Clin Rheumatol (2012) 31:133–137 DOI 10.1007/s10067-011-1790-6
ORIGINAL ARTICLE
Indirect comparison of etanercept, infliximab, and adalimumab for psoriatic arthritis: mixed treatment comparison using placebo as common comparator Alberto Migliore & Emanuele Bizzi & Serena Broccoli & Bruno Laganà
Received: 9 March 2011 / Revised: 10 May 2011 / Accepted: 31 May 2011 / Published online: 24 June 2011 # Clinical Rheumatology 2011
Abstract Psoriatic arthritis (PsA) is a chronic inflammatory joint disorder that is commonly associated with skin psoriatic lesions and can lead to severe disability. Current pharmacologic therapy for PsA includes TNFα-blocking agents for patients who are intolerant of or have an inadequate response to conventional disease-modifying antirheumatic drugs. Currently, there are no published randomized controlled trials providing a head-to-head comparison of the effectiveness of the three TNFαblocking agents used most often to treat patients with PsA (adalimumab, etanercept, and infliximab). In this study, we have performed the first indirect comparison among these biologic agents, in this setting, using a mixed treatment comparison analysis of the data from pivotal trials regarding efficacy profiles of adalimumab, etanercept, and infliximab evaluated as American College of Rheumatology (ACR) 20 response. Our results suggest that etanercept is expected to provide the greatest ACR20 response among the anti-TNFα agents compared A. Migliore (*) : E. Bizzi UO of Rheumatology, S. Pietro Fatebenefratelli Hospital, Via cassia 600, 00189 Rome, Italy e-mail:
[email protected] A. Migliore Research Center, S.Pietro Fatebenefratelli Hospital, Rome, Italy S. Broccoli Bioikos Pharma, s.r.l, Bologna, Italy B. Laganà Rheumatology, 2nd Medical School, La Sapienza University, Rome, Italy
with placebo in the treatment of patients with PsA unresponsive to conventional treatments. This analysis may be relevant for clinical decision-making, hence improving the management of PsA patients. Keywords Adalimumab . Efficacy . Etanercept . Infliximab . Psoriatic arthritis . TNFα blockers
Introduction Psoriatic arthritis (PsA) is a chronic inflammatory joint disorder commonly associated with psoriatic skin lesions. When appropriate treatment is not provided, many patients become severely disabled. The goal of an appropriate treatment is to arrest disease progression or to slow it down and achieve clinical remission. Current pharmacologic therapy includes disease-modifying antirheumatic drugs (DMARDs) and biologic agents. Patients who are intolerant of or have an inadequate response (IR) to DMARDs (DMARD-IR) are often treated with a biologic agent associated with a DMARD, such as methotrexate (MTX). Although disease control has been achieved with current treatments, further improvements and a better understanding of the precise therapeutic role of each drug are still necessary. Among the tumor necrosis factor alpha (TNFα) blockers currently approved for the treatment of DMARD-IR patients with PsA, adalimumab, etanercept, and infliximab are the biologic agents most widely prescribed. Adalimumab is a recombinant human immunoglobulin G1 (IgG1) monoclonal antibody, containing only human peptide sequences. Etanercept is a soluble dimeric TNF receptor fusion protein consisting of the extracellular ligand-binding domain of the TNF receptor linked to the Fc portion of human IgG1. Infliximab is a chimeric IgG1 monoclonal antibody composed of human constant and murine variable regions.
134
Clin Rheumatol (2012) 31:133–137
Informed treatment decisions require that the effectiveness of all available therapies be compared, preferably in randomized controlled trials (RCTs). In the absence of an RCT providing a head-to-head comparison of the biologic agents used to treat patients with PsA, mixed treatment comparison (MTC) is a valid statistical alternative [1–3]. MTC, an extension of traditional meta-analysis, enables multiple pairwise comparisons across a range of different interventions [4, 5]. The results from MTC provide an objective approach to the difficult choice of treatment when relevant data are unavailable. The use of the Bayesian approach for MTC allows for a probabilistic interpretation of the clinical data and ranking of the treatments. The estimated size of the treatment effect, as well as the associated uncertainty of each intervention, is translated into one measure: the probability that a certain treatment among the treatments compared may provide the greatest efficacy. This measure may be helpful to the physicians making treatment decisions. In 2007, Nixon estimated the efficacy of anti-TNFα treatments in rheumatoid arthritis by indirectly comparing them through MTC analysis [6]. More recently, Bergman et al. [7] used this method to compare patterns of American College of Rheumatology (ACR) response between tocilizumab and other biologic agents in DMARD-IR patients. No MTC studies are available comparing the efficacy of anti-TNFα agents in the management of PsA. The aim of this study is to compare the patterns of ACR criteria for 20% clinical improvement (ACR20) responses to different antiTNFα biologic agents in patients with PsA who have inadequate response to DMARDs.
Methods Identification of eligible studies and data extraction A literature search was performed in order to identify RCTs evaluating efficacy of biologic agents (adalimumab, etanercept, and infliximab) for treating patients suffering from PsA. MEDLINE and EMBASE databases were both searched using a combination of free-text and thesaurus terms relevant to the therapy for PsA. During the search period from 1990 to 2008, only fully published reports in English were considered (letters and abstracts were excluded). The search strategy is outlined in Table 1. The pivotal RCTs have been included [8–11].
Table 1 Research terms and search strategy used to search the EMBASE and MEDLINE literature databases
Inclusion criteria We have included RCTs with primary outcomes of ACR20 response after a minimum of 3 months on therapy. The ACR20 criteria were 20% reduction in tender and swollen joint count, and 20% improvement in at least three of the following five disease parameters: patient's global assessment of disease activity, physician's global assessment of disease activity, patient's assessment of pain, patient's assessment of physical disability (measured by the Health Assessment Questionnaire–Disability Index), and level of acute-phase reactants [12]. Exclusion criteria RCTs having continuous clinical values as primary endpoints, such as DAS28 or radiological outcome, were excluded as were the studies comparing different dosing regimens (not used in clinical practice) of the same agent. Intervention The intervention of at least one study group included one of the following anti-TNFα agents: adalimumab, etanercept, and infliximab. Statistical analysis The evaluations of the primary trial endpoint ACR20 were conducted in order to identify the possible existence of treatment differences. The number of responders in the treatment and placebo groups of each eligible trial was analyzed simultaneously using a fixed effect Bayesian MTC [13–16]. MTC enables multiple pairwise comparisons across a range of different treatments if analyzed as connected networks of RCTs. For example, in a dataset consisting of the pairwise comparisons AB, AC, BC, and AD, the A, B, C, and D treatment groups are connected. For each study, we extracted the details of the study design, patient characteristics, treatment interventions, endpoints, and length of follow-up. Treatments were compared in terms of odds ratios (OR) estimated in a Bayesian framework as the mean of the posterior distributions. The posterior distributions are a formal combination of a prior probability distribution of the unknown parameters with a distribution of the effects based
“Adalimumab” OR “etanercept” OR “infliximab” AND 1069 “psoriatic arthritis”
AND “double blind” 72
AND “ACR20” 9
Clin Rheumatol (2012) 31:133–137
135
on the observed data. For each pairwise comparison, the 95% credible intervals (CrI, the Bayesian version of the confidence intervals) were reported together with the OR. For each treatment, the probability of being the best treatment is also presented in the results. Analyses were performed using WinBUGS 1.4 statistical software (MRC Biostatistics Unit, Cambridge, UK).
Results Four placebo-controlled RCTs investigating the efficacy of anti-TNFα on PsA and meeting the selection criteria were identified [8–11]. Overall, the studies included more than 820 patients. The average age of patients was similar across the trials. In all of the studies, about 50% of patients received MTX at recommended doses. Disease duration was similar between studies ranging from 7.2±7.0 to 9.8± 8.3 years. ACR20 was the primary endpoint of analyzed studies. Follow-up times were either 12 weeks [8] or 24 weeks [9–11]. In Table 2, demographic characteristics are presented by intervention group for the original studies. In Table 3, the corresponding ORs are presented. As expected, each anti-TNFα agent performs significantly better than placebo (ORs significantly higher than one). On the contrary, no statistically significant differences are observed when the anti-TNFα agents are compared. However, there is a 71% probability that, of the three biological drugs, etanercept results in the greatest achievement of ACR20. The results of ranking suggest that etanercept has the greatest probability of being the best treatment followed by infliximab (24.7%) and adalimumab (4.3%).
Discussion This is the first study, of which we are aware, to use MTC to compare the efficacy of anti-TNFα agents in patients affected by PsA that is unresponsive to DMARDs. Our results suggest that for treating PsA, etanercept is expected to provide the greatest ACR20 response among antiTNFα agents compared with placebo. Etanercept has a 71% probability of being the most efficacious of all treatments examined at achieving an ACR20 response. The lack of a statistical significant difference in the direct comparison between etanercept and the other agents may be due to the low number of trials examined (only four) and to the sample dispersion. Anyway, both adalimumab and infliximab show statistically significant differences vs. placebo. Physicians may find this result helpful considering that the 71% probability for etanercept to provide ACR20 response corresponds to a 29% probability of making the wrong decision by prescribing etanercept to a PsA patient with inadequate response to DMARDs. Considering this in clinical practice, in the absence of direct evidence, i.e., a head-to-head comparison, each of the three treatments would have a 33% chance of providing the greatest ACR20 response, meaning the association with a 66% probability of making the wrong decision. The primary efficacy endpoint of the analysis was ACR20 as pivotal trials for biologic DMARDs used ACR20 response as their primary endpoints. Even if RCTs can suggest good evidence for the relative treatment effect of a single pairwise comparison, RCTs are often not able to compare all interventions available in the real clinical experience. This MTC includes placebo-controlled trials having the same endpoint since no head-to-head evidence is
Table 2 Demographic characteristics of treatment groups in the original trials Clinical study Agent Duration (weeks) Number of randomized patients Patients completing the study
Genovese et al. [8] Adalimumab 12 100
Mease et al. [11] Adalimumab 24 200
Mease et al. [10] Etanercept 24 205
Antoni et al. [9] Infliximab 24 315
Active treatment (50 of 51) 53.1% 56.9% 50.4±11.0 7.5±7.0 47.1%
Taking placebo (46 of 49) 49% 51% 47.7±11.3 7.2±7.0 46.9%
Active treatment (85 of 100) 29% 71% 47.1±12.8 8.4±7.2 47%
Taking placebo (47 of 100) 49% 51% 46.5±11.3 7.5±7.8 45%
Active treatment (93 of 101) 43% 57% 47.6±? 9.0±? 42%
Taking placebo (72 of 104) 55% 45% 47.3±? 9.2±? 41%
Active treatment (140 of 153) 43.7% 56.3% 48.6±12.5 9.8±8.3 51%
Taking placebo (149 of 162) 45.1% 54.9% 49.2±11.1 9.2±8.7 50%
1.6±1.7 1.0±1.0
1.1±0.6 1.9±2.1
1.1±0.6 2.3±3.4
ND ND
ND ND
1±0.6 1.4±2.1
1±0.7 1.4±1.7
Female Male Age (years) Disease duration Patients taking MTX at baseline HAQ at baseline 1.0±1.0 C-reactive protein level 1.6±1.7 (mg/dl) at baseline
136 Table 3 Results of mixed treatment comparison (fixed effects model) comparing the efficacy of etanercept, adalimumab, and infliximab vs. placebo on ACR20 response
Clin Rheumatol (2012) 31:133–137 Fixed effects mixed treatment comparisons Comparisons
Odds ratio
95% CrI lower
95% CrI upper
Probability of being the best treatment, %
Rank
Adalimumab vs. placebo Etanercept vs. placebo Infliximab vs. placebo Adalimumab vs. etanercept Adalimumab vs. infliximab Etanercept vs. infliximab
6.423737 10.27794 6.35982 0.625002 1.01005 1.616074
4.0552 5.697343 3.287081 0.286505 0.440432 0.644036
10.38124 19.29797 12.67967 1.32313 2.2705 4.01485
8 79 13
3 1 2
available. They were similar with regard to average age, disease duration, and rate of concomitant therapy with MTX, excluding these factors as sources of heterogeneity across trials and, therefore, bias in the indirect estimates. There does not appear to be a difference in disease severity among treated populations in the four studies. However, there is a difference in the length of follow-up across trials: in Genovese et al. [8], patients treated with adalimumab were followed for only 12 weeks, while patients in the other trials were followed for 24 weeks. This could implicate a reduction of results in the adalimumab group. Compared to traditional meta-analysis, where multiple studies are used to estimate the efficacy of a single agent relative to placebo or other reference treatments, MTC has additional advantages as it provides combined effectiveness estimates of studies with similar comparisons and grants the possibility of obtaining relative efficacy estimates among treatments for which no direct comparisons are available [1– 4]. Application of the mixed treatment comparison approach is increasing in the analysis of several medical issues, e. g., stroke prevention [13], antidepressants [14], psychological interventions in heart disease [15], and prevention of vertebral fractures in women with postmenopausal osteoporosis [16]. Because our main objective was to determine practical parameters for use in real-world medical decisions, we concentrated our efforts on the relative efficacy of currently licensed doses of commonly used treatments for PsA. It is of great interest among physicians and decision-makers to identify the most effective treatment from among those available. However, in the context of decision-making, differences in compliance, tolerability, and safety of each treatment also have to be considered. Moreover, in the case of arthritis, differences in the ability not only to improve clinical surrogate measures, such as ACR20, but also to arrest radiological progression of disease could be considered. There is hope that radiological damage may be preventable, but it remains to be shown whether patients benefit from long-term anti-TNF therapy, and whether radiological progression and ankylosis can be stopped; in this case it will be the best outcome to evaluate. Currently, data on possible differences in the ability to stop radiological
progression among the three agents are not available. In a recent meta-analysis of six RCT including 982 patients affected by PsA [17], Saad et al. found that there were no significant differences between TNFα inhibitors and placebo in the proportions of patients experiencing withdrawal for any reason, or withdrawal due to either serious and mild adverse events or upper respiratory tract infections. In a systematic review and meta-analysis of 21 randomized, placebo-controlled trials, Wiens et al. [18] noted no statistically significant differences in the safety of any of the three drugs compared with placebo, but infliximab had the highest RR for withdrawing from the study due to adverse events. Also, data proceedings from the British Society of Rheumatology Biologics Register [19] between 2002 and 2006 confirmed that for patients starting antiTNFα therapy there was better persistence with etanercept when compared with infliximab; these data are also in accordance with the data published earlier from the Swedish [20] and the Spanish [21] biologic registers. Another factor influencing decisions is the cost-effectiveness of each agent. For instance, patients treated with infliximab overtime need dose increases and reductions in the administration interval [22], resulting in increased expenditures. The results of this study are subject to limitations. Some difference in trial procedures and populations exist; however, meta-analytic techniques are able to recognize this possibility and to assess the uncertainty of estimated ORs. One of the limitations of this study is the lack of possibility to perform a randomized effect model MTC, due to the small sample size. In fact, in order to consider unknown or unmeasured differences of covariates that act as effect across trials, a random effects approach should be performed to detect possible heterogeneity across comparisons. In this analysis, only a fixed effects model has been used. Other limitation is the low number of trials and biological agents to compare.
Conclusions This analysis may be relevant for clinical decision-making, contributing to improve management of PsA patients
Clin Rheumatol (2012) 31:133–137
because even though the mixed treatment comparisons between three currently used TNFα blockers do not show a statistically significant difference, our results suggest that etanercept comparing to placebo is expected to provide the greatest ACR20 response in patients affected by PsA who are not responsive to DMARDs. Acknowledgments The entire study was conducted with unrestricted funds from ANTIAGE (a nonprofit national association for ultrasoundguided intra-articular hip therapy) whose president is Prof. Alberto Migliore. Prof. Alberto Migliore and Prof. Bruno Laganà received grants as consultants from Pfizer, Abbott, and Merck for national and international studies and courses.
137
10.
11.
12.
Disclosures None. 13.
References 1. Jansen JP, Crawford B, Bergman G et al (2008) Bayesian metaanalysis of multiple treatment comparisons: an introduction to mixed treatment comparisons. Value Health 11:956–964 2. Ades AE, Sculpher M, Sutton A et al (2006) Bayesian methods for evidence synthesis in cost-effectiveness analysis. PharmacoEconomics 24(1):1–19 3. Ades AE, Welton N, Lu G (2007) Introduction to mixed treatment comparisons. University of Bristol, Bristol, UK. https://www.bris. ac.uk/cobm/docs/intro%20to%20mtc.doc 4. Caldwell D, Ades A, Higgins J (2005) Simultaneous comparison of multiple treatments: combining direct and indirect evidence. Br Med J 331(7521):897–900 5. Lu G, Ades AE (2004) Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 23(20):3105–3124 6. Nixon RM, Bansback N, Brennan A (2007) Using mixed treatment comparisons and meta-regression to perform indirect comparisons to estimate the efficacy of biologic treatments in rheumatoid arthritis. Stat Med 26(6):1237–1254 7. Bergman GJ, Hochberg MC, Boers M, Wintfeld N, Kielhorn A, Jansen JP (2010) Indirect comparison of tocilizumab and other biologic agents in patients with rheumatoid arthritis and inadequate response to disease-modifying antirheumatic drugs. Semin Arthritis Rheum 39(6):425–441 8. Genovese MC, Mease PJ, Thomson GT, Kivitz AJ, Perdok RJ, Weinberg MA, Medich J, Sasso EH (2007) Safety and efficacy of adalimumab in treatment of patients with psoriatic arthritis who had failed disease modifying antirheumatic drug therapy. J Rheumatol 34(5):1040–1050 9. Antoni CE, Kavanaugh A, Kirkham B, Tutuncu Z, Burmester GR, Schneider U, Furst DE, Molitor J, Keystone E, Gladman D, Manger B, Wassenberg S, Weier R, Wallace DJ, Weisman MH, Kalden JR, Smolen J (2005) Sustained benefits of infliximab therapy for dermatologic and articular manifestations of psoriatic arthritis: results from the infliximab multinational psoriatic
14.
15.
16.
17.
18.
19.
20.
21.
22.
arthritis controlled trial (IMPACT). Arthritis Rheum 52(4):1227– 1236, Erratum in: Arthritis Rheum 52 (9):2951 Mease PJ, Kivitz AJ, Burch FX, Siegel EL, Cohen SB, Ory P, Salonen D, Rubenstein J, Sharp JT, Tsuji W (2004) Etanercept treatment of psoriatic arthritis: safety, efficacy, and effect on disease progression. Arthritis Rheum 50(7):2264–2272 Mease PJ, Gladman DD, Ritchlin CT, Ruderman EM, Steinfeld SD, Choy EH, Sharp JT, Ory PA, Perdok RJ, Weinberg MA (2005) Adalimumab for the treatment of patients with moderately to severely active psoriatic arthritis: results of a double-blind, randomized, placebo-controlled trial. Arthritis Rheum 52 (10):3279–3289 Felson DT, Anderson JJ, Boers M, Bombardier C, Furst D, Goldsmith C, Katz LM, Lightfoot R Jr, Pauls H, strand V et al (1995) American College of Rheumatology. Preliminary definition of improvement in rheumatoid arthritis. Arthritis Rheum 38 (6):727–735 Cooper NJ, Sutton AJ, Lu G, Khunti K (2006) Mixed comparison of stroke prevention treatments in individuals with non rheumatic atrial fibrillation. Arch Intern Med 166:1269–1275 Cipriani A, Furukawa TA, Salanti G, Geddes JR, Higgins JPT, Churchill R, Watanabe N, Nakagawa A, Omori IM, McGuire H, Tansella M, Barbui C (2009) Comparative efficacy and acceptability of 12 new-generation antidepressants: a multiple-treatments metaanalysis. Lancet 373:746–758 Welton N, Caldwell D, Adamopoulous E, Vedhara K (2009) Mixed treatment comparison meta-analysis of complex interventions: psychological interventions in coronary heart disease. Am J Epidemiol 169:1158–1165 Jansen JP, Bergman GJ, Huels J, Olson M (2009) Prevention of vertebral fractures in osteoporosis: mixed treatment comparison of bisphosphonate therapies. Curr Med Res Opin 25(8):1861–1868 Saad AA, Symmons DP, Noyce PR, Ashcroft DM (2008) Risks and benefits of tumor necrosis factor-alpha inhibitors in the management of psoriatic arthritis: systematic review and metaanalysis of randomized controlled trials. J Rheumatol 35(5):883– 890 Wiens A, Venson R, Correr CJ, Otuki MF, Pontarolo R (2010) Metaanalysis of the efficacy and safety of adalimumab, etanercept, and infliximab for the treatment of rheumatoid arthritis. Pharmacotherapy 30(4):339–353 Saad AA, Ashcroft DM, Watson KD, Hyrich KL, Noyce PR, Symmons DP (2009) Persistence with anti-tumour necrosis factor therapies in patients with psoriatic arthritis: observational study from the British Society of Rheumatology Biologics Register. Arthritis Res Ther 11(2):R52 Kristensen LE, Gulfe A, Saxne T, Geborek P (2008) Efficacy and tolerability of anti-tumour necrosis factor therapy in psoriatic arthritis patients: results from the South Swedish Arthritis Treatment Group register. Ann Rheum Dis 67:364–369 Gomez-Reino JJ, Carmona L (2006) Switching TNF antagonists in patients with chronic arthritis: an observational study of 488 patients over a four-year period. Arthritis Res Ther 8:R29 Kalb RE, Gurske J (2005) Infliximab for the treatment of psoriasis: clinical experience at the State University of New York at Buffalo. J Am Acad Dermatol 53(4):616–622
