Graves’ disease (GD) is a thyroid autoimmune disorder in which anti-thyrotropin (TSH) receptor autoantibodies (TRAb) with stimulating activity induce hyperthyroidism. Graves’ orbitopathy (GO), characterized by inflammation and remodelling of orbital tissues, is the main extra-thyroidal manifestation of GD. Moderate-to-severe GO determines major facial disfigurement, psychosocial impact and, rarely progression to sight loss (1, 2). Typically, GO has a diphasic course (3) with an inflammatory and progressive phase, followed by stabilization and subsequent scarring due to tissue fibrosis. Medical treatment, based on immunosuppression, is indicated during the active phase of GO. Intravenous steroids have been shown to be effective in about 70% of patients, although unresponsiveness and disease relapse are observed in a good proportion of patients (20-30%) (4).

Rituximab (RTX) is a human/mouse chimeric monoclonal antibody that binds the CD20 antigen on the surface of B cells and induces rapid depletion of B cells in both periphery and lymphoid organs. Clinically approved indications for the use of RTX in autoimmune disease are currently rheumatoid arthritis and ANCA-related vasculitis (5). In previous studies (6), RTX has been shown to be effective in inactivating moderate-to-severe GO when used early in the disease, but may have little effect in longer duration disease (7). The clinical effect of RTX in GO is quite rapid, since disease generally inactivates within 4 – 6 weeks after the first infusion (8). RTX induces B-cell lysis and may activate anaphylatoxins and other inflammatory cytokines with subsequent recruitment of phagocytes within the orbital tissue and depletion of inflammatory cells yielding tissue fibrosis (9, 10). This may explain why, after B cells repopulation, reactivation of GO was only observed rarely.

RTX therapy is not devoid of side effects: infections may arise due to hypogammaglobulinemia, although they are also commonly associated to therapy with steroids or other immunosuppressive drugs (11). Acute infusion reactions related to the release of cytokines produced by macrophages, monocytes, lymphocytes and NK cells, can be observed in 10-30% of patients at the first infusion (12). These effects may be associated to larger doses of RTX and significant lymphocytic infiltrates in target organs (13).

The optimal dosage of RTX in GO has never been studied in randomized, dose finding clinical trials. Standard dosing in rheumatoid arthritis and other autoimmune diseases consists of two i.v. administrations of 1000 mg with a 2-week interval. This dose was also initially used in most open studies (8) and in individual case reports (10) in patients with GO. Rapid and complete B lymphocyte depletion was shown to occur already after very low dosages (25-100 mg) of i.v. RTX (9). In a prospective randomized clinical study it was shown that the efficacy of a single 500 mg RTX dose was comparable to two 1000 mg doses (7) in inactivating GO and more effective than i.v. methylprednisolone (6). Later observation that in GO total peripheral (and orbital) B cell depletion was occurring even after low doses of RTX (100 mg) (9), has allowed us to reduce the therapeutic dose to a single infusion of 100 mg RTX (14).

Aim of the present study was to compare the efficacy and safety of different doses of RTX, employed prospectively in two open label studies (14, 15) and in one trial in which RTX was randomized to i.v. steroid therapy (6), in active GO.

The first and most important finding of this study is that doses of RTX ranging from 100 to 2000 mg invariably induce inactivation of moderate-to-severe GO with only some differences in the clinical response, especially on diplopia, quality of life assessment and CD 19+ cells repopulation. This study is based on a post-hoc analysis of data collected from one prospective RCT and two open label studies that have employed different protocols and dosing of RTX administration.

Our findings show that any of the three RTX dosages used (100 mg, 500 mg and 1000 mg x 2) is equally effective in inactivating the disease in >95% of patients at 12 and 24 weeks after infusion. Efficacy of RTX has been known to be associated to the duration of B cell depletion in the circulation (14). In this study we show that peripheral CD19+ cells begin repopulating only slightly earlier than 24 weeks in patients treated with the very low dose (100 mg), but later in those administered either 500 or 2000 mg. B cell return in the circulation, though, does not modify the disease outcome or favor disease relapse during follow-up (up to 12 months) (6, 14, 20). Based on these observations, we believe that in moderate-severe GO lower doses of RTX, which are generally burdened by a lower risk of adverse events such as reactivation of infections (12) or the induction of other autoimmune processes after depletion of B cells (21), should be preferred.

We studied whether RTX induced inactivation may depend on GO duration of less or more than 6 months, as previous studies have suggested that in GO of longer duration RTX might be less effective (7, 17). We did not find differences in the therapeutic response based on the duration of disease with any of the doses employed. We suggest that efficacy of RTX treatment is likely more related to the early inflammatory phase of GO, than to disease duration. We observed no difference in the response of patients with newly diagnosed GO or of those with relapsing disease after a complete cycle of intravenous steroids, who invariably had longer disease duration (up to 66 months). This findings, besides confirming that disease activity rather than duration is important for the effect of RTX in GO, also suggest that this modality of therapy can be proposed as second line treatment when steroids fail (1, 2).

The therapeutic outcome on parameters of disease severity, especially those related to muscle involvement, appear to be of importance, as the presence of diplopia in inactive GO has an impact on the quality of life of patients. While there were no significant changes in the degree of proptosis with any of the three dosages used, we did observe a significant improvement of the Gorman score for diplopia when higher RTX doses (Group 2 and Group 3) were employed. A possible explanation for this is that higher doses might be more effective than lower doses in preventing the orbital tissue fibrosis that follows the active disease phase.

The QoL questionnaire used in the study has been validated and is specific for assessing quality of life changes in GO patients (18). It is divided into two sections, the first one assessing the impact of the aesthetic changes (appearance) on patients, the second one that of the eye dysfunction. The perception of self-appearance was found to improve at 24 weeks only in patients treated with 100 mg RTX in a single administration, while improvement of function was reported by those treated with a single larger dose of 500 mg. Improvement of inflammatory signs were in fact more rapidly observed after treatment with the low dose (>90% of patients inactivated at 12 weeks), whereas the impact of RTX on diplopia was found to be associated to the use of larger doses. This suggests that high doses of RTX might be preferable in patients with significant extraocular muscle involvement, as it would reduce the need of rehabilitative surgery. Unfortunately, we could not assess QoL outcome in Group 3 because of the small number of questionnaires returned, and this is a limitation of this study. Differences in the quality of life assessment were not associated to differences in the number of rehabilitative surgeries required, whichever dose of RTX was employed.

In two patients treated with the low dose, RTX therapy failed to inactivate GO and to prevent progression to DON. These patients may have had a subclinical form of DON at the time of treatment which might have been further exacerbated by the increase of intraorbital edema induced by an acute release of cytokines mediated by RTX as previously reported (9). In other studies RTX has been used in patients with DON (10) resulting in improvement of vision. In addition, the observation of the syndrome cytokine release in two patients receiving only 100 mg RTX, suggests that this adverse event is not related to the dose employed and can be prevented only by pretreating patients with fairly high intravenous methylprednisolone doses (up to 500 mg) (22). Caution is therefore suggested for the use of RTX in patients with suspected subclinical DON (1).

Restoration of euthyroidism after RTX may be due to either its effect on antibodies stimulating the TSH receptor (23) or to remission of hyperthyroidism after prolonged antithyroid treatment (24). While some authors have hypothesized that RTX does induce a decrease of serum TRAb in patients with GD (25, 26), others have shown that the levels of antibodies stimulating the TSH receptor (TSAb) are not modified in GO patients (23). RTX causes depletion of B lymphocytes in the peripheral blood but also in the thyroid (27) and in the orbit (9) and histological analysis of orbital tissues of patients who received RTX showed absence of both B and T lymphocytes (8). This observation supports the hypothesis that RTX blocks B cell function as antigen presenting cells and interaction with helper T cells, required for the initiation of the autoimmune process (28). Interestingly, we observed that the number of NK cells increased following RTX infusion in Group 1 and 2, compared to baseline values. Previous studies on rheumatoid arthritis and idiopathic membranous nephropathy have shown a rise in the NK number, associated with an altered NK cell function. Although, the clinical significance of this observation needs to be established, it has been associated with a better response to RTX treatment (29, 30).

This may explain why after CD19+ cells repopulation, inflammation does not relapse and the CAS continues to decrease until stable disease inactivation (31). We in fact did observe that although CD19+ cells return earlier (<24 weeks) in patients treated with the 100 mg dose, when compared to Groups 2 and 3, this does not result in GO reactivation.

In analyzing the use of RTX in GO, we also looked at the cost of therapy in relation to the administered dose. When compared to first line treatment with intravenous steroids, the cost of a full cycle of treatment with RTX in Europe is about 4.8, 24 and 61 times more expensive for the 100, 500 and 1000 x 2 mg dose, respectively. Given the more favorable response/relapse rate and the clinical evidence shown by this study, we believe that a single RTX 500 mg dose is best indicated as second line treatment for the majority of GO patients.

The results of this study have to be seen in light of some limitations: in first instance the retrospective study design, the small sample size and its heterogeneity. Indeed, cases had a different duration of GO (which was longer in patients belonging to Group 3 compared to Group 1 and 2) and were enrolled among Graves’disease, Hashimoto’s thyroiditis and euthyroid GO patients. Furthermore, some patients received previous corticosteroid treatment that may have had influenced the final outcome.

In this study, in conclusion, we sought to understand, from the limited available evidence, which dose of RTX is optimal as second line therapy of active moderate-to-severe GO. In this analysis we considered, both the effectiveness of the drug in inactivating the disease, the possibility of changing its natural course, the effect on disease severity and the impact on the patients’ quality of life. Prospective randomized studies specifically addressing RTX dosing in GO are warranted.

The datasets presented in this article are not readily available because they contain some sensitive personal data. Requests to access the datasets should be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Fondazione IRCCS Ca’ Granda, Ospedale Policlinico, Milan, Italy. The patients/participants provided their written informed consent to participate in this study.

Data of each trial were collected by IC, DC, GV, MS, EL, LF, NC, LP and CG, BM and MD performed some serological assays and collected biochemical and cytofluorimetric data, the final database was produced by IM and EL, results were analyzed by IC and EL, the draft of the manuscript was written by IM and MS, the final manuscript was reviewed and revised by MS and MA. All authors contributed to the article and approved the submitted version.

This work was supported by funds of Fondazione Ca’ Granda, IRCCS, Milano, Italy to MS (RC 2020).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.