Older anti-TNF-α agents: why not group them for common indications in the EU?

Key words:

• anti-TNF-α
• rheumatoid arthritis
• psoriasis
• economic evaluation
• biosimilars

Introduction

Tumor necrosis factors (TNF) are a group of cytokines that can cause cell death by apoptosis. TNF-α, the best-known member of this class, has a central role in the pathogenesis of various immune-mediated diseases such as rheumatoid arthritis (RA), ankylosing spondylitis, psoriasis (Ps), psoriatic arthritis and Crohn’s disease.[1] Consequently, in the last decades, several TNF-α antagonists have been developed for therapeutic application in a wide range of disorders in rheumatology, dermatology and gastroenterology. TNF-α can be inhibited with two main types of immunosuppressant bio-agents: a monoclonal antibody such as infliximab (IFL) [2] and adalimumab (ADA),[3] or a circulating receptor fusion protein such as etanercept (ETN).[4]

IFL, a chimeric human/mouse antibody, can be considered the “pioneer” of this class of drugs. First approved by the European Medicines Agency (EMA) for the treatment of RA in 1999, it gradually achieved positive recommendations for many other indications (Table 1). It is administered by intravenous infusion lasting 2 h at a dosage based on body weight at first, then a maintenance regime every 8 weeks.[2]

Table 1. Therapeutic indications for each mature TNF-α inhibitor in Europe.

Therapeutic indication	Infliximab	Adalimumab	Etanercept
Rheumatoid arthritis	1999	2003	2000
Psoriatic arthritis	2004	2005	2006
Psoriasis	2005	2007	2004
Crohn’s disease	2006	2007	-
Juvenile Idiopathic arthritis	2006	2008	2006
Ulcerative colitis	2006	2012	-
Ankylosing spondylitis	2007	2006	2006
Total	7	7	5

TNF: Tumor necrosis factor.

ETN and ADA are the first “followers” of IFL: both were first approved for the same indication (RA) by the EMA, respectively, in 2000 and 2003. While ETN has shown no real benefit for gastroenteric disorders, ADA has currently the same indications as IFL (Table 1). Both are injected subcutaneously, weekly or twice-weekly for ETN and every 2 weeks for ADA.[3,4]

Two other TNF-α antagonists (certolizumab and golimumab) have now been approved. Although no comparative clinical trial has ever evaluated the relative efficacy of these bio-agents, their “place in therapy” appears to be substantially equivalent in their common indications [5,6] and their safety profiles too can be considered similar.[7] When the IFL patent expired in 2013, EMA approved two specialties containing the same biosimilar,[8] with the same therapeutic indications as the originator and the same dosing regimens, pharmaceutical forms and strengths.[9] Biosimilars of ETN and ADA should follow in the next few years.[10]

Here we critically review the full economic evaluations conducted in the European Union (EU) settings on the three older TNF-α inhibitors in their main common indications, to assess whether they can provide useful information to health authorities for purchasing strategies in the light of the changing “competitive arena”. Right now, the three agents are still priced separately in most EU countries.

Economic evaluations in the EU

We searched the PubMed international database to select economic evaluations conducted in EU countries, in which the older anti-TNF-α were compared in their indications for RA and Ps.¹ Five of the seven studies finally selected focused on Ps,[11–15] two on RA [16,17] (Table 2). How the alternatives were evaluated varied a lot in Ps: three studies compared the three anti-TNF-α antagonists against placebo [13,14] or supportive care,[15] one directly with each other,[12] the last only ADA with ETN.[11] In the two RA studies, ADA, ETN and IFL—alone [16] or in co-therapy with methotrexate [16,17]—were compared with traditional antirheumatic drugs. Three studies on Ps included a cost-effectiveness analysis,[12–14] the two on RA [16,17] and one more on Ps [11] a cost-utility, the remainder on Ps both techniques.[15] Five studies took a third-party payer’s viewpoint,[12–15,17] the other two a societal perspective.[11,16]

Table 2. Main characteristics of the selected studies.

Study	Alternatives with TNF-i	Type of study and perspective	Type of model (time horizon, discount)	Efficacy end points and sources	Direct cost items	Main conclusion	Most influential variables in SA
Knight Sweden	ADA vs ETA	CEA, CUA Society	Markov [10 years, 3%]	PASI-75 (RCTs and a 120 weeks open-label study)	-Drugs and their administration -Hospitalizations -Outpatient consultations	ETA is cost-effective compared with ADA.	PASI-75 (CEA), related utility values (CUA) (one-way)
De Portu Italy	INF vs ETA INF vs ADA	CEA TPP	Decision tree (24 and 48 weeks)	PASI-75 (RCT and assumptions)	-Drugs and their administration -Outpatient consultations	INF seems to be more cost-effective than ADA and ETA.	INF price (one-way and extreme scenario)
Ferrandiz Spain	ADA/ETA/INF vs placebo	CEA TPP	Decision tree (24 weeks)	PASI-75 (RCTs)	-Drugs	ADA is the most cost-effective TNF-α.	Unspecified (extreme scenario)
Schmitt-Rau Germany	ADA/ETA/INF vs placebo	CEA TPP	Decision tree (12 weeks)	PASI-75 (RCT)	-Drugs -Medical consultations -Monitoring and laboratory tests	INF (3 mg/kg dosage) is the most cost-effective agent.	PASI-75, drug prices (one-way and extreme scenario)
Sizto The UK	ADA/ETA/INF vs supportive care	CUA TPP	Markov (1 average year)	PASI-50/75/90 (RCTs*)	-Drugs and their administration -Hospitalizations -Laboratory tests -Outpatient visits	ADA is the most cost effective biologic.	Days of stay, ETA dosage (one-way and probabilistic analysis)
Lekander Sweden	ETA/(ADA/INF with or without MTX vs tDMARD	CUA Society	Markov (20 years, 3%)	DAS28, HAQ (registry)	-Drugs and their administration -Medical services -Adverse events -Transportation and informal care -Devices and investments	ETA is associated with a slightly higher QALY gained and lower incremental cost than the two other TNF-α.	Age at start of treatment, utility weights attributed to disease states (one-way and probabilistic analysis)
Bansback Sweden	(ADA+MTX)/(INF+MTX)/(ETA+MTX)/ vs tDMARD	CUA TPP	Markov (lifetime, 3%)	ACR20 and ACR50 (RCTs and observational studies)	-Drugs -Monitoring -Adverse events -Joint replacement	ADA appears to be at least as cost effective as the other TNF-α antagonists.	Baseline age, SMR, HAQ-DI (one-way and multi-way)

ACR: American College of Rheumatology response criteria; ADA: Adalimumab; ; CEA: Cost-effectiveness analysis; CUA: Cost-utility analysis; DMARD: Disease-modifying antirheumatic drug; ETA: Etanercept HAQ-DI: Health Assessment Questionnaire Disability Index; INF: Infliximab; PASI: Psoriasis Area and Severity Index; RCT: Randomized clinical trial; SA: Sensitivity analysis; SMR: Standardized mortality ratio; TPP: Third-party payer.

*None of the RCTs were cited in the reference list.

The two studies on RA [16,17] and the most recent on Ps [11] adopted long-term Markov models; three studies on Ps short-term horizons (max. 48 weeks) based on “decision trees” [12–14] referred to the clinical trials from which they sourced efficacy; the remaining study on Ps [15] was based on a Markov model and adopted an unclear “average year” derived from a mix of trials.

The majority of studies sourced drug efficacy mainly from clinical trials conducted on one of the three anti-TNF-α antagonists,[11–15,17] only one from a patients’ registry.[16] Two studies on Ps [11,15] included all health-care costs except those due to adverse events, two [12,14] limited their analysis to the outpatient costs, the remainder [13] to drugs only; the two studies on RA [16,17] included all health-care costs, one of them [16] the direct non-health-care costs too.

Two studies on Ps [12,14] did not declare any funding and concluded in favor of IFL,[12,14] the other three in favor of the sponsored anti-TNF-α inhibitor.[11,13,15] The conclusions in the two studies on RA conflicted depending on the sponsor.[16,17] Drug prices and dosages were the variables most likely to affect the baseline results in short-term studies,[12–14] efficacy and utility values in long-term models.[11,16,17]

Policy implications

Our review of European economic evaluations on the three older TNF-α antagonists in their main common indications showed the scant utility of these studies for public decision-making. Their main ex ante limitation is the lack of clinical evidence based on “head-to-head” trials so many years after the launch of these agents.[6,18] Therefore, despite the efforts of indirect comparison studies, their comparative effectiveness can only be inferred through arguable estimates and uncertain assumptions.

Starting from such a major limit, economic evaluations can hardly add valuable information for health decision-makers, by definition. In fact, they are likely to add only further uncertainty when based on sophisticated long-term models populated by doubtful data,and poor information when based on short-term “decision trees” derived from heterogeneous and inconclusive clinical trials. In our review, the former were sensitive to discretionary estimates of efficacy and quality-of-life outcomes, the latter to slight differences in health-care costs and/or doubtful “trade-offs” between their items (particularly drug administration and day-of-stay costs). Last but not least, funding always seemed to play a fundamental role in the final results of sponsored studies.

Against this disappointing background, we wonder what lessons can be drawn from these studies on behalf of health authorities. The most pragmatic one could be that, since it is hard to demonstrate any evidence-based difference between these drugs in common indications,[19–21] it is equally hard to consider them not substantially equivalent after so many years on the market, as the many economic evaluations conducted on the TNF-α inhibitors as a whole group indirectly confirm.

Since it appears that not even the potential “switch” between the three agents should be so problematic,[22] the only potentially significant difference from the health authorities’ viewpoint might be the mode of administration, IFL being intravenous and thus necessarily implying a clinical access, whereas ADA and ETN are given subcutaneously. This might also influence patient’s preferences, according to the different indications. However, the much lower frequency of injections and the (by definition) better patient’s compliance with IFL should limit the importance of this potential inconvenience; in other terms, we believe that it can be hardly considered an issue of public health, as it has not been so far.

To sum up, having taken into account the different therapeutic regimens, what really matters for health authorities is the “cost per patient” treatment. The best way to minimize it can vary a lot from one European country to another, depending on the distribution channels (community pharmacies or hospitals) and purchasing procedures (tenders or private negotiations), which are generally influenced by the type of health-care system (e.g., “Beveridgean” like in Italy and the UK or “Bismarckian” like in Germany and the NL). However, a general, sound recommendation would be to group at least the older anti-TNF-α agents under a sort of common reference price “umbrella” (as in the NL), trying to standardize their “cost per patient” despite the different regimens, and eventually make it more sustainable for health authorities. We feel this decision should be taken very soon, to prevent a very likely fast switch of prescription patterns to “me too” in-patent biologics in the next few years, as it has often happened in the past decades with chemical entities.

To conclude, if pharmaceutical companies have compared these drugs for years, striving unsuccessfully to finding differences, we wonder why health authorities should not group them for purchasing strategies.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Notes

1. We used “infliximab OR adalimumab OR etanercept” AND “rheumatoid arthritis OR psoriasis” and “costs OR cost” as search terms. From the 386 articles published in English from January 2000 until June 2015 initially identified, 328 were immediately discarded, being epidemiological or clinical studies (139); reviews (85); letters or comments (29); policy articles (15) and partial economic evaluations (60). Of 58 potentially eligible economic evaluations, we further excluded studies not conducted in EU settings (22) and studies focused on other diseases, other biologics or without any direct or indirect comparisons between the three older anti-TNF-α (29). We finally selected and screened seven articles to assess their main methodological features, using a common checklist.