in access to specialized healthcare.
IVGC therapy is the first-line treatment for moderate-to-severe and active patients with TED. 37 However, it is only effective in 30–50% cases. 38,39 Our nomogram examining IVGC responsiveness in Chinese patients revealed that lid aperture, CAS, and exophthalmos were the three most influential variables in the model.
Limitations
This retrospective study has inherent limitations, including potential data collection variations and incomplete records. Also, this study did not explore the underlying biological mechanisms driving clinical differences between patients from inland and coastal regions or those with and without I131 treatment. The absence of detailed treatment data also limits our ability to assess treatment efficacy across different risk factors. The analysis of IVGC responsiveness included only 126 patients, as data on this outcome were missing for most patients in the study. Furthermore, disease severity in TED may evolve over time, and our analysis did not account for the timing of peak severity, which may better reflect the relationship between risk factors and disease progression. Future research could adopt a prospective design, incorporating longitudinal data to evaluate treatment efficacy and disease trajectories. Expanding to international multicenter studies and investigating the mechanistic basis of risk factors would strengthen the scientific rigor and clinical applicability of these findings.
Conclusions
This study provides an overview of TED clinical characteristics and treatment responsiveness in a large Chinese cohort, emphasizing the need for tailored diagnostic and management strategies in Asian populations. It is the first to reveal the association between geographic factors and TED progression, while examining novel associations of variables with sight-threatening TED and IVGC therapy responsiveness. These findings offer innovative clinical insights and a valuable framework for future prospective research to inform personalized treatment planning.
Footnotes
Acknowledgments
Appreciation is extended to Hongyi Zhu, Yonglin Guo, Xuanjie Lin, Jialu Liu, Yifan Gao, all regional officers, and the research staff for their data collection efforts. The authors express their gratitude to Haixia Guan for her constructive advice on the whole research. The authors are grateful to Ziwei Du for her help with the preparation of figures in this article. The authors thank the editors and the reviewers for their useful feedback that has greatly improved this article.
Authors’ Contributions
H.Z., R.S.B., Z.Y., and C.L. conceived and designed the study. C.L., X.L., Y.R., and D.W. led the data collection efforts. C.L., X.L., R.L., X.S., and H.Z. organized on-site investigation projects across various provinces. X.L., Y.R., and S.Z. cleaned and validated the data, performed the statistical analysis with support from Z.Y. C.L., X.L., and Y.R. wrote the initial draft of the report. All authors contributed to the acquisition, analysis, or interpretation of data, revised the report, and approved the final version for submission.
Data Sharing Statement
Study components, including the protocol and statistical analysis methods, will be accessible to researchers upon approval from the Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China. Request petitions for data can be addressed to the corresponding author for consideration. Access to anonymized data may be granted after a review process.
Author Disclosure Statement
All the authors declare no competing interests related to the present work.
Funding Information
This work was supported by the National Natural Science Foundation of China (82388101, 82271122, and 82201166), National Key R&D Program of China (2024YFB4710200 and 2024YFB4710205), Science and Technology Commission of Shanghai Municipality (20DZ2270800), Shanghai Key Clinical Specialty, Shanghai Eye Disease Research Center (2022ZZ01003), Shanghai Municipal Commission of Health and Family Planning Project (2022XD006), Shanghai Jiao Tong University 2030 Initiative (2030-B23), Shanghai Three-Year Plan for the Inheritance and Innovative Development of Traditional Chinese Medicine (2-5-1), Hainan Provincial Key Research and Development Projects (ZDYF2024LCLH004), Shanghai Jiao Tong University Hainan Institute’s Independent Research Initiative (HRSJ-ZSZX-009). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the article; and decision to submit the article for publication.