Economic Evaluation of Sarilumab in the Treatment of Adult Patients with Moderately-to-Severely Active Rheumatoid Arthritis Who Have an Inadequate Response to Conventional Synthetic Disease-Modifying Antirheumatic Drugs

Rheumatoid arthritis (RA) is a chronic, progressive autoimmune condition characterized by joint damage, stiffness and swelling [1] that affects up to 1.5 million adults in the US [2]. The chronicity of RA and the need for long-term treatment create significant lifetime humanistic burden to patients, with up to 43.5% of patients with RA facing activity limitations due to their condition [3]; epidemiologic and economic burdens of RA add up to $39 billion in annual costs to society [4‐7].

While appropriate management of RA is critical for modifying the course of joint damage and reducing patient morbidity, and subsequently, economic and humanistic burden, the focus of RA treatment has moved away from managing symptoms to a strategy that targets the underlying inflammation to prevent disease progression [8]. Beginning with early disease onset, treatment with one or more conventional synthetic disease-modifying antirheumatic drugs (csDMARDs; e.g., methotrexate, sulfasalazine, leflunomide, hydroxychloroquine) in addition to glucocorticoids forms the standard of care in RA [9, 10]. For patients with an inadequate response or intolerance to csDMARDs (csDMARD-IR), the addition of a targeted DMARD to a csDMARD is a global recommendation [9, 10]. Biologic DMARDs (bDMARDs) comprise tumor necrosis factor-α inhibitors (TNFi), T cell costimulatory inhibitors, anti-B cell agents and anti-interleukin-6 receptor (anti-IL-6R) monoclonal antibodies. Targeted synthetic DMARDs (tsDMARDs) of the janus kinase inhibitor class are also available for the management of RA.

Sarilumab (Kevzara^®) is a human monoclonal antibody directed against the anti-IL-6Rα. The efficacy and safety of sarilumab have been evaluated in both monotherapy and combination therapy for the treatment of moderate-to-severely active RA in patients who have had an inadequate response or intolerance to one or more DMARDs [11‐14].

In developing optimal treatment pathways for RA with the availability of sarilumab, evidence of both the clinical effectiveness and economic consequences of this treatment against relevant comparators is considered by payers and clinicians. The present study evaluated the cost-effectiveness of sarilumab subcutaneous (SC) 200 mg in combination with methotrexate, in accordance with its approved indication in the treatment of adult patients with moderate-to-severely active RA who have had an inadequate response to methotrexate. This study was conducted from a US healthcare payer perspective and aimed to provide decision makers information on the cost-effectiveness of sarilumab in combination with methotrexate. Sarilumab in combination with methotrexate was compared against csDMARD active treatment or approved bDMARDs including adalimumab SC, certolizumab SC, etanercept SC, golimumab SC, tocilizumab SC and the tsDMARD, tofacitinib, all administered in combination with methotrexate.

The cost-effectiveness analysis was conducted using a lifetime Markov state transition model with individual patient simulation (IPS). The model was developed in Microsoft Excel^® (version 2013). Health outcomes were measured in quality-adjusted life-years (QALYs), since RA affects both mortality and morbidity. The target population was patients with an inadequate response to methotrexate (i.e., csDMARD-IR population) and therefore eligible for first-line combination treatment with a targeted DMARD, including sarilumab.

Individual profiles of the patients simulated in the model were based upon profiles of patients enrolled in MOBILITY, a pivotal phase 3 trial of sarilumab SC in combination with methotrexate [11]. In MOBILITY, adult patients fulfilling the 1987 American College of Rheumatology (ACR) classification criteria for RA [1, 15] with moderate-to-severe RA were included; other inclusion and exclusion criteria are reported elsewhere [11]. Baseline demographic information of the individual patient simulation cohort is presented in Table 1. In total, 1197 eligible patients were randomized (1:1:1) to sarilumab SC 150 mg or 200 mg or placebo SC every 2 weeks (q2w) added to methotrexate. Patient ages ranged from 19 to 75 years (mean 50.6 ± 11.6); 81.6% were female and 86.1% were White/Caucasian. Duration of RA ranged from 0.3 to 44.7 years (mean 9.0 ± 7.8), and baseline Health Assessment Questionnaire Disability Index (HAQ-DI) scores ranged from 0 to 3.0 (mean 1.6 ± 0.6) [11].

This article does not contain any studies with human participants or animals performed by any of the authors.

Assuming all treatment classes were ascribed equivalent time to treatment discontinuation after 6-month response, average time on first-line treatment based on response ranged from 3.5 years (tofacitinib) to 5.0 years (etanercept) and was lowest at 2.3 years for the csDMARD active treatment sequence. Sarilumab patients experienced a longer time on first-line therapy (4.3 years) versus all comparators except etanercept (5.0 years) because of a relatively higher response rate (Table 4).

QALYs ranged from 3.43 for the active csDMARD treatment sequence to 5.79 for sarilumab and 5.94 for etanercept treatment sequences. Lifetime costs ranged from $115,019 for the active csDMARD to $524,832 for etanercept treatment sequences. By achieving lower costs and higher effectiveness in first-line treatment, sarilumab dominated adalimumab, certolizumab, golimumab and tofacitinib treatment sequences. Sarilumab was associated with ICERs of $84,079 and $134,286 compared with tocilizumab and csDMARD active treatment sequences, respectively. The substantially higher costs and minimal gain in effectiveness of etanercept versus sarilumab would result in an ICER > $500,000 (Table 4). csDMARD active treatment, sarilumab and etanercept were the three drug sequences that were not dominated or extendedly dominated, and therefore were positioned on the cost-effectiveness frontier (Table 4, Fig. 2a).

From a societal perspective, the sarilumab sequence would result in fewer workdays lost because of better treatment response and longer treatment duration versus most of the comparator sequences (differences ranging from − 10 to − 21 days versus the other bDMARDs and tofacitinib; − 370 days versus csDMARD active treatment); however, versus etanercept, sarilumab was associated with 25 additional workdays lost (Table 4). Therefore, the ICERs remained stable in the scenario analyses that considered societal costs (i.e., workdays lost) (Table 5). Across other scenarios, results were sensitive to the model time horizon and treatment discontinuation (Table 5). OWSA of sarilumab versus tocilizumab revealed that results were most sensitive to sarilumab dose and dosing intensity as well as tocilizumab dose (Fig. 2b).

Simulations from PSA for sarilumab versus tocilizumab were slightly to the right-hand side of the origin on the cost-effectiveness plane, indicating comparable costs and slightly better QALYs for sarilumab (Fig. 2c). CEAC indicated that, irrespective of threshold, the probabilities of sarilumab and tocilizumab SC being cost-effective were close to equivalent (Fig. 2d).

Findings from this cost-effectiveness analysis indicate that a treatment sequence initiated with sarilumab can be cost-effective compared with sequences beginning with any other bDMARD, the tsDMARD tofacitinib or csDMARD active treatment for moderate-to-severely active RA adults who have had an inadequate response to methotrexate. The sarilumab + methotrexate treatment sequence was on the efficiency frontier and compared with tocilizumab resulted in an ICER within acceptable ranges as per thresholds suggested by the World Health Organization of up to three times the per capita gross domestic product [36] across a range of diseases and populations with thresholds ranging from $129,000 to $150,000 [33, 34] and per an updated threshold suggested for the US [37] of up to $150,000.

Versus the sequence that begins with csDMARD active treatment, higher response rates and longer time on first-line treatment for sarilumab (63.6% on ACR20; 4.3 years, respectively) combined with methotrexate resulted in lower HAQ-DI scores. As such, patients in the sarilumab sequence attained reduced mortality and higher levels of health-related quality-of-life. Sarilumab, compared with adalimumab, certolizumab, golimumab and tofacitinib, each combined with methotrexate, was the dominant strategy (i.e., more effective and less costly). This outcome was due to increased QALYs for sarilumab derived from higher response rates coupled with lower drug and administration costs, directly and indirectly from the shorter time subsequently spent on some of the more costly treatment lines.

The various scenario analyses indicated that the model was robust and economic efficiency of sarilumab was upheld from the societal perspective in terms of reduced workdays lost associated with treatment relative to most comparators. Only the scenarios with short time frames and discontinuation rates based on drug class were of consequence. Reasonably long model horizons are necessary for differences in duration of treatment to fully accrue; therefore, shorter horizons did not allow the benefits associated with increased time on sarilumab treatment to fully emerge, reducing the QALYs and the economic efficiency of sarilumab. However, the scenario that applied discontinuation rates ascribed to the IL-6 class compared with the relatively lower duration of treatment for the TNFi drug class enhanced the treatment duration for sarilumab and provided increased benefits and reduced costs. It is notable that a very conservative assumption to the model was applied in the base case whereby an equivalent rate of treatment discontinuation was applied for all treatment classes, despite the potential increased benefit on this parameter specifically for IL6 inhibitors according to the RHUMADATA analysis.

In terms of cost-effectiveness, sarilumab was dominant versus certolizumab, golimumab, tofacitinib and adalimumab. While etanercept was more effective, it was also more expensive, producing an ICER of $609,545 versus sarilumab. These findings are in distinct contrast to the recent evidence appraisal of RA biologics published in the report Targeted Immune Modulators for Rheumatoid Arthritis: Effectiveness and Value issued by the Institute for Clinical and Economic Review [18]. The discrepancy in results may be explained by the following two factors. First, the economic models in RA are highly sensitive to the manner in which the ACR response is converted to a utility score, the main driver of lifetime QALYs. The Institute for Clinical and Economic Review model employed a published utility function [31], but deviated from the original equation by putting all coefficients into exponential form, an approach that diverges from other published equations [38, 39]. This reduced the utility benefit associated with an HAQ-DI decrease resulting from, for example, ACR20 response from approximately 0.29 in the original equation to approximately 0.09 in the revised exponential equation, resulting in a substantially lower QALY value compared with the present analyses.

Second, the ICER model assumed a relatively slow HAQ-DI progression among patients treated with csDMARD palliative treatment of 0.0269 per year; furthermore, this increase was only applied in the first 15 years, at which point RA was assumed to no longer be a progressive disease. Patients in the csDMARD active treatment sequence were ascribed an HAQ-DI score of 1.7 and could only progress to a maximum score of 2.1 (1.7 + 15 × 0.0269). Given that the HAQ-DI score in turn informs both costs and QALYs, a lower than expected lifetime HAQ-DI for csDMARD active treatment artificially enhanced its economic value compared with sequences beginning with targeted DMARDs. This approach also diverges from other published RA models [38‐40]. In contrast, the present study assumed a constant HAQ-DI progression at a rate of 0.045 per year with no lifetime limits as applied in previous health technology appraisals [40]. As a result, some control patients included in the IPS reached a maximum HAQ-DI score of 3.0, which may be more reflective of the expected progressive disease burden of RA.

As for all cost-effectiveness models, the use of multiple sources and assumptions leads to important limitations. Similar to other long-term CEAs in RA [40], the limitations of the present study stem from the uncertainty inherent to the limited, and often conflicting, evidence regarding short-term treatment response rates and long-term benefits of bDMARDs. For example, the 6-month response rates are based on an NMA, which can be affected by a heterogeneous trial population and placebo arm response rates that have proven to be heterogeneous across trials. In the long-term, due to lack of long-term disease activity data, after the initial 6 months, long-term treatment discontinuation from real-world evidence (RHUMADATA registry) was used as a proxy for determining the efficacy of treatment in reducing disease activity. Data from the Canadian RHUMADATA registry were used in lieu of US data. In addition, the present model assumed a linear progression for csDMARDs, while separate studies have assumed non-linear progression [21]. Because of linear progression, worse HAQ-DI scores were modeled in the long-term, therefore resulting in relatively worse QALYs and higher costs for csDMARDs compared with a non-linear approach. This is an important limitation compared with csDMARDs; however, since the population in this study is csDMARD-IR, a comparison is only conducted to compare results with prior publications [41], most of which use linear progression. If non-linear progression was used here, the results for this comparison would most likely change.

Consistent with previously published cost-effectiveness models in RA [21], the initial 6-month treatment response was based on ACR 20/50/70 response criteria and assumed to be independent from the line of treatment. This was a conservative assumption ascribed to the model to address the limited evidence on efficacy per line of treatment; however, some evidence supports that TNFi have lower efficacy when used after an initial TNF [42, 43]. While the ACR 20/50/70 response criteria have limited clinical utility in RA, they are only included to determine if a treatment is active or not in the initial 6 months, in line with other RA cost-effectiveness studies. Additionally, extensive comparative data were only available for ACR response from the trials included.

The strengths of the current CEA include its conduct according to best practice guidelines [44]. Results of the CEA remained robust in various sensitivity analyses.

In patients with moderate-to-severe RA who are naïve to treatment with targeted DMARDs, sarilumab 200 mg SC q2w + methotrexate can be considered a cost-effective treatment option, as it results in lower costs and greater health benefit than multiple licensed bDMARD and tsDMARD treatments (adalimumab, certolizumab, golimumab and tofacitinib) and falls below commonly accepted cost-effectiveness thresholds compared with tocilizumab + methotrexate and csDMARD active treatment.