Rituximab (RTX) use in open-label series has been associated with very encouraging responses in patients with active and moderate-to-severe Graves’ orbitopathy (GO). Recently, randomized controlled trials of RTX have been performed in such patients to answer the question of clinical efficacy and the safety profile of this agent. That data, reported separately, focused on Clinical Activity Score (CAS) and indicated in one trial a strong benefit of RTX in comparison with IV glucocorticoids, whereas the other trial noted the absence of a benefit by comparison with placebo. The outcome was reanalyzed post hoc here, using EUGOGO criteria, and the results were not significantly different. The authors comment further on the differences between the two trials regarding populations treated, methodology, analysis of outcomes and the adverse effect profile of RTX. The populations treated appear different with younger patients, lower TRAb and shorter duration of disease prevalent in the Italian trial, all elements favoring a better response. Smoking, usually diminishing a response, was also more prevalent in some patients. The combined outcome proposed by EUGOGO revealed similar results with CAS regarding RTX efficacy; yet, it might be a more comprehensive outcome. The adverse events of concern relate mainly to the risk of DON, which seems to be increased by the use of RTX in a certain subset of patients. Based on available data, a multicenter trial using the EUGOGO-proposed outcomes might be the next best step to define the role of RTX in GO therapy.
Invited Author’s profile
Marius N Stan, MD, is a consultant in endocrinology at Mayo Clinic in Rochester, MN, USA. He is the current Chair of the Thyroid Core Group and Assistant Professor in Medicine. He combines clinical work with clinical research. He is primarily working in the area of autoimmune thyroid diseases, with a specific interest in Graves’ disease and Graves’ orbitopathy. A second area of interest is centered on thyroid nodules, thyroid surgery and acute post-surgical hypoparathyroidism. Other areas of research involve drugs that alter thyroid function, mainly the thyroid dysfunction caused by amiodarone (amiodarone-induced thyrotoxicosis). Dr Stan’s research aims to identify novel treatments as well as methods of prevention for these conditions. He is involved in multiple prospective clinical trials as well as retrospective cohort studies.
Rituximab (RTX), a chimeric mouse-human monoclonal antibody directed against the CD 20 antigen on B lymphocytes, has been associated with very encouraging responses in treating patients with active moderate-to-severe GO since 2006 (1, 2, 3). B cell depletion induced by RTX may affect pathogenic TSH-R autoantibody or directly interfere with B cells’ antigen-presenting cell function (4, 5). Several uncontrolled studies have suggested that RTX is especially useful in the control of the early inflammatory phase of GO (reviewed in detail by Salvi et al. (6)), eventually affecting the subsequent natural course of the disease. Recent randomized controlled trials of RTX in active and moderate-to-severe GO have been designed to answer this question. In addition, as RTX induces systemic immunosuppression, the safety profile of this biological agent was needed to be studied before we can trustfully use it in patients who are affected by a progressive and often disfiguring disease, but in fact, it was self-limiting in its course.
The potential mechanistic actions of rituximab in GO are discussed here based on studies targeting GO as well as non-GO pathologies. RTX is approved for clinical use in rheumatoid arthritis and antineutrophil cytoplasmic antibody-associated (ANCA) vasculitis. When administered, RTX depletes more than 95% of CD 20-expressing B cells from the stages of immature to mature B cells, but not antibody-producing plasma cells (7). Decrease of TSH-R antibodies after RTX has been reported inconsistently in uncontrolled studies (6). It has been suggested in one study that the subpopulation of TSH-R-stimulating antibodies are affected by B cell depletion (8), thereby inducing stable remission of hyperthyroidism, but this finding has not been confirmed in patients treated for active GO (9). Alternatively, RTX has been shown to interfere with the production of inflammatory cytokines and with B cell-driven antigen presentation (7), suggesting that RTX might be indirectly responsible for the depletion of autoreactive T cells also. Lack of T cell infiltration after B cell depletion with RTX has been reported in the orbital tissues of patients with GO (3, 10), suggesting that this might be a likely mechanism for its effect. Preliminary studies have suggested that RTX can be effective either as first-line treatment for moderate-to-severe GO or as additional therapy in patients resistant to intravenous steroids (6).
Based on these promising data, 2 randomized control trials (RCT) have tested the benefit of RTX in GO patients, and they have generated significantly different results (11, 12). We will reference them by their geographic location as the Italian trial (11) and the US trial (12). These trials were both designed to determine the benefit of RTX in patients with active and moderate-to-severe GO. Although the US trial used placebo as the comparator, the Italian trial used intravenous methylprednisolone (IVGC) for the control group. The differences between the results of the two trials were significant. The Italian study used a…’ and correct if necessary. Italian study used a decrease in Clinical Activity Score (CAS) by 2 points or reaching a CAS <3 as a primary outcome and thus showed RTX to offer a ‘better therapeutic outcome in active moderate-to-severe GO, compared to ivMP, even after a lower RTX dose’ (11), whereas the US trial using CAS improvement as a continuum and separately as improvement by ≥2 points as primary outcome found that RTX ‘offered no additional benefit over placebo to our patients with active and moderate-to-severe GO and carried with it non-negligible adverse effects’ (12). Therefore, we have decided to reanalyze some of the data from the two trials to better understand those differences.
We have evaluated the data from the 2 RCTs and reanalyzed their outcomes using the outcomes utilized by EUGOGO when testing the differences between the various doses of IV glucocorticoids (13). They defined improvement in a composite ophthalmic score as a change in 2 or more of the following outcomes without deterioration of any of those in any of the eyes: (1) improvement in palpebral aperture by ≥3 mm; (2) improvement in NOSPECS class II signs by ≥2 grades; (3) improvement in Bahn–Gorman diplopia score or improvement of ≥8° in motility; (4) improvement in proptosis by ≥2 mm; (5) improvement in CAS by ≥2 points. Deterioration was defined as DON development or occurrence of two of the following: (1) increase in palpebral aperture by at least 3 mm; (2) increase of any of the class 2 signs of NOSPECS by at least 2 grades; (3) increase of proptosis by at least 2 mm; (4) decrease of ≥8° in motility or worsening of diplopia score and (5) deterioration of the CAS. If changes were smaller than the previously defined parameters, then the outcome was considered as ‘no change’. Similar outcomes have been employed in other recent large multicenter trials in GO (13, 14) and are supported as most useful outcome by most experts in the field though not yet formally endorsed by the specialty societies, EUGOGO or International Thyroid Eye Disease Society.
Both trials provided additional information on the follow-up of patients after trial completion (data on additional therapeutic requirements in the Italian study are also presented in the original paper (11).
We have revisited the data from both trials and have reanalyzed it based on the outcome definitions proposed by EUGOGO and detailed in the ‘Methods’ section. The analysis was performed on an intention-to-treat approach.
The US trial found that at 24 weeks, the time frame set for the primary outcome, improvement was present in 1/13 (7.6%) RTX-treated patients, stability was present in 8/13 (61.5%) patients and 4 patients (30.7%) deteriorated. This compares with improvement in 1/12 (8.3%) patients in the placebo group where the other 11 patients (91.7%) remained stable (P = 0.11 for comparison between groups) and none deteriorated. On intention-to-treat analysis at 52 weeks, there were 6 patients that improved in the RTX group (46%), 5 that were stable (38%) and 2 patients that met the criteria for deterioration (15%). By comparison using intention-to-treat analysis, there were 3 patients (25%) that improved in the placebo group, 8 that remained stable (67%) and 1 that deteriorated (8%) (P = 0.37 for comparison between RTX and placebo groups). However, 5 patients did not continue beyond 24 weeks. Thus, when we analyzed the patients that continued per protocol beyond 24 weeks, we had a total of 20 patients (10 in each group), of which one later dropped out (12). Analyzing the composite ophthalmic score in these groups at 52 weeks, we found that 6 improved and 4 remained stable in RTX group, compared with 3 patients that improved, 6 that remained stable and one patient that deteriorated in the placebo group (P = 0.30).
When we looked at proptosis changes during the trial, we identified 6 patients that deteriorated in their proptosis measurement from baseline to 24 weeks – 4 in the RTX group and 2 in the placebo group. Of these cases, 3 patients returned to baseline proptosis values by 52 weeks (all 3 in the RTX group), whereas 1 patient treated with RTX and 2 treated with placebo remained with same exophthalmos as noted at 24-week measurement at the end of the trial. In contrast, there were 2 patients that developed dysthyroid optic neuropathy (DON), both were treated with RTX, and in both cases, proptosis measurements did not change from baseline to the time the diagnosis of DON was made.
We have also surveyed the participants at a mean of 44.3 (s.d. 15.7) months after completion of the US trial to define the additional therapeutic burden that they experienced after the intervention. We have found that of the 11 patients in the RTX group that returned the survey, 5 required additional therapies. In these 5 patients, there were in total 6 orbital decompressions, 13 strabismus surgeries and 6 eyelid surgeries. Of the 9 patients in the placebo group that responded, there were also 5 that required additional therapy: 1 received IV glucocorticoids, 1 received oral prednisone, there were 2 orbital decompressions, 11 strabismus surgeries and 2 eyelid surgeries. Thus, we did not identify a difference regarding additional therapeutic burden after the trial was completed (P = 0.65). In total, 7 of the 10 patients that required therapy had more than one intervention during follow-up.
The same analysis was performed in the Italian trial at 24 weeks. In patients treated with RTX, disease improvement was present in 9/15 (60%), stability in 5/15 (33.3%) and deterioration in only 1/15 (6.7%), whereas in those treated with IVGC, improvement was observed in 6/16 (37.5%), stability in 2/16 (12.5%), but worsening in 8/16 (50%) (P = 0.026), thus showing a significant reduction of disease recurrence in the RTX group when compared to standard steroid therapy. This analysis was not performed at 52 weeks.
Patients were followed up to 76 weeks after the beginning of therapy with the aim of reporting the long-term benefits of either treatment. In the 15 RTX-treated patients, elective orbital decompression was performed in one patient affected with unilateral GO who additionally required strabismus and lid surgery, strabismus surgery in 2/15 patients and lid surgery in 2/15 patients. In the 16 patients treated with IVGC, the following surgical procedures were performed: urgent decompression for DON in 2 patients, elective decompression in 4 patients, strabismus surgery in 3 patients and lid surgery in 3 patients. Overall, surgical procedures were carried out in 12/16 IVGC patients and in 5/15 RTX patients (P = 0.049; see Supplementary Table 2 in (11)). No other additional therapy was required by these patients who continue to be seen periodically in the investigator’s clinic.
As for proptosis measurements, there was an increase (≥2 mm) in proptosis in 3 RTX patients, in one of whom the increase was recorded at 24 weeks, but subsequently returned to baseline values. In the IVGC group, there were 2 patients that had persistent increase in proptosis.
Dosing and side effects of rituximab
In rheumatoid arthritis, RTX is generally employed at the dose of 1000 mg administered twice at two-week interval and possible retreating is considered in general after 6 months, based on the duration of B cell depletion and the clinical course of the disease. In GO, the doses used have been similar in most studies (10, 15, 16), but other dosing regimens have been employed from 100 mg once (17) to 375 mg/m2 × 42. Infections have been reported as a consequence of the decrease of immunoglobulin levels after repeated RTX doses, although recent large retrospective studies in autoimmune disease have shown that severe infections are more often associated to concurrent therapy with steroids or other immunosuppressants and not to hypogammaglobulinemia per se (18). Likewise, progressive multifocal leukoencephalopathy (PML) has rarely been reported in patients receiving RTX, who were previously treated with other immunosuppressive drugs including cyclophosphamide, azathioprine and even oral or intravenous steroids (19). In the RCTs performed for GO, there have been a number of adverse effects as described in Table 1. This is in the context where the doses utilized have been the same as those for rheumatoid arthritis (1 g × 2) in both trials at the beginning and then the dose was decreased in one (the Italian trial after 5 patients) to 0.5 g × 1, while it was kept constant in the US trial. Although it is uncertain that RTX played a role in the development of some of the adverse effects (i.e. DON), as patients on steroids also showed progression of disease to sight-threatening forms of GO, a number of these adverse effects are of significant concern (both DON and vasculitis) that specific attention will be required when considering the use of RTX in GO.
|Side-effect type||Placebo (events/treated patients)||Rituximab (events/treated patients)||IV GC (events/treated patients)|
|Skin (rash, itching)||0/12||2/28||0/16|
|Infectious (bronchitis, conjunctivitis)||1/12||1/28||0/16|
|Transient loss of vision with cytokine release syndrome||0/12||2/28||0/16|
|Severe eye tearing||0/12||1/28||0/16|
|Blood pressure fluctuation||0/12||0/28||1/16|
|Liver test abnormalities||0/12||0/28||2/16|
|GI (Tongue pain, abdominal pain, diarrhea)||1/12||2/28||1/16|
|Mild infusion reaction||0/12||13/28||0/16|
|Total severe events||0/12||2/12||2/12|
|Total patients affected||3/12||8/13 + 13/15||9/16|
The reanalysis of the data generated by the 2 trials did not produce any change in the previously reported conclusions regarding the use of RTX in GO – the US trial did not find a benefit in the population treated in Rochester, Minnesota, USA, whereas the Italian trial found a strong benefit for GO patients as enrolled in Milan, Italy. Why would there be these differences?
First, we will review briefly the pretrial data generated by open-label case series, the protocols used and the actual GO benefits noted in those series.
Rituximab in GO: pretrial experience
In earlier case reports reviewed by Salvi et al. (6), RTX has been used with a very encouraging response in patients with active moderate-to-severe GO unresponsive to previous intravenous steroids. To date, 49 patients treated with RTX have been reported in uncontrolled studies. Overall, significant GO inactivation was observed in about 90% of patients and infusion-related side effects (both minor and major) were reported in about 30% of the patients treated with different treatment protocols and doses administered. Of note, disease relapse has not been observed in any of the reported patients, even after a long follow-up period (52–76 weeks), although we are aware that non-responders or patients with disease relapse after RTX may have not been reported in these uncontrolled studies. The dose of RTX in these studies varied according to the different study protocols performed (see above).
Total peripheral B cell depletion was reported even after a very low dose of RTX (100 mg) by Salvi et al. (17) in two patients in whom RTX infusion had to be discontinued due to the occurrence of cytokine release reaction. GO inactivation occurred within a few weeks, despite the administration of a dose about 20 times less than the standard dose used in systemic autoimmune disease. Monitoring of peripheral B cell depletion to titrate the RTX dose has been suggested by some authors in autoimmune renal disease (20). On the other hand, the variable response to RTX noted in different systemic autoimmune disorders makes it difficult to translate the dosing regimens employed for those entities to GO research and management.
Failure of RTX was reported in one patient whose GO progressed to acute DON (21) and in one who already had DON (22) and failed to improve on RTX despite achieving peripheral B cell depletion. On the other hand, another ten patients with DON reviewed in the paper by Salvi et al. (6) have responded to RTX with improvement of their visual sight. No formal trial has addressed the use of RTX in DON, and given the DON changes noted in the RCTs discussed here, this issue will have to be carefully considered.
Analysis of the randomized trials characteristics
A number of elements were considered to understand the differences between the trials: population enrolled, drug protocol, trial methodology used and then we compared the trial results. We have looked at the populations treated in the 2 trials, particularly at the groups that received RTX and found that there were a number of important differences in the 2 populations (described in Table 2). The Italian population was younger and was composed almost exclusively of women. These characteristics are both associated with a higher likelihood of response to therapy by comparison with older man (23). Thus, from this perspective, the population treated in the US trial had a lower likelihood of response to therapy. Another element that can predict response to therapy seems to be the duration of disease (24). Here again, the Italian trial recruited a population that had GO by an average of only 4.5 months by comparison with almost 3 times longer disease duration in the US trial. This also could lead to the recruitment of patients that might have spontaneous improvement, as depicted in Rundle’s curve (25), although all but two patients, randomized in the steroid arm, were seen at least twice before enrolment and had consistently CAS >3, suggesting active and progressive disease. A final element that was in favor of a better response in the Italian trial is the lower TRAb noted in the RTX-treated group by comparison with the homologous group in the US trial, with TRAb values about 3× higher in the latter (P = 0.036). The influence that TRAb has in predicting GO outcome has been outlined by Eckstein et al. (26) and others (27), and it is likely linked to the role that TRAb is purported to have in GO pathophysiology (28). As expected, smoking prevalence was higher in the European study than that in the US trial. Smoking is usually associated to a worse therapeutic outcome after steroids (29). Although there are no data in GO, in patients with rheumatoid arthritis, response to RTX has been shown to be less satisfactory in smokers than in non-smokers (30). Therefore, one could presume that the response to RTX would be more significant in the US trial, but that was not the case.
Population differences between the two RCTs.
|Italian study (n = 15)||US study (n = 13)||Comments|
|Age (mean, years)||51.9||57.6||Possibly different|
|Gender (% women)||93||69||Possibly different|
|Smokers (%)||66.7||15.4||Likely to be different|
|GO duration (months)||Mean: 4.5 ± 2.9||Mean: 30 ± 47.6Median: 12 (8.2–27.2)||Likely to be different|
|CAS baselineCAS ≥4/CAS ≥6||Mean: 4.4/10 ± 0.714/15 and 2/15||Mean: 4.9/7 ± 1.013/13 and 3/13||Likely to be sameCAS 7 (US, Italy) and 10 (Italy)|
|GO severity||Moderate-to-severe||Moderate-to-severe ± progressive||Likely to be same|
|Previous steroid therapy||3/15 (20%)(≥12 weeks prior)||4/13 (31%)(≥8 weeks prior)||Likely to be same|
|TRAb (IU/L)||Mean: 10.7 ± 9.1||Mean: 28.1 ± 23.4Median: 20 (9–60)||Likely to be different|
|TRAB >20||4/15||7/13||Likely to be same|
By design, there is a difference regarding steroid use before the trial with the Italian patients having been off steroids for at least 3 months vs a minimum of 2 months for the US patients (minimum 12 weeks vs 8 weeks). Whether this might have led to a diminished benefit in the US patients is not yet known. A similar use of steroids was usually the case for the pretrial case series that reported benefit with the use of RTX (10, 15, 31), and there are no clinical data to indicate that a longer washout from steroids would in any way potentiate RTX response.
The protocol employed in the 2 RCTs is outlined in Table 3. The double-blinding approach used by the 2 trials was slightly different with the Italian investigators allowing the endocrinologists to be unblinded in an effort to ensure the safety of the participants. In the US trial, all investigators were blinded, and the patients’ safety was ensured by the use of a Data Safety Monitoring Board. The board was notified of all adverse events and was able to obtain all pertinent patient information, including randomization assignment, and make safety decisions independent of the investigative team. Both Italian and US investigators are convinced that no undesirable bias has permeated the analysis as a result of these processes. The other procedural difference relates to the dose of RTX used in the trials and described briefly earlier. Most of the pretrial data regarding RTX benefit in GO were obtained with the use of 1 g × 2, similar with the dosing employed in the rheumatoid arthritis (32). The US trial has kept this approach throughout. The Italian study design was amended based on the recognition that one single dose of 500 mg was as effective as 1000 mg twice in attaining total B cell depletion.
Protocol comparison between the two RCTs.
|Italian study||US study||Comments|
|Randomization||2 groups, blocks of 4||2 groups, block of 6||Stratified by smoking|
|Stratification||Not applicable||Smoking and CAS||Smoking and CAS balanced in both|
|Blinding||Patient + ophthalmologist||Patient + endo + ophthalmologist||Endo unblinded in Italian study (for patient safety)|
|Dose of RTX||1000 mg × 2 (5 patients)500 mg × 1 (10 patients)||1000 mg × 2||Changed dose based on 2 patients reported separately ((17) #3537)Protocol amended|
|Follow-up||At 52 and 76 weeks both groups||52 weeks in the clinic, 1 year later by mail|
|Analysis||Intention to treat||Intention to treat|
|Raw data provided||Baseline and outcome dataSurgical data||Baseline and outcome dataSurgical data||US post-trial surgical data reported here|
|Outcome||CAS @ 6 monthsSeverity – 6/12/18 monthsQOL – 6/12/18 months||CAS @ 6/12 monthsSeverity – 6/12 monthsQOL – 6/12 months|
The outcomes of the 2 trials as reported are compared in Table 4. We see here the significant difference between the trials regarding CAS and, implicitly, the prevalence of active/inactive disease after intervention, as per the original study design. When the response to therapy was analyzed by calculating the overall ophthalmic score, the Italian investigators found that only one patient had worsening of disease after RTX compared to 50% after IVGC at 24 weeks, suggesting a disease-modifying effect of RTX. That opinion is not endorsed by the US investigators who did not find this agent as better performing than placebo. The CAS as reported shows a significant difference in response between the 2 RTX-treated groups. As GO is self-limiting in its course, this has also been considered a good outcome for studies on the treatment of active moderate-to-severe GO. However, a more comprehensive score of disease improvement is likely to become a preferred primary outcome of clinical trials for GO. We have tested that instrument here, and the results are noted to be similar in significance but different in magnitude than those obtained by measuring success based on CAS. The two RTX-treated groups remain different on the EUGOGO response at 24 weeks, and the comparison with their respective control groups does not change based on this post hoc outcome. This might be informative for future trial planning.
Outcome data of RTX treatment.
|Italian studyRTX/IVGC||US studyRTX/Placebo||Comments|
|CAS 6 months (mean, s.d.)|
|6 months||0.6 (SE 0.3)/2.3 (0.5)||3.7 (1.9)/3.8 (1.4)||RTX better than IVGC (Italy)RTX same as placebo (US)|
|12 months||0.5 (SE 0.2)/1.1 (0.7)||2.0 (1.7)/2.4 (2.0)||RTX better than IVGC (Italy)RTX same as placebo (US)|
|At 6 months||100%/69%||31%/17%||RTX better than IVGC (Italy)RTX same as placebo (US)|
|At 12 months||100%/75%||60%/40%||RTX better than IVGC (Italy)RTX same as placebo (US)|
|Proptosis improved ≥2 mm||33%/6.2% (exophthalmometer)||30%/30% (exophthalmometer)20%/30% (CT)||RTX same as IVGC (Italy)RTX same as placebo (US)|
|24 weeks||Function: 38%/61% Appearance: 62%/46%||Physical: 36%/55%Mental: 46%/82%||GO-QOL is a validated instrument|
|52 weeks||Function: 75%/54% Appearance: 54%/46%||Physical: 56%/67%Mental: 89%/44%||SF-12 is not specific for GO|
|Success rate (%)|
|Trial criteria||100%/69% (24 weeks)||31%/25% (24 weeks) 52%/52% (52 weeks)||US trial criteria: ΔCAS ≥2 + no additional therapy; Italian trial criteria: disease inactivation|
|EUGOGO composite score criteria||60%/38% (24 weeks)||8%/8% (24 weeks) 46%/25% (52 weeks)||EUGOGO criteria: see ‘Methods’ section|
The Italian study used GO-QOL while the US study used SF-12. exo, exophthalmometer
As far as proptosis is concerned, there were no differences detected in either trial. It is interesting that there was a transient increase in proptosis in 3 RTX-treated patients in the US trial. That was noticeable at 24 weeks and it regressed thereafter. By contrast, no such case was noted in the placebo group. Also, one RTX-treated patient in the Italian trial had transient increase of proptosis, with return to baseline at 52 weeks. This suggests that in some patients, RTX may induce an increase in the intra-orbital tissue volume that in turn can lead to anterior displacement of the eye globe with increase in proptosis. In cases where the orbit, due to local anatomy and existing orbital changes, cannot accommodate this ‘natural decompression’, the likelihood of DON increases, perhaps even more if DON is present at a subclinical stage and cannot be recognized easily. It is possible that this was the mechanism in the 2 patients that developed DON in the US trial as they were both in the RTX group. It is interesting that 2 patients in the Italian trial, both affected with unilateral GO, developed transient visual loss right after the infusion. We can hypothesize that RTX can cause rapid edema of orbital tissue with consequent volume shift due to massive lysis of B cells mostly occurring in patients with a large intra-orbital lymphocytic infiltration. The fact that the two Italian patients had unilateral orbital involvement is consistent with this hypothesis. On the other hand, the occasionally transient nature of this phenomenon poses questions on the interpretation of this as a ‘classical DON’, at least in some patients. The results of the 52-week data on the patients that continued per protocol suggest a possible subgroup that would benefit. If we were able to identify the individuals that are unlikely to develop DON and/or side effects to RTX therapy, we might be able to increase the chance of improvement with RTX therapy to 60% for this population, which in many cases has already failed alternative therapies.
The populations treated in the 2 trials were different with younger patients, lower TRAb and shorter duration of disease prevalent in the Italian trial, all elements favoring a better response. On the other hand, smoking, usually associated with a weaker response, was also more prevalent in the Italian patients. The use of CAS might not be an ideal outcome for an intervention trial, but the combined outcome proposed by EUGOGO has revealed similar results regarding RTX efficacy, and it might be a more comprehensive outcome from a clinical perspective. The adverse events of concern relate mainly to the risk of DON, which seems to be increased by the use of RTX in a certain subset of patients. Based on available data, a multicenter trial using the EUGOGO proposed outcomes might be the next best step to define the role of RTX in GO. At the moment, RTX might be considered as second-line intervention in patients with moderate-to-severe, early and active disease that has failed initial therapy, as long as this treatment is administered in center with GO expertise where the risks and benefits of this therapy can be properly discussed, as also suggested by EUGOGO in the recently published guidelines (33).
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this review.
Dr Marius N Stan did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector. Dr Mario Salvi’s work was funded by Fondazione Cà Granda, IRCCS, Milan, Italy.
SalviMVannucchiGCampiICurroNDazziDSimonettaSBonaraPRossiSSinaCGuastellaCTreatment of Graves’ disease and associated ophthalmopathy with the anti-CD20 monoclonal antibody rituximab: an open study. European Journal of Endocrinology200715633–40. (doi:10.1530/eje.1.02325)
SalviMVannucchiGCampiIRossiSBonaraPSbrozziFGuastellaCAvignoneSPirolaGRatigliaREfficacy of rituximab treatment for thyroid-associated ophthalmopathy as a result of intraorbital B-cell depletion in one patient unresponsive to steroid immunosuppression. European Journal of Endocrinology2006154511–517. (doi:10.1530/eje.1.02119)
VannucchiGCampiIBonomiMCovelliDDazziDCurroNSimonettaSBonaraPPersaniLGuastellaCRituximab treatment in patients with active Graves’ orbitopathy: effects on proinflammatory and humoral immune reactions. Clinical and Experimental Immunology2010161436–443. (doi:10.1111/j.1365-2249.2010.04191.x)
SalviMVannucchiGCurroNCampiICovelliDDazziDSimonettaSGuastellaCPignataroLAvignoneSEfficacy of B-cell targeted therapy with rituximab in patients with active moderate to severe Graves’ orbitopathy: a randomized controlled study. Journal of Clinical Endocrinology and Metabolism2015100422–431. (doi:10.1210/jc.2014-3014)
BartalenaLKrassasGEWiersingaWMarcocciCSalviMDaumerieCBournaudCStahlMSassiLVeronesiGEfficacy and safety of three different cumulative doses of intravenous methylprednisolone for moderate to severe and active Graves’ orbitopathy. Journal of Clinical Endocrinology and Metabolism2012974454–4463. (doi:10.1210/jc.2012-2389)
van VollenhovenRFEmeryPBinghamCO3rdKeystoneECFleischmannRMFurstDETysonNCollinsonNLehanePB. Long-term safety of rituximab in rheumatoid arthritis: 9.5-year follow-up of the global clinical trial programme with a focus on adverse events of interest in RA patients. Annals of the Rheumatic Diseases2013721496–1502. (doi:10.1136/annrheumdis-2012-201956)
EcksteinAKPlichtMLaxHNeuhauserMMannKLederbogenSHeckmannCEsserJMorgenthalerNG.Thyrotropin receptor autoantibodies are independent risk factors for Graves’ ophthalmopathy and help to predict severity and outcome of the disease. Journal of Clinical Endocrinology and Metabolism2006913464–3470. (doi:10.1210/jc.2005-2813)
MitchellALGanEHMorrisMJohnsonKNeohCDickinsonAJPerrosPPearceSH.The effect of B cell depletion therapy on anti-TSH receptor antibodies and clinical outcome in glucocorticoid-refractory Graves’ orbitopathy. Clinical Endocrinology201379437–442. (doi:10.1111/cen.12141)
BartalenaLBaldeschiLBoboridisKEcksteinAKahalyGJMarcocciCPerrosPSalviMWiersingaWM.The 2016 European thyroid association/European group on Graves’ orbitopathy guidelines for the management of Graves’ orbitopathy.European Thyroid Journal201659–26. (doi:10.1159/000443828)