Economic consequences of sequencing biologics in rheumatoid arthritis: a systematic review

Abstract

Background and objectives:

Tumor necrosis factor-alpha (anti-TNF) blocking agents are effective for the treatment of rheumatoid arthritis (RA), with mean response rates of 60–70%. Patients with incomplete response to initial anti-TNF treatment often are switched to other biologic treatments with some success. However, little is known about whether or not switching to anti-TNF or other non-TNF biologic treatments is cost-effective. This study sought to review the economic evidence of sequencing various biologic treatments in RA.

Methods:

A systematic review was conducted of published and unpublished literature (January 2000 to October 2012) on the cost-effectiveness of sequencing biologic treatments in RA after failure of an initial biologic treatment. It included modeling and other economic studies that assessed cost-effectiveness of one or more sequences of biologics. Studies were excluded that evaluated non-biologic sequencing.

Results:

This review of the available evidence suggests that there is limited evidentiary support favoring the cost-effectiveness of switching from one anti-TNF agent to another within the anti-TNF category of biologics. This is due, in large part, to the limited clinical evidence base supporting the incremental efficacy of second- and third-line anti-TNF treatments and to variation on how and when to assess non-response to the first-line biologic. When compared to anti-TNF agents, biologic treatments with a different mechanism of action are more cost-effective as second-line agents.

Limitations:

Not all sequences and patterns of switching, either within or outside of therapeutic class, have been evaluated for clinical benefit and cost-effectiveness, limiting the interpretation of these findings.

Conclusions:

Switching from one anti-TNF agent to another after first-line treatment failure may not be a cost-effective treatment strategy. However, when non-TNF biologics are included in the sequence they are likely to be more cost-effective than anti-TNF specific cycling sequences.

Keywords::

• TNF Inhibitor
• Cost-effectiveness
• Rheumatoid arthritis
• Sequencing
• Biologics

Introduction

Rheumatoid arthritis (RA) is a chronic disease that leads to inflammation of joints, connective tissue, and may involve additional vital organs. Patients with persistently active RA (≥6 months duration that has continued despite the use of methotrexate or other disease modifying anti-rheumatic drugs) are often managed with antagonists to tumor necrosis factor-alpha (anti-TNF), a strategy supported by American College of Rheumatology guidelines1. The most widely available anti-TNF agents in the US and Europe are adalimumab, certolizumab pegol, etanercept, infliximab, and golimumab. Other biologic agents commonly used include rituximab (targeting B lymphocytes), anakinra (an interleukin-1 inhibitor), abatacept (an antibody that targets T-cell interactions), and tocilizumab (an interleukin-6 inhibitor). These biologic agents have been shown to be similarly effective when used after failure of methotrexate or other first line non-biologic disease modifying anti-rheumatic drugs (DMARDs) for the treatment of RA. Evidence of benefit of biologic treatments consistently shows a mean response rate of 60–70%2–4.

Despite reasonable first-line response rates, patients often require adjustments to their treatment regimens to achieve effective disease control. Joints may become irreversibly damaged if inflammation is not controlled, and early treatment with both non-biologic and biologic DMARDS may result in improvements in both short-term and long-term disease control5.

Patients with an incomplete response to initial anti-TNF treatment often are switched to other biologic treatments with some success6,7. Although an inadequate response to one anti-TNF agent does not predict resistance to other agents in this class, treatment failure rates may increase with successive switches, and switching to a treatment with a different mechanism of action may result in an improved response8–10. However, little is known about whether or not switching amongst anti-TNFs or other non-TNF biologic treatments is a cost-effective strategy for patients who previously failed anti-TNF-initiated therapy.

We sought to review and report the economic evidence for sequencing various biologic treatments in patients with persistently moderate-to-severe active RA who have an incomplete response to first line anti-TNF therapy.

Patients and methods

We conducted a systematic assessment of published and unpublished literature (January 2000 to October 2012) on the cost-effectiveness of sequencing biologic treatments in RA after failure of initial treatment with a TNF inhibitor. Databases searched included MEDLINE, EMBASE, the Cochrane Library, and Google Scholar. We also searched the websites of the National Institute of Clinical Excellence (NICE), the Agency for Healthcare Research and Quality (AHRQ), the Canadian Agency for Drug and Technology in Health (CADTH), and the Blue Cross and Blue Shield Association Technology Assessment Center for relevant reviews, reports, and citations.

The following search terms were used: Arthritis, Rheumatoid; Economics; Cost-Effectiveness Analysis; Quality-Adjusted Life; Receptors, Tumor Necrosis Factor; Tumor Necrosis Factor-alpha/Antagonists & Inhibitors; Failure; Inadequate Response; Switch; and Cycling. We reviewed the reference lists of papers to identify additional studies that may have been missed.

We included trial-based cost-effectiveness studies, decision-analytic or other deterministic modeling studies, and other economic studies that assessed cost-effectiveness of one or more sequences of biologics. We excluded studies that evaluated non-biologic (DMARD) sequencing as these were not relevant to our research question. From identified studies, we extracted information on treatments, dose, duration, study time horizon, clinical and cost end-points, method of economic evaluation, country of study origin, and a point estimate of the incremental cost-effectiveness ratio.

Results

Studies identified

A total of 78 articles were identified using the search criteria described previously. After reviewing the title and abstract, 28 papers were selected for a full text review. Of these, 15 articles met the full inclusion and exclusion criteria for data extraction and assessment. One of the papers11 presented the results of a systematic review of clinical and economic evidence intended to inform EULAR treatment recommendations. Because this paper did not provide new cost-effectiveness evidence, it was not included in our review. Of the remaining 14 papers, two were Health Technology Assessment (HTA) reports commissioned for the UK National Institute of Clinical Excellence (NICE)12,13. These two reports provide previously unreported cost-effectiveness estimates and are, therefore, included in this review.

One of the papers reports on a budget impact analysis of biologic sequencing in France14. While this is not a traditional cost-effectiveness analysis, we think the reader might find the results interesting. The remaining papers represent de novo evaluations of the cost-effectiveness of biologic sequencing in patients with persistently active moderate-to-severe RA who experienced an inadequate response to a trial of TNF inhibitor therapy15–25. Table 1 includes detailed information about each of the 14 reviewed studies, arrayed by date of publication12–25.

Table 1.  Summary of cost-effectiveness and budget impact of biologic treatment sequencing in adult RA.

Reference, Country	Method	Intervention sequences	Results	Economic metric
Chen et al.12, UK	Systematic review of randomized controlled trials and CEA	ADA + MTX → ADA ADA + MTX → IFX + MTX ETN + MTX → IFX + MTX ETN + MTX → ADA + MTX ETN + MTX → ETN ETN → ADA	£52,000 £140,000 £47,000 £190,000 £240,000 £240,000	ICER–Cost/QALY
Walsh et al.15, Ireland	Open label study (n = 19)	IFX → ADA	$15,570	Total costs of switching from IFX to ADA
Launois et al.14, France	Budget Impact Model	RTX only   TNFa inhibitor and RTX   TNFa Inhibitor only	€260,882,640 €290,532,754 €378,513,468	Budget impact of RTX availability
Vera-Llonch et al.16, US	Markov Model	TNFa → DMARD   TNFa → RTX	Referent $50,576	ICER–Cost/QALY
Davies et al.17, US	Patient-Level Simulation Model	DMARD ETN alone ADA + MTX IFX + MTX ADA + MTX/ETN	Referent Extended Dominance $47,157 Extended Dominance $42,727	ICER–Cost/QALY
Lindgren et al.18, Sweden	Discrete Event Simulation and Registry Data	TNF → TNF   TNF → RTX	Referent RTX dominates TNF as second line after failure of 1st Line TNF	ICER–Cost/QALY
Russell et al.19, Canada	Decision Analytic Model	Low Disease Activity (LDAS) ETN → IFX → ADA ETN → ABT → IFX ABT → ETN → IFX Remission Outcome (RS) ETN → IFX → ADA ETN → ABT → IFX ABT → ETN → IFX	Referent $12,514 Dominant Referent $16,829 Dominant	ICER–Cost/LDAS or RS
Hallinen et al.20, Finland	Markov Model	Best Supportive Care RTX → ADA RTX → ETN RTX → IFX RTX → ABT RTX → IFX RTX → IFX → ADA RTX → IFX → ETN RTX → IFX → ABT	Referent €38,235 €38,938 €32,621 €46,367 Referent €54,701 €54,836 €70,616	ICER–Cost/QALY
Saraux et al.21, France	Decision Analytic Models and Expert Opinion	Expected Number of Days (TEND) ETN → ABT → ADA ETN → RTX → ADA ETN → ADA → ABT ETN → ADA → IFX Low Disease Activity (LDAS) ETN → ABT → ADA ETN → RTX → ADA ETN → ADA → ABT ETN → ADA → IFX	€276 €310 €500 €810 €534 €741 €1,517 €2,672	Mean Cost Effectiveness Ratio
Merkesdal et al.22, Germany	Markov Model	ETN → TNFa ETN → RTX	Referent €24,517	ICER–Cost/QALY
Beresniak et al.23, Spain	Decision Analytic Model	Remission Outcome (RS) ETN → ABT → ADA ETN → RTX → ADA Low Disease Activity (LDAS) ETN → ABT → ADA ETN → RTX → ADA	Referent Dominated by referent Referent Dominated by referent	ICER–Cost/LDAS or RS
Malottki et al.13, UK	Systematic review of randomized controlled trials and CEA using BRAM	TNFa → ADA/IFX/ETN TNFa → RTX TNFa → RTX TNFa → ABT	Referent Dominant vs any TNFa Referent €130,600	ICER–Cost/QALY
Cimmino et al.24, Italy	Decision Analytic Model	Remission Outcome (RS) ETN → ABT ETN → RTX Low Disease Activity (LDAS) ETN → ABT ETN → RTX	Referent Dominated by referent Referent Dominated by referent	ICER–Cost/LDAS or RS
Puolakka et al.25, Finland	Decision Analytic Model	Remission Outcome (RS) TNFa → ABT TNFa → RTX Low Disease Activity (LDAS) TNFa → ABT TNFa → RTX	Cost-effectiveness ratio’s not reported in the manuscript	ICER–Cost/LDAS or RS

BSC, Best Supportive Care; ABT, abatacept; ADA, adalimumab; ETN, etanercept; IFX, Infliximab; MTX, methotrexate; RTX, rituximab; TNFa, anti-TNF agents; ICER, Incremental cost-effectiveness ratio; Referent, the comparison strategy; Dominant, The strategy is more effective and less costly when compared to the referent group; Extended Dominance, The principle of extended dominance is applied in incremental cost-effectiveness analysis to eliminate from consideration strategies whose costs and benefits are improved by a mixed strategy of two other alternatives; BRAM, Birmingham Rheumatoid Arthritis Model.

Methods employed

With some exception, the majority of studies involved specification and use of deterministic disease-based models. The Birmingham Rheumatoid Arthritis Model (BRAM) is a commonly used simulation tool for assessing biologics in adult RA26, particularly for NICE assessments of new drugs. This model was used in both of the NICE HTA reports.

A short-term (2 year) decision analytic model, developed by Beresniak et al.23, has been used to evaluate the cost-effectiveness of biologic sequencing in Canada, Italy, Finland, and Spain. Unlike the BRAM model, the Beresniak et al. model does not report cost-effectiveness ratios in terms of cost per quality-adjusted life years gained. Rather, they report cost-effectiveness in terms of improvements in clinical outcomes, making cross-country comparisons to the remaining literature difficult.

The results from these reports are rarely comparable because of differences in the patient population, severity of disease, setting and perspective, treatment sequences, and comparators selected.

Sequences evaluated

The middle column of Table 1 displays the intervention sequences evaluated in the published economic reports. For example, the sequence ETN + MTX → ADA + MTX translates to: a trial of first line etanercept plus methotrexate → if unsuccessful, the patient is switched to second line adalimumab plus methotrexate.

The strategies and sequences evaluated in the published reports are varied. The specific sequences in each report often reflect either a newly introduced agent in the jurisdiction represented in the analysis or the emerging treatment patterns of existing agents.

Because of our inclusion criteria, all studies begin with adult RA patients who have experienced an inadequate response to a trial of TNF inhibitor therapy. Some of the reports evaluated the switch from one TNF inhibitor (adalimumab, etanercept, infliximab) to another. Others evaluate a switch from a TNF inhibitor therapy to a non-TNF inhibitor biologic (abatacept or rituximab). The astute reader of the table will note that not all possible RA biologics are evaluated. There are gaps in the evidence base, particularly for the newer agents (certolizumab, golimumab, tocilizumab). There simply are no economic data on these new treatments when used after failure of first line TNF inhibitors.

Outcomes reported

In seven of the papers, cost-effectiveness is reported as incremental cost per QALY gained. In the four papers utilizing the Beresniak et al.23 model, cost-effectiveness is reported as either incremental cost per remission achieved or per low disease activity status achieved.

Summary of findings

The evidence from the existing literature suggests that when patients fail an adequate trial of TNF inhibitor therapy, switching to another TNF inhibitor may not be cost-effective. That is, the ICER of switching between TNF inhibitors exceeds a willingness to pay threshold beyond what is considered acceptable in the specific jurisdiction. This is primarily because the clinical evidence to support switching between TNF inhibitor therapies is not clear.

On the contrary, the economic evidence from the collected reports is generally consistent and suggests that switching from a TNF inhibitor to a non-TNF inhibitor (abatacept or rituximab) is likely to be cost-effective and may, under certain conditions, be cost-saving.

Discussion

We sought to review and report on the economic evidence and cost-effectiveness of sequencing biologic treatments in patients with persistently active moderate-to-severe RA. Our review of the evidence suggests that there is limited evidence to support the cost-effectiveness of switching from one anti-TNF agent to another within the anti-TNF class of biologics. However, switching to a biologic with a different mechanism of action may be cost-effective as a second-line treatment. Under certain conditions, non-TNF inhibitor biologic treatments may be cost-saving in certain jurisdictions. These findings are in-line with comparative effectiveness studies that have reported improved clinical response for patients switching to a different class of biologic after initial incomplete response to an anti-TNF9,10.

Our findings should be interpreted in the context of the following limitations. The clinical evidence base that supports the effectiveness of switching between biologic treatments is limited. There are very few adequately powered head-to-head studies that directly evaluate the efficacy and safety of second- or third-line biologic treatments. In addition, not all sequences and patterns of switching, either within or outside of therapeutic class, have been evaluated for cost-effectiveness.

None of the included studies evaluated how and when to assess non-response to the first-line biologic treatment, including the selection of sequences and comparators in the analysis. In addition, most publications used economic models populated with information from randomized clinical trials, despite the lack of availability of direct head-to-head evidence on effectiveness and safety of anti-TNF cycling.

When non-TNF biologics are included in the sequence they seem to be more cost-effective than anti-TNF specific cycling sequences. Reimbursement authorities who have used these analyses to inform coverage and pricing decisions seem to agree with this conclusion. Among treatments with a non-TNF mechanism of action, HTA agencies such as NICE consider rituximab to be a cost-effective treatment alternative compared to switching to another anti-TNF as a second-line biological treatment. The favorable recommendation of HTA agencies such as NICE in regards to the cost-effectiveness of rituximab seem more likely to be driven by the lower acquisition costs of rituximab rather than a superior clinical profile. These clinical and economic evidence gaps impact the veracity and availability of the cost-effectiveness evidence.

Conclusions

This systematic review of the available evidence suggests that switching from one anti-TNF agent to another due to incomplete response may not be a cost-effective treatment strategy. However, relative to within-class switching, including a non-TNF biologic in the sequence may be more likely to be cost-effective.

Transparency

Declaration of funding

This research was supported by a grant from Novartis Pharmaceuticals. J C and SR are AHRQ-funded K-12 Scholars in Comparative Effectiveness Research (K12HS019482).

Declaration of financial and other relationships

SDS is a member of the Novartis Global Experts Board. UM is an employee of and holds shares in Novartis. RA-C, JC, and SR have no financial or other conflicts to report.

Acknowledgments

The authors thank Paul Kraegel, MSW, for editorial support.