Efficacy and safety of tripterygium glycosides for active moderate to severe Graves’ ophthalmopathy: a randomised, observer-masked, single-centre trial

Background: Tripterygium glycosides (TG) has been used to treat a spectrum of inflammatory and autoimmune diseases. Our preliminary studies have shown that TG is effective in the treatment of active Graves’ ophthalmopathy (GO). Objective: We aimed to compare the efficacy and tolerability of TG with intravenous methylprednisolone (iv.MP) in patients with active moderate-to-severe GO. Methods: This study was an observer-masked, single-centre, block-randomised trial. Patients with active moderate-to-severe GO were randomly assigned to receive iv.MP (500 mg once per week for 6 weeks followed by 250 mg per week for 6 weeks) or with TG (20 mg tablet three times per day for 24 weeks). The primary endpoints were the overall response rate and the patients’ quality of life at 12 and 24 weeks. Results: In this study, 161 patients were enrolled and randomised from 2015 to 2019. A total of 79 were randomly assigned to receive iv.MP and 82 to receive TG. A greater overall response rate was found in the TG group compared with the iv.MP group at week 24 (90.2% vs 68.4%, P = 0.000). Similarly, the patients’ quality of life of the TG group showed a significantly higher response than the iv.MP group at week 24 (89.02% vs 72.15%, P = 0.001). The TG therapy showed a better CAS response than the iv.MP (91.5% vs 70.9% improved, P < 0.05), and up to 91.2% of patients were inactive. Also, the TG group showed a significantly higher improved rate of diplopia, proptosis, visual acuity, soft tissue involved and the decrease of eye muscle motility than the iv.MP group at week 24. Significantly more patients in the iv.MP group than the TG group experienced adverse events. Conclusion: Compared with iv.MP treatment, TG therapy is more effective and safer for patients with active moderate to severe GO.


Introduction
Graves' orbitopathy (GO) is an inflammatory autoimmune disorder of the orbit which is associated with autoimmune thyroid disease, especially Graves' disease (1). The inflammatory phase of GO is characterized by T cells infiltration, often accompanied by B cells, mast cells and macrophages. European guidelines recommend intravenous methylprednisolone (iv.MP) as first-line treatment for active and severe Graves' orbitopathy; however, it is common for patients to have no response, or have a number of unfavourable side effects such as Cushing's syndrome, impaired liver function, gastrointestinal and cardiovascular side effects, or have relapsed after discontinuation of treatment, and progression to dysthyroid optic neuropathy in non-responders does not seem to be prevented by this approach (2,3,4,5). Therefore, there is still a need to identify a new non-steroidal immunosuppressive agent that directly target the pathogenic mechanisms of GO with higher efficacy and fewer side effects (6).
In traditional Chinese medicine, tripterygium glycosides (TG) is an effective component extracted from the root bark of Tripterygium wilfordii Hook F (TwHF) (also known as 'lei gong teng' in China). Functions like antiinflammation and immunosuppression of TG have been well proved, It has been widely used in China and other Asian countries for the treatment of various autoimmune and inflammatory diseases, such as rheumatoid arthritis (7,8), systemic lupus erythematosus (SLE) (9), Crohn's disease (10), idiopathic IgA nephropathy (11), psoriasis (12), and organ transplants (13). The immunosuppressive action of TG has been generally attributed to its suppression of T or B-lymphocyte functions, including T-cell or B-cell apoptosis induction, as well as inhibition of lymphocyte proliferation, DNA synthesis in alloreactive T lymphocytes, interleukin (IL)-2 and IL-2 receptor expression, interferon (IFN)-γ, TNF (TNF-α) production, and inflammation triggered by these cells (14,15,16,17,18,19,20,21). TG has been demonstrated to inhibit IL-2 production by inhibiting activation of the purine box regulator of the nuclear factor of activated T cells target DNA sequence in IL-2 enhancer (20). The activation of nuclear factor -kappa B (NF-κB), which is central to the pro-inflammatory cascade, has also been shown to be inhibited by TG (22,23). In another study, Luk et al. demonstrated that TG exerted a spectrum of immunomodulatory actions, including inhibition of mitogen-stimulated proliferation of lymphocytes in the allogenic mixed lymphocyte reaction (24). Moreover, TG can inhibit the maturation, antigen processing, and presentation of dendritic cells (25,26). Importantly, TG provides a strong immunosuppressive effect, strongly inhibited activated T-cells or B-cells, but with a weak action on resting cells and without causing serious damages to the normal immune system (14,15,16,17,18,19,20,21,22). These encouraging findings showed promise for TG in the treatment of several inflammatory autoimmune conditions in which activated leukocytes play a pivotal role in the inflammatory and autoimmune response.
TG was associated with good safety and tolerability. The major side effects of TG reported in the literature are gastrointestinal complaints, haematological disorders, skin rash, hepatotoxicity, a decrease in creatinine clearance in elderly patients, and dysfunction of the male and female reproductive systems, such as reversible sterility, dysmenorrhoea, or irregular menstruation (27,28,29). Our preliminary findings from the previous clinical study have shown an unexpectedly good therapeutic effect of TG observed in 15 patients after therapy (30), but the number of patients was small. This study aimed to compare the efficacy and tolerability of TG extract with intravenous methylprednisolone in patients with active moderate-to-severe GO.

Design overview
This randomized, controlled, observer-masked, open-label TG, 24-week trial was conducted in Jingling hospital. All eligible patients diagnosed as active moderate-to-severe GO from 2015 to 2019 in our hospital were included in this study. The trial was approved by Jingling hospital ethical committee and written informed consent was obtained from all patients. The study was funded by the Natural Science Foundation of Jiangsu Province of China (BK20180295). We aimed to determine whether therapy with TG, 60 mg/day, was statistically significantly better than iv.MP therapy, over 24 weeks in patients with active moderate-to-severe GO by using standard outcome measures.

Setting and participants
Eligible patients aged 18-70 years had established active GO (clinical activity score (CAS) of 3-7) (2), of moderateto-severe degree (moderate-to-severe active soft tissue involvement (according to the EUGOGO colour atlas evaluation), proptosis ≥21.6 mm (the upper limit of normal Chinese in our study was 18.6 mm) (31), eye muscle involvement with mono-ocular ductions in any direction of gaze of less than 30° or evident dysmotility, or diplopia (Gorman score of grade 1-3)). All eligible patients had euthyroidism for at least 8 weeks with antithyroid drugs or after thyroidectomy, or 6 months after radioiodine administration; with evidence of disease progression during the previous 8 weeks or lack of improvement in the prior 24 weeks, and normal heart, liver and kidney function. To be eligible, patients could not have received previous immunosuppressive treatment for GO, or received corticosteroids or other immunosuppressive agents within the past 12 weeks for any reason. Main exclusion criteria were as follows: evidence of dysthyroid optic neuropathy and impaired heart, acute or chronic viral hepatitis, any relevant malignancy, or chronic renal failure, contraindications to therapy with TG or iv.MP, prior orbital radiotherapy or decompression surgery, the recent improvement in the disease, or the patients who received any immunosuppressive therapy in the previous 12 weeks.
Patients were evaluated clinically and by laboratory measures at baseline, and every 4 weeks for a total of 24 weeks. Patients underwent clinical endocrinological assessment, biochemical testing (serum thyroid-related hormones were measured using an electrochemiluminescent immunoassay (Roche Diagnostics), thyrotrophin receptor antibody (TRAb) was measured using a commercially available electro-chemiluminescence assays based on the M22 MAB, with a cut off of 1.75 U/L (Roche Diagnostics GmbH)).
All patients underwent complete ophthalmic and endocrine assessment and filled out the Graves' ophthalmopathy quality of life questionnaire (GO-QOL) at baseline and 12 and 24 weeks after starting treatment. The overall ophthalmic assessment was performed by an ophthalmologist masked to the treatment received according to a recent European Group on Graves' Orbitopathy consensus statement (32,33). A complete ophthalmological assessment included Hertel exophthalmometer and a study of ductions by the Foerster-Goldman perimeter. The clinical activity and severity (judged on the NOSPECS scale) of GO were assessed in accordance with the EUGOGO recommendations. Proptosis, eyelid width, diplopia, intraocular pressure (IOP), visual acuity and CAS were recorded by using the EUGOGO case record form (32,33). The questionnaire consists of two subscales: one for visual function (eight questions referring to limitations attributable to decreased visual acuity, diplopia, or both) and one for appearance (eight questions referring to limitations in psychosocial functioning attributable to changes in appearance). The questions are scored as severely limited (1 point), a little limited (2 points), or not limited at all (3 points). The two raw scores (8-24 points) can be transformed to total scores (0-100) using the formula: total score = 100 (raw score − x)/2x, where x stands for the number of completed questions. For both scores, higher scores indicate better quality of life.

Randomization and interventions
We used a computer-generated, pseudo-random code (with random, permuted blocks) to assign patients to treatment groups. We assigned eligible patients at a 1:1 ratio to receive either orally TG extract (Gyzz32021007, Jiangsu Meitong Pharmaceutical Co., Ltd.), 20 mg three times daily for 24 weeks, or intravenous methylprednisolone (500 mg once per week for 6 weeks followed by 250 mg per week for 6 weeks). All iv.MP patients were treated with proton pump inhibitors for the prevention of gastric bleeding. In the event of gastrointestinal intolerance, the protocol allowed for a temporary dose reduction of 50%. All ophthalmologists and the statistician were masked to group assignment.

Clinical outcome measures
The objective primary endpoint was the overall response at the 12th and 24th weeks. The response was defined as at least four of the following outcome measures: (i) improvement in CAS by 2 or more points or disease inactivation (CAS ≤ 3); (ii) improvement in soft tissue involvement by one grade in any of the following: eyelid swelling, eyelid erythema, conjunctival redness or conjunctival oedema (according to the EUGOGO colour atlas evaluation, available at www.eugogo.eu); (iii) reduction in proptosis by at least 2 mm; (iv) improvement of at least 8° in eye muscle motility (disappearance or reduction in the severity of decreased eye movements, assessed with orthoptic measurements of mono-ocular ductions (in degrees) in four directions of gaze (perimeter arc)); (v) improvement in diplopia (disappearance or reduction in severity in the Gorman diplopia score); (vi) increase in visual acuity ≥2/10 (using the Snellen chart in decimals); (vii) reduction in lid width by at least 2 mm (measured with a ruler in the primary gaze position); (viii) reduction in intraocular pressure by at least 2 mmHg. Deterioration (recurrence or relapse relative to baseline) was defined as a change in two of the following outcome measures in at least one eye: (i) increase in CAS by at least two points; (ii) increase in soft tissue involvement by one grade; (iii) increase in proptosis by at least 2 mm; (iv) increase in the severity of eye muscle motility; (v) increase in diplopia (new onset or upgrade in degree); (vi) decrease in visual acuity ≥ 2/10; (vii) increase in lid width by at least 2 mm; (viii) increase in intraocular pressure by at least 2 mmHg; (ix) occurrence of dysthyroid optic neuropathy. Unchanged was defined as no change or changes smaller than those defined in any of the previously mentioned parameters. The unchanged or deteriorate was defined as nonresponse.
The subjective primary endpoint were changes in the patients' quality of life at the 12th and 24th weeks. Improvement or deterioration in the quality of life was defined as an increase or decrease of 6 points or more in total score on either of the two quality of life scales, respectively. No change was defined as a change of fewer than 6 points in either direction.
The second outcome was changes in CAS and other parameters including evaluation of soft tissue changes, measurement of proptosis, eye muscle motility, diplopia, visual acuity, lid width, intraocular pressure and the patients' quality of life, adverse events and retreatment.

Adverse effects
Safety assessments consisted of all patients documenting adverse events in their drug diaries in accordance with the standardised medical dictionary for regulatory affairs (MedDRA) (34), as recommended by the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) (35). Vital signs and safety laboratory measures, including a complete blood count and a chemistry profile (electrolyte, glucose and liver and kidney function tests), were recorded at each visit and body weight were assessed every 2 weeks, so as to monitor treatment-associated impaired fasting glucose (IFG)/impaired glucose tolerance (IGT)/diabetes mellitus (DM), impaired liver function, and weight gain.
Adverse events were assessed for any alternative cause while judging relatedness to intake of TG (i.e. events deemed side-effects if related). Adverse events were graded by severity according to the seriousness criteria defined in the ICH E6 guideline for clinical practice (36). Adverse events were followed up until a stable outcome could be documented or until the patient was lost to follow-up. Patients exiting the study because of adverse events or side-effects or worsening of GO were kept in the primary analysis.

Statistical analysis
Continuous variables were expressed as mean ± s.d.; categorical variables were expressed as frequency and percentage, as specified. Analysis by a Fisher exact test, Wilcoxon test, repeated-measures ANOVA (Bonferroni test), or Mann-Whitney test was applied, as appropriate, and performed using SPSS 19 (SPSS Inc) software. Statistical significance was defined as P < 0.05.

Patients
A total of 200 patients with active moderate to severe GO were eligible to participate in the study. However, 18 did not meet inclusion criteria, 2 withdrawn consent and 2 loss to follow-up. Of the eligible 178 patients enrolled, 91 were allocated to the TG group and 87 were allocated to the iv.MP group. During the study, 17 discontinued and leaving 82 patients in the TG group and 79 in the iv.MP group (Fig. 1). Baseline demographics and clinical characteristics were summarised in Table 1. There were no significant differences between the two groups. Most patients were women (71.4%), and the mean age ± s.d. was 39.9 ± 12.8 years. The rate of patients with Graves' hyperthyroidism was as high as 86.9%, and most of them were treated with methimazole. Discontinuation occurred because of withdrawn consent (n = 6), loss to follow-up (n = 10), and lack of efficacy (n = 1). Reasons for discontinuing the study were similar between groups (Fig. 1). During the treatment period, dysthyroid optic neuropathy was not registered in patients in the TG group, whereas it occurred early in five patients in the iv.MP group between weeks 8 and 12. The five cases were regarded as treatment failure.

Patient response
After 12 weeks of treatment, patients who received TG showed a significantly higher overall clinical response rate than those patients treated with iv.MP (69.5% vs 51.8%, P = 0.000). Moreover, in the 24th week, the overall clinical response rate of the TG group significantly increased to 90.2%, while the patients in the iv.MP group marginally increased to 68.4% (Fig. 2 and Table 2). Similarly, the patients' quality of life of the TG group showed a significantly higher response than the iv.MP group at week 24 (89.02% vs 72.15%, P = 0.001).Relapses occurred in none patient in the TG group and five patients in the iv.MP group (Fisher; P = 0.033), one at the 12th weeks, two at the 16th weeks and two at the 20th weeks. Because patients were euthyroid at the time of GO reactivation, the reactivation was not related to thyroid dysfunction. Overall, the ophthalmic improvement was noted at week 12 and 24 in both study groups ( Fig. 2 and Table 2). The results showed that there were more patients having improvements in the TG group than those in the iv.MP group at week 12 and 24. CAS decreased significantly more in the TG group at week 12 than in the iv.MP group (P = 0.000) ( Table 2). Significant differences from baseline and significantly larger improvements in the TG group compared with the iv.MP group were apparent at 12 weeks of therapy and persisted throughout the study. Although CAS responses were significantly improved in both groups at week 12 and 24, patients receiving TG showed a trend towards better CAS response than those receiving iv.MP, and this difference was significant (Fig. 3A, B and Table 2). Furthermore, the disease inactivation (CAS < 3) occurred in 91.2% patients of TG group vs 69.2% of iv.MP group (Fisher, P = 0.001) ( Fig.  3C and D). Similarly, soft tissue involvement improved significantly in both groups, but the TG group showed a marginally higher response than the iv.MP group at week 12 and 24 (Table 2).
Compared to the baseline, significant improvements in diplopia occurred in both groups at week 12 and 24. The TG group showed a significantly higher diplopia response than the iv.MP group, but there was no significant difference at the 12th but the 24th weeks ( Table 2). Similar findings were also found in the proptosis and visual acuity, which all improved significantly in both groups at week 12 and 24. Although the proptosis and visual acuity response in TG group were higher than the iv.MP group, there was a significant difference only at week 24 (Table 2).
In addition, compared to the baseline, the decrease of eye muscle motility was significantly improved in the TG group at week 12 and 24, there was a significant difference. And we observed a trend towards a greater response to the TG than the iv.MP both at week 12 and 24, this difference was statistically significant ( Table 2). Concerning the lid width and intraocular pressure, significant improvements were only noted in the TG group, and the TG group showed a higher response than the iv.MP group both at week 12 and 24, this difference was statistically significant ( Table 2).

Safety and tolerability
More patients in the iv.MP group experienced adverse events than the TG group, the difference was statistically significant. Adverse events occurred in 8 of 82 patients receiving TG (9.8%) and 22 of 79 of those patients receiving iv.MP (27.8%), without any suspected unexpected serious adverse events. Seven of the eight patients (87.5%) with adverse effects of TG had menstrual disorders. Other adverse events that occurred in the TG group included gastrointestinal events in 2 patients, impaired liver function in 2 patients and rash in 1 patient. No severe adverse effects were recorded in the TG group. Seven patients developed mild menstrual disorders included hypomenorrhoea, delayed menorrhoea and amenorrhoea during the treatment and recovered after the dose of TG decreased to one half (30 mg/day). Two patients showed mildly impaired liver function (liver enzymes little more than the upper limit) in the 11th and 20th weeks of treatment and recovered after the hepato-protective drug treatment.
In the iv.MP group, weight gain was the most common adverse events, followed by glucose intolerance and menstrual disorders. A total of 21.5% (17/79) patients receiving iv.MP gained body weight over 0.5 kg by the end of therapy. Cushingoid features were present in 53.2% (42/79) of patients treated with iv.MP. One serious adverse event was observed in a patient treated with iv.MP, who showed a major impairment in liver function (the liver enzymes was more than three times the upper limit of normal) but with no history of liver disease. Apart from this case, no other severe adverse effects were recorded in the iv.MP group. The prevalence of new infections was 1.27% with iv.MP group (bronchitis in 1 patient) and 0% with the TG group; this difference was statistically significant. No patients in both groups discontinued the trial because of drug-related toxic effects.

Discussion
In this randomized, observer-masked trial, the prespecified primary outcomes were positive, there are statistically significant differences between the two treatment groups both at week 12 and 24.
Our results indicate that patients with active moderate to severe GO can be effectively treated with TG. During the 24 weeks study, treatment with TG resulted in rapid improvement in clinical signs and symptoms of GO, including CAS, soft tissue involvement, proptosis, diplopia, eye muscle motility, visual acuity, lid width, intraocular pressure and the patients' quality of life. As expected, TG showed the greatest effectiveness on CAS and soft tissue involvement, this may suggest TG is more efficient in controlling local inflammation than iv.MP. It may be because TG not only has potent antiinflammatory and immunomodulatory effects but also inhibits the transcription of cyclooxygenase-2 (37), which may result in the reduced production of prostaglandin E2 at inflammatory sites and therefore have a direct analgesic effect. Moreover, TG only inhibited PGE2 production and downregulated expression of the cyclooxygenase (COX)-2 gene at the inflammatory site without interfering with the COX-1-regulated PGE2 production in the noninflammatory organs (37).
Patients who received TG showed a significantly higher overall clinical response rate than those patients treated with iv.MP, and the overall clinical response rate significantly increased from 71.2% at week 12 to 90.2% at week 24, indicating that a longer duration of TG therapy might lead to a better response. Another piece of evidence is that improvements in proptosis, diplopia and lid retraction were statistically significantly greater in the TG group than in the iv.MP group starting from 12 weeks of therapy, and this suggests that TG therapy is superior to iv.MP in terms of disease severity. The low response rate of diplopia and proptosis in patients treated with iv.MP was in agreement with the data of previous literatures (5,38,39,40). Furthermore, there was no one retreatment event in the TG group. Also, no patient in the TG group developed to dysthyroid optic neuropathy. Thus, we speculate that TG can prevent or delay the occurrence of dysthyroid optic neuropathy.
Importantly, TG was associated with good safety and tolerability. In our study, the most frequent complains in the TG group were menstrual disorders. However, all menstrual disorders recovered after the dose of TG decreased to one half. This effect of reversible menstrual disorders, which has also been reported in  other studies (41,42), may make this drug more attractive in the treatment of postmenopausal women. No serious adverse events during our trial, and no patient had to stop treatment because of treatment-related adverse effects. The TG extract at dosages up to 570 mg/day appeared to be safe, and doses >360 mg/day were associated with clinical benefits in the patients with rheumatoid arthritis (43). The frequency and severity of treatment-associated adverse effects are dose dependent (41,42,43). In our study, we chose a low dose of TG (60 mg/day), which based on previous work in autoimmunity (30,43). From our previous pilot study, we observed that even low doses of TG (30 mg/day) were effectively suppressing orbital inflammation (30). With a lower TG dose, patients are exposed to lower risks of potentially severe side effects, such as gastrointestinal events and irregular menstruation. In this study, we found significantly greater response rate and less severe toxicity for patients treated with TG. The results demonstrated that the treatment with a standardised extract from the peeled roots of TG administered (20 mg three times daily) within 24 weeks may be both effective and safe in the treatment for patients with active moderate to severe GO. P value a, baseline vs 12th week; b, baseline vs 24th week; c, improvement at 12th week between the two groups; d, improvement at 24th week between the two groups.

Figure 3
Improvement of CAS in the 12th weeks and 24th weeks in the two treatment group (A and B). Improvement was defined as decrease in CAS by 2 or more points or disease inactivation (CAS ≤ 3). Outcome of the activity in GO patients after TG or iv.MP treatment in the 12th weeks and 24th weeks (C and D).
In our study, cushingoid features were present in more than half of the patients treated with iv.MP. This may be because our study population is relatively young and nearly overweight. BMI ≥ 24 was defined as overweight in Chinese people. Although the patients in both groups have not reached the overweight standard, their BMI is very close to 24 (23.89 ± 3.13, 23.96 ± 4.61, respectively).
In conclusion, when compared with iv.MP therapy, TG might be a more effective and safer treatment in patients with active moderate to severe GO. This efficacy may be due to greater and more sustained suppression of local inflammation. TG could be considered a suitable treatment option for certain patient populations, specifically patients who have contraindications to other therapies and postmenopausal women. However, the sample size was relatively small and the study was singlecenter, only observer-masked are the main limitations of this trial. Further high-quality multicenter, double-blind, large-sample clinical trials are needed to evaluate the longterm effects of TG for the treatment of active moderate to severe GO patients.