What is the best biological treatment for rheumatoid arthritis? A systematic review of effectiveness

Abstract

AIM: To evaluate the effectiveness of the biological disease-modifying antirheumatic drugs (bDMARD) in the treatment of rheumatoid arthritis through a systematic review of observational studies.

METHODS: The studies were searched in the PubMed, EMBASE, Cochrane Controlled Trials Register and LILACS databases (until August 2014), in the grey literature and conducted a manual search. The assessed criteria of effectiveness included the EULAR, the disease activity score (DAS), the Clinical Disease Activity Index, the Simplified Disease Activity Index, the American College of Rheumatology and the Health Assessment Questionnaire. The meta-analysis was performed with Review Manager^® 5.2 software using a random effects model. A total of 35 studies were included in this review.

RESULTS: The participants anti-tumor necrosis factor inhibitors (TNF) naïve, who used adalimumab (P = 0.0002) and etanercept (P = 0.0006) exhibited greater good EULAR response compared to the participants who used infliximab. No difference was detected between adalimumab and etanercept (P = 0.05). The participants who used etanercept exhibited greater remission according to DAS28 compared to the participants who used infliximab (P = 0.01). No differences were detected between adalimumab and infliximab (P = 0.12) or etanercept (P = 0.79). Better results were obtained with bDMARD associated with methotrexate than with bDMARD alone. The good EULAR response and DAS 28 was better for combination with methotrexate than bDMARD monotherapy (P = 0.03 e P < 0.00001). In cases of therapeutic failure, the participants who used rituximab exhibited greater DAS28 reduction compared to those who used anti-TNF agents (P = 0.0002). The participants who used etanercept achieved greater good EULAR response compared to those who did not use that drug (P = 0.007). Studies that assessed reduction of the CDAI score indicated the superiority of abatacept over rituximab (12.4 vs +1.7) and anti-TNF agents (7.6 vs 8.3). The present systematic review with meta-analysis found that relative to anti-TNF treatment-naïve patients, adalimumab and etanercept were more effective when combined with methotrexate than when used alone. Furthermore, in case of therapeutic failure with anti-TNF agents; rituximab and abatacept (non anti-TNF) and etanercept (as second anti-TNF) were more effective. However, more studies of effectiveness were found for the rituximab.

CONCLUSION: The best treatment for treatment-naïve patients is adalimumab or etanercept combined with methotrexate. For anti-TNF therapeutic failure, the best choice is rituximab, abatacept or etanercept.

Key Words: Systematic review; Meta-analysis; Effectiveness; Biological disease-modifying antirheumatic drugs; Rheumatoid arthritis

Core tip: Rheumatoid arthritis is a chronic, progressive, systemic inflammatory disease that preferentially affects the synovial membranes of joints, eventually leading to bone and cartilage destruction. Its worldwide prevalence is estimated to be 0.3% to 1%. Observational studies could provide relevant information for deciding the choice of treatments, the elaboration of clinical protocols, and the formulation of health policies. The present systematic review of biological disease-modifying antirheumatic drugs included cohort observational studies that reported treatment results applied in real-life conditions; thus, these studies are able to fill in gaps in knowledge left by clinical trials.

INTRODUCTION

Rheumatoid arthritis (RA) is a chronic, progressive, systemic inflammatory disease that preferentially affects the synovial membranes of joints, eventually resulting in destruction of bone and cartilage[1]. Its worldwide prevalence is estimated to be 0.3% to 1%[2].

Treatment of RA includes non-steroidal anti-inflammatory drugs, corticoids and synthetic (sDMARD) and biological [biological disease-modifying antirheumatic drugs (bDMARD)] disease-modifying antirheumatic drugs. bDMARD are indicated for individuals with persistent disease activity despite the use of sDMARD[3-5]. Tumor necrosis factor inhibitors (anti-TNF) are inhibitors of tumor necrosis factor alpha, rituximab is depleting B lymphocyte, abatacept is blocking of costimulation of T lymphocyte and tocilizumab is a blocking interleukin-6 receptor. Among the bDMARD, anti-TNF represent the first choice after failure of regimens that included sDMARD, and there is more evidence of the post-marketing efficacy and safety for anti-TNF agents[4,5]. Nevertheless, anti-TNF could eventually exhibit therapeutic failure, in which case another anti-TNF drug or another class of bDMARD might be used[6,7].

Appropriate knowledge of the effectiveness profiles of all of these strategies is relevant for choosing the best option for each patient. In this regard, observational studies are particularly interesting, as they seek to understand treatments in the actual practice setting. Thus, this type of study could contribute to decide the choice of treatments, the elaboration of clinical protocols, and the formulation of health policies. The present systematic review selected cohort observational studies. These types of studies more accurately represent real-life conditions (actual practice setting) and are able to provide complementary data to the results of randomized clinical studies conducted in controlled conditions[8].

The aim of the present study was to assess the effectiveness of the anti-TNFs adalimumab, etanercept, infliximab, golimumab and certolizumab pegol and of the non anti-TNF rituximab, tocilizumab and abatacept, in the treatment of active RA by means of a systematic review with meta-analysis.

MATERIALS AND METHODS

This systematic review followed the recommendations in the Cochrane Collaboration Handbook and was elaborated using Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA)[9,10].

Eligibility criteria

We included prospective and retrospective cohort studies and database records of patients with RA whose diagnoses were confirmed based on the ACR 1987 and the more recent ACR/EULAR 2010 criteria. Studies that accessed the effectiveness of adalimumab, etanercept, infliximab, golimumab, certolizumab pegol, rituximab, tocilizumab and abatacept between themselves, in monotherapy or combined with sDMARD were evaluated for inclusion.

Study search

We performed an electronic search of relevant articles published before August 2014 in the PubMed, EMBASE, Cochrane Controlled Trials Register and LILACS databases. Several combinations of terms corresponding to the disease, interventions and type of study were used in the search strategy (Table 1).

Table 1 Search strategies.

PubMed

(((Arthritis, Rheumatoid[Text Word] or "Arthritis, Rheumatoid"[Mesh]) and (((((((((((rituximab[Text Word] or Mabthera[Text Word]) or Rituxan[Text Word]) or IDEC-C2B8 antibody[Text Word]) or "rituximab"[Supplementary Concept]) or (((((((TNFR-Fc fusion protein[Text Word] or TNR 001[Text Word]) or TNR-001[Text Word]) or TNF receptor type II-IgG fusion protein[Text Word]) or recombinant human dimeric TNF receptor type II-IgG fusion protein[Text Word]) or Enbrel[Text Word]) or etanercept[Text Word]) or "TNFR-Fc fusion protein"[Supplementary Concept])) or ((((infliximab[Text Word] or monoclonal antibody cA2[Text Word]) or MAb cA2[Text Word]) or Remicade[Text Word]) or "infliximab"[Sup plementary Concept])) or ((adalimumab[Text Word] or Humira[Text Word]) or "adalimumab"[Supplementary Concept])) or (((((certolizumab[Text Word] or CDP870[Text Word]) or CDP 870[Text Word]) or Cimzia[Text Word]) or certolizumab pegol[Text Word]) or "certolizumab pegol"[Supplementary Concept])) or ((((((((((((abatacept[Text Word] or BMS 188667[Text Word]) or BMS-188667[Text Word]) or nulojix[Text Word]) or CTLA-4-Ig[Text Word]) or cytotoxic T lymphocyte-associated antigen 4-immunoglobulin[Text Word]) or CTLA4-Fc[Text Word]) or CTLA4-Ig[Text Word]) or LEA29Y[Text Word]) or Orencia[Text Word]) or BELATACEPT[Text Word]) or BMS-224818[Text Word]) or "abatacept" [Supplementary Concept])) or (((tocilizumab[Text Word] or atlizumab[Text Word]) or Actemra[Text Word]) or "tocilizumab"[Supplementary Concept])) or ("golimumab"[Supplementary Concept] or (Simponi[Text Word] or golimumab[Text Word]))))) and (("Cohort Studies"[Mesh]) or (((cohort*[Text Word]) or controlled clinical trial[Publication Type]) or epidemiologic methods))

EMBASE

"golimumab"/exp and [embase]/lim or ("cnto$148" and [embase]/lim) or ("simponi" and [embase]/lim) or ("tocilizumab"/exp and [embase]/ lim) or ("actemra" and [embase]/lim) or ("actemra 200" and [embase]/lim) or ("atlizumab" and [embase]/lim) or ("r$1569" and [embase]/lim) or ("roactemra" and [embase]/lim) or ("abatacept"/exp and [embase]/lim) or ("bms$188667" and [embase]/lim) or ("ctla4$ig" and [embase]/ lim) or ("ctla4 immunoglobulin" and [embase]/lim) or ("ctla4 immunoglobulin g" and [embase]/lim) or ("orencia" and [embase]/lim) or ("certolizumab pegol"/exp and [embase]/lim) or ("cdp$870" and [embase]/lim) or ("cimzia" and [embase]/lim) or ("pha$738144" and [embase]/ lim) or ("adalimumab"/exp and [embase]/lim) or ("humira"/exp and [embase]/lim) or ("monoclonal antibody d2e7" and [embase]/lim) or ("trudexa" and [embase]/lim) or ("infliximab"/exp and [embase]/lim) or ("avakine" and [embase]/lim) or ("inflectra" and [embase]/lim) or ("remicade" and [embase]/lim) or ("remsima" and [embase]/lim) or ("revellex" and [embase]/lim) or ("etanercept"/exp and [embase]/lim) or ("embrel" and [embase]/lim) or ("enbrel" and [embase]/lim) or ("recombinant tumor necrosis factor receptor fc fusion protein" and [embase]/lim) or ("tnr$001" and [embase]/lim) or ("tumor necrosis factor receptor fc fusion protein" and [embase]/lim) or ("rituximab"/exp and [embase]/lim) or ("idec c2b8" and [embase]/lim) or ("mabthera" and [embase]/lim) or ("monoclonal antibody idec c2b8" and [embase]/lim) or ("reditux" and [embase]/lim) or ("rituxan" and [embase]/lim) or ("rituxin" and [embase]/lim) and ("rheumatoid arthritis"/exp and [embase]/lim or ("arthritis, rheumatoid" and [embase]/lim)) and ("cohort analysis"/exp and [embase]/lim or ("longitudinal study"/exp and [embase]/lim) or ("prospective study"/exp and [embase]/lim) or ("follow up"/exp and [embase]/lim) or ("cohort$" and [embase]/lim))

Cochrane Controlled Trials Register

#1 MeSH descriptor: [Arthritis, Rheumatoid] explode all trees

#2 Rheumatoid Arthritis in Trials

#3 golimumab in Trials

#4 tocilizumab in Trials

#5 abatacept in Trials

#6 certolizumab pegol in Trials

#7 adalimumab in Trials

#8 infliximab in Trials

#9 etanercept in Trials

#10 rituximab in Trials

#11 #1 or #2 in Trials

#12 #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 in Trials

#13 #11 and #12

LILACS

(tw:((mh:(arthritis, rheumatoid)) or (tw:(artrite reumatoide)) or (tw:(artritis reumatoide)) )) and (tw:((tw:(adalimumab)) or (tw:(etanercept)) or (tw:(infliximab)) or (tw:(rituximab)) or (tw:(golimumab)) or (tw:(tocilizumab)) or (tw:(abatacept)) or (tw:(certolizumab pegol))))

In addition, we conducted a manual search in the 2012 and 2013 editions of four rheumatology journals (Journal Rheumatology, Rheumatology, Rheumatology International and the Brazilian Journal of Rheumatology) and in the abstracts of the ACR and the EULAR meetings. Also, we searched for grey literature in the Digital Library of Theses and Dissertations of University of São Paulo, and ProQuest Dissertation and Theses Database.

Study selection and data collection processes

We performed the study selection in duplicate by four independent examiners (JBS, JOC, HAOJ, LLPL). The steps included analysis of titles, abstracts, and analysis of the full-texts of articles. Divergences were analyzed by another reviewer (VEA). Data collection was performed by four investigators (JBS, JOC, HAOJ, LLPL). The authors were contacted for additional information whenever needed. We assessed effectiveness as indicated by the rate of response to bDMARD according to the criteria of ACR and EULAR. We also analysed the Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), Disease Activity Score (DAS28) and the Health Assessment Questionnaire (HAQ).

Assessment of methodological quality

The methodological quality of each study was assessed by four examiners (JBS, JOC, HAOJ, LLPL); divergences were solved by consensus. For that purpose, we used the Newcastle-Ottawa scale, as recommended by the Cochrane Collaboration in the case of observational studies[11]. This scale assesses studies in three major domains: selection of the study groups, comparability of groups, and ascertainment of exposure and of results of interest. The maximum total score is nine stars, and scores above six stars are indicative of high methodological quality.

Funding sources were identified to establish potential sources of bias. Publication bias was assessed by funnel plot analysis of the results of EULAR responses and DAS28.

Statistical analysis

We used the Software Review Manager^® 5.2 to perform the meta-analyses. The results are expressed as relative risks (dichotomous variables) or means differences (continuous variables) with the corresponding 95%CIs. Values of I² > 40% and P < 0.10 on the χ² test were considered as indicative of significant heterogeneity. The causes of heterogeneity were investigated by excluding one study at a time and checking the changes in I² and P values.

RESULTS

Study inclusion

A total of 5422 articles were found in the investigated electronic databases, and a further four after manual search. Following the exclusion of duplicates, 5089 articles were selected for title analysis, from which 1724 were selected for abstract analysis, and finally 131 for full-text reading. Following full-text reading, 35 studies were included in the review, corresponding to 30 full-text articles[12-42] and five abstracts[43-47] (Figure 1). No observational study assessed the medicines golimumab or certolizumab pegol.

Figure 1 Study flow diagram.

Characteristics of the studies

Among the 35 observational studies included, 16 were registry studies and 19 were cohort studies; eight were retrospective, and 27 were prospective. The study duration varied from 15 to 80 mo, though this information was not provided by some authors. The participants were followed from three to 48 mo. Five studies were funded by pharmaceutical companies, two studies were not funded by the pharmaceutical industry, and 16 had mixed funding; in the remainder articles the authors did not disclose the funding source. Nine studies assessed anti-TNF naïve participants, and 11 studies assessed cases of therapeutic failure with at least one anti-TNF agent; the remainder of the studies did not inform whether therapeutic failure had occurred or did not separate patients into subgroups (Table 2). Disease duration varied from 6 to 20 years. Approximately 50% of the participants used glucocorticoids and the use of sDMARD varied from 31% to 100%. In most of the studies, the DAS28 score was > 5.1, which indicates high disease activity. The HAQ score varied from 0.4 to 2.2 (Figure 2).

Table 2 Characteristics of included studies.

Nstudy	Ref.	Type of study	Time horizon	Patient	Intervention	Country conducting the study	Funding Sources	Duration of the study (mo)	Follow-up (mo)
1	Geborek et al[12]	Cohort	Prospective	Naive	ETA vs IFX vs LEF	Sweden	NR	24	12
2	Van Vollenhoven et al[13]	Registry	Prospective	NR	ETA vs ETA + MTX	Sweden	Mixed	NR	12
3	Cohen et al[14]	Cohort	Retrospective	Therapeutic failure	IFX vs ETA	France	NR	48	3
4	Finckh et al[15]	Registry	Prospective	Mixed	ADA vs ETA vs IFX	Switzerland	Mixed	80	12
5	Heiberg et al[16]	Cohort	Prospective	Mixed	ADA monotherapy vs ADA + MTX	Norway	Mixed	NR	12
6	Hyrich et al[17,18]	Registry	Prospective	NR	ETA monotherapy vs ETA + MTX vs ETA + DMARD and ADA monotherapy vs ADA + MTX vs ADA + DMARD	England	Pharmaceutical industry	NR	6
7	Kristensen et al[19]	Cohort	Prospective	Naive	ETA vs IFX	Sweden	Mixed	55	36
8	Bernal Rivera et al[20]	Cohort	Prospective	Naive	ADA vs ETA vs IFX	Spain	NR	24	12
9	Kristensen et al[19]	Cohort	Prospective	Naive	ETA vs IFX	Spain	NR	72	6
10	Radstake et al[23]	Cohort	Prospective	NR	IFX vs ADA	The Netherlands	Mixed	NR	6
12	Bazzani et al[24]	Registry	Prospective	Mixed	ADA vs ETA vs IFX	Italy	Pharmaceutical industry	25.29	36
13	Greenwood et al[47]	Cohort	Retrospective	NR	ADA vs ETA vs IFX	England	NR	NR	12
14	Laas et al[25]	Cohort	Prospective	Naive	ETA vs ADA	Finland	No pharmaceutical industry	36	3
15	Arenere Mendoza et al[26]	Cohort	Retrospective	Mixed	ADA vs ETA vs IFX	Spain	NR	80	12
16	Buch et al[46]	Cohort	Prospective	Therapeutic failure	RTX vs anti-TNF	England	NR	NR	6
17	Canhão et al[27]	Registry	Prospective	Naive	ADA vs ETA vs IFX	Portugal	Mixed	NR	12
18	Hetland et al[28]	Registry	Prospective	Naive	ADA vs ETA vs IFX	Denmark	Mixed	86	12
19	Blom et al[29]	Registry	Prospective	Therapeutic failure	RTX vs anti-TNF	The Netherlands	Mixed	NR	12
20	Chatzidionysiou et al[30]	Registry	Prospective	Mixed	RTX monotherapy vs RTX + MTX vs RTX + LEF	Europe	Pharmaceutical industry	NR	12
21	Gotenberg et al[45]	Registry	Prospective	Mixed	RTX vs ABAT	France	NR	NR	6
22	Iannone et al[32]	Registry	Prospective	NR	ADA vs ETA vs IFX	Italy	NR	NR	48
23	Leffers et al[33]	Registry	Prospective	Mixed	ABAT vs TOCI	Denmark	Mixed	NR	48
24	Martínez-Pérez et al[44]	Cohort	Retrospective	Mixed	RTX vs IFX	Spain	NR	NR	12
25	Wakabayashi et al[34]	Cohort	Retrospective	Therapeutic failure	TOCI vs ETA	Japan	No pharmaceutical industry	60	12
26	Finckh et al[36]	Cohort	Prospective	Therapeutic failure	RTX vs anti-TNF	Switzerland	Mixed	NR	24
27	Gomez-Reino et al[35]	Cohort	Prospective	Therapeutic failure	RTX vs anti-TNF	Spain	Pharmaceutical industry	36	12
28	Greenberg et al[37]	Registry	Prospective	Naive	ADA vs ETA vs IFX	Unied States	Mixed	74	24
29	Kekow et al[38]	Cohort	Retrospective	Therapeutic failure	RTX vs anti-TNF	Germany	Pharmaceutical industry	NR	6
30	Schabert et al[39]	Cohort	Retrospective	NR	ADA vs ETA vs IFX	Unied States	Mixed	15	12
31	Chatzidionysiou et al[40]	Registry	Prospective	Therapeutic failure	Anti-TNF vs ETA vs ADA	Stockholm	NR	NR	6
32	Keystone et al[43]	Cohort	Retrospective	Therapeutic failure	ABAT vs TOCI	Canada	NR	NR	12
33	Emery et al[41]	Cohort	Prospective	Therapeutic failure	RTX vs anti-TNF	Multicentre	Mixed	NR	12
34	Flouri et al[42]	Registry	Prospective	Mixed	ADA vs ETA vs IFX	Greece	Mixed	60	12
35	Harrold et al[31]	Registry	Prospective	Therapeutic failure	ABAT vs TOCI	Unied States	Mixed	NR	12

ADA: Adalimumab; ETA: Etanercept; IFX: Infliximab; RTX: Rituximab; ABAT: Abatacept; TOCI: Tocililizumab; LEF: Leflunomid; MTX: Methotrexate; sDMARD: Synthetic disease-modifying antirheumatic drugs; NR: Not reported.

Figure 2 Patient characteristics of included articles. ADA: Adalimumab; ETA: Etanercept; IFX: Infliximab; RTX: Rituximab; ABAT: Abatacept; TOCI: Tocililizumab; MTX: Methotrexate; sDMARD: Synthetic disease-modifying antirheumatic drugs; NR: Not reported; NS: Not significant; NA: Not applicable.

Methodological quality

From the 35 analyzed studies, two achieved the highest score on the Newcastle-Ottawa scale, nine stars; 14, eight stars; seven, seven stars; 10, six stars; and two, five stars (Table 3). The funnel plot did not exhibit asymmetry relative to outcomes in the DAS 28 and EULAR response, which indicated the absence of publication bias, and thus of overestimation of the intervention effects calculated in the meta-analysis (data not shown).

Table 3 Quality assessment of articles for Newcastle Ottawa scale.

N study	Ref.	Selection				Comparability	Results			Total
		Representati-veness of the cases	Selection of controls	Ascertain-ment of exposure	Demonstration that outcome of interest was not present at start of study	Comparability of cohorts on the basis of the design or analysis	Assessment of outcome	Was follow-up long enough for outcomes to occur	Adequacy of follow up of cohorts
1	Geborek et al[12]	1	1	1	1	2	0	1 (12 mo)	1	8
2	Van Vollenhoven et al[13]	0	1	1	1	2	0	1 (24 mo)	0	6
3	Cohen et al[14]	1	1	1	1	1	0	1 (3 mo)	0	6
4	Finckh et al[15]	1	1	1	0	1	1	1 (12 mo)	1	7
5	Heiberg et al[16]	1	1	1	1	2	0	1 (6 mo)	1	8
6	Hyrich et al[17,18]	1	1	1	1	2	0	1 (6 mo)	1	8
7	Kristensen et al[19]	1	1	1	0	2	0	1 (36 mo)	0	6
8	Bernal Rivera et al[20]	1	1	1	0	2	0	1 (12 mo )	1	7
9	Fernández-Nebro et al[21]	1	1	1	1	2	0	1 (6 mo)	1	8
10	Radstake et al[23]	0	1	1	1	2	0	1 (6 mo)	0	6
11	Kievit et al[22]	1	1	1	1	2	0	1 (6 mo)	0	7
12	Bazzani et al[24]	1	1	1	0	2	1	1 (6 mo)	1	8
13	Greenwood et al[47]	0	1	1	1	2	0	1 (12 mo)	0	6
14	Laas et al[25]	1	1	1	1	2	1	1 (3 mo)	1	9
15	Arenere Mendoza et al[26]	1	1	1	1	2	0	1 (12 mo)	0	7
16	Buch et al[46]	1	1	1	0	2	0	1 (6 mo)	0	6
17	Canhão et al[27]	1	1	1	0	2	0	1 (12 mo)	1	7
18	Hetland et al[28]	1	1	1	1	2	0	1 (12 mo)	1	8
19	Blom et al[29]	1	1	1	1	2	1	1 (12 mo)	1	9
20	Chatzidionysiou et al[30]	1	1	1	1	2	0	1 (12 mo)	0	7
21	Gotenberg et al[45]	1	1	1	1	2	0	1 (6 mo)	1	8
22	Iannone et al[32]	1	1	1	0	2	0	1 (48 mo)	0	6
23	Leffers et al[33]	1	1	1	1	2	0	1 (12 mo)	0	7
24	Martínez-Pérez et al[44]	0	1	0	1	2	0	1 (12 mo)	0	5
25	Wakabayashi et al[34]	0	1	1	1	2	1	1 (12 mo)	1	8
26	Finckh et al[36]	1	1	1	1	1	1	1 (24 mo)	1	8
27	Gomez-Reino et al[35]	1	1	1	0	2	1	1 (12 mo)	1	8
28	Greenberg et al[37]	1	1	1	1	2	0	1 (24 mo)	1	8
29	Kekow et al[38]	1	1	1	1	2	0	1 (6 mo)	1	8
30	Schabert et al[39]	1	1	1	1	1	1	1 (12 mo)	1	8
31	Chatzidionysiou et al[40]	1	1	1	0	2	0	1 (6 mo)	0	6
32	Keystone et al[43]	0	1	1	0	2	0	1 (12 mo)	0	5
33	Emery et al[41]	1	1	1	0	2	0	1(12 mo)	0	6
34	Flouri et al[42]	1	1	1	1	2	0	1 (12 mo)	1	8
35	Harrold et al[31]	1	1	1	0	2	0	1 (12 mo)	0	6

Data synthesis

A total of 22 studies assessed the drugs adalimumab, etanercept and infliximab; nine studies assessed anti-TNF naïve patients only and seven anti-TNF naïve participants and cases of therapeutic failure; six studies did not inform whether therapeutic failure had occurred. Nineteen of those studies were included in the meta-analyses of EULAR responses, DAS28, remission acoording to DAS28, CDAI, SDAI, ACR20, 50 and 70, and HAQ (Table 4).

Table 4 Meta-analysis of the outcomes for patients with treatment-naïve and therapeutic failure.

Intervention	Outcomes	Studies (references)	Partici-pants	Relative risk (95%CI) or other mesure	I2(%)	P value
IFX vs ETA	EULAR good response	10 (18,19,21,22,24, 26,27, 28, 32,42)	7247	0.86 [0.72-1.02]	76	< 0.0001
	EULAR moderate response	9 (18,19,21,22,24, 26, 28, 32,42)	6791	0.98 [0.84-1.15]	78	< 0.0001
	EULAR no response	9 (18,19,21,22,24, 26, 28, 32,42)	6791	1.20 [1.05-1.38]	46	0.06
	DAS 28 remission	7 (21,26,27,28,32,37,42)	2868	0.70 [0.59-0.84]	0	0.51
	DAS 28	2 (21,26)	196	0.40 [-0.27- 1.07]	59	0.12
	DAS 28 reduction	2 (15,22)	1321	0.40 [0.04-0.77]	77	0.04
	CDAI remission	4 (27,28,37,42)	2293	0.90 [0.74-1.09]	0	0.89
	SDAI remission	2 (27,42)	840	0.87 [0.61-1.26]	0	0.9
	HAQ	3 (21,26,39)	495	0.14 [0.00-0.27]	0	0.51
	ACR 20	2 (19,37)	1309	0.95 [0.86-1.06]	0	0.47
	ACR50	3 (19,28,37)	2315	0.92 [0.81-1.03]	10	0.33
	ACR70	3 (19,28,37)	2315	0.88 [0.57-1.36]	79	0.009
ADA vs ETA	EULAR good response	8 (20,22,24,26,27,28,32,42)	2492	0.97 [0.79-1.20]	73	0.0005
	EULAR moderate response	7 (20,22,24,26,28,32,42)	2080	1.00 [0.89-1.12]	0	0.48
	EULAR no response	7 (20,22,24,26,28,32,42)	2080	0.90 [0.62-1.32]	76	0.0003
	DAS 28 remission	6 (26,27,28,32,37,42)	2412	0.93 [0.68-1.26]	80	0.0001
	DAS 28	2 (20,26)	180	-0.09 [-0.25-0.06]	0	0.73
	DAS 28 reduction	2 (15,22)	1392	0.17 [-0.19-0.52]	68	0.08
	CDAI remission	4 (27,28,37,42)	1883	1.16 [0.77-1.74]	70	0.02
	SDAI remission	2 (27,42)	641	1.40 [0.76-2.59]	55	0.13
	HAQ	2 (26,39)	339	-0.15 [-0.39-0.10]	49	0.16
	HAQ reduction	2 (22,25)	653	-0.07 [-0.16-0.03]	0	0.92
	ACR 20	2 (20,37)	445	0.89 [0.71-1.12]	0	0.68
	ACR 50	3 (20,28,37)	1217	1.09 [0.91-1.31]	18	0.3
	ACR 70	3 (20,28,37)	1436	1.15 [0.92-1.43]	0	0.82
IFX vs ADA	EULAR good response	8 (22,23,24,26,27,28,32,42)	3025	1.25 [1.06-1.47]	57	0.02
	EULAR moderate response	7 (22,23,24,26,28,32,42)	2657	0.91 [0.79-1.04]	31	0.19
	EULAR no response	7 (22,23,24,26,28,32,42)	2657	0.77 [0.56-1.05]	75	0.0006
	DAS 28 remission	6 (26,27,28,32,37,42)	2760	1.15 [0.91-1.46]	63	0.02
	DAS 28 reduction	2 (15,22)	1097	-0.24 [-0.96-0.48]	91	0.001
	CDAI remission	4 (27,28,37,42)	2332	1.30 [0.90-1.88]	68	0.02
	SDAI remission	2 (27,42)	765	1.66 [0.94-2.93]	61	0.11
	HAQ	2 (26,39)	182	-0.33 [-0.53-0.13]	0	0.92
	ACR 50	2 (28,37)	1458	1.14 [0.71-1.84]	79	0.03
	ACR 70	2 (28,37)	1458	1.41 [0.81-2.44]	72	0.06

A value of I² > 40% indicates statistical heterogeneity between the studies. A value of P < 0.10 from the chi-square test indicates statistical heterogeneity between the studies. CI: Confidence interval; ADA: Adalimumab; ETA: Etanercept; IFX: Infliximab; EULAR: European League Against Rheumatism; DAS 28: Disease activity score; SDAI: Simplified disease activity index; CDAI: Clinical disease activity; HAQ: Health Assessment Questionnaire; ACR: American College Rheumatology.

The good EULAR response for the participants who used etanercept was no different as that for the participants who used infliximab (P = 0.08) (Figure 3). However, the meta-analysis exhibited high heterogeneity. Following exclusion of the studies by Kristensen et al[19] (2006) and Hyrich et al[17] (2006), the heterogeneity was lowered, and the results became favorable to etanercept (P < 0.0001). No difference was found between adalimumab and etanercept (P = 0.80) (Figure 4). That meta-analysis also exhibited high heterogeneity; and after the exclusion of the study by Iannone et al[32] (2011), no heterogeneity was detected (P = 0.05). The participants who used adalimumab presented higher good EULAR response compared to those who used infliximab (P = 0.009) (Figure 5). However, that meta-analysis exhibited high heterogeneity. Following exclusion of the study by Iannone et al[32] (2011), the heterogeneity was lowered (P < 0.00001). Comparison of etanercept vs infliximab, adalimumab vs etanercept, and adalimumab vs infliximab found similar results relative to moderate EULAR response (P > 0.05). Regarding the EULAR no response, the results were favorable to infliximab compared to etanercept (P = 0.01), while no difference was detected between adalimumab and infliximab (P = 0.09) or etanercept (P = 0.60). The study by Gottenberg et al[45] (2011), which was not included in the meta-analysis due to the lack of studies comparing abatacept and rituximab, did not detect a difference in the EULAR responses between the two drugs (P > 0.05). Additionally, the study by Leffers et al[33] could not be included in the meta-analysis for the same reason and did not detect a difference in the EULAR responses between abatacept and tocilizumab (P > 0.05).

Figure 3 EULAR good response - Etanercept vs Infliximab.

Figure 4 EULAR good response - Etanercept vs Adalimumab.

Figure 5 EULAR good response - Infliximab vs Adalimumab.

The participants who used etanercept exhibited greater remission according to DAS28 compared to the participants who used infliximab (P < 0.0001). Comparison of adalimumab and infliximab did not reveal a significant difference (P = 0.23). However, that meta-analysis exhibited moderate heterogeneity. Following exclusion of the study by Iannone et al[32], the heterogeneity was lowered, and the result became favorable to adalimumab (P = 0.001). No significant difference was detected between adalimumab and etanercept (P = 0.63). However, the meta-analysis exhibited high heterogeneity. Following exclusion of the study by Iannone et al[32], heterogeneity was lowered (P = 0.21). The participants who used etanercept exhibited greater reduction in the DAS28 score compared to the participants who used infliximab (P = 0.03). Significant differences were not detected between adalimumab and etanercept (P = 0.36) or infliximab (P = 0.52). Comparison of etanercept vs infliximab or adalimumab did not reveal any statistically significant differences relative to DAS28 (P > 0.05). The study by Arenere Mendoza et al[26] (2010), which was not included in the meta-analysis due to the lack of studies that analyzed the DAS28 outcome, did not report differences between adalimumab and infliximab (P > 0.05). The study by Greenwood et al[47] (2009), which was not included in the meta-analysis due to lack of data, did not report significant differences in the DAS28 response when comparing adalimumab vs etanercept, infliximab vs adalimumab, and infliximab vs etanercept (P > 0.05). The study by Gottenberg et al[45] (2011), which was also not included in meta-analysis, did not report a difference relative to DAS28 outcome between abatacept and rituximab (P > 0.05). The study by Leffers et al[33] (2011) also did not report a difference between abatacept and tocilizumab relative to DAS28 remission (P > 0.05).

In regard to the ACR20 outcome, the comparison of etanercept vs adalimumab or infliximab presented similar results (P > 0.05). Comparisons of etanercept vs infliximab, etanercept vs adalimumab, and infliximab vs adalimumab did not reveal differences relative to the outcomes of CDAI and SDAI remission, ACR50 and ACR70 (P > 0.05).

The HAQ scores of the participants who used adalimumab (P = 0.0009) and etanercept (P =0.04) were better compared to the participants who used infliximab. Adalimumab and etanercept were not different in regard to that outcome (P = 023). Adalimumab and etanercept were also not different in terms of the HAQ reduction outcome (P = 0.16). The study by Martinez-Pérez et al[44] (2011), which was not included in the meta-analysis due to the lack of studies comparing infliximab and rituximab, did not report a difference with respect to HAQ (P > 0.05). The study by Leffers et al[33] (2011), which was also not included in the meta-analysis, did not report a significant difference in HAQ between abatacept and tocilizumab (P > 0.05).

Anti-TNF naïve patients: Nine studies assessed anti-TNF-naïve individuals only, from which seven were included in the meta-analysis that assessed the outcomes of EULAR, remission according to DAS28 and CDAI, and ACR20, 50, 70 (Table 5).

Table 5 Meta-analysis of the outcomes for anti-tumor necrosis factor naïve patients.

Intervention	Outcomes	Studies (references)	n	RR (95%CI)or other mesure	I2(%)	P value
IFX vs ETA	EULAR good response	5 (19, 21, 22, 27, 28)	2822	0.82 [0.62-1.09]	82	0.0001
	EULAR moderate response	4 (19, 21, 22, 28)	2366	0.90 [0.61-1.33]	90	< 0.00001
	EULAR no response	4 (19, 21, 22, 28)	2366	1.29 [1.09-1.53]	27	0.25
	DAS 28 remission	4 (21, 27, 28, 37)	1804	0.82 [0.70-0.95]	0	0.4
	ACR 20	2 (19, 37)	1309	0.95 [0.86-1.06]	0	0.47
	ACR50	3 (19, 28, 37)	2315	0.92 [0.81-1.03]	10	0.33
	ACR70	3 (19, 28, 37)	2315	0.88 [0.57-1.36]	79	0.009
	CDAI remission	3 (27, 28, 37)	1876	0.88 [0.72-1.08]	0	0.93
ADA vs ETA	EULAR good response	4 (20, 22, 27, 28)	1590	1.11 [1.00-1.23]	0	0.4
	EULAR moderate response	3 (20, 22, 28)	1178	1.01 [0.83-1.24]	19	0.29
	EULAR no response	3 (20, 22, 28)	1178	0.69 [0.53-0.89]	11	0.32
	DAS 28 remission	3 (27, 28, 37)	1380	1.03 [0.82-1.29]	37	0.21
	ACR 20	2 (20, 37)	445	0.89 [0.71-1.12]	0	0.68
	ACR 50	3 (20, 28, 37)	1217	1.09 [0.91-1.31]	18	0.3
	ACR 70	3 (20, 28, 37)	1436	1.15 [0.92-1.43]	0	0.82
	HAQ reduction	2 (22, 25)	653	-0.07 [-0.16-0.03]	0	0.92
	CDAI remission	3 (27, 28, 37)	1601	1.02 [0.67-1.56]	72	0.03
IFX vs ADA	EULAR good response	3 (22, 27, 28)	1706	1.42 [1.18-1.72]	42	0.18
	EULAR moderate response	2 (22, 28)	1338	0.96 [0.58-1.59]	80	0.03
	EULAR no response	2 (22, 28)	1338	0.56 [0.45-0.69]	0	0.88
	DAS 28 remission	3 (27, 28)	1648	1.23 [0.95-1.59]	48	0.15
	ACR 50	2 (28, 37)	1458	1.14 [0.71-1.84]	79	0.03
	ACR 70	2 (28, 37)	1458	1.41 [0.81-2.44]	72	0.06
	CDAI remission	3 (27, 28, 37)	1875	1.17 [0.75-1.82]	75	0.02

A value of I² > 40% indicates statistical heterogeneity between the studies. A value of P < 0.10 from the chi-square test indicates statistical heterogeneity between the studies. CI: Confidence interval; ADA: Adalimumab; ETA: Etanercept; IFX: Infliximab; EULAR: European League Against Rheumatism; DAS 28: Disease activity score; CDAI: Clinical disease activity; HAQ: Health assessment questionnaire; ACR: American College Rheumatology.

The good EULAR response for the participants who used etanercept were the same as those of the participants who used infliximab (P = 0.17). However, that meta-analysis exhibited high heterogeneity. Following exclusion of the study by Kristensen et al[19] (2006), the heterogeneity was low, and the results became favorable to etanercept (P = 0.0006). No difference was detected between adalimumab and etanercept (P = 0.05). The participants who used adalimumab exhibited greater good EULAR response compared to the participants who used infliximab (P =0.0002). The results relative to the moderate EULAR response outcome were similar in the comparisons of etanercept vs infliximab, adalimumab vs etanercept, and adalimumab vs infliximab (P > 0.05). With regard to the EULAR no response, the results were favorable to infliximab compared to etanercept (P = 0.004) or adalimumab (P < 0.00001) and to etanercept compared to adalimumab (P = 0.004).

The participants who used etanercept exhibited greater remission according to DAS28 compared to the participants who used infliximab (P = 0.01). No differences were detected between adalimumab and infliximab (P = 0.12) or etanercept (P = 0.79).

With regard to the ACR20 outcome, the results of etanercept vs adalimumab or infliximab were not different (P > 0.05). No differences were detected for the outcomes of ACR50 and 70 and the remission according to CDAI between etanercept and infliximab, etanercept and adalimumab, and infliximab and adalimumab (P > 0.05). Only the study by Greenberg et al[37] (2012) compared the ACR20 outcome between adalimumab and infliximab (P > 0.05). The study by Geborek et al[12] (2002), which was not included in the meta-analysis because it reported graphical data without numerical values; showed that the ACR 20 response was better with etanercept compared to infliximab (P < 0.05), while no differences were detected relative to ACR50 and 70.

The study by Kievit et al[22] (2008) assessed the HAQ reduction outcome and found that the results were better with adalimumab compared to infliximab (P < 0.05). Kievit et al[22] (2008) and Laas et al[25] (2009) did not detect a difference between adalimumab and etanercept (P = 0.16).

Patients with anti-TNF therapeutic failure: Eleven studies assessed anti- TNF therapeutic failure only, from which nine were included in the meta-analyses that assessed the outcomes of EULAR response and DAS28 reduction (Table 6).

Table 6 Meta-analysis of the outcomes for patients with anti-tumor necrosis factor therapeutic failure.

Intervention	Outcomes	Studies (references)	n	Relative risk (95%CI)or other mesure	I2(%)	P value
RTX vs anti-TNF	EULAR good response	4 (35, 38, 40, 46)	1608	0.96 [0.60-1.54]	74	0.009
	EULAR moderate response	5 (29, 35, 38, 40, 46)	1706	1.02 [0.79-1.32]	66	0.02
	EULAR no response	3 (35, 38, 40)	1406	1.00 [0.53-1.89]	85	0.001
	DAS 28 reduction	6 (35, 36, 38, 40, 41, 46)	1584	0.42 [-0.65--0.20]	62	0.02
ETA vs control	EULAR good response	2 (14, 40)	173	2.11 [1.23-3.62]	0	0.48
	IFX		38	1.60 [0.63-4.09]
	RTX		135	2.42 [1.25-4.68]
	DAS 28 reduction	2 (34, 40)	152	0.15 [-0.65-0.95]	77	0.04
	RTX		113	-0.22 [-0.64-0.20]
	TOCI		39	0.60 [-0.05-1.25]

A value of I² > 40% indicates statistical heterogeneity between the studies. A value of P < 0.10 from the chi-square test indicates statistical heterogeneity between the studies. CI: Confidence interval; RTX: Rituximab; ETA: Etanercept; IFX: Infliximab; TOCI: Tocilizumab; EULAR: European League Against Rheumatism; DAS 28: Disease activity score.

The participants who used rituximab exhibited greater DAS28 reduction compared to those who used anti-TNF agents (P = 0.0002); however, the EULAR responses did not differ between these groups (P > 0.05). In addition, all of the corresponding meta-analyses exhibited high statistical heterogeneity. The study by Blom et al[29] (2011), which was not included in the meta-analysis because it did not report the DAS28 absolute scores, also detected lower scores for the participants who used rituximab compared to those who used anti-TNF (P = 0.004). Among four studies that assessed HAQ, only the Finckh et al[36] study (2012) found that the participants who used rituximab exhibited greater score reductions compared to those who used anti-TNF, which did not represent a clinically significant improvement (e.g., a reduction of 0.22 points in the HAQ score).

The participants who used etanercept achieved greater good EULAR response compared to those who did not use that drug (P = 0.007); that difference resulted from one study that compared etanercept vs rituximab[37]. With regard to the DAS28 score reduction, no differences were reported between the groups (P = 0.71).

The abstracts of studies that assessed reduction of the CDAI score indicated the superiority of abatacept over rituximab (12.4 vs +1.7) and anti-TNF agents (7.6 vs 8.3); those results could not be assessed in the meta-analysis due to the lack of data[28,41]. Harrold et al[31] (2014) also assessed ACR20 and 50 outcomes and did not detect any differences between the groups treated with abatacept or anti-TNF agents [0.87 (0.59; 1.29) and 0.86 (0.58; 1.27), respectively].

Patients who used bDMARD in monotherapy or in combination with methotrexate: Four studies assessed individuals treated with bDMARD monotherapy or in combination with methotrexate, and three were included in the meta-analysis that assessed EULAR response, DAS28 and HAQ outcomes (Table 7).

Table 7 Meta-analysis of the outcomes for patients in treatment with biological monotherapy vs biological in combination with methotrexate.

Intervention	Outcomes	Studies (references)	Participants	Relative risk (95%CI) or other mesure	I2(%)	P value
bDMARD monotherapy vs bDMARD + MTX	EULAR good response	3 (16,17, 30)	3000	0.57 [0.34-0.95]	82	0.0008
	DAS 28	3 (17, 20, 30)	2913	0.25 [-0.02-0.52]	69	0.01
	HAQ	2 (165, 30)	655	0.13 [0.03-0.22]	0	0.43

A value of I² > 40% indicates statistical heterogeneity between the studies. A value of P < 0.10 from the χ² test indicates statistical heterogeneity between the studies. CI: Confidence interval; bDMARD: Biological disease-modifying antirheumatic drugs; MTX: Methotrexate; EULAR: European League Against Rheumatism; DAS 28: Disease activity score; HAQ: Health Assessment Questionnaire.

Regarding the good EULAR response, combination with methotrexate was better than bDMARD monotherapy (P = 0.03) (Figure 6). No difference was found relative to DAS28 between bDMARD monotherapy and combination with methotrexate (P = 0.07). However, this meta-analysis exhibited high heterogeneity. Following exclusion of the study by Chatzidionysiou et al[30] (2012)[30], no heterogeneity was detected, and the results became favorable to the combination with methotrexate (P < 0.00001). The study by van Vollenhoven et al[13] (2003), which was not included in the meta-analysis because it reported graphical data without numerical values, reported a difference in DAS28 favorable to combination with methotrexate compared to bDMARD monotherapy (P < 0.05). The study by Heiberg et al[16] (2006), which was not included in the meta-analysis due to the lack of studies that assessed the DAS28 reduction outcome, reported a difference favorable to bDMARD in combination with methotrexate (P < 0.05). With regard to the HAQ score outcome, the best results were exhibited by bDMARD in combination with methotrexate (P = 0.009).

Figure 6 EULAR good response - Biologic monotherapy vs Biologic + MTX.

DISCUSSION

Patients who used adalimumab and etanercept presented similar results among them and better outcomes compared to patients under infliximab therapy. The analysis of subgroup of anti-TNF naïve participants showed better results for adalimumab and etanercept compared to infliximab. The results were similar to the group with all patients (anti-TNF naïve and/or therapeutic failure), probably because most of the participants under treatment were anti-TNF naïve. The use of bDMARD in combination with methotrexate exhibited greater results than bDMARD monotherapy. Rituximab, etanercept and abatacept proved to be effective therapeutic options following therapeutic failure with anti-TNF agents. However, most of the studies on therapeutic failure assessed rituximab; thus, more studies comparing other drugs are needed to contribute to the choice of third-line agents in actual clinical practice.

Systematic reviews that performed indirect comparison meta-analyses of randomized clinical trials that assessed the efficacy of the anti-TNFs adalimumab, etanercept and infliximab reported similar results[48-50]. One meta-analysis found that the efficacy of etanercept was lower compared to that of infliximab and adalimumab; however, the patients selection for the study was different: the study divided patients by those on etanercept (methotrexate-naïve individuals) and other drugs (patients resistant to methotrexate), which makes the comparison of the results between the medicines difficult[51]. The difference of these studies relative to ours might be most likely due to the characteristics of the participants and the low dose of infliximab (3 mg/kg). Some studies reported that patients using infliximab required dose escalation more often compared to those who used etanercept and adalimumab[52,53]. Dose escalation might increase the cost of treatment with infliximab[52] and might thus result in moderate effectiveness[54]. In addition, Pascual-Salcedo et al[55] observed that the production of anti-infliximab antibodies is associated with loss of clinical response[55].

The superiority of the combination of bDMARD and sDMARD compared to bDMARD monotherapy was also reported in other recent meta-analyses[51,56]. In particular, the same pattern was reported for etanercept in combination with methotrexate in a randomized clinical trial[57,58]. The fact that infliximab should be administered in combination with methotrexate is well established[59].

Despite the publication of recent studies on the subject, the definition of the best strategy for patients who exhibit therapeutic failure to at least one anti-TNF agent still poses a challenge[60]. Some studies assessed subgroups in an attempt to identify profiles of patients who will benefit from treatment with rituximab. Thus, whereas testing positive for rheumatoid factor did not induce significant changes in the results[61], rituximab proved to be more effective in individuals who tested positive for rheumatoid factor and for anti-cyclic citrullinated peptide antibody[39].

One of the limitations of systematic reviews with meta-analysis of cohort studies concerns selection bias, which is intrinsic to the design of such studies, as the participants are not randomized but might be allocated to a given treatment based on their patient and physician preferences. A consequence of that limitation was the difference noted among the groups at the onset of treatment that, as a whole, manifested as poorer prognosis in the participants from the rituximab group relative to the numbers of anti-TNF and sDMARD previously used, older age, and greater DAS28 and HAQ scores at baseline[22,36,41].

One further limitation is related to the fact that observational studies are conducted under real-life non-controlled conditions. For that reason, differences were detected in the number of participants among the groups, in the disease activity, and in the lack of dose standardization, especially in the case of infliximab. Moreover, observational studies have the advantage of recruiting large numbers of participants. These types of studies more accurately represent real-life conditions and are able to provide complementary data to the results of randomized clinical studies. Some studies reported that the participants in randomized controlled clinical trials exhibited greater disease activity and fewer associated comorbidities compared to those patients treated in the actual practice setting. The practice of prescribing has been modified over time in real-life. bDMARDs (specially in clinical trials) were prescribed only when patients presented high activity of disease and, now, the medicines are prescribed when the activity is moderate or high[62-64]. Kievit et al[65] (2007) called attention to the reduction of the external validity of randomized clinical trials[65], while another study found similar rates of response in both randomized clinical trials and clinical practice[62].

Nevertheless, all of the assessed therapies were effective to reduce the disease activity and might be considered as therapeutic alternatives as they are proven to exhibit benefits such as greater comfort, less adverse effects and lower cost.

The results of the observational studies included in this review, which reflect the “real-life” use of bDMARD. The best choice for bDMARD treatment-naïve individuals are adalimumab or etanercept in combination with methotrexate. In cases of therapeutic failure with anti-TNF agents rituximab or abatacept (non anti-TNF) or etanercept (as second anti-TNF) might be used; however, more studies of effectiveness were found for rituximab.

COMMENTS

Background

Observational studies could provide relevant information for deciding the choice of treatments, the elaboration of clinical protocols, and the formulation of health policies. The present systematic review of biological disease-modifying antirheumatic drugs included cohort observational studies that reported treatment results applied in real-life conditions; thus, these studies are able to fill in gaps in knowledge left by clinical trials.

Research frontiers

This study evaluate, by systematic review and metanalysis, the effectiveness of biological disease-modifying antirheumatic drugs (bDMARD) for treatment of rheumatoid arthritis for naïve and therapeutic failure patients.

Innovations and breakthroughs

The innovations of this study is that do not exist others studies evaluating the direct comparison between bDMARD adalimumab, etanercept, infliximab, golimumab, certolizumab pegol, rituximab, tocilizumab and abatacept in the real-world. Furthermore, the study assess naïve and therapeutic failure patients groups.

Applications

That study is able to fill in gaps in knowledge left by clinical trials with bDMARD adalimumab, etanercept, infliximab, golimumab, certolizumab pegol, rituximab, tocilizumab and abatacept. Furthermore, do not exist others studies evaluating the direct comparison. Then provide relevant information for deciding the choice of treatments.

Terminology

The Disease Activity Score (DAS) is a clinical index of rheumatoid arthritis (RA) disease activity that combines information from swollen joints, tender joints, the acute phase response and general health. The EULAR response criteria is a classified response criteria which classifies the patients individual as non, moderate or good responders dependent on the change and the level of the DAS and DAS28. The ACR score represents a percentage. An ACR20 score means that a person’s RA has improved by 20%, an ACR50 score means it has improved by 50%, and an ACR70 score means it has improved by 70%. The CDAI is a clinical index of RA disease activity that combines information from swollen joints, tender joints and general health. The HAQ is one of the first self-report functional status (disability) measures.

Peer-review

The authors present an extensive revision about the effectiveness of the biological treatment for rheumatoid arthritis. The paper is well written. It explores the best treatment options for patients with DMARDs failure and provides useful and practical information for clinicians involved in the care of patients with this disease.