Rheumatol Int (2008) 28:553559 DOI 10.

1007/s00296-007-0475-6

O R I G I N A L A R T I CL E

Meta-analysis of the combination of TNF inhibitors plus MTX compared to MTX monotherapy, and the adjusted indirect comparison of TNF inhibitors in patients suVering from active rheumatoid arthritis
Young Ho Lee Jin Hyun Woo Young Hee Rho Seong Jae Choi Jong Dae Ji Gwan Gyu Song

Received: 28 June 2007 / Accepted: 6 October 2007 / Published online: 18 October 2007 Springer-Verlag 2007

Abstract This present study was designed to examine (1) whether a combination therapy of TNF (tumor necrosis factor) blockers and methotrexate (MTX) is better than MTX monotherapy, and (2) if the TNF inhibitors such as etanercept, inXiximab and adalimumab are all same for treating patients with active rheumatoid arthritis (RA). We performed meta-analysis of a combination therapy of TNFblockers and MTX compared to MTX monotherapy and we performed adjusted indirect comparison of the TNF-blocking agents for their eYcacy and toxicity. Three studies met the inclusion criteria for the analysis. Meta-analysis showed that the combination of MTX with anti-TNF inhibitors was more eVective than MTX monotherapy and this indicated that combination therapy of anti-TNF inhibitors and MTX was comparable with MTX monotherapy in terms of withdrawal due to the side eVects (RR: 1.05, 95% CI: 0.52 2.09, P = 0.86). The adjusted indirect comparison did not show any diVerences between inXiximab and adalimumab. However, there was a signiWcant diVerence for clinical eYcacy and side eVects between etanercept, adalimumab and inXiximab. The RRs for achieving ACR20, ACR50 and ACR70 responses and withdrawal due to the side eVect in the etanercept group were lower when compared with the adalimumab group. The RR for achieving an ACR20 response in the etanercept group was lower when compared with the inXiximab group. The adjusted indirect comparison analysis suggests that the TNF-blocking agents all are

not the same with respect to eVectiveness and toxicity for the treatment of active RA. Keywords Rheumatoid arthritis TNF inhibitors Adjusted indirect comparison Meta-analysis

Introduction Tumor necrosis factor- (TNF- ) plays a key role in the pathogenesis of rheumatoid arthritis (RA) [1]. The advent of anti-TNF- therapy is an important advance in the management of RA, and the introduction of TNF- inhibitors has revolutionized the treatment of RA. There are presently three TNF-blocking agents that have been granted approval by the US Food and Drug administration for the treatment of RA; they are etanercept, inXiximab and adalimumab. Etanercept is a soluble dimeric TNF-receptor fusion protein that consists of the extracellular ligand-binding portion of human p75 TNF-R that is linked to the Fc portion of human IgG1, which can bind to both TNF- and TNF- (lymphotoxin- ) [2]. InXiximab is a chimeric IgG1 monoclonal antibody and is composed of the constant human (75%) and variable murine (25%) regions. InXiximab speciWcally binds to human TNF- , but not to TNF- . Adalimumab is a recombinant human IgG1 monoclonal antibody (human-derived heavy-chain and light-chain variable regions and human IgG1 constant regions). As with inXiximab, adalimumab does not neutralize TNF- , but both (but not etanercept) Wx complement and therefore can lyse cells that express TNFon their surface [3]. While it is clear that the TNF antagonists diVer in terms of their structure, pharmacokinetics and binding abilities, it is not clear whether these characteristics contribute to the eYcacy, tolerability and safety proWles of the TNF inhibitors for the treatment of RA [4].

Y. H. Lee (&) J. H. Woo Y. H. Rho S. J. Choi J. D. Ji G. G. Song Division of Rheumatology, Department of Internal Medicine, Korea University Medical Center, 126-1 Ka, Anam-Dong, Seongbuk-Ku, Seoul 136-705, South Korea e-mail: lyhcgh@korea.ac.kr

123

554

Rheumatol Int (2008) 28:553559

TNF inhibitors are recommended for the treatment of active RA after an adequate trial with the disease modifying anti-rheumatic drugs (DMARDs) has failed [5]. Several randomized, placebo-controlled studies have been performed with TNF blockers plus methotrexate (MTX) for the patients with refractory RA, despite that they were Wrst treated with DMARDs [68]. In the randomized, double blind, placebo-controlled studies using the TNF blockers with MTX for the treatment of RA, signiWcantly more of the patients who were treated with etanercept, inXiximab or adalimumab at the approved doses achieved American College of Rheumatology (ACR) scores of 20, 50 and 70 when compared with MTX alone at 5254 weeks of treatment. Hard data regarding the comparative long-term eYcacy and toxicity of these agents, as well as their variable rates of response, will be important for their rational clinical use. However, there are currently no published data from headto-head trials to support using one of these drugs over another. In the absence of randomized trials that makes head-to-head comparisons impossible, an adjusted indirect comparison is possible using a common comparator and this may provide useful information about the relative eYcacy of these drugs [911]. Thus, we aimed here to examine whether combination therapy with the TNF blockers and MTX is eVective compared to MTX monotherapy and if TNF inhibitors such as etanercept, inXiximab and adalimumab have the same eVect in patients suVering from active RA by performing meta-analysis and using an adjusted indirect comparison method.

of trials and (4) it compared TNF inhibitors plus MTX with MTX alone in patients with active RA. The following information was sought from each article: the authors identity, year of publication, the patients age and gender, duration of RA, the proportion of RF positive patients, concomitant treatment (use of glucocorticoids and the MTX dose), the tender joint count, the swollen joint count, the health assessment questionnaire (HAQ) score, name of the TNF inhibitors and the outcome measures, the ACR criteria for 20% (ACR20), 50% (ACR 50) and 70% (ACR 70) improvement, patient withdrawal due to lack of eYcacy and side eVect at 5055 weeks duration. Several dosage regimens for the TNF inhibitors were tested in the articles. We used the recommended dosages of TNF inhibitors for the meta-analysis and we adjusted for indirect comparison; etanercept 25 mg twice weekly, inXiximab 3 mg/kg intravenously (iv) q 8 weeks and adalimumab 40 mg subcutaneously (sc) q 2 weeks. We used the Jadad score to assess the quality of the trials [12]. Evaluation of statistical associations Meta-analysis We assessed the within-study variation and the betweenstudy variation, or heterogeneity, by testing with Cochrans Q-statistic [13]. This heterogeneity test assesses the null hypothesis that all the studies were evaluating the same eVect. A signiWcant Q-statistic (P value < 0.10) indicates heterogeneity across the studies. A random eVect model using the DerSimonian and Laird method was selected for this meta-analysis since it accounts for the potential interstudy heterogeneity [14]. Outcome measures for the metaanalysis were the number of patients achieving ACR20, 50 and 70 responses and the number of patients who withdrew due to lack of eYcacy and side eVects at 5255 weeks duration. Pooled RRs in the meta-analysis were obtained by weighting the individual RRs via the inverse of their variance. Statistical manipulations were undertaken using the program RevMan 4.28 (Oxford, England). Data analysis for adjusted indirect comparison The RR and its 95% CI for a combination of TNF inhibitors plus MTX compared with MTX monotherapy were estimated. The comparison of TNF inhibitors was performed using the method of adjusted indirect comparisons as described by Bucher et al. and Song et al. [911]. In brief, comparison of the result of group A with the result of group B within a randomized controlled trial gives an estimate of the eYcacy of intervention A versus B. The indirect comparison of intervention A and B is adjusted by the results of their direct comparisons with a common intervention C as

Materials and methods IdentiWcation of eligible studies and data extraction We searched the literature from MEDLINE and the Cochrane Library to identify the available articles, and we limited our searches to the randomized-controlled trials. The following key words and subject terms were searched: Tumor necrosis factor, TNF- , TNF blocker, etanercept, inXiximab, adalimumab and rheumatoid arthritis. The language of the papers was not restricted. References in the studies were reviewed to identify any additional studies that were not indexed by the electronic databases. MTX has been considered the standard against which newer DMARDs should be evaluated. We included the randomized controlled studies that have been performed with TNF inhibitors plus MTX for the patients suVering with active RA despite treatment with DMARDs, including MTX. A study was included in the analysis if: (1) it was published before February 2006; (2) it was original data (independence among the studies); (3) it was a double blind, randomized and controlled trial that completed 5055 weeks

123

Rheumatol Int (2008) 28:553559

555

follows: lnRRAB = lnRRAC lnRRBC. Because RRAC and RRBC are estimated from diVerent studies, they are statistically independent; hence, the variance of lnRRAB can be obtained from var(lnRRAB) = var(lnRRAC) + var(lnRRBC) where var indicates the square of the standard error (variance). From these variables we calculated a 95% CI for lnRRAB. All the values are back transformed to give the estimate of RRAB with a 95% CI, and the test of interaction is the ratio of lnRRAB to its standard error.

Results Meta-analysis Of the 120 studies identiWed by the literature search, 22 were chosen for detailed review. Among them, a total of three articles met the inclusion criteria [68]. Table 1 shows the baseline characteristics of the studies and Table 2
Table 1 Baseline characteristics of the patients enrolled in the MTX group as a control

shows eYcacy and toxicity of the TNF inhibitors in each study. The remainder of the studies was excluded for the following reasons: (1) they were a review article, (2) the data had been published more than once, (3) the study period was less than 50 weeks, and (4) there was no control group for MTX alone. We included each study on the use of etanercept, inXiximab and adalimumab, respectively. All the three studies had Jadad scores of 4. The eYcacy of TNF inhibitors in combination with MTX for treating RA was better than MTX monotherapy in all the individual studies. Meta-analysis showed that combination of TNF inhibitors/MTX therapy was superior to MTX alone at 5254 weeks, based on achieving the ACR 50/ACR 70 response criteria and also based on the withdrawal due to lack of eYcacy (Fig. 1). However, meta-analysis indicated that the combination therapy of anti-TNF inhibitors and MTX was comparable with MTX monotherapy in terms of withdrawal due to side eVects (RR: 1.05, 95% CI: 0.522.09, P = 0.86) (Fig. 1).
MTX group InXiximab 88 54 51 (12) 80 11 (8) 77 64 NA 16
a

Characteristics TNF inhibitors Number (MTX group) Follow-up period (weeks) Age [mean (SD), years] Women (%) Disease duration [mean (SD), years] RF positive (%) Corticosteroid use (%) Number of previous DMARDs use (mean) MTX dose (mg/week) Tender joint count [mean (SD), years] Swollen joint count [mean (SD), years] HAQ [mean (SD), years]

Etanercept 228 52 53 (12.8) 79 6.8 (5.5) 71 64 2.3 10

b

Adalimumab 207 52 56.1 (12.0) 73 10.9 (8.8) 89.5 NA 2.4 15b 28.1 (13.8) 19.0 (9.5) 1.48 (0.59)

SD standard deviation a Mean b Median

31 (18) 21 (12) 1.7 (0.6)

33.1 (13.4) 22.6 (10.7) NA

Table 2 Randomized controlled studies included in the meta-analysis for evaluating the eYcacy and toxicity of TNF inhibitors for patients with active rheumatoid arthritis Study TNF inhibitor Number Achieving number ACR 20 Lipskey et al. [7] InXiximab 3 mg/Kg 8 weekly, i.v. + MTX MTX Klareskog et al. [6] Etanercept 25 mg twice weekly, s.c. with MTX MTX Keystone et al. [8] Adalimumab 40 mg every second week with MTX MTX i.v. Intravenous, s.c. subcutaneous 86 88 231 228 207 200 42 17 196 171 122 48 ACR 50 21 8 159 98 86 19 ACR 70 10 2 99 43 48 9 Withdrawal due to lack of eYcacy 17 32 6 21 6 23 Withdrawal due to adverse eVects 5 7 24 32 26 13

123

556
Review: Comparison: Outcome: Study or sub-category Lipsky Keystone Klareskog Comparisons of TNF blockers 01 TNF blockers with MTX vs. MTX 01 ACR 20 Treatment n/N 42/86 122/207 196/231 524
2

Rheumatol Int (2008) 28:553559

Control n/N 17/88 48/200 171/228 516

RR (random) 95% CI

Weight % 30.96 33.83 35.21 100.00

RR (random) 95% CI 2.53 [1.57, 4.08] 2.46 [1.87, 3.22] 1.13 [1.03, 1.24] 1.89 [0.89, 4.00]

Total (95% CI) Total events: 360 (Treatment), 236 (Control)

2

Test for heterogeneity: Chi = 53.08, df = 2 (P < 0.00001), I = 96.2% Test for overall effect: Z = 1.65 (P = 0.10) 0.1 0.2 0.5 1 2 5 10

Favours control

Favours treatment

Review: Comparison: Outcome: Study or sub-category Lipsky Keystone Klareskog

Comparisons of TNF blockers 01 TNF blockers with MTX vs. MTX 02 ACR 50 Treatment n/N 21/86 86/207 159/231 524
2

Control n/N 8/88 19/200 98/228 516

RR (random) 95% CI

Weight % 28.17 34.00 37.84 100.00

RR (random) 95% CI 2.69 [1.26, 5.73] 4.37 [2.77, 6.91] 1.60 [1.35, 1.90] 2.61 [1.20, 5.66]

Total (95% CI) Total events: 266 (Treatment), 125 (Control)

2

Test for heterogeneity: Chi = 19.65, df = 2 (P < 0.0001), I = 89.8% Test for overall effect: Z = 2.42 (P = 0.02) 0.1 0.2 0.5 1 2 5 10

Favours control

Favours treatment

Review: Comparison: Outcome: Study or sub-category Lipsky Keystone Klareskog

Comparisons of TNF blockers 01 TNF blockers with MTX vs. MTX 03 ACR 70 Treatment n/N 10/86 48/207 99/231 524
2

Control n/N 2/88 9/200 43/228 516

RR (random) 95% CI

Weight % 15.07 35.32 49.61 100.00

RR (random) 95% CI 5.12 [1.15, 22.67] 5.15 [2.60, 10.22] 2.27 [1.67, 3.09] 3.43 [1.74, 6.75]

Total (95% CI) Total events: 157 (Treatment), 54 (Control)

2

Test for heterogeneity: Chi = 5.57, df = 2 (P = 0.06), I = 64.1% Test for overall effect: Z = 3.56 (P = 0.0004) 0.1 0.2 0.5 1 2 5 10

Favours control

Favours treatment

Review: Comparison: Outcome: Study or sub-category Lipsky Keystone Klareskog

Comparisons of TNF blockers 01 TNF blockers with MTX vs. MTX 04 W ithdrawal due to lack of efficacy Treatment n/N 17/86 6/207 6/231 Control n/N 32/88 23/200 21/228 516 RR (random) 95% CI Weight % 48.65 25.91 25.44 100.00 RR (random) 95% CI 0.54 [0.33, 0.90] 0.25 [0.10, 0.61] 0.28 [0.12, 0.69] 0.38 [0.22, 0.64]

524 Total (95% CI) Total events: 29 (Treatment), 76 (Control) 2 Test for heterogeneity: Chi = 3.17, df = 2 (P = 0.20), I2 = 36.9% Test for overall effect: Z = 3.59 (P = 0.0003)

0.1

0.2

0.5

10

Favours control

Favours treatment

R eview: C om pa rison: O u tcom e: Study or sub-category Lipsky K eystone K lareskog

C om pariso ns of T N F 01 T N F blockers with M T X vs. M T X 05 W ith drawa l due to side e ffect Treatment n/N 5/ 86 26/ 207 24/ 231 Control n/N 7/ 88 13/ 200 32/ 228 51 6 RR (random) 95% CI Weight % 22. 39 36. 33 41. 29 100. 00 RR (random) 95% CI 0.7 3 [ 0.2 4, 2.2 1] 1.9 3 [ 1.0 2, 3.6 5] 0.7 4 [ 0.4 5, 1.2 2] 1.0 5 [ 0.5 2, 2.0 9]

52 4 T otal (95% C I) T otal events: 55 (T reatm ent), 52 (C on trol) 2 2 T est for heterogen eity: C hi = 5 .8 3, df = 2 (P = 0.05), I = 6 5.7% T est for overall effe ct: Z = 0 .1 3 (P = 0.90)

0.1

0.2

0.5

10

Favours control

Favours treatment

123

Rheumatol Int (2008) 28:553559

557

Fig. 1 Meta-analysis evaluating the eYcacy and toxicity of TNF

inhibitors combined with MTX as compared with MTX alone for the patients suVering with active RA. All of TNF inhibitors were superior to MTX with respect to achieving ACR 20, ACR 50 and ACR 70 responses and patient withdrawal due to lack of eYcacy, and all the TNF inhibitors were comparable to MTX in terms of withdrawal due to side eVects

group. The RR for achieving an ACR20 response for the etanercept group was signiWcantly low when compared with the inXiximab group (RR: 0.45, 95% CI: 0.270.73, P = 0.001).

Discussion Adjusted indirect comparison of the anti-TNF inhibitors At baseline, the patients in all the three groups had active RA despite the DMARDs therapy. The baseline characteristics of the MTX groups as controls were generally comparable (Table 1). The number of subjects ranged from 88 to 228, the disease duration varied from 6.8 years (years) to 11 years, and the RF positive proportion varied from 71 to 89.5%. The patients age, corticosteroid use, DMARDs use, health assessment questionnaire (HAQ), tender and swollen joint counts were comparable among the three groups. Table 3 shows the relative risks (RRs) and 95% conWdence intervals (CIs) for the adjusted indirect comparisons of the TNF inhibitors for their eYcacy and side eVects. Adjusted indirect comparison did not show any diVerences between inXiximab and adalimumab on achieving ACR20, ACR50 and ACR70 responses and withdrawal due to lack of eYcacy and side eVects. However, there was a signiWcant diVerence in clinical eYcacy and side eVects between etanercept and adalimumab. The RRs for achieving ACR20, ACR50 and ACR70 responses and withdrawal due to side eVects in the etanercept group were signiWcantly low when compared with the adalimumab group. The RR for withdrawal due to lack of eYcacy, which was derived from the adjusted indirect comparisons of the two TNF-blockers, did not signiWcantly diVer from unity (etanercept vs. adalimumab; RR: 1.12, 95% CI: 0.323.94, P = 0.86). Indirect comparison of etanercept and inXiximab revealed no diVerence except for the ACR20 response. The RRs for achieving ACR50 and ACR70 responses and withdrawals due to lack of eYcacy and side eVects were not statistically signiWcant between the etanercept group and the inXiximab Our meta-analysis supports that the combination of TNF blockers and MTX is more eVective than MTX monotherapy without increasing the side eVects, and the adjusted indirect comparison suggests that there is a diVerence in eVectiveness and side eVects among the TNF inhibitors for the treatment of active RA. TNF inhibitors are currently indicated for the treatment of patients with active RA after an adequate trial of DMARDs and MTX is most commonly used. This metaanalysis shows that for patients with an incomplete response to DMARDs including MTX, the addition of TNF inhibitors improves the eYcacy without signiWcant side eVects. The TNF blockers are a diverse group of biologic agents and they diVer in their chemical structure and speciWc mechanism of action. Whereas all the three available TNF inhibitors have demonstrated eYcacy for the treatment of RA, they have varying eYcacy for treating granulomatous diseases and other inXammatory arthritides [15]. Why do these agents exhibit diVering activity and eYcacy proWles across certain disease states? The diVerences in the eYcacy of the drugs are likely associated with the pathophysiology of the diseases as well as the drugs characteristics such as their pharmacokinetics and dosing and their ability to block lymphotoxin or Wx complement. Etanercept and inXiximab have diVering eVects in granulomatous diseases, and inXiximab appears to be more eVective for the treatment of these conditions than etanercept [16, 17]. Etanercept and inXiximab have diVerent binding characteristics, with inXiximab binding to both soluble and membrane-bound TNF and etanercept binding primarily to soluble TNF. InXiximab may lyse the in vitro TNF-producing cells via activation of

Table 3 Adjusted indirect comparison of etanercept versus inXiximab versus adalimumab for eYcacy and side eVects Etanercept versus. InXiximab RR ACR 20 ACR 50 ACR 70 Withdrawal due to lack of eYcacy Withdrawal due to side eVect 0.45 0.59 0.44 0.52 1.01 95% CI 0.270.73 0.271.29 0.102.03 0.191.42 0.303.42 P value 0.001 0.19 0.29 0.20 0.98 Etanercept versus. Adalimumab RR 0.46 0.37 0.44 1.12 0.38 95% CI 0.340.61 0.220.60 0.210.93 0.323.94 0.170.86 P value <0.0001 <0.0001 0.03 0.86 0.02 InXiximab versus. Adalimumab RR 1.03 0.62 0.99 2.16 0.37 95% CI 0.591.78 0.251.49 0.195.13 0.776.07 0.111.36 P value 0.92 0.28 0.99 0.14 0.14

123

558

Rheumatol Int (2008) 28:553559

complement [3]. Etanercept and inXiximab have diVerent pharmacokinetic proWles. The preliminary data indicate that switching among these agents to overcome inadequate eYcacy or poor tolerability is beneWcial for some patients [18]. It may be important to recognize the diVerences among the agents that could potentially inXuence the patients outcomes. A direct head-to-head comparison study of the TNF blocking agents would give answers if the TNF inhibitors were all same for treating RA, but there has not been a direct comparison study. Although the results of any indirect comparisons should be interpreted with caution, the adjusted indirect method may provide useful information about the relative eYcacy of diVerent, competing interventions when there is no direct randomized evidence available [9]. In this adjusted indirect comparison analysis, we did not Wnd any signiWcant diVerence in eVectiveness and toxicity between inXiximab and adalimumab. However a lower drop-oV rate and lesser eYcacy with regard to achieving ACR20, 50, and 70 responses were found for the patients taking etanercept when compared with the patients taking adalimumab, and the rate for achieving an ACR 20 score for the patients taking etanercept was low when compared with inXiximab. These results suggest that TNF inhibitors may diVer in their eYcacy and toxicity when they are used to treat RA. Understanding the diVerences between these drugs will be an important challenge for the future, not just to explain their diVerences in clinical settings, but also to stimulate new programs for drug discovery and development. This studys limitations deserve mention. First, three trials included were heterogeneous by deWnition. The etanercept trial was a study in MTX-naive patients, whereas the other trials were in patients with an insuYcient response to MTX. It may be no surprise that the MTX response is better in the Wrst trial compared to the other two. This might lead to a smaller diVerence in eYcacy between the treatment groups in the Wrst trial, suggesting reduced eYcacy of etanercept compared to the other two drugs. This may aVect our adjusted indirect comparison results. Our data do not mean that etanercept is less eVective than inXiximab or adalimumab. Second, the adjusted indirect comparison here may have low power due to the small number of studies. More data would give more accurate results. Third, adjusted indirect comparisons usually, but not always, agree with the results of head-to-head randomized trials, and the adjusted indirect comparison has lower precision when recognizing the limited strength of inference. Interpretation of indirect comparison should be done with caution. Fourth, the validity of adjusted indirect comparisons depends on the internal validity and similarity of the involved trials. It is assumed that the patient samples in the diVerent trials were drawn from the same population of

patients. The validity of this assumption can be assessed in part by examining the demographics and baseline clinical variables of the patients enrolled in the trials. Although there were subtle diVerences in the patients baseline characteristics, they are likely to be comparable. We compared the proportion of patients who were randomly allocated to receive MTX alone as a control group. The proportion of patients randomly allocated to receive MTX alone and who completed the 5254 weeks of the trials, developed ACR20, 50 and 70 responses and then withdrew due to side eVects, was comparable. We do not believe that the subtle diVerences in baseline characteristics aVected the adjusted indirect comparison results. Fifth, there was a diVerence in the clinical outcomes based on the ACR response rate and the withdrawal due to lack of eYcacy. This may be the reason that using patients withdrawal due to lack of eYcacy may be less accurate than using the ACR response criteria. In conclusion, this meta-analysis conWrms that the combination of MTX with TNF inhibitors is more eVective than MTX monotherapy without increasing the side eVects. Although our result should be interpreted with caution, the adjusted indirect comparison analysis suggests that all TNF-blocking agents may not be the same with respect to their eVectiveness and toxicity for the treatment of active RA. Further studies are needed to better deWne the diVerences and to achieve optimized utilization of the TNF inhibitors.

Reference
1. Choy EH, Panayi GS (2001) Cytokine pathways and joint inXammation in rheumatoid arthritis. N Engl J Med 344:907916 2. Mohler KM, Torrance DS, Smith CA, Goodwin RG, Stremler KE, Fung VP, Madani H, Widmer MB (1993) Soluble tumor necrosis factor (TNF) receptors are eVective therapeutic agents in lethal endotoxemia and function simultaneously as both TNF carriers and TNF antagonists. J Immunol 151:15481561 3. Scallon BJ, Moore MA, Trinh H, Knight DM, Ghrayeb J (1995) Chimeric anti-TNF-alpha monoclonal antibody cA2 binds recombinant transmembrane TNF-alpha and activates immune eVector functions. Cytokine 7:251259 4. Mpofu S, Fatima F, Moots RJ (2005) Anti-TNF-alpha therapies: they are all the same (arent they?). Rheumatology (Oxford) 44:271273 5. Furst DE, Breedveld FC, Kalden JR, Smolen JS, Burmester GR, Bijlsma JW, Dougados M, Emery P, Keystone EC, Klareskog L, Mease PJ (2005) Updated consensus statement on biological agents, speciWcally tumour necrosis factor {alpha} (TNF{alpha}) blocking agents and interleukin-1 receptor antagonist (IL-1ra), for the treatment of rheumatic diseases, 2005. Ann Rheum Dis 649(Suppl 4):iv2iv14 6. Klareskog L, van der HD, de Jager JP, Gough A, Kalden J, Malaise M, Martin ME, Pavelka K, Sany J, Settas L, Wajdula J, Pedersen R, Fatenejad S, Sanda M (2004) Therapeutic eVect of the combination of etanercept and methotrexate compared with each treatment alone in patients with rheumatoid arthritis: double-blind randomised controlled trial. Lancet 363:675681

123

Rheumatol Int (2008) 28:553559 7. Lipsky PE, van der Heijde DM, St Clair EW, Furst DE, Breedveld FC, Kalden JR, Smolen JS, Weisman M, Emery P, Feldmann M, Harriman GR, Maini RN (2000) InXiximab and methotrexate in the treatment of rheumatoid arthritis. Anti-tumor necrosis factor trial in rheumatoid arthritis with concomitant therapy study group. N Engl J Med 343:15941602 8. Keystone EC, Kavanaugh AF, Sharp JT, Tannenbaum H, Hua Y, Teoh LS, FischkoV SA, Chartash EK (2004) Radiographic, clinical, and functional outcomes of treatment with adalimumab (a human anti-tumor necrosis factor monoclonal antibody) in patients with active rheumatoid arthritis receiving concomitant methotrexate therapy: a randomized, placebo-controlled, 52-week trial. Arthritis Rheum 50:14001411 9. Song F, Altman DG, Glenny AM, Deeks JJ (2003) Validity of indirect comparison for estimating eYcacy of competing interventions: empirical evidence from published meta-analyses. BMJ 326:472 10. Altman DG, Bland JM (2003) Interaction revisited: the diVerence between two estimates. BMJ 326:219 11. Bucher HC, Guyatt GH, GriYth LE, Walter SD (1997) The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials. J Clin Epidemiol 50:683691 12. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ (1996) Assessing the quality of reports

559 of randomized clinical trials: is blinding necessary? Control Clin Trials 17:112 Davey SG, Egger M (1997) Meta-analyses of randomised controlled trials. Lancet 350:1182 DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177188 Hochberg MC, Lebwohl MG, Plevy SE, Hobbs KF, Yocum DE (2005) The beneWt/risk proWle of TNF-blocking agents: Wndings of a consensus panel. Semin Arthritis Rheum 34:819836 Targan SR, Hanauer SB, van Deventer SJ, Mayer L, Present DH, Braakman T, DeWoody KL, Schaible TF, Rutgeerts PJ (1997) A short-term study of chimeric monoclonal antibody cA2 to tumor necrosis factor alpha for Crohns disease. Crohns Disease cA2 Study Group. N Engl J Med 337:10291035 Sandborn WJ, Hanauer SB, Katz S, Safdi M, Wolf DG, Baerg RD, Tremaine WJ, Johnson T, Diehl NN, Zinsmeister AR (2001) Etanercept for active Crohns disease: a randomized, double-blind, placebo-controlled trial. Gastroenterology 121:10881094 Buch MH, Conaghan PG, Quinn MA, Bingham SJ, Veale D, Emery P (2004) True inXiximab resistance in rheumatoid arthritis: a role for lymphotoxin alpha? Ann Rheum Dis 63:13441346

13. 14. 15.

16.

17.

18.

123