Relapse Risk and Safety of Long-Term Tocilizumab Use Among Patients With Giant Cell Arteritis: A Single-Enterprise Cohort Study

DISCUSSION

We present the largest North American single-enterprise cohort of patients with GCA treated with TCZ. Outside of randomized, placebo-controlled trials evaluating TCZ in GCA, there have been several reports of the efficacy of TCZ through case series, small prospective studies, and a few real-world cohorts. Nevertheless, the majority of these reports include 60 patients or fewer,^{12-14,16,23-32} with few studies monitoring to 2 years of TCZ treatment.^13,14,29 Comparable large cohorts of patients with GCA receiving TCZ have been reported in Spain¹⁵ and Switzerland,³³ with 134 and 186 patients, respectively. However, median time of observation on TCZ among the European cohorts was only 11 months to 12 months. Therefore, the size and duration of follow-up presented in the current report provides critical insight into longer-term safety and efficacy data of TCZ in GCA within a real-world setting and highlights the variability of clinical practice with regard to patients included and the dose and route of TCZ used.

Clinical trials evaluating TCZ in GCA have shown a 4-fold improvement in sustained remission⁷ and relapse-free survival⁶ at 52 weeks, in comparison to placebo. Uncontrolled observational studies have shown similar, but attenuated, results. The current study reaffirms the efficacy of TCZ in real-world management of GCA as evidenced by a 3-fold reduction in the ARR while receiving TCZ compared to prior to TCZ initiation. Although the frequency of relapses prior to TCZ initiation were lower than other reported studies, our findings are comparable to the ARR observed in a real-world multicenter retrospective analysis in France¹³ (n = 43), where the ARR decreased from 1.26 to 0.44 relapses/year (2.86-fold reduction), as well as comparable to a single-institution American cohort (n = 60), where the relapse rate decreased from 1.4 to 0.6 relapses/year (2.3-fold reduction).¹⁶

Tapering and discontinuation of GCs is a primary goal in GCA management. Although TCZ has demonstrated a significant steroid-sparing effect, only slightly more than half the patients in our cohort were able to stop GCs while receiving TCZ. This has been similarly observed in other cohorts, with 46% to 60% of patients able to discontinue GC therapy during TCZ treatment.^13,16,32 Given the uncontrolled and retrospective nature of this study, it is not feasible to determine the rationale for ongoing use. Possible reasons include provider/patient preference, PMR symptoms, relative adrenal insufficiency, and smoldering disease activity, among others. Our group has shown previously that GC discontinuation among patients receiving GC monotherapy was low with only 6% discontinuing at 1 year and 24% at 2 years after GCA diagnosis.³ Although TCZ improves likelihood of GC discontinuation, patients and providers should be aware that approximately half of patients may require ongoing use of GCs despite initiation of TCZ.

Because of the retrospective nature of this study, the dosing and route of TCZ was determined by the treating rheumatologist. Until recently, the only FDA-approved formulation of TCZ for GCA management was 162 mg SC weekly. However, only 42% of patients received this dose and route at TCZ initiation. This study was completed prior to the approval of the IV infusion formulation of TCZ (6 mg/kg IV monthly). Consequently, this dose and route was the least frequently used in the current cohort. Similar variability in TCZ dosing has been demonstrated in other observational cohorts,^14-16 raising the possibility that patient factors, provider preference, and medication formulation availability likely play a role in the decision of TCZ dosing. High-dose and low-dose regimens were compared to see if the decision to start TCZ was influenced by prior steroid-sparing agents, prior relapses, presence of LVV, symptoms (cranial, visual, and PMR), prednisone dose, prednisone duration, or inflammatory markers. Interestingly, none of these features appeared to be associated with the chosen dosing regimen. High-dose and low-dose regimens were also assessed for risk of relapse after the start of TCZ and after discontinuation. Neither dose nor route was found to be a risk factor for relapse at either point; unlike the results of the Giant-Cell Arteritis Actemra (GiACTA) trial, which showed weekly SC TCZ had a lower risk of flare in the patients with relapsing disease^7,11 compared to every-other-week dosing. Similar to our findings, Rossi and colleagues showed significant efficacy of lower-dose TCZ in the elderly population with relapsing disease³⁴ and a comparably large multicenter Spanish cohort has also shown clinical benefit regardless of TCZ administration route or disease duration.¹⁵ Further research into which subtypes of patients may benefit from lower-dose regimens is needed before routine use is considered.

The optimum duration of TCZ treatment remains unknown. Among those stopping TCZ, the median duration of treatment was 16.8 months. Following TCZ discontinuation, 58% of patients had a subsequent relapse within 12 months. These findings are commensurate to frequencies reported among both clinical trials and observational studies, with relapse rates of 33% to 62% after TCZ discontinuation.^{9,11,13,23,26,27,33} The time from TCZ stoppage to first flare is also similar, with a median of 8.4 months in the current cohort, compared to ranges of 2 months to 9 months in other studies.^13,23,26,27 Long-term follow-up of patients after TCZ stoppage is limited; consequently, ARR analysis following TCZ discontinuation has not been previously reported. In the current study, we observed the ARR following TCZ discontinuation to be 0.64 relapses/year, which is > 2 times higher than the ARR on TCZ and is close to pre-TCZ initiation rates. Of importance, in the current study, the duration of TCZ therapy prior to discontinuation was not associated with subsequent risk of relapse. Taken together, these findings affirm the suppressive effect of TCZ but confirm that treatment extension beyond 12 months prior to discontinuation does not necessarily increase the likelihood for prolonged treatment-free remission. Therefore, it is imperative that future research be focused on identifying the optimal duration of TCZ therapy in patients with GCA in order to determine which patients are best suited for discontinuation. Among those for which TCZ therapy is discontinued, close observation at least to 6 months to 12 months off therapy is strongly suggested to assess for disease recurrence.

Clinical features at the time of GCA diagnosis and the time of TCZ initiation have not been reliable predictors of future relapse. Prior studies have failed to demonstrate any specific baseline clinical, radiographic, or laboratory variable that predicts risk of relapse either during TCZ treatment or after TCZ discontinuation.^9,33 In the current study, the only feature associated with risk of relapse after TCZ start was the presence of cranial symptoms at time of GCA diagnosis. The pertinence of this finding is uncertain since it has not been observed in other cohorts and directly contrasts with the observation noted by Clemént and colleagues, where an absence of ischemic signs (jaw claudication, scalp tenderness/necrosis, blindness, and peripheral arterial disease) was associated with increased risk of relapse after TCZ start.¹³ Use of statin at GCA diagnosis was the only factor associated with reduced risk for relapse after TCZ discontinuation. The relevance of this is unknown as the effect of statin use and relapse has provided conflicting results and statins are not currently recommended specifically for treatment of newly diagnosed GCA unless a patient’s cardiovascular risk warrants statin initiation.⁸ The ability to prognosticate relapse risk has also been elusive in the pre-TCZ era, with individual groups identifying potential risk factors that have failed to be consistent across cohorts.^3,5,35-37 This further highlights the need for research collaboration to establish methods of novel multivariable analysis of large international cohorts and ongoing efforts to identify novel biomarkers suitable for risk stratification.

Overall, TCZ was generally well tolerated within this cohort, with only 14.9% stopping TCZ because of a study-defined AESI, despite long-term follow-up and median TCZ treatment duration of 2.3 years. Total AE and serious AE (SAE) rates are notably variable among cohorts, which is attributable to differences in study definitions.^{6,7,13,16,30,32,38} SAEs resulting in permanent discontinuation of TCZ appear to be a more reasonable indicator of overall long-term safety, with apparent greater uniformity and consensus among reporting cohorts. Reassuringly, discontinuation because of SAEs is uncommon in observational trials reporting on cohorts with ≥ 20 patients with GCA with rates between 6% to 15%.^{14-16, 30, 38} Vitiello and colleagues noted a 25% discontinuation rate because of SAEs; however, this study has limitations of generalizability because of the small cohort size (n = 12) and greater frequency of MTX use (40%) while receiving TCZ.³² Understandably, long-term biologic therapies in the elderly raise concerns for both patients and providers. We did not observe any new safety signals during our follow-up period beyond what has been previously noted. Colonic perforation was seen in only 1 patient in our cohort, and although it is important to consider and discuss this risk with patients, the low rate highlights the rarity of this event. This rarity has been reinforced by healthcare claim analysis reviewing 4804 patients with GCA showing gastrointestinal perforation rates of 0.55/100 person-years of observation.²² We did not observe age (> 80 yrs), route, or dose of TCZ or disease-modifying antirheumatic drugs as risk factors for AESIs. Visual symptoms at time of TCZ initiation was the only item found to be a risk factor for AESIs, which likely reflects a subgroup of patients treated with a higher dose of prednisone, as vision changes often require pulse dose IV GCs upfront.

To our knowledge, the present study is the largest real-world single-enterprise cohort of patients in North America with GCA treated with TCZ with extended follow-up. Nevertheless, this study must be viewed in context of its limitations. Since this was a cohort from a single enterprise, findings may not be applicable to other regions. However, patients were recruited from campuses in the midwest, southeast, and southwest United States, so we expect this effect to be limited. Given the observational, retrospective nature of this cohort, not all patients had the same follow-up duration; we were also unable to confirm proper administration and adherence to prescribed medications but rather relied on documentation of the treating clinician and confirmation of infusion records. Because of the referral nature of the practice, patients may represent those with more refractory disease and therefore may not be representative of the general GCA population. Seventeen patients with study-defined radiographic confirmation of LVV did not fulfill either 1990 ACR or 2022 ACR/EULAR classification criteria; further evaluation is needed to determine whether the treatment response to TCZ differs in this particular patient subset. Finally, because of the inclusion of patients before FDA approval of TCZ, there may be selection bias in what patients were treated with TCZ prior to this date and what dose and route was chosen.

In conclusion, this large real-world cohort with an extended duration of follow-up further affirms the safety and efficacy of TCZ in GCA. No differences were found between the groups of patients that were prescribed low- and high-dose TCZ. Additionally, neither the dose nor the route was found to be associated with risk for relapse or AESIs. Relapse rates following TCZ discontinuation were similar to other series with shorter duration of treatment. Duration of TCZ prior to discontinuation was not associated with reduced risk of relapse. Further studies are needed to determine the optimal duration of TCZ therapy. Additional research is necessary to identify which subsets of patients are more likely to remain in prolonged remission following TCZ discontinuation.