Incidence and risk factors for Graves’ orbitopathy in patients who underwent anti-inflammatory and immunosuppressive treatment during medical treatment for Graves’ disease: investigation of 1,553 cases with newly diagnosed Graves’ disease and proposal of a predictive score

Abstract

Appropriate administration of anti-inflammatory and immunosuppressive treatment (AIIST) is important for patients with Graves’ orbitopathy (GO). This study aimed to clarify the incidence and risk factors for GO treated with AIIST and propose a predictive score, among newly diagnosed Graves’ disease (GD) patients in Japan. A total of 1,553 GD patients who were newly diagnosed during the year 2011 were investigated. AIIST included local and/or systemic glucocorticoid administration and retrobulbar irradiation. A multivariable Cox proportional hazards model was used to investigate the risk factors for GO underwent AIIST during medical treatment, including at diagnosis, of GD. Then, a GO score was created by summing each point assigned to risk factors based on their coefficient obtained in the Cox model. AIIST was administered to 107 patients (6.9%). The risk factors and hazard ratios for GO underwent AIIST were: age (per 10 years), 1.32 (95% confidence interval: 1.16–1.50), p < 0.0001; TSH binding inhibitory immunoglobulin (TBII) (per 10 IU/L), 1.33 (1.15–1.54), p = 0.0001; and thyroglobulin antibody (TgAb) negativity, 2.98 (1.96–4.59), p < 0.0001. The GO score, ranging from 0 to 8 points, showed moderate performance (area under the curve: 0.71, cut-off value: 5 points, sensitivity: 0.76, specificity: 0.59, positive predictive value: 0.12, negative predictive value: 0.97). AIIST was performed for patients with active manifestations of GO in 6.9% of newly diagnosed GD patients. The risk factors for GO underwent AIIST were higher age, higher TBII, and TgAb negativity. The GO score based on these factors may be useful in managing GO.

GRAVES’ ORBITOPATHY (GO) is an important complication of Graves’ disease (GD) that requires proper management, because it can have a negative impact on patients’ quality of life [1]. In recent years, effective new therapies have been emerging, such as teprotumumab [2, 3], whereas anti-inflammatory and immunosuppressive therapy (AIIST) is a long-established, but still useful, therapy that can reduce the severity of active GO. After the active phase, the effects of AIIST are diminished, and disease severity cannot be decreased [4, 5]. Therefore, appropriate AIIST should be administered during the active phase. The risk factors that have been reported for GO include smoking [6], high TSH binding inhibitory immunoglobulin (TBII) [7], high thyroid-stimulating antibody (TSAb) [8], low thyroglobulin antibody (TgAb) [9], low thyroid peroxidase antibody (TPOAb) [9, 10], elderly [11], and the clinical activity score (CAS) [7]. However, a risk factor analysis focused on AIIST for active findings has yet to be performed. In addition, there have yet to be any extensive studies of Japanese GD cases; therefore, the clinical features of GO in Japan are not well understood. Therefore, in conjunction with the Olympia Eye Hospital, which is a hospital that specializes in thyroid orbitopathy [12], a joint research project that examined a large number of first-visit untreated GD patients was performed. This study aimed to clarify the incidence and risk factors for cases that required AIIST for active manifestations, and to propose a predictive score. A recent survey reported that approximately 40% of GO patients are found during treatment for hyperthyroidism of GD, and that the difficulty in predicting onset is a very important concern in GO management [13]. We hope that this study will help endocrinologists, thyroidologists, and ophthalmologists specializing in GO, as well as general internists, to manage GO during the treatment of GD.

Materials and Methods

Patient selection and clinical records

This study involved 1,697 GD patients who were newly diagnosed between January 2011 and December 2011. Of these, 144 subjects who underwent first blood collection for GD at another medical institution were excluded, with a total of 1,553 patients subsequently included in the analysis. Their electronic medical records were retrospectively reviewed. The AIIST group consisted of patients who received AIIST for active findings of GO during medical treatment, including at diagnosis, of GD. Other patients were categorized as the no-AIIST group.

Institutional Review Board Approvals

This study was approved by the Ethics Committee of Ito hospital (approval number 291) and was conducted according to the Declaration of Helsinki and current legal regulations in Japan. The patients were given the opportunity to refuse to participate in the study via opt-out.

Evaluation and treatment for GO

For the active manifestations of GO, AIIST was started. At the start of AIIST, severity and activity were evaluated based on the NOSPECS classification [14], consensus statement of the European Group on Graves’ orbitopathy [15] and CAS (with a score of 3 or higher on a 7-point scale defined as active) [15, 16]. Activity was also evaluated by magnetic resonance imaging (MRI) of the orbit [17], even in patients with low CAS, where necessary. Since many GO patients with low CAS have highly active orbitopathy and require AIIST in Japan [18], the Guidance on Thyroid Eye Disease by the Japan Endocrine Society and the Japan Thyroid Association suggests that MRI should be used to evaluate the activity of the patient’s orbitopathy and to select treatment according to the condition [19]. AIIST included local and/or systemic glucocorticoid administration and retrobulbar irradiation [19]. The indication and method for applying each treatment at Olympia Eye Hospital were as follows: 1) subcutaneous triamcinolone acetonide (TA) injection for upper eyelid retraction and swelling with inflammation on MRI, i.e., high intensity of the levator palpebrae superioris muscle on T2-weighted fat-suppression images (Supplemental Fig. A) (eyelid injection) [20]; 2) sub-Tenon TA injection for single extraocular muscle hypertrophy with inflammation on MRI (Supplemental Fig. B) (sub-Tenon injection) [21]; 3) systemic glucocorticoid administration for moderate to severe active GO with hypertrophy of more than one extraocular muscle with inflammation on MRI (Supplemental Fig. C) (systemic administration), but when intravenous glucocorticoid administration was contraindicated due to old age, occult hepatitis, etc., glucocorticoid was given orally; and 4) retrobulbar irradiation for moderate to severe active GO with extraocular muscle hypertrophy with inflammation, when systemic glucocorticoid administration was not effective in reducing inflammation or was not indicated.

Assays

Serum concentrations and antibody titers of free T₃ (FT₃), free T₄ (FT₄), TSH, TBII, TPOAb, and TgAb were determined using a commercial rapid electrochemiluminescence immunoassay kit [ECLusys FT₃, FT₄, TSH, TRAb (third generation), TPOAb, and TgAb, Roche Diagnostics GmbH, Basel, Switzerland]. The reference ranges of the assays were 2.2–4.3 pg/mL, 0.80–1.60 ng/dL, 0.20–4.50 μU/mL, ≤2.0 IU/L, ≤28 U/mL, and ≤40 U/mL, respectively.

Statistical analysis

Non-normally distributed data were expressed as medians and ranges and were analyzed by the Wilcoxon test, whereas Student’s t-test was used for continuous variables with a normal distribution. The significance of differences in frequencies was calculated using the χ² test or Fisher’s exact test. The cumulative rate of AIIST for GO was calculated by the Kaplan-Meier method. Data on patients who changed treatment for GD, from medical treatment to radioiodine therapy or thyroidectomy, were censored. The risk factors for GO during medical treatment for GD were assessed using a multivariable Cox proportional hazards model with baseline variables at the time of GD diagnosis. GO predictive scores were created by assigning scores with reference to parameter estimates in a multivariable Cox model [22]. The cutoff value was determined as the Youden Index that maximizes sensitivity – (1 – specificity). A Cochran-Armitage trend test was performed for the frequencies of GO treatment based on the predictive score and risk classification. A p-value <0.05 was regarded as significant. JMP version 14.0.0 (SAS Institute Inc., Cary, NC, USA) was used for all statistical analyses.

Results

Patients’ characteristics

The baseline characteristics of the 1,553 patients with newly diagnosed GD are listed in Table 1. Their median age was 38 years (range 8–84 years), and 1,250 patients (80%) were females (Table 1). The initial treatment for GD was medical therapy for all 1,553 patients. Radioiodine therapy was given to 227 patients, and thyroidectomy was performed for 24 patients after initial medical treatment for GD (Fig. 1).

Table 1

Baseline Characteristics at diagnosis of GD

Variables	ALL	AIIST group	no-AIIST group	p value
No. of patients	1,553	107	1,446	—
Age (median [IQR]) (y)	38 [30–48]	44 [34–56]	38 [29–47]	0.002
Sex (male/female) [number (%)]	303/1,250 (20/ 80)	25/82 (23/77)	278/1,168 (19/81)	0.31
Smoker [number (%)]	386 (25)	25 (23)	361 (25)	0.42
Laboratory tests
FT3 (median [IQR]) (pg/mL)	15.6 [9.0–24.0]	15.3 [7.2–26.5]	15.7 [9.0–23.9]	0.73
FT4 (median [IQR]) (ng/dL)	4.6 [3.0–7.2]	4.5 [2.5–7.7]	4.7 [3.0–7.1]	0.45
TSH (median [IQR]) (μU/mL)	0.01 [0.01–0.01]	0.01 [0.01–0.01]	0.01 [0.01–0.01]	0.45
TBII (median [IQR]) (IU/L)	10.9 [5.5–22.9]	15.5 [7.7–32.0]	10.5 [5.3–22.4]	0.0007
TgAb (median [IQR]) (U/mL)	108.0 [15.1–443.3]	15.2 [10.4–178.7]	120.7 [15.9–453.8]	<0.0001
TPOAb (median [IQR]) (U/mL)	90.3 [11.4–327.4]	27.0 [9.2–156.5]	95.2 [11.8–339.6]	0.0008
TgAb negative [number (%)]	620 (40)	71 (67)	549 (38)	<0.0001
TPOAb negative [number (%)]	586 (38)	54 (51)	532 (37)	0.004
TgAb and TPOAb negative [number (%)]	360 (23)	46 (43)	314 (22)	<0.0001

GD, Graves’ disease; AIIST, anti-inflammatory and immunosuppressive treatment; IQR, interquartile range; FT₃, free T₃; FT₄, free T₄; TSH, thyroid stimulating hormone; TBII, TSH binding inhibitory immunoglobulin; TgAb, thyroglobulin antibody; TPOAb, thyroid peroxidase antibody

Fig. 1

Flow chart for the cases

GD, Graves’ disease; GO, Graves’ orbitopathy; RAI, radioactive iodine therapy; AIIST, anti-inflammatory and immunosuppressive treatment

Incidence and treatment for GO

Of the 1,553 patients with newly diagnosed GD, 107 (6.9%) received AIIST for GO (AIIST group) (Fig. 1). The median interval between the first visit and AIIST was 5 months (range 0.1–44 months) (Table 2). The cumulative rate of AIIST for GO during medical treatment for GD was 5.9% at 1 year, 7.0% at 2 years, 7.4% at 3 years, and 7.6% at 4 years, after which the rate reached a plateau. Fig. 2 shows the severity and activity at the most severe and the most active times in each patient. The severity of GO in those who underwent AIIST was mild in 29 patients (1.9%, 29/1,553), moderate to severe in 76 cases (4.9%, 76/1,553), and the most severe, that is, sight-threatening was observed in only 2 cases (0.1%, 2/1,553) (Fig. 2). As AIIST, 103 patients were given glucocorticoids, and 26 patients received irradiation therapy. Of these patients, 44 received combination therapy, and 63 patients were treated with monotherapy. Of the 77 patients who underwent eyelid injection, 29 were administered monotherapy despite moderate to severe GO. The reason for this was because of moderate to severe findings only in the eyelids (n = 7) or because of moderate to severe proptosis, but with only active manifestations in the eyelids to be treated (n = 22). As for non-anti-inflammatory GO treatment, 13 cases underwent surgery, and the type of surgery was strabismus in 8 cases, eye lid surgery in 1 case, and orbital decompression in 4 cases.

Table 2

Interval and frequency of AIIST for GO

Type of treatment	No.	Interval (month) [median (range)]	Number [median (range)]
Glucocorticoid	103*
Eyelid injection	77*	4 (0.1–40)	2 (1–9)
Sub-Tenon’s injection	38*	8 (0.3–44)	2 (1–6)
Systemic administration	23*	4 (0.6–34)	1 (1–2)
Irradiation	26*	4 (0.4–18)	1#
Total	107	5 (0.1–44)	2 (1–9)

The interval was between the first visit and the GO treatment; GO, Graves’ orbitopathy; AIIST, anti-inflammatory and immunosuppressive treatment; *, There is an overlap. Some patients received more than one treatment. ; ^#, All cases was received external beam retrobulbar irradiation of 15Gy divided 10 times to the orbit lesion only one time.

Fig. 2

Severity and activity of GO at the AIIST

The degree of disability, severity and CAS was calculated based on the findings at the time of the most severe symptoms. ♯, Degree of disability according to the NOSPECS classification [14]; ¶, Severity according to the European Group on Graves’ Orbitopathy [15]. The median measurement of proptosis was 18 mm, and the range was 12–24 mm. Because of ethnic variations, proptosis was determined based on the classification of the Japan Thyroid Association, which uses the following categories: absent, <15 mm; mild, ≥15 and <18 mm; moderate, ≥18 and <21 mm; and severe, ≥21 mm. GO, Graves’ orbitopathy; AIIST, anti-inflammatory and immunosuppressive treatment.

Comparisons between the AIIST group and the no-AIIST group

Comparing the AIIST group and the no-AIIST group, the age of patients with AIIST was significantly older, and the TBII value was significantly higher. The titers of both TgAb and TPOAb were significantly lower, and the negative rate of TgAb and TPOAb was significantly higher in the AIIST group (Table 1). The rate of smokers was not different between the groups (Table 1). However, compared to nonsmokers, smokers had higher TBII levels [median 13.2 (interquartile range, 5.9–28.2) vs. 10.2 (5.4–22.3) IU/L, p = 0.01], lower TgAb levels [76.1 (15.1–407.8) vs. 115.7 (14.8–457.0) IU/L, p = 0.11], a higher TgAb negative rate [44.3 vs. 38.5%, p = 0.04], higher TPOAb levels [160.1 (24.7–403.4) vs. 68.9 (10.2–297.6) IU/L, p < 0.0001], and a lower TPOAb negative rate [25.9 vs. 41.7%, p < 0.0001].

Risk factor analysis and GO score

The risk factors for GO during medical treatment for GD were analyzed using Cox proportional hazards models based on the variables at the time of GD diagnosis (Table 3). On univariable analysis, higher age, higher TBII, lower TPOAb, negative TgAb, negative TPOAb, and both negative TgAb and TPOAb were significant risk factors for the requirement of AIIST for GO. On multivariable Cox proportional hazards analysis, the significant risk factors were older age, higher TBII, and negative TgAb. Based on the coefficient of each factor obtained in the multivariable model, scores were assigned to each factor, as shown in Table 4. The predictive score for the requirement for AIIST for GO (GO score) was then calculated by summing these points, ranging from 0 to 8 points (Table 4). On receiver-operating characteristic curve analysis, the area under the curve (AUC) was 0.71 (Fig. 3A). At the cut-off value of 5 points, sensitivity was 0.76, specificity was 0.59, positive predictive value (PPV) was 0.12, and negative predictive value (NPV) was 0.97. Based on the GO score, risk groups were classified as follows: low risk for up to 4 points; moderate risk for 5–6 points; and high risk for 7–8 points. The frequency of AIIST for each GO risk classification increased with the risk grade (Cochran-Armitage trend test p < 0.0001) (Fig. 3B).

Table 3

Risk factors for GO underwent AIIST

Variables	Univariable analysis	p value	Multivariable analysis	p value
Age (/10 y)	1.27 [1.12–1.46]	0.002	1.32 [1.16–1.50]	<0.0001
Sex (male to female)	1.31 [0.83–2.03]	0.23	—
Smoker	0.96 [0.60–1.48]	0.77	—
FT3 (/pg/mL)	1.00 [0.98–1.02]	0.73	—
FT4 (/ng/dL)	0.96 [0.87–1.04]	0.32	—
TBII (/10 IU/L)	1.23 [1.07–1.41]	0.003	1.33 [1.15–1.54]	0.0001
TgAb (/100 U/mL)	0.96 [0.90–0.99]	0.03	—
TPOAb (/100 U/mL)	0.82 [0.72–0.91]	0.007	—
TgAb negative	3.16 [2.12–4.79]	<0.0001	2.98 [1.96–4.59]	<0.0001
TPOAb negative	1.73 [1.18–2.54]	0.005	1.50 [0.99–2.28]	0.06
TgAb and TPOAb negative	2.63 [1.79–3.86]	<0.0001	—

GO, Graves’ orbitopathy; AIIST, anti-inflammatory and immunosuppressive treatment; FT₃, free T₃; FT₄, free T₄; TSH, thyroid stimulating hormone; TBII, TSH binding inhibitory immunoglobulin; TgAb, thyroglobulin antibody; TPOAb, thyroid peroxidase antibody

Table 4

GO Score

	Variables	points
TBII (IU/L)	less than 5	0
	5 or more and less than 10	2
	10 or more and less than 20	3
	20 or more	4
Age	under 30 years old	0
	30 to 49 years old	1
	50 years old or older	2
TgAb	positive	0
	negative	2
Total		0–8

GO, Graves’ orbitopathy; TBII, TSH binding inhibitory immunoglobulin; TgAb, thyroglobulin antibody

Fig. 3

Verification of GO score. A) Receiver operating characteristic curve for GO score and AIIST for GO

B) AIIST ratio for each GO risk group. GO, Graves’ orbitopathy; AIIST, anti-inflammatory and immunosuppressive treatment.

AUCs of the GO score were 0.72 (cutoff value: 5) for patients who started AIIST in the first year of GD treatment (n = 85), 0.65 (cutoff value: 5) for patients who started AIIST in the second year of GD treatment (n = 15), and 0.69 (cutoff value: 7) for those who started in the third year or later (n = 7). GO scores for local injection, with median values of 6 [interquartile range, 5–7], were not significantly different from the scores for systemic glucocorticoid (5 [5–7], p = 0.79) and irradiation (6 [4–7], p = 0.87). The no-AIIST group (n = 1,446) included the following after ophthalmological examination including MRI: A) mildly active GO that resolved spontaneously and no AIIST was performed (no-AIIST-A, n = 10); B) GO in the inactive phase (no-AIIST-B, n = 353); C) no GO (no-AIIST-C, n = 86); and also included D) GD patients with no need for a detailed ophthalmological evaluation and no ophthalmological examination performed (no-AIIST-D, n = 997). About half of the patients in the no-AIIST-B and C groups underwent ophthalmological examination as part of the evaluation before radioiodine therapy for GD. GO scores in the AIIST group (median 6 [interquartile range 5–7] ) were not significantly different compared to the GO scores in the no-AIIST-A group (5 [3–6] ) (p = 0.15), but they were significantly higher compared to the GO scores in the no-AIIST-B, C, and D groups; 5 [3–6] (p < 0.001), 4 [3–5] (p < 0.001), and 4 [3–5] (p < 0.001), respectively. Although GO scores were slightly higher in the no-AIIST-B group than in the no-AIIST-D group (p < 0.0001), GO scores were not significantly different between the no-AIIST-C and D groups (p = 0.18).

Discussion

This study analyzed the incidence, risk factors, and predictive scores of GO underwent AIIST in newly diagnosed GD patients. GO was treated with AIIST in 6.9% of newly diagnosed GD patients. Consistent with the previously reported risk factors for GO, this study demonstrated that higher age, higher TBII, and negative TgAb were associated with treatment for GO. The GO score from scoring of these three factors had moderate performance for GO requiring AIIST (AUC: 0.71).

GO remains a challenging disease for clinicians to evaluate and manage. A recent survey reported that approximately 40% of GO patients are found during the management of hyperthyroidism of GD [13]. Therefore, clinicians treating GD, such as endocrinologists, thyroidologists, and general internists, play an important role in the diagnosis, initial management, and selection of patients requiring referral to an ophthalmologist specializing in GO [23]. In addition, the difficulty in predicting onset has been reported as a very important concern regarding GO [13]. It is significant for all clinicians treating GD that predictive scores be established to help manage GO.

In a multicenter, prospective, observational study by Wiersinga et al., 348 patients were analyzed for new-onset and exacerbation of GO during treatment for newly diagnosed GD, and a predictive score called PREDIGO was proposed [7]. For PREDIGO, the independent baseline GO predictors that were assigned included the CAS, TBII levels, duration of hyperthyroidism, and current smoking. The model predicts the occurrence of GO when a patient’s score exceeds 6 of 15 in the scoring system; the AUC was 0.71, and the NPV was high, at 0.91. Recently, Ahn et al. presented a predictor for the occurrence and exacerbation of GO in a retrospective analysis of a large number of patients (1,074 patients) with new-onset GD at a single institution [24]. In their study, the predictors for the occurrence of GO were female sex, smoking history, younger age, family history of GD, high cholesterol level, and high TBII level, and based on these 6 factors, a nomogram predicting the occurrence of GO was developed with an AUC of 0.71. Factors for exacerbation of GO were age (>45 years) and CAS. The current study was a retrospective analysis, included a large number of new-onset GD (1,553 patients), and defined GO requiring AIIST as an event, differing from the two aforementioned prediction scores that defined onset and exacerbation of GO as events. The GO score proposed in the current study also showed a high NPV (0.97), comparable to that of PREDIGO, suggesting that the prediction of patients not requiring AIIST would be more feasible. Interestingly, the AUCs of these prediction scores, including the two aforementioned previous studies and ours, were all 0.71, which is only moderately accurate and indicates limited predictive power. In the current analysis, smoking was not identified as a risk factor. This could be due to a change in smoking habits after the diagnosis of GD, but this was unclear, because information on changes in smoking habits could not be obtained from the medical records. For more precise prediction, incorporating modifiable dynamic factors such as smoking, hypercholesterolemia [24], and TBII levels, in addition to unmodified GO risks such as age and sex, into the analysis is a challenge for the future.

It has been reported that as much as 50% of cases of GD are associated with GO [25], whereas approximately 60% occur spontaneously [26] and 70% of patients with GO do not require treatment [27]. In addition, recent reports of decreased activity and severity of GO [28], and racial differences, such as the reportedly milder GO in Asians [29] may explain the lower frequency of patients actually requiring treatment in the present study, i.e., only 6.9% of newly diagnosed GD patients received AIIST. Local glucocorticoid injections into the eyelid were frequently used in the present study. This is consistent with Japanese guidance that local injection of triamcinolone should be considered when eyelid swelling and/or upper eyelid retraction are present, and MRI shows inflammation in the eyelid adipose tissue and/or the levator palpebrae superioris muscle [19]. Future prospective studies are needed to determine whether these local treatments prevent the progression of GO.

Newer treatments, such as teprotumumab [2, 3], an IGF-I receptor antibody approved in the United States in 2020, have been introduced and reported to be effective. In North America, where clinical use of teprotumumab began, 37% of respondents to the survey conducted by the American Thyroid Association among its members already recognized the drug as a first-line treatment for moderate to severe active orbitopathy [13]. In contrast, intravenous glucocorticoids were preferred in Europe and other regions where teprotumumab was not approved at the time of the survey (73% and 42%, respectively) [13]. Although glucocorticoid therapy has been the most proven effective treatment, there are concerns regarding adverse events and the presence of non-responder or recurrent cases. In order to achieve optimal treatment for GO, it is necessary to continuously discuss strategies, including the selection of novel agents and local glucocorticoid therapy, taking into account individual pathophysiology, treatment tolerability, and cost-effectiveness, etc.

GD and Hashimoto’s disease share many similarities in genetic background [30-32]. Clinically, although the presence of TgAb and/or TPOAb is characteristic of Hashimoto’s thyroiditis, both antibodies are also shared in the majority of GD patients [33, 34]. The clinical implications of TgAb and/or TPOAb-positive GD remain unclear, and there are discrepancies in the results. Concerning GO, Khoo et al. reported that TPOAb-negative rates were significantly higher in patients with GO than in those without GO and were an independent predictor of GO [9]. Goh et al. also reported lower TPOAb and TgAb levels in ophthalmic-dominant as opposed to thyroid-dominant patients [10]. Similar to these previous studies, in the present analysis, both TgAb and TPOAb levels were low in the AIIST group, suggesting that differences in antibody patterns are associated with a predisposition to GO.

There were several limitations to the current study. First, the current results were determined based on retrospectively collected data from a thyroid specialty hospital and a thyroid ophthalmology specialty hospital. Further validation in a multicenter, prospective study is warranted. Second, the analysis did not include TSAb levels. The TSAb level reflects the stimulatory activity of TBII on TSH receptors and correlates well with GO [35, 36]. New TSAb assays with excellent rapidity and quantification [8, 37] might be more useful as a prognostic factor. Finally, future work to be undertaken includes creation of a score to predict the treatment response of GO. Glucocorticoids have been the gold standard GO treatment, but in recent years, new molecular-targeted agents such as teprotumumab have begun to be used clinically [2, 3], and promising antibodies K1-70 specifically inhibiting the TSH receptor are being developed [38, 39]. Prediction of treatment response for each treatment is desirable to enable avoidance of ineffective treatment and appropriate selection of the various therapies, such as by, for example, assessing fibrosis on MRI to predict response to glucocorticoid administration [40].

In conclusion, AIIST was performed for patients with active manifestations of GO in 6.9% of newly diagnosed GD patients. The risk factors for GO treatment were higher age, higher TBII, and TgAb negativity. The GO score based on these factors may be useful in managing GO patients.

Acknowledgments

The authors would like to thank medical secretary Nobuko Imai for data collection.

Author Disclosure Statement

Natsuko Watanabe and Kiminori Sugino are members of Endocrine Journal’s Editorial Board. Other authors declare no conflicts of interest in association with this manuscript.

Supplemental Fig.

A: high intensity of levator palpebrae superioris muscle on T2-weighted fat-suppression images (yellow arrow).

B: high intensity of inferior rectus muscle in T2-weighted fat-suppression images (yellow arrow).

C: high intensity of inferior, medial, superior rectus muscles in T2-weighted fat-suppression images (yellow arrow).

References

• 1   Kahaly  GJ,  Petrak  F,  Hardt  J,  Pitz  S,  Egle  UT (2005) Psychosocial morbidity of Graves’ orbitopathy. Clin Endocrinol (Oxf) 63: 395–402.
• 2   Smith  TJ,  Kahaly  GJ,  Ezra  DG,  Fleming  JC,  Dailey  RA, et al. (2017) Teprotumumab for thyroid-associated ophthalmopathy. N Engl J Med 376: 1748–1761.
• 3   Douglas  RS,  Kahaly  GJ,  Patel  A,  Sile  S,  Thompson  EHZ, et al. (2020) Teprotumumab for the treatment of active thyroid eye disease. N Engl J Med 382: 341–352.
• 4   Rundle  FF,  Wilson  CW (1945) Development and course of exophthalmos and ophthalmoplegia in Graves’ disease with special reference to the effect of thyroidectomy. Clin Sci 5: 177–194.
• 5   Rundle  FF (1957) Management of exophthalmos and related ocular changes in Graves’ disease. Metabolism 6: 36–48.
• 6   Cawood  TJ,  Moriarty  P,  O’Farrelly  C,  O’Shea  D (2007) Smoking and thyroid-associated ophthalmopathy: a novel explanation of the biological link. J Clin Endocrinol Metab 92: 59–64.
• 7   Wiersinga  W,  Zarkovic  M,  Bartalena  L,  Donati  S,  Perros  P, et al. (2018) Predictive score for the development or progression of Graves’ orbitopathy in patients with newly diagnosed Graves’ hyperthyroidism. Eur J Endocrinol 178: 635–643.
• 8   Lytton  SD,  Ponto  KA,  Kanitz  M,  Matheis  N,  Kohn  LD, et al. (2010) A novel thyroid stimulating immunoglobulin bioassay is a functional indicator of activity and severity of Graves’ orbitopathy. J Clin Endocrinol Metab 95: 2123–2131.
• 9   Khoo  DH,  Ho  SC,  Seah  LL,  Fong  KS,  Tai  ES, et al. (1999) The combination of absent thyroid peroxidase antibodies and high thyroid-stimulating immunoglobulin levels in Graves’ disease identifies a group at markedly increased risk of ophthalmopathy. Thyroid 9: 1175–1180.
• 10   Goh  SY,  Ho  SC,  Seah  LL,  Fong  KS,  Khoo  DH (2004) Thyroid autoantibody profiles in ophthalmic dominant and thyroid dominant Graves’ disease differ and suggest ophthalmopathy is a multiantigenic disease. Clin Endocrinol (Oxf) 60: 600–607.
• 11   Laurberg  P,  Berman  DC,  Bulow Pedersen  I,  Andersen  S,  Carle  A (2012) Incidence and clinical presentation of moderate to severe graves’ orbitopathy in a Danish population before and after iodine fortification of salt. J Clin Endocrinol Metab 97: 2325–2332.
• 12   Watanabe  N,  Noh  JY,  Kozaki  A,  Iwaku  K,  Sekiya  K, et al. (2015) Radioiodine-associated exacerbation of graves’ orbitopathy in the Japanese population: randomized prospective study. J Clin Endocrinol Metab 100: 2700–2708.
• 13   Brito  JP,  Nagy  EV,  Singh Ospina  N,  Zˇarković  M,  Dosiou  C, et al. (2022) A survey on the management of thyroid eye disease among American and European thyroid association members. Thyroid 32: 1535–1546.
• 14   Werner  SC (1977) Modification of the classification of the eye changes of Graves’ disease. Am J Ophthalmol 83: 725–727.
• 15   Bartalena  L,  Baldeschi  L,  Dickinson  A,  Eckstein  A,  Kendall-Taylor  P, et al. (2008) Consensus statement of the European Group on Graves’ orbitopathy (EUGOGO) on management of GO. Eur J Endocrinol 158: 273–285.
• 16   Mourits  MP,  Koornneef  L,  Wiersinga  WM,  Prummel  MF,  Berghout  A, et al. (1989) Clinical criteria for the assessment of disease activity in Graves’ ophthalmopathy: a novel approach. Br J Ophthalmol 73: 639–644.
• 17   Yokoyama  N,  Nagataki  S,  Uetani  M,  Ashizawa  K,  Eguchi  K (2002) Role of magnetic resonance imaging in the assessment of disease activity in thyroid-associated ophthalmopathy. Thyroid 12: 223–227.
• 18   Tachibana  S,  Murakami  T,  Noguchi  H,  Noguchi  Y,  Nakashima  A, et al. (2010) Orbital magnetic resonance imaging combined with clinical activity score can improve the sensitivity of detection of disease activity and prediction of response to immunosuppressive therapy for Graves’ ophthalmopathy. Endocr J 57: 853–861.
• 19  (2020) Treatment guide for thyroid-associated ophthalmopathy. The Japan Thyroid Association, The Japan Endocrine Society (eds). Medical Review Co., Ltd., Tokyo, Japan (In Japanese). http://www.j-endo.jp/uploads/files/edu/koujyousengansyo_digest.pdf accessed on July 27, 2023.
• 20   Kozaki  A,  Nakamura  H,  Inoue  T (2018) Clinical efficacy of transcutaneous triamcinolone acetonide injection for upper eyelid retraction and swelling in patients with thyroid eye disease. Int Med Case Rep J 11: 325–331.
• 21   Ito  G,  Kozaki  A,  Inoue  R,  Funaki  C,  Mizuno  K, et al. (2013) Effect of sub-Tenon’s triamcinolone acetonide injection for hypertrophic inferior rectus muscle in thyroid ophthalmopathy. Rinsho Ganka (Jpn J Clin Ophthalmol) 67: 1367–1371 (In Japanese).
• 22   Sullivan  LM,  Massaro  JM,  D’Agostino RB  Sr (2004) Presentation of multivariate data for clinical use: The Framingham Study risk score functions. Stat Med 23: 1631–1660.
• 23   Burch  HB,  Perros  P,  Bednarczuk  T,  Cooper  DS,  Dolman  PJ, et al. (2022) Management of thyroid eye disease: a consensus statement by the american thyroid association and the european thyroid association. Thyroid 32: 1439–1470.
• 24   Ahn  HY,  Lee  J,  Lee  JK (2022) Predictive model for Graves’ ophthalmopathy in patients with new-onset graves’ disease. Thyroid 32: 1559–1567.
• 25   Lazarus  JH (2012) Epidemiology of Graves’ orbitopathy (GO) and relationship with thyroid disease. Best Pract Res Clin Endocrinol Metab 26: 273–279.
• 26   Perros  P,  Crombie  AL,  Kendall-Taylor  P (1995) Natural history of thyroid associated ophthalmopathy. Clin Endocrinol (Oxf) 42: 45–50.
• 27   Bartley  GB,  Fatourechi  V,  Kadrmas  EF,  Jacobsen  SJ,  Ilstrup  DM, et al. (1996) The treatment of Graves’ ophthalmopathy in an incidence cohort. Am J Ophthalmol 121: 200–206.
• 28   Perros  P,  Zarkovic  M,  Azzolini  C,  Ayvaz  G,  Baldeschi  L, et al. (2015) PREGO (presentation of Graves’ orbitopathy) study: changes in referral patterns to European Group On Graves’ Orbitopathy (EUGOGO) centres over the period from 2000 to 2012. Br J Ophthalmol 99: 1531–1535.
• 29   Tellez  M,  Cooper  J,  Edmonds  C (1992) Graves’ ophthalmopathy in relation to cigarette smoking and ethnic origin. Clin Endocrinol (Oxf) 36: 291–294.
• 30   Tomer  Y (2014) Mechanisms of autoimmune thyroid diseases: from genetics to epigenetics. Annu Rev Pathol 9: 147–156.
• 31   Davies  TF,  Latif  R,  Yin  X (2009) Inheriting autoimmune thyroid disease. Endocr Pract 15: 63–66.
• 32   Ueda  S,  Oryoji  D,  Yamamoto  K,  Noh  JY,  Okamura  K, et al. (2014) Identification of independent susceptible and protective HLA alleles in Japanese autoimmune thyroid disease and their epistasis. J Clin Endocrinol Metab 99: E379–E383.
• 33   Schott  M,  Eckstein  A,  Willenberg  HS,  Nguyen  TB,  Morgenthaler  NG, et al. (2007) Improved prediction of relapse of Graves’ thyrotoxicosis by combined determination of TSH receptor and thyroperoxidase antibodies. Horm Metab Res 39: 56–61.
• 34   McLachlan  SM,  Feldt-Rasmussen  U,  Young  ET,  Middleton  SL,  Dlichert-Toft  M, et al. (1987) IgG subclass distribution of thyroid autoantibodies: a ‘fingerprint’ of an individual’s response to thyroglobulin and thyroid microsomal antigen. Clin Endocrinol (Oxf) 26: 335–346.
• 35   Diana  T,  Brown  RS,  Bossowski  A,  Segni  M,  Niedziela  M, et al. (2014) Clinical relevance of thyroid-stimulating autoantibodies in pediatric graves’ disease-a multicenter study. J Clin Endocrinol Metab 99: 1648–1655.
• 36   Noh  JY,  Hamada  N,  Inoue  Y,  Abe  Y,  Ito  K, et al. (2000) Thyroid-stimulating antibody is related to Graves’ ophthalmopathy, but thyrotropin-binding inhibitor immunoglobulin is related to hyperthyroidism in patients with Graves’ disease. Thyroid 10: 809–813.
• 37   Hiromatsu  Y,  Eguchi  H,  Matsuo  Y,  Kato  T,  Tani  J, et al. (2020) Role of a new bioassay for thyroid-stimulating antibodies (aequorin TSAb) in Graves’ ophthalmopathy. Endocr J 67: 347–352.
• 38   Ryder  M,  Wentworth  M,  Algeciras-Schimnich  A,  Morris  JC,  Garrity  J, et al. (2021) Blocking the thyrotropin receptor with K1-70 in a patient with follicular thyroid cancer, Graves’ disease, and Graves’ ophthalmopathy. Thyroid 31: 1597–1602.
• 39   Furmaniak  J,  Sanders  J,  Sanders  P,  Li  Y,  Smith  BR (2022) TSH receptor specific monoclonal autoantibody K1-70TM targeting of the TSH receptor in subjects with Graves’ disease and Graves’ orbitopathy-Results from a phase I clinical trial. Clin Endocrinol (Oxf) 96: 878–887.
• 40   Matsuzawa  K,  Izawa  S,  Kato  A,  Fukaya  K,  Matsumoto  K, et al. (2020) Low signal intensities of MRI T1 mapping predict refractory diplopia in Graves’ ophthalmopathy. Clin Endocrinol (Oxf) 92: 536–544.