How can we prevent disease relapse in Graves’ orbitopathy after immunosuppressive treatment?

KEYWORDS:

• Graves’ orbitopathy
• Radioiodine
• Glucocorticoids
• Disease relapses
• Rehabilitative Surgery
• monoclonal antibodies

1. Introduction

GO is an inflammatory process associated with autoimmune thyroid diseases and characterized by increased adipogenesis and enlargement of the extraocular muscles resulting in the expansion of the retro-orbital content.

The optimal management of GO relies on the assessment of disease activity and severity, as anti-inflammatory/immunosuppressive treatments are effective on active GO, whereas long-term consequences of inactive GO require rehabilitative surgery, including orbital decompression, squint, and lid surgery [1,2].

GO activity is assessed with the Clinical Activity Score (CAS) (Table 1a) [3], whereas severity is graded, based on the EUGOGO score, as mild, moderate-to-severe, and sight-threatening [4]. This classification mirrors the treatment needs (Table 1b), as milder cases are managed conservatively while moderate-to-severe GO requires immunosuppressive treatments, aimed at inactivating GO, in an attempt to avoid rehabilitative surgery.

Table 1. (a) Classifications of GO activity according to clinical activity score (CAS) [3]; b) Classifications of GO severity according to the EUGOGO score [4]; c) Composite index for the assessment of clinical response to treatment [6].

A. Clinical activity score (CAS) This score is calculated as the sum of each item has the same value of 1 GO is defined as active when CAS ≥3 at first assessment and CAS ≥3 at follow-up	B. EUGOGO score Mild GO: conservative treatment, lifestyle changes, selenium supplementation Moderate-to-severe GO: immunosuppression (if active) or elective surgical intervention (if inactive). Sight-threatening (very severe) urgent medical and/or surgical decompression	C. Composite index (2020 revision) Objective measures of treatment outcome Positive response to treatment: improvement in ≥2 features in one eye without deterioration in the other eye
Spontaneous orbital pain Gaze evoked orbital pain Eyelid swelling * Eyelid erythema* Conjunctival redness* Chemosis Inflammation of caruncle OR plica Follow-up (1–3 months) (8) Increase of >2 mm in proptosis (9) Decrease in ocular ductions in any one direction (>8°) (10) Decrease of acuity equivalent to 1 Snellen line	Mild GO lid retraction (<2 mm) mild soft-tissue involvement exophthalmos (<3 mm above normal for race and gender) absent/intermittent diplopia corneal exposure responsive to lubricants Moderate-to-severe GO lid retraction ≥ 2 mm moderate or severe soft-tissue involvement exophthalmos (≥3 mm above normal for race and gender) inconstant or constant diplopia Sight-threatening (very severe) GO Patients with dysthyroid optic neuropathy and/or corneal breakdown	≥2-mm reduction of lid aperture ≥1 point reduction in five-item CAS (item 3 to 7) ≥2 mm reduction in exophthalmos ≥8° increase of eye muscle duction
A color atlas for objective assessment is avaliable at the website of the EUGOGO group: https://www.eugogo.eu/en/home/		A positive response to treatment is defined as the improvement in 2 or more items in one eye without deterioration in the fellow eye

In Europe, intra-venous methylprednisolone (ivMP) ± sodium mycophenolate for three months, in weekly infusions is the recommended first-line treatment for active GO [5]. To maximize benefits and minimize adverse effects, the ivMP route is modulated based on GO severity: the highest doses (7.5 g) are prescribed to patients with severe forms [5] while intermediate doses (4.5 g) to the milder cases. Sight-threatening GO requires medical decompression, followed by urgent bony orbital decompression, if unresponsive in a few weeks. Medical decompression consists of three ivMP pulses (500–1000 g) given on consecutive or alternate days [6], repeated after one week and followed by weekly infusions not exceeding 8 g.

Over the last two decades, the therapeutic toolkit for active GO has been enriched with new biological disease-modifying drugs including rituximab, tocilizumab, and teprotumumab. Currently, these molecules are recommended as second-line treatments alternatively to orbital radiotherapy (OR) or to a second course of ivMP. Nonetheless, randomized controlled trials (RCT), meta-analysis, reviews, and statistical analyses on pooled data suggest that these drugs are effective and will probably enter into clinical practice more and more frequently in the near future. Indeed, in the US, following FDA approval, teprotumumab is now used by several physicians as first-line treatment for GO. Nevertheless, the proportion of relapses after teprotumumab is similar to that reported after ivMP (15–30%) and some safety concerns have been raised following the observation of increased risk of hearing loss with teprotumumab [7–10].

Response to treatment can be objectively measured by scores such as the composite index, which assess changes of lid retraction, inflammatory signs, exophthalmos, and eye muscle ductions (Table 1c) [11].

Unfortunately, available therapies are still inadequate since a successful treatment measured as the change of biological parameters, does not imply per se an improvement of patients’ quality of life (Qol). Indeed, GO patients exhibit long-term visual, appearance, and psychosocial disability and suffer from emotional stress and occupational impairment even many years after being treated with immunosuppressants and surgery [12,13].

2. Incidence of ‘relapsing’ GO

The natural history of GO is self-limiting according to the Rundle curve, characterized by an initial ‘active phase’ in which the severity increasingly worsens, followed by an ‘inactive phase’ of improvement of inflammation and subsequent tissue remodeling and fibrosis.

Rarely, the course of the disease diverges from this monophasic curve because of relapses during the follow-up as outlined in Figure 1.

Figure 1. The biphasic course of GO according to the model of Rundle (Panel A). The shaded area represents the severity of GO graded according with the EUGOGO score (right axes). The black continuous line represents GO activity measured by the clinical activity score (CAS) (left axes). Dotted arrows and text in regular font outline the recommended therapies for each stage (active/inactive) and grade (mild, moderate to severe and sight-threatening). In italics fonts the treatments that might be useful in the future, but for which data are insufficient to support a recommendation. The gray arrow is the 6- weeks-follow up assessment and thereafter not-responders undergo a second line treatment. Patients developing sight-threatening GO (DON, corneal breakdown and sublussation of the eye globe) require urgent medical/surgical decompression.

Panel B A GO patient is fully responsive to treatment. The successful immunosuppressive therapy reduced either activity or the degree of residual disease, avoiding rehabilitative surgery. Patients who relapse undergo to an early (Panel C) or late recurrence (Panel D) after a variable period of inactive disease and require additional treatments. Known (bold regular font) or supposed (italic font) risk-factors/triggers are represented by arrows pointing up, while protective measures are outlined by arrows pointing down. TFTs = thyroid function tests; Se = selenium; RAI radioactive iodine therapy; GD Graves’ disease; GO = Graves’ orbitopathy; ivMP intravenous methylprednisolone; RX = orbital radiotherapy; CYA cyclosporine; OC oral corticosteroids; AZP = azathioprine; RTX = rituximab; MMF mycophenolate mofetil; DON dysthyroid optic neuropathy.

GO relapses following the first diagnosis can be defined as the recurrence of inflammatory signs and symptoms after a period of remission or improvement of variable length.

To simplify this issue, we can grossly divide these patients into two groups: early and late relapsing.

Early recurrences occur after few months (e.g. <6 months) of inactive disease. This depends on a prolonged active phase due to an asynchronous/asymmetric involvement of the orbits and to a variable response to first-line therapy with a 15–30% of the patients not responding/relapsing after ivMP [14].

Late recurrences may occur in a minority of patients who flare after a period of quiescent disease lasting several months/years. The reported frequency is variable, ranging from 5 to 16% according to the diagnostic criteria used [15,16]. In a retrospective study, Patel et al. reported a recurrence rate as high as 16%, in most of the cases occurring from 2 to 5 years after first diagnosis [17]. This percentage is likely overestimated as late relapses are observed very rarely in a real-life clinical context. Late recurrences must be differentiated from chronic congestive GO, in which the impaired orbital vascular flow causes edema, eyelid/conjunctival erythema, and chemosis mimicking active inflammation with high CAS. In this condition, only surgical treatment is expected to improve GO.

3. Hypothetical risk factor/triggering events associated with GO relapses

High TRAb levels, older age, male gender, longer duration of GD/GO symptoms, smoking habit, oxidative stress, poor control of the underlying thyroid dysfunctions and radioiodine treatment are known risk factors for GO worsening and poor response to treatments. The same factors are also associated with an increased risk of GO in newly diagnosed GD patients [2,17], and hypothetically could influence GO relapse after treatment.

Orbital rehabilitative surgery but also cataract surgery have been associated with late GO recurrences, being smokers more at risk. In the study of Patel, periocular surgery was the most frequent (12%) trigger event associated with GO recurrences, followed by thyroid dysfunctions (5%). Nonetheless, the overall risk related with surgery is quite low as Baldeschi and coworkers reported GO recurrences after bony decompression in about 1% (3/239) of patients [18].

Recently, additional risk factors have been identified opening new perspectives in personalized risk-assessment and targeted prevention.

GD patients with high LDL levels have an increased risk to develop GO, and the administration of atorvastatin to hypercholesterolemic GO patients undergoing ivMP improved the clinical outcomes. Nevertheless, it is unclear whether the enhanced response is due to the anti-inflammatory or lipid lowering properties of atorvastatin [19].

Gut microbiota composition seems to be different in GO/GD patients compared to controls with increasing levels of Prevotella, Lactobacillus, and Veillonella, although it is still debated if this is a cause or a consequence of the disease. In animal models of GO, differences in gut microbiota influenced the heterogeneity of thyroid and ocular disease being Firmicutes associated with orbital-adipogenesis. The transplant of fecal material from GO donor to these mice increased the severity of GD/GO. Unpublished data of the study named ‘Investigation of Novel biomarkers and Definition of the role of the microbiome In Graves’ Orbitopathy (INDIGO)’ (https://cordis.europa.eu/project/id/612116/reporting/it) suggested that the administration of a probiotic (LAB4, Cultech Ltd.) to GD patients reduced the risk of hyperthyroidism relapse, without influencing GO severity/progression.

4. Can we avoid GO flares?

The majority of the studies focused on preventing progression of GO from mild to severe forms and on the treatment moderate-to-severe active GO, while scarce data are available on how to avoid GO relapses after first-line treatment. Although hypothetical, it seems reasonable that the same supportive therapies and lifestyle changes useful to prevent GO progression may also reduce the relapse rate after first-line therapy.

Cautious use of radioiodine (RAI) therapy, smoking cessation, proper treatment of thyroid dysfunction, and selenium supplementation, have a proven effectiveness and have been extensively reviewed elsewhere [20].

The most obvious modifiable risk factor is cigarette smoking, which is associated with a poorer outcome after medical treatments and orbital radiotherapy, higher risk of GD relapse, increased likelihood of GO worsening/occurrence after radioiodine and increased risk of early and late GO recurrences [21]. Both active and passive smokings are associated with tissue hypoxia and increased generation of reactive oxygen species and free radicals, enhanced production of inflammatory cytokines, imbalance of Th1/Th2/Th17 responses, alteration of intestinal microbiome [19]. These effects are thought to be dose-dependent and reversible upon cessation. Noteworthy, more than a half of quit attempts fails within few weeks when patients are left without any assistance. Thus, specialized behavioral support and pharmacotherapy to relieve the symptoms of nicotine withdrawal should be offered to GO patients [22].

Stable restoration of euthyroidism, avoiding phases of uncontrolled hypo and hyperthyroidism, is a crucial factor to improve response to treatment and prevent early and late recurrences. In these patients, whether the treatment of hyperthyroidism would be conservative (anti-thyroid drugs) or definitive (radioiodine, thyroidectomy, or both) is still matter of debate. RAI treatment can trigger GO in about 15–20% of cases being this event prevented by a short-term course of low-dose oral/iv glucocorticoids prophylaxis [2,23,24]. RAI is contraindicated in patients with active GO but can be considered in those with burn-out GO after immunosuppressive treatment. One advantage of thyroidectomy is the faster and permanent remission of hyperthyroidism, and this would be preferable in patients with severe residual GO in order to speed up the rehabilitative surgery.

Selenium has been reported to decrease the ocular involvement and to reduce GO progression of mild Graves’ orbitopathy even after the treatment is stopped [25]. Nevertheless, in this study selenium levels were not measured, and it is thus unknown whether these beneficial effects were due to its antioxidant effects or to the correction of an associated selenium deficiency. Currently, selenium supplementation is not recommended in moderate-to-severe GO, except in cases of evident dietary deficiency.

The effectiveness of lipid-lowering therapies on relapse rates is currently unknown, however we believe it is appropriate recommending statin at least in all GO patients meeting the criteria for primary and secondary cardiovascular prevention.

Observations concerning gut microbiota open a number of interesting scenarios as diet, physical exercise, antibiotics, probiotics, nutritional n-3 polyunsaturated fatty acids, and antioxidants that may modulate its composition. More research and carefully designed clinical trials are needed before suggesting probiotics or changes in diet, even though these practices have gained popularity over the last few years among patients, such as self-prescribed dietary restrictions (e.g. vegan or gluten-free diet).

Finally, the appropriate management of first-line treatments has an impact on the likelihood of subsequent GO flares [Table 2].

Table 2. Efficacy of available treatments for GO. The response and relapse rate and efficacy in preventing dysthyroid optic neuropathy (DON) are reported.

Treatment	N of patients	Type of study	Responders	Not responders	DON occurrence	Relapses	Long term outcome	Reference
ivMP alone	>250	5 RCT	70–88%	10–20% (dose dependent)	Variable (<10%)	15–30%	Effect on proptosis and diplopia variable	[30]; Kahaly GJ, Lancet Diabetes Endocrinol. 2018 and J Clin Endocrinol Metab 2005; Bartalena J Clin Endocrinol Metab 2012; Marcocci J Clin Endocrinol Metab 2001
Mycophenolate +ivMP	73	RCT vs ivMP alone	>75%	≈20%	<7%	<10%	Residual disease significantly less severe than after IVMP alone	Kahaly GJ, Lancet Diabetes Endocrinol. 2018
Teprotumumab	84	pooled data of 2 RCT	>80%	<10%	≈1%	≈20%	Durability of the response still unclear	[9]
Rituximab	141 (24 steroid resistant)	3 RCT (1 vs placebo; 2 vs ivMP) and 3 open studies	>85%	≈10%	<3%	<1%	Effect on proptosis and diplopia variable (probably equal to ivMP)	[30,31]; Li J, Pharmacology 2017; Khanna D, Ophthalmology, 2010; Du Pasquier-Fediaevsk L Ocul Immunol Inflamm. 2019 and Vannucchi G Thyroid. 2021 (low dose)
Tocilizumab	63	RCT vs placebo 1 open study	≈90%	≈10%	Not reported	Not reported	Ongoing definition	Perez-Moreiras JV, Am J Ophthalmol. 2018; Sánchez-Bilbao L, J Clin Med. 2020
RX alone	92	RCT	50–60%	30–40%	Unclear	Unclear	Effective on recent onset ocular motility impairment, less effective on proptosis	Mouritz MP Lancet 2000, Prummel MF J Clin Endocrinol Metab. 2004, Marcocci J Endocrinol Invest. 1991
RX + OC	56	2 RCT	≈65%	≈35%	Unclear	Unclear	Combination therapy improves final outcome	Marcocci J Endocrinol Invest. 1991, Marcocci J Clin Endocrinol Metab 2001
RX + ivMP	41	1 RCT	>85%	<15%

RCT = randomized controlled trials, ivMP = intravenous methylprednisolone, OC = oral corticosteroids, RX = orbital radiotherapy

Time is increasingly recognized as a key factor in treatment planning, and the choice of treatment may depend on disease duration. For instance, the efficacy of ivMP and rituximab is greater in highly active and recent GO, whereas of teprotumumab it is effective also in patients with a longer disease duration [22]. Teprotumumab is associated with a greater improvement in proptosis as compared with ivMP, however data from the OPTIC and OPTIC-x trials showed a relapse rate of nearly 30% after a first course of teprotumumab with one patient developing dysthyroid optic neuropathy (DON) [7]. The whole of these data suggests that ivMP is still appropriate in the majority of patients affected with GO and it seems unlikely that steroids will be rapidly supplanted by this or other novel treatment [8,9].

Since GO recurrences, including DON, may occur at ivMP withdrawal, patients should continue ivMP treatment for several weeks. To reduce the frequency of GO flares, some clinicians administer oral prednisone between iv pulses and during the tapering period after pulses discontinuation, even though this practice seems not to increase the overall response rate [26]. Some retrospective studies suggested that OR may be employed in preventing GO flares after steroids titration [27], but prospective trials are not available to fully support its routine use.

As the majority of patients with active GO responsive to ivMP improve as early as 6 weeks, it has been suggested that non-responders can be switched to other treatments alone or in combination with steroids without excessive delay [5,28]. Nevertheless, a shared decision-making discussion with the patient should be encouraged, since some of these cases may become responders at 12 weeks of treatment [29].

In RCTs or open studies, both rituximab and tocilizumab improved GO activity and severity in patients who were non-responsive or flared after steroids. Albeit with conflicting evidence [30,31], the monoclonal anti CD20 antibody rituximab seems associated with a lower recurrence rate compared to steroids and teprotumumab, although at present, it is still recommended as second-line therapy [Table 2].

5. Conclusions

The course of GO can be influenced by preventive measures that might reduce the progression rate of GO to more severe forms. Several risk factors for GO occurrence and poor response to treatments have been unequivocally acknowledged (smoking, uncontrolled thyroid dysfunction, administration of radioiodine without a proper prophylaxis, not timely treatment of the active phase), while others remain unclear (dyslipidaemia, intestinal dysbiosis). Although very reasonable, it is unclear if the correction of these risk factors may also be helpful in preventing recurrences of GO after immunosuppressive treatments.

Given the large heterogeneity of GO manifestations, a more tailored approach is desirable. Further studies are warranted to identify the characteristics of subgroups patients with the highest chances to respond to a certain treatment.

Until more effective and specific drugs are discovered in the future, long-term surveillance and counseling are recommended for all GO patients.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was partially supported by the Italian Ministry of Health.