Abstract
Objective Tocilizumab (TCZ) is effective for giant cell arteritis (GCA). However, little is known regarding treatment modification and clinical outcomes after unfavorable events such as GCA relapses or TCZ discontinuation due to adverse events (AEs).
Methods This multicenter retrospective study included patients with GCA who initiated TCZ from 2008 to 2021 at 5 Japanese hospitals. GCA relapses and TCZ-related AEs were monitored for 2 years after TCZ initiation. In patients with GCA relapses, subsequent clinical courses, including relapse symptoms and treatment modification, were followed for 90 days after the relapses. Similarly, patients who discontinued TCZ because of AEs were additionally followed until 1 year after the TCZ discontinuation to evaluate AEs, relapses, and treatment changes.
Results Of 62 eligible patients, 10 patients (16%) relapsed after initiating TCZ therapy. Most relapses (8 of 10) occurred after extending TCZ intervals or discontinuing TCZ. Combinations of adjusting TCZ intervals, adjusting glucocorticoid (GC) dose, and/or adding or increasing methotrexate (MTX) therapy could manage the relapses without serious complications. In the entire cohort, AEs occurred in 28 patients (45%), and 8 patients (13%) discontinued TCZ because of AEs. After AE-related TCZ discontinuation, 6 patients attempted to taper GCs without other immunosuppressive therapy (IST), and 4 subsequently relapsed. In contrast, 2 patients who used other IST or biologic therapy could decrease GCs without relapses.
Conclusion Although GCA relapses can occur after initiating TCZ therapy, most relapses can be safely managed by adjusting TCZ, GC, and/or MTX doses. Adding IST or biologic treatments may potentially be related to preventing relapses when patients discontinue TCZ because of AEs.
Giant cell arteritis (GCA) is a large-vessel vasculitis that potentially leads to irreversible visual impairment.1,2 Glucocorticoids (GCs) have been a primary treatment for GCA. Combination therapies with immunosuppressants or biologics are employed to taper GCs and mitigate long-term GC-related adverse events (AEs).3-5
The effectiveness of tocilizumab (TCZ), an interleukin-6 receptor inhibitor, for GCA was first shown in a retrospective multicenter study in 2015.6 Then, the phase III Giant-Cell Arteritis Actemra (GiACTA) trial demonstrated the efficacy of TCZ in patients with GCA.7 TCZ has been approved in many countries and is becoming a mainstay of treatment for GCA, in addition to GCs.5 Previous observational studies also demonstrated the effectiveness and safety of TCZ in daily practice.8-11 Despite this, these studies found up to 30% of patients may have a GCA relapse under TCZ treatment, and > 10% of patients need to discontinue TCZ because of AEs.7,9,11,12
However, in patients with GCA who initiated TCZ therapy, little information has been available regarding treatment modification and subsequent clinical courses after unfavorable events, such as GCA relapses and TCZ discontinuation because of AEs. Further, only a few studies have been conducted on the effectiveness and safety of TCZ in a real-world setting. Long-term evidence on TCZ has been particularly limited in Japanese patients with GCA because the GiACTA trial did not include Japanese patients,7 and the observational period of the Japanese phase IV study was only up to 1 year.11
Hence, we conducted a multicenter observational study in patients with GCA who initiated TCZ to reveal the clinical outcomes and treatment modifications after GCA relapses and TCZ discontinuation due to AEs. In addition, we aimed to provide longer-term effectiveness and safety data on TCZ in Japanese patients with GCA in a real-world setting.
METHODS
Study design, setting, and patients. We conducted a retrospective multicenter study at 5 secondary or tertiary care hospitals in Japan (St. Luke’s International Hospital, Kameda Medical Center, Suwa Central Hospital, Chubu Rosai Hospital, and Daido Hospital). We retrospectively reviewed the electronic medical records of patients with GCA who were treated with TCZ from January 2008 to July 2021. The index date was the day of TCZ initiation. To evaluate relapses and AEs, the patients were followed for 2 years after TCZ initiation, until the visit closest to 2 years from the index date. Patients were additionally followed if they had unfavorable events, such as GCA relapses and TCZ discontinuation because of AEs, to reveal subsequent clinical courses. The patients with GCA relapse were additionally followed for 90 days after the relapses. Similarly, patients who discontinued TCZ because of AEs were followed for 1 additional year from TCZ discontinuation.
We included patients who were diagnosed with GCA and who initiated TCZ in our institutions from January 2008 to July 2021. The diagnosis of GCA was based on clinical diagnosis by treating rheumatologists; board-certified rheumatologists confirmed the diagnosis during a retrospective chart review. Patients were excluded from this study if their follow-up after TCZ initiation was < 6 months. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.13
Data collection. We retrospectively reviewed medical records, and rheumatologists at each institution collected the patient information. At baseline, the information included age, sex, GCA symptoms (headache, eye involvement, ischemic optic neuropathy, jaw claudication, scalp tenderness, fever, weight loss, upper or lower limb claudication, polymyalgia rheumatica [PMR], and other symptoms the treating physicians considered related to GCA), results of temporal artery biopsy, large-vessel involvement on thoracic and abdominal/pelvic computed tomography (CT) and positron emission tomography/CT (PET/CT), reasons for initiating TCZ, details of TCZ therapy (route of administration, dose, and intervals), previous and concomitant immunosuppressive therapy (IST; GC doses equivalent to prednisolone and immunosuppressants). After baseline, relapses and AEs were monitored for 2 years after TCZ initiation.
Relapse of GCA and subsequent clinical course. A relapse of GCA was defined as the recurrence of clinical symptoms and/or elevated inflammatory markers that a treating physician attributed to GCA and necessitated intensified treatment. In the patients experiencing relapse, the following information was additionally collected on the visit of the relapse and at days 30, 60, and 90 after relapse: symptoms of relapsed GCA, details of TCZ therapy, GC dose, and concomitant medications.
AEs and TCZ discontinuation. The AEs of interest included death, severe infection (requiring inpatient care or intravenous [IV] antibiotics), nonsevere infection, hematologic abnormalities, injection-site reactions, and cardiovascular events. Hematologic abnormalities included neutropenia (neutrophil counts < 1000/μL), severe neutropenia (neutrophil counts < 500/μL), thrombocytopenia (thrombocyte counts < 100,000/μL), and severe thrombocytopenia (thrombocyte counts < 50,000/μL). Other AEs that the treating physician considered important were also recorded. We evaluated whether patients discontinued TCZ because of the AEs. In patients who discontinued TCZ because of AEs, we collected additional information, including treatment modification, subsequent AE courses, and relapses, on the visit of TCZ discontinuation and at days 30, 60, 90, 180, and 365 after TCZ discontinuation.
Statistical analyses. Patient characteristics at TCZ initiation were described using summary statistics,. including the number with percentage and median with IQR. The Kaplan-Meier curve was used to examine the relapse-free survival after TCZ initiation. Patients were followed from the index date until relapse or last visit during the 2-year follow-up, whichever occurred first. In patients with GCA relapse, we summarized individual clinical courses after the relapses, including information on treatment modification. Then, we summarized the total AEs and AEs that led to TCZ discontinuation using descriptive statistics. In patients who discontinued TCZ due to AEs, we described individual clinical courses in detail up to 1 year after TCZ discontinuation. All analyses were performed in complete cases using Stata (Version 17.1, StataCorp).
RESULTS
Characteristics of patients at TCZ initiation. Sixty-two eligible patients were identified (Table 1). At TCZ initiation, the median age (IQR) was 74 years (69-79 yrs), and 40 patients (65%) were female. Temporal artery biopsy was performed on 45 patients (73%), of which 36 patients (58%) were positive. Among 56 patients (90%) who underwent thoracic and abdominal/pelvic CT and/or PET/CT at the time of GCA diagnosis, large-vessel involvement was detected in 29 patients (47%). At the time of GCA diagnosis, 56 patients (90%) fulfilled the 2022 American College of Rheumatology/European Alliance of Associations for Rheumatology classification criteria for GCA.14 The median (IQR) time from GCA diagnosis to TCZ therapy was 53 days (15-220 days). TCZ was initiated in 45 patients (73%) during remission induction therapy or during GC tapering before relapses. The remaining 17 patients (27%) added TCZ because of GCA relapse. In these patients, the median (IQR) GC dose at relapse was 10.0 mg/d (3.0-15.0 mg/d). Before TCZ, 37 patients (60%) had received IST in addition to GCs.
Patient characteristics at TCZ initiation (N = 62).
At TCZ initiation, the main clinical features included PMR (n = 16 [26%]), headache (n = 6 [10%]), and jaw claudication (n = 5 [8%]). Fifty-six patients (90%) received subcutaneous (SC) TCZ, and 6 (10%) received IV TCZ. The median (IQR) dose of IV TCZ was 8.0 mg/kg (7.7-8.2 mg/kg) and the median (IQR) interval of IV TCZ was 4 weeks (4-4 wks). All patients received concomitant GCs, with a median (IQR) dose of 25.0 mg/d (12.5-45.0 mg/d). Concomitant IST included methotrexate (MTX) in 17 patients (27%) and mizoribine (MZR) in 7 patients (11%). MZR is an immunosuppressant with a mechanism of action similar to that of mycophenolate mofetil. MZR is effective in autoimmune diseases, including PMR,15-18 with a favorable safety profile in elderly patients.19 Five patients (8%) received abatacept (ABA) prior to TCZ. At that time, neither drug was approved for GCA, but treating physicians deemed ABA safe for elderly patients. One patient (2%) was treated with upadacitinib before TCZ as part of a clinical trial. The patients were followed for a median (IQR) of 726 days (394-744 days) after TCZ initiation.
GCA relapse and AE-related TCZ discontinuation. Of the 62 patients, 14 (23%) had unfavorable events, including GCA relapse in 10 (16%) and AE-related TCZ discontinuation in 8 (13%). Four patients overlapped between the 2 groups, and the remaining 48 patients (77%) did not experience these unfavorable events (Figure 1).
The number of patients with GCA relapse and TCZ discontinuation due to AEs. AE: adverse event; GCA: giant cell arteritis; TCZ: tocilizumab.
GCA relapse and subsequent clinical courses. Among the 10 patients (16%) with GCA relapse after TCZ therapy (Figure 2), 2 patients relapsed while using TCZ weekly, and the remaining 8 patients relapsed after TCZ discontinuation or extending TCZ dose intervals. The median GC dose (IQR) at relapse was 3.5 mg/d (0-6.5 mg/d). Symptoms and signs at relapse included PMR in 60% (n = 6), ocular symptoms (amaurosis fugax) in 10% (n = 1), scalp and temporal pain in 10% (n = 1), cough in 10% (n = 1), and elevated C-reactive protein (CRP) in 10% (n = 1). Among the 17 patients with prior relapse history, 3 (18%) relapsed after TCZ therapy. Of the 45 without prior relapse history, 7 (16%) experienced a relapse after initiating TCZ. At relapse, 6 of 10 patients (cases 4-6 and 8-10 in Table 2) had elevated CRP levels in addition to clinical symptoms. In 2 patients (cases 5 and 7 in Table 2), relapses were confirmed through joint ultrasound and thoracic and abdominal/pelvic CT scans. In the remaining 3 cases, relapses were diagnosed based solely on clinical symptoms and physical examination. In the 6 patients diagnosed with PMR at relapse, 4 (cases 5 and 7-9 in Table 2) exhibited PMR symptoms along with elevated CRP levels and/or consistent imaging findings.
Relapse-free survival after tocilizumab initiation in patients with giant cell arteritis.
Clinical courses after relapses of giant cell arteritis.
Treatment modifications after the relapses were as follows (Table 2). In 2 patients on weekly SC TCZ, MTX was added or increased in addition to increasing GC doses (cases 1 and 2 in Table 2). In 2 patients who extended the SC TCZ interval because of well-controlled disease, shortening the TCZ intervals without resuming GC led to an improvement in relapses (cases 3 and 4 in Table 2). In another 2 patients who held TCZ because of other reasons (fracture or leukopenia by other medication), resuming TCZ with increasing GC managed relapses (cases 5 and 6 in Table 2). In the 4 patients who were unable to use TCZ because of AEs, relapse symptoms improved after increasing GC doses with or without adding IST (cases 7-10 in Table 2). However, 1 patient subsequently experienced a relapse, which led to the initiation of ABA. Despite treatment, the patient experienced recurrent relapses, requiring additional IST (case 7 in Table 2). In most cases, symptoms or signs of GCA improved within 30 days after treatment modification. Most patients (8 of 10 patients) achieved low-dose GC (≤ 5 mg/d) during the 90-day follow-up after relapses.
AEs, TCZ discontinuation, and subsequent clinical courses. There were 39 AEs in 28 patients (45%). Infections (n = 14 [23%]) and hematologic abnormalities (n = 12 [19%]) were common AEs. No deaths occurred during the study period. The following 8 AEs (13%) led to TCZ discontinuation: 2 severe infections, 2 hematologic abnormalities, 1 injection-site reaction, and 3 other AEs (Table 3). Because of AEs, 4 patients discontinued TCZ within 6 months, and the remaining 4 patients discontinued TCZ between 6 months and 12 months. Patients who were able to continue TCZ for 12 months thereafter tolerated TCZ.
Adverse events after initiating TCZ.
We examined individual clinical courses of the 8 patients who discontinued TCZ because of AEs (Table 4). AEs improved by discontinuing TCZ in 7 patients. Elevated liver enzymes in 1 patient did not improve for at least 1 year after discontinuing TCZ. The patient had originally been diagnosed with fatty liver, which could be a cause of elevated liver enzymes. After TCZ discontinuation due to AEs, 6 patients (cases 7-12 in Table 4) tapered GCs without other ISTs. However, 4 of the 6 patients (cases 7-10 in Table 4) relapsed with a mean GC dose of 4.5 mg/d (range 3-8 mg/d). Another patient (case 11 in Table 4) who slowly decreased GC from 10 mg/d to 7 mg/d over 1 year had no relapse during the follow-up. Another patient (case 12 in Table 4) successfully discontinued GC without adding ISTs. In the remaining 2 patients (cases 13 and 14 in Table 4), alternative biologic therapy or concomitant IST was used after TCZ discontinuation because of AEs. One patient (case 13 in Table 4) switched TCZ to sarilumab because of injection-site reactions and thereafter successfully discontinued GC without relapses. The remaining patient, who used concomitant MTX (case 14 in Table 4), decreased GCs from 6 mg to 4 mg over 1 year without relapse.
Clinical courses of patients who discontinued TCZ due to AEs.
DISCUSSION
This retrospective multicenter study examined the longer-term effectiveness and safety of TCZ over 2 years in Japanese patients with GCA in a real-world setting. Of 62 patients, 14 (23%) had unfavorable events, such as GCA relapse and/or AE-related TCZ discontinuation. GCA relapse occurred in 10 patients (16%) after TCZ therapy, but we found most relapses (8 of 10 relapses) occurred after TCZ dose spacing or TCZ discontinuation for any reason. Combinations of adjusting TCZ, GC, and/or MTX doses managed the relapses without serious consequences, and most relapsed patients thereafter achieved remission with low-dose GC (≤ 5 mg/d) therapy. Eight patients (13%) discontinued TCZ due to AEs. After that, GCA relapses subsequently occurred in 4 of 6 patients who decreased GC without alternative ISTs. In contrast, the remaining 2 patients who used alternate biologics or ISTs had no GCA relapses during a year of follow-up.
The rates of relapses and AEs after TCZ initiation in this study were comparable to those reported in previous studies. Our present study found that 16% of patients with GCA relapsed during 2 years of follow-up after initiating TCZ therapy. Our results are similar to those in previous studies (5.9-30%).9,11 Another previous report showed a relapse rate of 60%, higher than in this study, possibly because of more prior relapses, lower TCZ doses, and different relapse definitions.20 PMR was the most common manifestation of GCA relapses and was observed in 6 of 10 (60%) relapsed patients in this study. Previous reports also noted PMR was a common manifestation of GCA relapse with or without TCZ therapy.10,11,21 Our present study detected AEs in 28 patients (45%), and infections and hematologic abnormalities were common AEs. The proportion of patients with severe infections (11%) in this study was consistent with previous reports (6-19%).7,10-12 The proportion of patients with AE-related TCZ discontinuation in the current study (13%) was in line with the proportion of 12.6% to 14.9% in other observational studies,12,20 but higher than the proportion of 6% in the GiACTA trial.7 This greater discontinuation rate of TCZ can be explained by the observational study setting and the older age of included patients in our study and in other observational studies.
This study uniquely examined treatment modifications and clinical courses after GCA relapses and TCZ discontinuations due to AEs. This information is clinically relevant but has not been fully studied.22 Current guidelines have no clear recommendations for managing these patients.5 Our present study found GCA relapses could be managed by combinations of adjusting GC, TCZ, and/or MTX doses without serious complications. Only 1 patient had recurrent relapses, and most relapsed patients achieved remission with low-dose GC (≤ 5 mg/d) therapy. In patients with GCA relapses after TCZ therapy, subsequent treatment differed based on the treatment at GCA relapse. Patients who relapsed while on full-dose TCZ were successfully treated by increasing GCs and adding MTX. This result aligns with a previous study that demonstrated the effectiveness of MTX in 134 patients with GCA.23 The previous study found that adding MTX to TCZ appears to allow a higher rate of sustained remission, even in patients with longer disease duration, greater involvement of large extracranial vessels, and elevated acute-phase reactants.23 Impressively, only shortening TCZ intervals could manage relapses in those who had extended TCZ intervals and discontinued GCs because of well-controlled disease. Resuming TCZ with increasing GC could be a treatment option in patients who had discontinued but can tolerate TCZ. For those who discontinued TCZs because of AEs, physicians may increase GCs while adding immunosuppressive agents, such as MTX. As for GCs, most patients were treated with a mild GC increase of up to 10 mg/d for GCA relapses, except for 1 patient who required a GC of 30 mg/d for amaurosis fugax.
In patients with GCA who discontinued TCZ because of AEs, this study found that attempting to taper the GC without other ISTs frequently led to GCA relapses. Approximately two-thirds (4 of 6 cases) of such patients had a subsequent GCA relapse. In contrast, 2 patients who used sarilumab or MTX successfully decreased GCs without relapses. Although the number of patients was limited, these results might suggest a potential benefit of adding ISTs or alternative biologic treatments to prevent relapses when TCZ is discontinued due to AEs. In this study, 1 patient switched biologics from TCZ to sarilumab because of an injection-site reaction, and the patient successfully discontinued GCs without GCA relapses. Sarilumab showed its efficacy and safety in GCA in a randomized controlled trial.24 Sarilumab can be a potential treatment option for patients with GCA who have TCZ-related AEs. However, to our knowledge, the safety of sarilumab in patients with previous TCZ-related AEs has not been thoroughly investigated. Further studies on this topic are warranted.
The overall number of patients in this study was relatively small, but we conducted a multicenter study. Further, GCA is very rare in Asian countries,1 with an estimated prevalence of 0.003% in Japan.25 The number of GCA relapses and TCZ discontinuations due to AEs was also limited. However, we collected additional data and summarized the detailed clinical courses of the patients individually. We used data from 2008 to 2021, but little is known about the use of TCZ for GCA prior to 2017. Consequently, some heterogeneity in clinical practice may exist in this study. However, the majority of patients (81%) initiated TCZ treatment after 2017. GCA was clinically diagnosed by treating physicians, and board-certified rheumatologists confirmed the diagnosis by chart review. Of the 10 cases deemed to relapse, 3 patients were diagnosed based only on clinical symptoms and physical examination, with CRP levels tested but not elevated, and without imaging studies.
In conclusion, in this multicenter retrospective study, GCA relapses and AEs in Japanese patients with GCA who were treated with TCZ were comparable with previous studies. Most GCA relapses occurred after extending TCZ dose intervals or TCZ discontinuation. These relapses could be safely managed by combinations of adjusting GC, TCZ, and/or MTX doses. When patients with GCA discontinued TCZ due to AEs, tapering GCs without other ISTs frequently led to GCA relapses. Using other ISTs or biologic treatments can potentially prevent these relapses. More studies are needed to find potential treatment strategies in patients with GCA who had unfavorable events after TCZ therapy, such as GCA relapses and AE-related TCZ discontinuation. In particular, prospective studies with different treatment strategies to manage these patients are warranted in the future.
Footnotes
↵F.N. Nagase and S. Fukui contributed equally to this study as co-first authors.
CONTRIBUTIONS
SF, NT, TY, NO, HI, TI, MW, MS, YH, RR, MM, AN, and HT collected the data. FNN, SF, YS, AN, and EU proposed the creation of figures and tables. All authors analyzed and interpreted the patient data. All authors reviewed and approved the final manuscript.
FUNDING
This study was conducted without any grants or financial support from industries.
COMPETING INTERESTS
NT received personal fees from AstraZeneca, Sanofi, Chugai, AbbVie, Tanabe-Mitsubishi, GSK, Pfizer, Eisai, Eli Lilly, Taisho, and Novartis. AN received personal fees from Asahi Kasei, Eisai, and Tanabe-Mitsubishi. HT received personal fees from AstraZeneca, Kissei, Asahi Kasei, Sanofi, Eisai Chugai, AbbVie, Ono, Kyowa-Kirin, Takeda, Astellas, Tanabe-Mitsubishi, GSK, Pfizer, Daiichi Sankyo, Eisai, Eli Lilly, and Ayumi. The remaining authors declare no conflicts of interest relevant to this article.
ETHICS AND PATIENT CONSENT
This study was approved by the institutional review board of St. Luke’s International Hospital (number: 20-R005) and each participating institution. We obtained informed consent by using an opt-out approach.
- Accepted for publication October 3, 2024.
- Copyright © 2025 by the Journal of Rheumatology
