Clinical Effectiveness Of Tnf A Inhibitors Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

I searched electronic databases such as EMBASE, Medline, Cochrane Library and NHS Database of Review of Effectiveness from 1998 to 2009. The main purpose of literature review was to identify all published articles in peer-reviewed journals and this articles reported that the clinical effectiveness of TNF α inhibitors such as infliximab, adalimumab and etanercept in rheumatoid arthritis patients who had previously been treated with at least one TNF α inhibitor or DMARDs or biologic treatment. Then, for identified studies, reference lists were hand searched. (Suzanne 2010)

For the outcomes of these reviews, validated clinically meaningful binary measure of benefit and safety are ACR20 and ACR 50, means 20% and 50% improvement in swollen and tender joint counts plus 50% improvement in American College of Rheumatology systematic 3 of the following 5 criteria such as patient global assessment, physician global assessment, pain score, physical function (Health Assessment Questionnaire Score) and laboratory acute phase reactants (erythrocyte sedimentation rate on C-reactive protein level) (Chung 2006)

For safety, we achieved to include specific adverse events, which is a measure of patients' tolerance. So, we have chosen to include withdrawals because of adverse events. (Chung 2006)

Moreover, both the short-term efficacy and the toxic effects of new drug for RA are evaluated in clinical trials of 6-12months' duration. (Fries 1982, Felson 1995 and 1998)

Inclusion and Exclusion Criteria

Based on the titles and abstract, identified studies were selected for review and selected articles were assessed against inclusion criteria on the basis of the report.

The following selection criteria were used: (a) randomised controlled trials of MTX monotherapy versus MTX combined with other DMARDs of at least 12 weeks of trial duration; (b) patients with a clinical diagnosis of RA >18 years old; (c) data available on one or more of following pre-specified outcomes: ACR core set; ACR 20, 50 or 70 responses; ACR remission; Disease Activity Score (DAS); EULAR response; withdrawal due to lack of efficacy or adverse events (AEs); number of total or individual AEs. (Pinals 1981, Felson 1993 and 1995, Van der Heijde 1992, Prevoo 1995, Van Gestel AM 1996, Wells GA 2009)

In addition to the criteria given above, several studies were excluded at this stage because they replicated data reported in other studies included in the review.(Suzanne 2010)The following selection criteria of open-label extensions were excluded in studies comparing DMARDs not currently used for example, oral gold.(W Katchamart 2008) Moreover, rheumatoid arthritis patients could not be distinguished in the results because studies of patients with other condition such as juvenile arthritis, Crohn's disease, PsA (psoriatic arthritis) or SpA (spondyloarthropathies) were excluded. (Suzanne 2010) (Appendix.2)

Data Extraction

By using the Assessment of Multiple Systematic Reviews (AMSTAR) and quality assessment instrument, the data were extracted the quality of review benefits and safety results. (M C Hochberg 2003)Data extracted included baseline patient characteristics (age, sex, race, duration of RA, proportion rheumatoid factor positive), concomitant treatment (use of DMARDs), study treatment (study drug and regimen, methotrexate dose and duration), eligibility criteria (definition of active disease), baseline values for the WHO/ILAR core set variables (number of painful/tender joints, number of swollen joints, pain scores, patient and global assessment of disease activity, acute phase reactants and physical disability scores) and outcome measures (ACR20, 50, and 70) at either the end of trial or 24-30 weeks' duration if the trial lasted for 52 weeks. (Boers 1994, Felson 1993 and 1998)

Statistical Method

When the mixed-treatment comparison synthesis was evidenced to extend traditional meta-analysis, indirect comparisons method was used, instead of comparing same treatment with same comparator. (Lu G 2004) Although similar analytic methods were utilized different measures of effect size for the categorical and continuous data, each of the four outcome measures such ACR, EULAR, DAS and HAQ were considered in the analysis. (Suzanne 2010)

Mixed-effects logistic regression was performed not only random-effects meta-analysis models that were used from the outset due to the known clinical heterogeneity but also fixed-effects meta-analysis models that were carried out to obtain the overall outcome for the whole cohort. (Platt RW 1999, Suzanne 2010) Inconsistency index(I2) means a statistic for quantifying inconsistency of the results in the individual reviews for each of the drugs compared with placebo and combines the x² test (chi-squared test) to assess between study heterogeneity and quantified I2 statistics, which can take values between 0 and 100% where greater value shows strong heterogeneity. (Sutton 2000, W Katchamart 2008, Suzanne 2010) In the placebo group, for the comparison of the individual odd ratio, the number of people were evaluated the treatment of benefit and harm, with 95% confidence intervals (CIs) so p-values less than 0.05 and 95% CIs. (Osiri M 2003)

Initially, if patients treated with a cytokine antagonist agent compared with placebo or MTX, this method was performed an analysis to estimate the odd ratio of an ACR20, 50 responses at 6 months. In background therapy, this mixed-treatment analysis allows the cytokine antagonist to be used with or without MTX and the placebo to be administered with or without MTX. This model examines not only the binary results in a meta-analysis on the odd ratio scale because these odd ratios are more heterogeneous than absolute measure but also how these odd ratios vary with important prognostic factor.(R. Nixon 2007)

For each of the three TNFα blocking agents, patients who are gaining an ACR20 and ACR50 response by using the rate ratio (RR) and 95% confidence interval (CI) for the active treatment group compared with placebo. In randomized trial, patients received placebo who achieved an ACR20, 50 or 70 responses was compared the treatment using x² test. The methods of adjusted indirect comparisons were used for the comparison of the three TNF blocking agents with each other. (Bucher 1997)

Meta-analysis was used to pool the results from multiple studies. Meta-regression models were developed to explain heterogeneity using the study-level demographic covariates of mean age, percentage of females and baseline characteristic that included the percentage of RF patients, mean disease duration, mean number of previous DMARDs and the mean duration of previous biologic treatment. (Sutton 2000, Suzanne 2010) (Appendix 3)


Meta-analysis results



Percent(95% CI)

ACR20 Responders

EULAR Responders


Van Vollenhoven











(51.78, 86.03)




( 55.04, 86.07 )






( 45.71, 80.91 )


(-2.11, -1.29)




(55.63, 84.35)


(-1.80, -1.25)




( 54.36,88.31)


( 50.18,85.41 )




( 56.58 , 63.32 )


( 72.86, 78.75 )






( 65.25, 84.79 )


( 49.45, 71.62 )



Dl Pol



( 48.10, 87.94 )






( 57.05 , 74.56 )










( 45.71, 56.34 )


( 72.86, 82.15 )

Van Der Bijl



( 31.85, 61.47 )


( 62.92, 88.16 )
















( 38.24, 52.38 )


( 55.10, 68.88 )






( 81.12, 88.06 )






( 36.55, 57.70 )










Table: Forest plots showing response rates on the ACR20 scale, response rates on EULAR scale, response to treatment recorded by change in DAS-28 scores compared with baseline scores and reduction in HAQ scores.

ACR20 (American College of Rheumatology)

Within the studies, the overall proportion of ACR 20 response rate was 60.8% (95%CI 53.8, 67.4) and ACR 20 responders ranged between 45 and 76% during the period from 2003 to 2009, so the analysis was indicating that at least half of patients switching to an alternative TNF α inhibitor because TNF α antagonists achieved a minimum of an ACR 20 response. Due to heterogeneity, there was a high degree of variability in the estimated response rate between the studies (I 2 = 77.5). For the reason for switching from one anti-TNF to another, there was no significant difference identified in results. (Suzanne 2010)

EULAR (European League Against Rheumatism)

Within the studies of period between 2003 and 2009, the overall effect size of EULAR was 70.5% (95%CI 63.7, 76.6) based on the EULAR responders ranged from 47 and 85%. Due to heterogeneity, there was a high degree of variability in the estimated proportion between studies (I 2 = 86.3%). Based on the reasons for switching that would show this heterogeneity, there was no significant effect of baseline characteristics, demographic variables and anti-TNF sequence. (Suzanne 2010)

DAS-28 (Disease Activity Score)

From the baseline, the overall value of the reduction in DAS-28 score was 1.53 (95%CI 1.25, 1.8) that was based on the ranged between −0.98 and −2.4 so the examination showed a high degree of variability in effect sizes (I 2 = 94.8%). In DAS-28 score, additional reduction was giving by −0.157 (95% CI −0.243, −0.072) per additional year so patients' DAS-28 responses had significant effect on disease duration.

HAQ (Health Assessment Questionnaire)

From the baseline, the overall result of the reduction in HAQ score was 0.25 (95% CI 0.11, 0.40) that was based on the ranged from increase by 0.15 to decrease by −0.48 and heterogeneity was I 2 = 93.5%. There had a significant effect on the change from baseline scores because there was no significant difference in the demographic variables, baseline characteristics or reason for switching or anti-TNF sequence. (Suzanne 2010)

Comparative efficacy of cytokine antagonists

Comparative efficacy of cytokine antagonists showed these expected odds ratio of ACR responses with average baseline disease duration and average baseline HAQ during 1998 to 2007. Among this efficacy of cytokine antagonists, etanercept mono-therapy has greater odd ratio in comparison with placebo (3.58 for ACR20 response and 4.21 for ACR50 response). During this analysis, TNF-α antagonists (etanercept, infliximab and adalimumab) was more benefit than anakinra because the odd ratio of anakinra was 1.7 (95%CI 0.9-3.19) for ACR20 response and the odd ratios of etanercept, infliximab and adalimumab were 3.58 (95%CI 2.09-6.91), 3.47 (95%CI 1.66-7.14) and 3.19 (95%CI 1.97-5.48) for AR20 response so the treatment of combination with TNF α antagonists and MTX were more efficacious than combination of anakinra and MTX. Moreover, there was no statistical difference between etanercept, infliximab and adalimumab because the odd ratio of three TNF α antagonists was almost similar so the confidence intervals were overlapping each other. (R. Nixon 2007) (Appendix 4)

During 2009, the overall benefit odd ratios for ACR50 response have significantly greater than placebo (3.35, 95% CI 2.62-4.29) and overall withdrawal (safety) odd ratios are 1.39, 95%CI 1.31-1.71. In the withdrawal odd ratios, the treatment of adalimumab, anakinra and infliximab were significantly increased than placebo (1.54, 1.67, and 2.21). But abatacept and rituximab were no significantly greater than placebo (1.24 and 1.34). Moreover, etanercept was significantly decreased withdrawal rates than placebo (0.82, 95%CI 0.56-1.19). (Jasvinder 2009)(Appendix 4)

On the other hand, in the benefit odd ratios, the odd ratio of etanercept was significantly increased than placebo (4.97, 95%CI 2.7-9.13). Moreover, the treatments of adalimumab, anakinra, rituximab have shown greater odd ratios than placebo (3.7, 1.68 and 4.1). But anakinra treatment had lower odd ratio than placebo (1.68, 95%CI 0.83-3.41). Furthermore, abatacept and infliximab were not significantly higher than placebo (2.98 and 2.91). So, the single biologic therapy was more efficacious than placebo in randomized controlled trials of short and immediate duration but not longer duration. (Jasvinder 2009)(Appendix 4)

Indirect comparisons of TNF α blocking agents

The adjusted indirect comparisons of TNF α blocking agents included relative risks and 95%CIs. For the efficacy, the data was no significantly difference between these comparison groups of TNF α blocking agents in ACR20 response. For ACR50 response, etanercept was more efficacious than both adalimumab and infliximab because relative risk for etanercept was 2.6 (95%CI 0.35-19). (G. W. Torrance) (Appendix 5)

Efficacy of DMARD

There were three groups of efficacy of DMARD such as DMARD naïve patients group, MTX inadequate response patients group and non-MTX DMARD inadequate response patients group.

In the efficacy of DMARD naïve patients group, ACR responses compared MTX monotherapies with MTX combination therapies. In this group, combination therapies were MTX+ciclosporin (CSA) and MTX+doxycycline (DOXY). For ACR70 response, one CSA trial showed significantly greater result with RR=2.41 (95%CI 1.07-5.44) so this result was favouring the MTX combination therapy. (Haagsma 1997, Dougados 1998, Marchesoni 2003, Hetland 2006, O'Dell 2006)(Appendix 7)

In MTX inadequate response patients group, ACR responses compared MTX monotherapies with MTX combination therapies. In this group, combination therapies were MTX+leflunomide (LEF), MTX+CSA, MTX+intramuscular gold (IM gold) and MTX+levofloxacin. The efficacy of these trial combination therapies were significantly more benefit than MTX monotherapy for ACR20, 50 and 70 responses with RR=2.51 (95%CI 1.92-3.28), RR=4.54 (95%CI 2.51-8.2), RR=5.59 (95%CI 2.08-15.01) so there were no data on ACR remission and EULAR response but the analysis was indicating that significantly fewer patient withdrawals in MTX combination therapies than in MTX monotherapy because there was significantly increased in toxicity. (Tugwell 1995, Kremer 2002, Lehman 2005, Ogrendik 2007) (Appendix 7)

In non-MTX DMARD inadequate response patients group, ACR responses compared MTX monotherapies with MTX combination therapies. In this group, combination therapies included MTX+Sulfasalazine (SSZ), MTX+bucillamine (BUC) but the results of MTX+SSZ and MTX+BUC for ACR20 response showed significantly efficacious than MTX montherapy with RR=1.85 (95%CI 1.21-2.83) so there were no data on ACR remission and no statistically difference between two groups for EULAR response. (O'Dell 1996, Felson 1998, Hanyu 1999, Ichikawa 2005, Capell 2007) (Appendix 6)


For the approved doses for rheumatoid arthritis, we systematically extracted from the existing peer-review journals and articles. Moreover, we performed network meta-analysis because these analytical techniques allowed the trials to be supplied and important study-level variables permitted that treatment can be compared indirectly identifying and adjusting for cytokine antagonists. (R. Nixon 2007, Jasvinder 2009) In this meta-analysis, although patients fail one TNF-α inhibitor, they may still benefit from switching to another drug from this group because initial usage of TNF-α inhibitor showed lower response rate so they can change from RCT to effectiveness of switching with new TNF-α inhibitors such as golimumab that showed 37% of patients achieved an ACR20 response compared with 18% receiving placebo. (Wordsworth 2009)

In these reviews, TNF α and IL-1 inhibitors were compared with placebo or MTX with or without combination MTX. In these network indirect comparisons, anakinra has lower rates of benefit than TNFα blockers and other biologics. For safety, infliximab, anakinra and adalimumab have higher rates of withdrawals related adverse events than etanercept. Moreover, in the comparison of safety between etanercept and adalimumab, etanercept lowers rates than adalimumab (Relative Risk 0.37, 95%CI 0.19-0.7). (Lee YH 2008) Therefore, examinations of all biologics, excluding anakinra, were equally efficacious in relative measures. So, physicians and patients' choices may depend on the cost of patients, health care systems, frequency of the administration and preferences for route of administration. (Jasvinder 2009)Among these comparison of drugs in TNFα inhibitors group, comparison of ACR 20 and ACR 50 responses in infliximab and etanercept was not statistically significant (0.97 (95%CI 0.34-2.33) for ACR20 response and 0.98 (95%CI 0.45-1.93) for ACR 50 response) so if patients with rheumatoid arthritis failed to treat the infliximab, they may not benefit from switching to etanercept or to continue infliximab. (Appendix 7)

In efficacy of DMARD studies, DMARD naïve or parallel strategy was availabel only ACR 70 responses indicated not only combination therapies were high significant improvement but also withdrawal risks were increasing due to toxicity. For MTX inadequate response or step-up strategy, combination therapy of ACR response was significantly more benefit than MTX mono-therapy of ACR response because MTX dosage was lower than current usage so doctors would increase the MTX dosage or change to add another DMARD. In non-MTX DMARDs inadequate response or step-up strategy, ACR responses were available from only two studies (Capell 2007 and Ichikawa 2005) but patients failed to receive MTX alone or MTX +SSZ in Capell study and bucillamine trial was not commonly used in North America or Europe so conclusion cannot be reached. (W Katchamart 2008)

Another factor predicting response to treatment in rheumatoid arthritis are disease duration, female sex, prior DMARD use, disease functional class and disease activity. The effect of disease duration was strongest in patients with early disease duration. When the comparison of treatment with patients with early disease and with longer disease duration of rheumatoid arthritis, the longer disease duration of patients do not respond this treatment very well. These factors have implications for not only trial interpretation but also the clinical expectation of rheumatoid arthritis patients. The relative efficacy of the cytokine antagonists vs. MTX increases with disease duration but only the cytokine antagonist's efficacy decrease slowly with the disease duration and the efficacy of MTX absolutely decrease much more rapidly with disease duration. (Jennifer 2000) For example, if patients were separated into three groups such as the number of previous DMARDs, have never attempted existing DMARDs and have only partially responded to DMARDs, so only patients have failed a specific number of previous treatment. Therefore, treatment of rheumatoid arthritis does the well respond not only in patients with early disease duration but also patients have never attempted existing DMARDs.


Based on BRAM (Broadcast Recognition Access Method), the results of economic evaluations were considered as the review of clinical effectiveness such as treatment ranking. TNF inhibitors are most cost-effectiveness and most appropriate choice depends on patient preferences and routes of administration because the ICER (Incremental Cost-Effectiveness Ratio) for etanercept (£28,000 per QALY (Quality-adjusted Life-year)) was significantly lower than ICER for infliximab (£38,000 per QALY) or adalimumab (£30,000 per QALY). In 2002 NICE (National Institutes for Health and Clinical Excellence) guidance, early rheumatoid arthritis effectiveness data was recommended the use of third-line TNF inhibitors cost-effectiveness which gives ICERs around 30,000 per QALY and late rheumatoid arthritis effectiveness data was recommended the use of first-line TNF antagonists which gives ICERs around £50,000 per QALY for mono-therapies of adalimumab and etanercept with high figures of combination with methotrexate. For either direct head-to-head trial of DMARD and TNF inhibitors or longer term follow-up and post-marketing surveillance is needed not only to establish with more certainty values for different agents but also to find out the true risk of side effects.(Choi HK 2002, Brennan 2004, Welsing 2004, Bansback 2005, Kobelt 2004 and 2005)


This analysis has some limitations because the duration of trials for these biologics were too short to assess the long-term benefits and harms. Moreover, in the controlled trials, delayed and rare effects would not be detected. But, most randomized controlled trials test for new therapies in patients who are taking methotrxate because methotrexate is the standard of care for the rheumatoid arthritis treatment. By using indirect randomized data, the mixed-treatment comparison is the best tool for comparing treatments and making recommendations but a head-to-head randomized study of cytokine antagonists is unlikely because direct comparison is a hierarchy of evidence. So, clinicians and patients may wish to make decisions to take radiographic progression, adverse events and extending out to quality of life measures and economic costs. (Phillips 2001, Fleischmann 2003, Gabriel 2003, Balsa 2004, Genovese 2005, R.Nixon 2007, Jasvinder 2009)


In summary, we have been using mixed-treatment comparison and meta-regression for systematic reviews and meta-analysis of cytokine antagonists for rheumatoid arthritis. The result of odd ratios of an ACR20 or ACR 50 responses at 6months: treatment with comparison of each biologic is better than placebo; the use of combination therapy of cytokine antagonists and MTX improves the response than MTX monotherapy; TNF α antagonists treatments are better than with the IL-1 inhibitor treatment; the efficacy of TNF α antagonists' class are no significant difference from each other. Indirect comparisons from the double-blind, placebo-controlled trials provide the benefits and safety of the biologics for rheumatoid arthritis that including the data about relative and absolute benefit and safety, the various levels of functional limitation and the nature of prior treatment. Therefore, from this information of methodology and evidence-based comparison between cytokine antagonists, physicians and patients make decisions to expand new areas and effective therapies for rheumatoid arthritis.