Article Text

This article has a correction. Please see:

Updated consensus statement on biological agents for the treatment of rheumatic diseases, 2007
  1. D E Furst1,
  2. F C Breedveld2,
  3. J R Kalden3,
  4. J S Smolen4,
  5. G R Burmester5,
  6. J Sieper6,
  7. P Emery7,
  8. E C Keystone8,
  9. M H Schiff9,
  10. P Mease10,
  11. P L C M van Riel11,
  12. R Fleischmann12,
  13. M H Weisman13,
  14. M E Weinblatt14
  1. 1
    David Geffen School of Medicine, Los Angeles, California, USA
  2. 2
    Department of Rheumatology, Leiden University Medical Centre, Leiden, The Netherlands
  3. 3
    Department of Internal Medicine III, Institut for Clinical Immunology, University of Erlangen-Nuremberg, Erlangen, Germany
  4. 4
    Institute of Rheumatology, Clinic for Internal Medicine III, Vienna General Hospital, Vienna, Austria
  5. 5
    Department of Rheumatology, and Clinical Immunology, Charité Hospital, Berlin, Germany
  6. 6
    University Clinic Benjamin Franklin, Germany
  7. 7
    Leeds University, Department of Rheumatology, Leeds General Infirmary, Leeds, UK
  8. 8
    Department of Rheumatology, Mount Sinai Hospitial, Toronto, Canada
  9. 9
    Denver Arthritis Clinic, Denver, Colorado, USA
  10. 10
    Seattle Rhumatology Associates, Seattle, Washington 98104, USA
  11. 11
    University Nijmegen Medical Centre, Nijmegen, The Netherlands
  12. 12
    University of Texas Southwestern Medical Center, Dallas, Texas USA
  13. 13
    Cedar Sinai Hospital, Los Angeles, California, USA
  14. 14
    Brigham and Women's Hospital, Boston, Massachusetts, USA
  1. D E Furst, David Geffen School of Medicine, UCLA – RM 32-59, 1000 Veteran Avenue, Los Angeles, CA 90025, USA; defurst{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

As in previous years, the consensus group to consider the use of biological agents was constituted by rheumatologists from the universities of Erlangen, Leiden and Vienna in Europe in cooperation with other universities in the USA, Canada and Europe. Pharmaceutical industry support was obtained from a number of companies, but these companies had no part in the decisions regarding the specific programme or about the academic participants at this conference. These sponsors participated in the initial small breakout groups with emphasis on supplying factual information. The companies, on the other hand, had no part in the larger, final consensus group or in the final consensus statement.

The perspective of this consensus is from the treating physician’s point of view.

The 160 rheumatologists and bioscientists who attended the consensus conference were chosen from a worldwide group of physicians and other scientists from 21 countries, with expertise in the use of biological agents for the treatment of rheumatic diseases. The number of attendees and participants was limited so that not everyone who might have been interested could be invited.

Based on the new data regarding TNF blocking agents, B cell-specific agents & IL1rα, an update of the previous consensus statement is appropriate.1 The consensus statement is annotated to document the credibility of the data supporting it as much as possible. This annotation is that of Shekelle et al. and is described in an appendix.2 We have modified the Shekelle annotation by designating all abstracts as “Category D evidence”, whether they describe well-controlled trials or not, based on a need to be able to describe details of the studies and results. As the number of possible references has become so large, reviews are sometimes used and, if they contain Category A references, will be referred to as Category A evidence. All participants reviewed relevant clinical published articles relating to tumour necrosis factor (TNF) and interleukin (IL) 1 blocking agents, and abatacept and rituximab. They were given a draft consensus statement and were asked to revise the document in small discussion groups; open discussion of the revisions led to a final document, representing this updated consensus statement.


Individual patients differ in the aggressiveness of their disease and its concomitant structural damage, the effect of their disease on their quality of life, and the symptoms and signs engendered by their disease. They also differ in their susceptibility to, and expression of, side effects to drugs. All these factors must be examined when considering biological treatment for a patient, as must the toxicity of previous and/or alternative disease-modifying antirheumatic drug (DMARD) use.

As increasing evidence has accumulated for treating psoriatic arthritis (PsA) and ankylosing spondylitis (AS) with biological agents, efficacy and clinical use for these diseases will be treated separately from rheumatoid arthritis (RA). Adverse reactions, however, will remain combined for all indications. In general, in RA, when measuring response to therapy or when following patients over time, the American College of Rheumatology (ACR) response criteria (as a combined index) should not be used in a clinical practice setting to monitor individual response, although some validated measure of response (such as those which follow) should be employed (Category B evidence3). Validated quantitative measures such as Disease Activity Score (DAS), Simple Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), Health Assessment Questionnaire disability index (HAQ-DI), visual analogue scales (VAS) or Likert scales of global response or pain by the patient or global response by the physician, other validated measures of pain for individual patient care, joint tenderness and/or swelling counts, and laboratory data all may be used and may be the most appropriate measures for individual patients (Category B evidence3 4). The physician should evaluate a patient’s response using the above measures to determine the patient’s status and improvement.

For PsA, measures of response such as joint tenderness and swelling, global and pain response measures, functional indices and acute phase reactants have been used and appear responsive (Category A evidence5). For AS, measures such as the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), and the Bath Ankylosing Spondylitis Functional Index (BASFI) have been used in a clinical trial setting but have not been validated for the routine clinical practice setting. Measures such as joint tenderness and swelling, spinal motion, global and pain response measures, functional indices and acute phase reactants have been used and are validated (Category A evidence610).

The use of biological agents will require physicians experienced in the diagnosis, treatment and assessment of RA, PsA, AS and other rheumatic diseases. These physicians will need to make long-term observations for efficacy and toxicity, including cohorts, registries and so on. Because these agents have toxicities, patients or their representatives should be provided with information about potential risks and benefits so that they may give informed consent for treatment.


TNF blocking agents differ in composition, precise mechanism of action, pharmacokinetics, biopharmaceutical properties and so on, but this document emphasises areas of commonality. Data that clearly have differentiated compounds will be discussed if such areas can be identified.


Rheumatoid arthritis

In most patients, TNF blockers are used in conjunction with another DMARD, usually methotrexate (MTX). TNFα blocking agents have also been used successfully with other DMARDs, including sulfasalazine and leflunomide (Category A/B evidence1137). There is evidence that TNF blockers are effective for the treatment of RA in MTX-naive patients (Category A evidence,1217 21 22 26 27 31 34 36 Category D evidence19 24 25). TNF blocking agents can be used as the first DMARD in some patients (Category A evidence,1218 22 2628 31 33 36 Category D evidence19 26 35 36). Adalimumab and etanercept are both approved as monotherapy for RA. Infliximab is only approved for use with MTX in RA. However, observational data indicate that infliximab, too, is sometimes used as monotherapy (Category C evidence11 23 37). Evidence from several randomised controlled trials suggests that the combination of a TNF blocking agent and MTX yields superior results for RA when compared with monotherapy, particularly with respect to excellent clinical responses (ACR 70, European League Against Rheumatism remission) and radiological outcomes (Category A evidence).1116 1820 2235 37

Evidence has become available which shows that TNF-blocking agents are cost-effective from a societal perspective (Category D evidence38 39).

Psoriatic arthritis

The three available anti-TNF medications, etanercept, infliximab and adalimumab, have been approved globally for the treatment of psoriatic arthritis (PsA) based on demonstration of control of signs and symptoms of joint and skin disease, improvement of function (HAQ), quality of life (SF-36), and inhibition of structural damage evidenced by x ray (Category A evidence4047).

A comparator trial of methotrexate versus anti-TNF therapy showed that the latter yielded superior results in joint and skin signs and symptoms, function and quality of life (Category B evidence47 48).

Preliminary data suggest that anti-TNF therapy can benefit cardiovascular risk in PsA (Category A evidence49). Anti-TNF therapy has been demonstrated to reduce healthcare resource utilisation (Category B evidence50), and improve employment status, time lost from work and productivity (Category B evidence44).

Ankylosing spondylitis

Adalimumab, etanercept and infliximab have been approved for the treatment of severe, active AS in Europe and North America (Category A evidence,610 5156 Category D evidence57). In the clinical trials, efficacy of the TNF blocking agents was demonstrated as monotherapy as well as with concomitant second-line agents including sulfasalazine or MTX (Category A evidence610 5156). TNFα blocking agents maintain efficacy over 1–5 years in open-label studies (Category B evidence55 56 58).

The approved doses of TNF blockers for treatment of AS are as follows: infliximab, 5 mg/kg intravenously every 6–8 weeks after induction; etanercept, subcutaneously 25 mg twice a week or 50 mg once a week; and adalimumab 40 mg subcutaneously every other week (Category A and B evidence610 38 51 59). No dose-ranging studies have been done with any of these drugs. There is no evidence that one TNF blocking agent is more effective than any other for the treatment of musculoskeletal manifestions of AS. TNF blocking agents used for AS may reduce economic resource utilisation (Category D evidence38).

Juvenile idiopathic arthritis of the polyarticular type

Adalimumab, etanercept and infliximab have been approved for juvenile idiopathic arthritis of the polyarticular type (Category A evidence;6062 Food and Drug Administration (FDA) Summary Basis of Approval).

Use in other rheumatic diseases or those with prominent rheumatic manifestations

Chronic inflammatory bowel disease

Infliximab has been approved to treat luminal and fistulising Crohn’s disease and ulcerative colitis in the USA (Category A evidence;63 64 FDA Summary Basis of Approval). Infliximab is approved for ulcerative colitis and Crohn’s disease in Europe (Category A evidence;65 66 European Medicines Agency Summary). Adalimumab is approved for Crohn’s disease in the USA.

Controlled trials that demonstrated a difference from placebo or positive control

  • Etanercept was effective for treating some of the mucocutaneous manifestations of Behçet’s syndrome compared with placebo over 4 weeks (Category A evidence67).

  • Etanercept improved hepatitis C viraemia but not symptoms, compared with placebo, when given on a background of interferon αand ribavirin (Category A evidence68).

  • Infliximab improved the signs and symptoms of ulcerative colitis (abstracts from double blind trials (Category D evidence, n = 364 patients in each of two trials)65 66).

  • Infliximab improved Pyoderma gangrenosum (abstract of a double blind trial (Category D evidence, n = 3069)).

  • Infliximab improved pulmonary sarcoidosis (Category A evidence, n = 8670).

Controlled trials that failed to demonstrate a difference from placebo

  • Etanercept and infliximab in Sjögren’s syndrome (Category A evidence7173) (see also anecdotal data in table 1).

  • Etanercept in Wegener’s granulomatosis (Category A evidence74) (see also anecdotal data in table 1).

  • Etanercept in autoimmune ear disease (Category A evidence,75 very small (n = 21)).

  • Infliximab in chronic obstructive pulmonary disease (Category A evidence,76 very small (n = 30)).

  • Infliximab for disc herniation (abstract of a double blind trial, Category D evidence,77 very small (n = 40)).

  • Infliximab in giant cell arteritis (GCA) (Category B evidence,78 very small (n = 44)).

  • Infliximab for polymyalgia rheumatica (PMR) (abstract of a double blind trial, Category B evidence,79 very small (n = 51)).

  • Anecdotal series or studies with promising results (see Appendix 1).

Clinical use

Rheumatoid arthritis and juvenile idiopathic arthritis

TNF blocking agents, when given up to the maximum approved dosing regimens for RA and juvenile idiopathic arthritis of the polyarthritic variety, are expected to lead to significant, documentable improvement in symptoms, signs and/or laboratory parameters within 12 weeks (Category A and B evidence610 1220 22 2429 3137 56 8093). There is no evidence that any one TNF blocking agent should be used before another one can be tried, just as there is no evidence that any TNF blocker is more effective than any other in RA (Category A and B evidence12 15 24 85 8890 94 95). Patients have been switched from one TNF blocking agent to another, but no well-controlled switch trials have been published (Category B and D evidence58 72 85 8890 94 95 168171). There is evidence that loss of response to a TNF blocking agent can occur and studies suggest that failure to respond to one TNF blocking agent does not preclude response to another (Category B and D evidence85 8890 94 95).

Initial observational data suggest the possibility that primary non-responding patients are less likely to respond to a second anti-TNF agent. Patients who have not tolerated one TNF blocker may respond to a second but are also less likely to tolerate a second TNF blocking agent (Category B and D evidence85 90 94 95). The optimal therapy of patients not responding to TNF blockers remains to be determined.

Individually important responses including patient-oriented measures (eg, HAQ-DI, patient’s global VAS, Medical Outcome Survey Short Form 36 (SF-36)) or physical measures (for example, joint tenderness)) should be demonstrated within 12 weeks for RA (Category A evidence610 1220 2529 3137 56 8084). If such improvement occurs, treatment should be continued. If patients show no response to these agents, their continued use should be re-evaluated. Observations with infliximab suggest that increasing the dose or reducing the dosing intervals may provide additional benefit in RA, as may the addition or substitution of other DMARDs (Category A evidence15 18 31 37). Raising the dose of etanercept does not seem to have an added benefit at 12 or 24 weeks in RA (Category D evidence8991). However, because regression to the mean may occur, caution is needed when interpreting apparent improvements following dose escalation in practice (Category C evidence92).

There are data showing that TNF blocking agents slow and/or inhibit radiographic progression in RA, even in some patients without a clinical response (Category A evidence12 15 19 21 22 27 34 40 82 84 93 96). Although some RA patients without measurable clinical response have slowing of radiographic progression (Category A evidence34 93 97), the long-term clinical implications of these changes still remain unknown.

Psoriatic arthritis

As noted above, the three available anti-TNF medications, etanercept, infliximab and adalimumab, have been approved globally for the treatment of psoriatic arthritis (PsA).4051 98101 In addition, statistically significant improvement of enthesitis and dactylitis has been demonstrated with infliximab.42 100 102 103 Also, improvement of fatigue has been noted with adalimumab.40 Improvement of signs and symptoms, function and quality of life occurs within 12 weeks.4051 A systematic evidence-based review of the various domains of PsA (joints, enthesitis, spondyloarthropathy, skin) and efficacy of their treatment has recently been published.99101 103108

Durability of clinical and radiographic data at 2 years in PsA has been demonstrated with etanercept and adalimumab.43 47 Similar durability at 1 year has been demonstrated with infliximab in PsA.44 46

Preliminary data suggest that one can achieve benefit for joint and skin signs and symptoms by switching to a different anti-TNF medication, even if efficacy from a previous anti-TNF agent was never achieved, was lost or if toxicity occurred.4850 109 110

Anti-TNF therapy is superior to methotrexate for joint and skin signs and symptoms, function and quality of life.48

An exploratory trial with anakinra in PsA failed to demonstrate significant clinical or synovial improvement as determined histologically and by MRI.111 An exploratory trial with efalizumab, approved for the treatment of psoriasis in the USA and Europe, failed to show statistically significant improvement in PsA.112 An exploratory trial with alefacept, approved for treatment of psoriasis in the USA, did show statistically significant improvement in joint and skin signs and symptoms in PsA.113


Safety and tolerability of anti-TNF therapy in PsA is no different than an RA clinical trial. Psoriasis, particularly pustular psoriasis, can rarely occur when using anti-TNF therapy.109

Ankylosing spondylitis

Generally, a reduction in signs and symptoms, and improvement in quality of life will be observed by 12–16 weeks in response to treatment with a TNF blocking agent.114 In clinical trials, improvement in signs and symptoms was assessed by patient-reported outcomes (BASDAI, BASFI, patient global VAS, SF-36), physical measures and laboratory parameters. The Assessment in Ankylosing Spondylitis working group has published recommendations for the use of TNF blocking agents in AS (Category A evidence54 115). If a 50% improvement or an absolute improvement of 2 points (on a 0–10 scale) of the BASDAI is not reached within 12 weeks, their continued use should be re-evaluated (Category A and B evidence6 7 9 10 53 54 57 81 115).

Several studies have demonstrated that active inflammation of the sacroiliac joints or spine, as shown by MRI, is significantly reduced by all three TNF blocking agents (Category C evidence).4143 57 Inhibition of radiographic progression, as measured by plain radiography, has not been evaluated in prospective randomised controlled trials in AS. In addition, changes in MRI findings have not yet been correlated with changes in plain radiographs in the spine or sacroiliac joints in patients with AS (Category C evidence).57 116 117

Other for the treatment of AS

Clinical trial data with conventional DMARDs are few but suggest efficacy of sulfasalazine for the peripheral, but not spinal, manifestations of AS.54 115 In those who fail to respond to sulfasalazine, a TNF blocking agent should be considered next (Category D evidence115).

The IL-RA antagonist anakinra has been investigated in two open-label studies in AS with no clear efficacy (see section on IL1 receptor blockade). No data are available at the moment regarding efficacy of other biologicals such as rituximab or abatacept in AS.

Warnings/adverse events

General reviews of TNF blocking agent safety have been published.20 51 58 95 118120


An increased susceptibility to tuberculosis (TB) or reactivation of latent TB should be considered a characteristic of TNF blocking agents. The risk of TB is increased by the use of DMARD per se and is increased further by TNF blocking agents.121123

The clinical picture of active TB may be atypical in these patients (for example, miliary or extrapulmonary presentations) as has been seen with other immunocompromised patients (Category C evidence118 119 123126). There have been more reported cases of TB as a proportion of the total number of individuals treated in patients using infliximab and adalimumab than using etanercept (Category C evidence118 120 122129). This may be due in part to differences in mechanism of action, biology or kinetics as compared to the soluble receptor (Category D evidence)123 126 128 but may also be, in part, due to the fact that populations treated with the various TNF blocking agents differ and the data come from registries and voluntary reporting systems. No head-to-head comparisons among TNF blocking agents have been carried out and thus no definitive data on comparisons between these agents are available regarding the incidence of reactivation of latent TB.

Screening of patients about to start TNF blocking agents has reduced the risk of activating TB for patients treated with these agents (Category B evidence118 119 123 129131). Every patient should be evaluated for the possibility of latent TB, including a history that includes evaluation for the risk of latent TB (Category B evidence58 119 120 126 127 129 132). This history should include seeking a history of prior exposure, prior drug addiction or active drug addiction, HIV infection, birth or extended living in a region of high TB prevalence and a history of working or living in TB high-risk settings such as jails, homeless shelters and drug rehabilitation centres (Category D evidence132). In addition, physical examination and screening tests such as skin tests and chest radiographs should be carried out before anti TNF-therapy is initiated, according to local recommendations (Category B, C and D evidence118 119 122 123 129131). Continued vigilance is required to prevent activation of latent tuberculosis or acquisition of new cases. The role of repeated or serial tuberculin skin testing during anti-TNF therapy is unclear. In the USA, repeat or annual testing has not been recommended in this or other settings of immune suppression unless patients are at risk for ongoing TB exposure.132

In treating latent TB, the optimal timeframe between starting preventive therapy for latent TB infection and starting TNF blockade is unknown. Given the low numbers of bacilli present in latent TB infection, it is likely that waiting long time periods between initiating preventive TB therapy and TNF blockade are unnecessary. While there are no prospective trials assessing this question, observational data from Spain suggest that initiating isoniazid therapy 1 month prior to TNF blockade substantially decreases the risk of latent TB reactivation (Category C evidence122 129131).

Treatment of TB in immunocompromised hosts is as effective as in non-immunocompromised patients, suggesting that long periods of preventive treatment prior to starting TNF blocking agents may not be as necessary as previously thought. Prior to initiating preventive anti-TB therapy in accordance with local guidelines, consulting with a TB expert should be considered.

Opportunistic infections have occurred in the setting of TNF blocking agent use (Category C evidence12 13 1517 1922 24 25 64 84 118120 132141). Particular vigilance is needed when considering those infections whose containment is macrophage/granuloma-dependent, such as listeriosis, coccidiomycosis or histoplasmosis, including reactivation of latent histoplasmosis, (Category C and D evidence118120 126 127 134 137 139142) but the incidence of opportunistic infections is extremely low (Category C and D evidence126 127 129).

Serious bacterial infections have been observed in patients receiving TNF blocking agents at rates between 0.07 and 0.09/patient year compared with 0.01–0.06/patient year in controls using other DMARDs but not TNF blocking agents (Category C evidence118120 143 144). Risk ratios of 1 to up to 3 were documented (Category B evidence144 145). High doses of TNF blocking agents may further increase the risk of serious infections (Category B evidence146). Other studies indicate that serious infections in certain sites are more common when using TNF blocking agents, such as the skin, soft tissues and joints (Category B evidence144 147).The possible contribution of corticosteroids to increasing the risk of infection should always be considered (Category C evidence148). The incidence of other infections (not designated as serious) may be increased when using TNF blocking agents (relative risk (RR) 2.3–3.0, 95% confidence interval (CI) 1.4 to 5.1) (Category C evidence143). The incidence of serious infections is higher when IL1rα and etanercept or abatacept with any of several biologicals are used in combination (3.9% in combination vs 1.0–1.6% in controls) (Category A evidence118120 149 150). Therefore, a combination of biologicals is not recommended. TNF blocking agents should not be started or their discontinuation should be considered when serious infections and/or opportunistic infections occur, including septic arthritis, infected prostheses, acute abscess, osteomyelitis, sepsis, systemic fungal infections and listeriosis (Category C evidence; FDA12 13 1517 1922 24 25 33 64 84 119 120 126 134 135 137 139142 145). Treatment with TNF blockers in such patients may be resumed if the infections have been treated adequately (Category D evidence; FDA 120 123126 134 137 139142 145).

TNF blocking agents do not significantly influence the development of protective antibodies after vaccination, although there is a small decrease in the prevalence of adequate protection and titre of response, especially in combination with methotrexate (Category B evidence151 152). Vaccination with live attenuated vaccines (eg, BCG, yellow fever) is not recommended (Category D evidence).

Injection site/infusion reactions

In placebo-controlled trials, injection site reactions, most of which were mild to moderate but some of which resulted in drug discontinuation, were more common with subcutaneously administered TNF blocking agents than with placebo (Category B evidence10 1218 24 25 3133 61 62 82 83 96 119). One study indicates that human antichimeric antibodies (HACA) against infliximab was associated with decreased response and increased infusion reactions (Category C evidence153). Infusion reactions after infliximab and adalimumab are uncommon and are usually mild to moderate, but may, rarely, be serious (Category A evidence,12 15 21 24 25 31 62 82 83 96 119 Category B and C evidence18 22 63 84).

Treatment limiting infusion reactions can be treated by the use of corticosteroids or antihistamines, or by slowing the infusion rate (Category C and B evidence 153 154).


The incidence of lymphoma is increased in chronic inflammatory diseases such as RA. This increase is associated with high disease activity (Category C evidence155 156). The risk for lymphoma is increased two to fivefold in patients with RA as compared to the general population (especially non-Hodgkin’s lymphoma). A similar risk is seen in RA patients who have received anti-TNF therapy (Category C evidence120 155 157). There are conflicting data about whether there is an increased risk for lymphoma and solid malignancies with anti-TNF therapies for RA. Several large observational databases and a case control study did not demonstrate an increased incidence of solid tumours in patients receiving TNF blocking agents compared with matched controls, while two meta-analyses of anti-TNF therapies (on infliximab and adalimumab) reported a higher rate of solid malignancies including skin cancers (Category A and C evidence156161). One population-based study showed that the incidence of both melanoma and non-melanoma skin cancer were slightly increased when TNF blocking agents were used (RR: 1.4–2.0; Category C evidence160). In patients at risk for malignancies (for example, smokers) or in patients with chronic obstructive pulmonary disease (COPD), there may be an increased risk of lung cancers. In a trial of patients with COPD assigned to infliximab versus placebo, nine developed lung cancers during the trial and another four lung cancers were found during open-label follow-up (Category A evidence162). In a study of Wegener’s granulomatosis, the use of etanercept with cyclophosphamide was associated with six solid malignancies versus none in the cyclophosphamide–placebo group (Category A evidence163). The concomitant use of azathioprine with infliximab in adolescents has been associated with the occurrence of rare hepatosplenic lymphomas (FDA). There is limited information about the risk of developing a future malignancy in patients receiving anti-TNF therapy who had a previous malignancy (Category D evidence164). Vigilance with respect to the occurrence of lymphomas and other malignancies, including recurrence of solid tumours, remains warranted in patients using these medications.


Rare instances of pancytopaenia and aplastic anaemia have been reported (Category A and C evidence18 31 33 82 119 165). If haematological adverse events occur, TNF blockers should be stopped and patients evaluated for evidence of other underlying disease or other causative medications (Category D evidence).


High-dose infliximab (10 mg/kg) appears to be associated with an increased relative risk of worsening congestive heart failure (CHF) and mortality, particularly in RA patients with New York Heart Association class III–IV CHF (Category B and D evidence118 120 166 167). There is presently no substantive evidence that infliximab 5 mg/kg, or etanercept at 25 mg twice a week increases the incidence of CHF or CHF-related mortality in patients with functional class I CHF (Category B and D evidence118 166 167). However, it should be noted that well-controlled RA studies have excluded patients with complicating illnesses, including CHF, and RA per se appears to be associated with increased atherosclerotic cardiovascular disease (ASCVD) and ASCVD-related mortality (Category C evidence118). One RA study using infliximab demonstrated an increase in total, HDL and LDL lipids (Category D evidence168). One cohort observational study in RA patients without overt CHF showed a possible decrease in myocardial infarction-related mortality when using TNF blocking agents (Category D evidence3 169). Each patient’s risk versus benefit should be carefully considered before TNF blocking agents are begun or continued (Category D evidence; FDA).


Rare instances of new onset interstitial lung disease have been documented in patients using infliximab both with and without baseline pulmonary findings (Category C171 and D evidence170 171).


The long-term safety of TNF blockers in patients with chronic viral hepatitis is not known. In Hepatitis C, observational studies and one controlled investigation revealed that etanercept did not have an effect on viral load and there was no increased incidence of adverse effects; in addition, there was evidence, in this study, that symptoms and liver function tests improved (Category C and D evidence167 172178).

With respect to hepatitis B, patients treated with all three agents have experienced increased symptoms and worsening of viral load (Category C evidence,173 Category D evidence.177 179). Elevations have been observed in liver function tests with therapeutic strategies employing the use of adalimumab, infliximab and etanercept with ALT-AST elevated in 3.5–4.9% and elevations of these liver enzymes >2X ULN in about 0.1–0.2%, although concomitant medications and other clinical circumstances confound the interpretation of this information (FDA; Category C evidence18 173 175177 180 181). As a result, specific warnings regarding hepatitis B reactivation have been added to the US label by the FDA (FDA). Therefore TNF blockers should not be used in patients with known hepatitis B infection; in the event that hepatitis B infection is discovered during use of TNF blockers, prophylactic antiviral therapy can be employed (Category C evidence;58 173 177 179 182 Canadian Regulatory Authorities).

The follow-up and monitoring for liver function test elevations should be governed by the patient’s concomitant medications, conditions and patient-related risk factors.

Conception and pregnancy

There is insufficient information at present to safely counsel the continuation of TNF blockers during pregnancy. Some women have become pregnant while being treated with TNF blocking therapy. A small pharmacovigilance study and a survey study comprising about 200 pregnancies in women being treated with TNF blockers187 revealed that the rates of normal live births, miscarriages and therapeutic terminations in patients taking TNF blockers did not differ from published rates for the normal population (Category D evidence183186 188 189). Pre-pregnancy planning is strongly recommended in patients who are either on TNF blockers or considering their use; the need for disease control with alternate methods should be the subject of these conversations. There are no data available on the subsequent safety of a pregnancy if the male partner is taking TNF blockers.

It is advised that women and physicians discuss the issue of TNF blocking therapy when pregnancy planning takes place or if pregnancy is discovered during ongoing TNF blocking therapy, and that this discussion is documented.

Using peak concentrations as a measure, only about 0.0001% of the peak plasma concentration of etanercept is found in the breast milk (Category D evidence190).

Autoimmune-like syndromes

Antiphospholipid syndrome, syndromes resembling drug-induced lupus and vasculitis have occurred in patients receiving TNF blocking agents, and treatment should be stopped if there is clinical evidence of a drug-induced lupus-like syndrome. These symptoms are highly likely to resolve on discontinuation of the TNF blocking agent (Category C and D evidence31 33 63 82 118 120 191199). There is an increased incidence of several autoantibodies (for example, antinuclear antibody (ANA), anti-double-stranded DNA) after infliximab and it is probably not a class effect (category C evidence191195 198 199). However, there is no evidence that patients with RA who had, or develop, a positive ANA, anticardiolipin antibodies and/or dsDNA are at significantly increased risk for the development of drug-induced lupus (FDA; Category C and D evidence18 31 82 120 191 192 196199).

Neurological diseases

The incidence of demyelinating-like syndromes, optic neuritis, transverse myelitis, multiple sclerosis and Parkinson’s disease is no greater than expected in the general population (Category C evidence31 82 120 165 200202). However, rare cases of these syndromes have been reported, more often with etanercept than with infliximab, most but not all improving or remitting after the TNF blocker was withdrawn (Category C evidence82 120 165 200203). There is insufficient evidence to suggest that TNFα blockers unmask latent disease, although this remains a possibility. These agents should be stopped if a demyelinating-like disorder or optic neuritis occurs. Patients with a history of definite demyelinating disease or optic neuritis should not receive TNF blocking agents (Category D evidence).

Psoriatic skin lesions

Some cases of new-onset psoriatic skin lesions or exacerbations of pre-existing psoriasis have been reported in patients with RA who used TNF blocking agents. (Category C evidence109).

Issues specific to PsA

Safety and tolerability data with anti-TNF medications in PsA have not demonstrated any adverse events that were significantly different from RA trials. However, because liver biopsy studies suggest that patients with psoriasis and PsA demonstrate a greater proclivity for hepatotoxicity with MTX therapy than RA patients (Category B evidence204), it is not known if the safety profile from RA trials is completely comparable with PsA.

Precautionary statements

The safety of TNF blockade is unknown or has not been established in the following situations:

  • HIV, and so on

  • during lactation.

Other areas where knowledge is lacking are highlighted in the consensus group’s recommendations for areas most urgently requiring further research.


Among a number of potential areas requiring action and/or further research, the consensus group felt the following projects or directions were most important in each of three areas: registries, efficacy and toxicity.


  1. Long-term registries continue to be needed to monitor the toxicity of biologicals and are strongly recommended, requiring a cooperative effort between payers, government, industry and rheumatologists.

  2. Registries of pregnancy outcomes under anti-TNF therapy (and after cessation of therapy) should be continued.


  1. What are the optimal dosing regimens when using TNF blocking agents?

  2. Are there predictors of toxicity for TNF blocking agents?

  3. Is there a correlation between radiological effect and long-term effectiveness for TNF blocking agents?

  4. What are the outcomes in patients treated with TNF blocking agents where disease activity persists without joint destruction and where joint destruction is observed with little disease activity?

  5. Can biologicals be administered at lower than currently used doses and/or at dosing intervals longer than currently employed to slow or halt radiographic progression of RA in the absence of an ACR 20 response?

  6. What is the effect of TNF blocking agents on growth in children with juvenile chronic arthritis?

  7. What, if any, dose response exists for the use of TNF blocking agents in PsA and/or AS?

  8. Do AS patients with advanced spinal fusion respond to TNF blocking agents?

  9. Do TNF blocking agents decrease the incidence of cardiovascular events or cerebrovascular incidents?

  10. What are the predictors of response to TNF blocking agents in early and advanced AS?

  11. Can TNF blocking agents be discontinued after an initial response?


  1. Can TNF blocking agents be used safely in pregnant women?

  2. What is the safety profile of TNF blocking agents during surgery? How does it compare with the safety profile of patients undergoing surgery without concomitant TNF blocker use?

  3. What duration of tuberculosis prophylaxis/treatment is necessary when patients are being treated with TNF blocking agents?

  4. Can TNF blocking agents be used in patients with a history of lymphoma and non-Hodgkin lymphoma or solid tumours? What is the time interval needed before TNFα blockers can be used after patients with malignancies have reached a full remission?

  5. What is the sensitivity and specificity of the quantiferon test compared with the purified protein derivative (PPD) in RA patients?


TNF blocking agents have proved to be effective DMARDs and are a major advance in the treatment of RA, PsA, AS and juvenile idiopathic arthritis. Their use is expanding to other rheumatic diseases. Studies in selected areas of efficacy, toxicity and general use of TNF blocking agents are needed to help further define the most appropriate use of these agents. Further considerations when using TNF blocking agents in these diseases should balance efficacy, toxicity and cost issues, and recognise that data in subpopulations are still being acquired. It is hoped that this statement, which is based on the best evidence available at this time and is modified by expert opinion, will facilitate the optimal use of these agents for patients with RA.


Only one IL1 blocking agent (anakinra) has been approved and the discussion below refers to this agent.


Anakinra may be used for treatment of active RA, alone or with MTX, at a dose of 100 mg per day subcutaneously (Category A evidence205210). In Europe, the anakinra label requires its use with methotrexate. Anakinra is recommended for the treatment of active RA after an adequate trial of another conventional DMARD, of which MTX is a common example (Category D evidence). The safety of anakinra has also been studied with other DMARDs (Category A evidence207 208).

The use of anakinra as the first DMARD for the treatment of RA should, at present, be limited because no trials in early RA have been published.

Anakinra has been investigated in two open-label studies in AS with no clear efficacy (see section on IL1 receptor blockade).211 212

Anakinra has been studied for a variety of off-label indications (see Appendix 2). It appears to be highly active in the auto-inflammatory syndromes (also called “periodic fever syndromes”), such as Muckle–Wells syndrome, neonatal-onset multisystem inflammatory disease (NOMID) and TNF receptor-associated periodic syndrome (TRAPS). It also appears to be active in systemic-onset juvenile arthritis and adult-onset Still’s disease. It has been used to treat osteoarthritis (Category D evidence213 214) and SLE (Category D evidence215 216).

Clinical use

Anakinra can lead to significant improvement in symptoms, signs and/or laboratory parameters within 16 weeks, and can slow the rate of radiographic progression (Category A evidence205 206 209). If improvement is not observed by 16 weeks, the continued use of anakinra should be re-evaluated.

Trials of patients failing TNF blocking agents demonstrate mixed responses (Category C evidence210). Anakinra did not inhibit anti-tetanus antibody response in a controlled trial (Category A evidence217). A dose-related incidence of injection site reactions, affecting up to 70% of patients, has occurred with the use of anakinra. These reactions often do not require treatment and seem to moderate with continued use in some patients (Category A evidence205207). There are no data to advise either termination or continuation of anakinra if a woman becomes pregnant.


Serious infections are increased in patients receiving anakinra, and its incidence is higher than in patients with RA using other DMARDs. This increased incidence was magnified by corticosteroid use (Category A evidence208). Anakinra should not be started or should be discontinued when serious infections occur (Category A evidence,12 36 61 96 Category D evidence34). Treatment with anakinra in such patients should only be resumed if the infection(s) has been adequately treated (Category D evidence). To date, there is no indication that anakinra is associated with an increased incidence of TB (Category D evidence).

In combination with etanercept, there was an increased rate of serious infections compared to either monotherapy, and no increase in efficacy. Therefore, the combination should not be used (Category A evidence150).

Precautionary statements

The safety of anakinra is unknown or has not been established in the following situations:

  • lymphoma, lymphoproliferative and other malignancies

  • during pregnancy and/or lactation.

Other areas where knowledge is lacking are highlighted in the consensus group’s recommendations for areas most urgently requiring further research.


Among a number of potential areas requiring action and/or further research, the consensus group felt the following projects or directions were most important in each of three areas: registries, efficacy and toxicity.


  1. Long-term registries to monitor the toxicity of biologicals are recommended, requiring a cooperative effort between payers, government, industry and rheumatologists.

  2. Registries of pregnancy outcomes under anakinra blocking therapy (and after cessation of therapy) should be continued.


  1. What is the efficacy of anakinra in polyarticular juvenile arthritis, systemic-onset juvenile and other rheumatic diseases including osteoarthritis?


  1. Can anakinra be used in patients who cannot be treated with TNF blocking agents because they have a history of TB or latent TB and cannot tolerate appropriate therapy for the latter, for some reason?


Anakinra is effective for the treatment of rheumatoid arthritis but its specific place (for example, before or after TNF blocking agents) in the rheumatological armamentarium is not yet defined. Publication of studies in selected areas of efficacy, toxicity and general use of anakinra is needed to help further define the most appropriate use of these agents. Further considerations when using anakinra in this disease must include cost issues and the recognition that data in subpopulations are still being acquired. It is hoped that this statement, which is based on the best evidence available at the time of its creation and is modified by expert opinion, will facilitate the optimal use of anakinra for patients with RA.


One agent which modulates T cell activation (abatacept) has been approved, and references are therefore to this product.


Abatacept is approved in North America for use alone or with background DMARDs for treatment of active RA. It is administered as intravenous infusions of up to 10 mg/kg (500 mg for weights less than 60 kg; 750 mg for weights of 60–100 kg and 1000 mg for weights over 100 kg) at 0, 2, 4 weeks then monthly (FDA product label218 219). However, it is not recommended for use with other biologicals (although there are no data regarding the use of rituximab in combination with abatacept). Abatacept is recommended for treatment of active RA after an adequate trial of MTX, another effective DMARD or TNFα blocking agent (Category A evidence218222).

Abatacept may be substituted directly for the next dose of a TNF blocking agent, when switching is undertaken (Category C evidence223). Abatacept has been used with MTX and other DMARDs (Category A evidence149). (Pharmacoeconomics studies are being carried out (Category D evidence).)

Clinical use

Abatacept is indicated to decrease signs and symptoms (including major clinical response), slow and/or inhibit progression of structural damage and improve physical function in adult patients with moderate to severely active RA who have had an inadequate response to one or more DMARDs such as MTX or TNF blocking agents (Category A evidence,218 219 223 FDA label). Improvement occurs within 16 weeks, although additional improvement can occur for up to 2 years (Category D evidence).224 225 If clinical improvement occurs, treatment should be continued (Category D evidence);where no meaningful improvement occurs within 12–16 weeks the continued use of abatacept should be re-evaluated.

Abatacept in combination with MTX inhibits radiographic progression in RA (category A evidence149 219 224). A study of abatacept for psoriasis vulgaris showed efficacy (Category C evidence226).

There is evidence for efficacy in juvenile idiopathic arthritis (Category C evidence227). Studies in early RA, undifferentiated arthritis and systemic lupus erythematosus are ongoing (Category D evidence).


An increased incidence of serious infection has been observed when comparing abatacept to placebo (3.0% with abatacept vs 1.9% with placebo). In combination with other biological agents, the rate of serious infections is 4.4% (vs 1.5% in controls).228 The use of abatacept with other biologicals is not recommended (Category A evidence221 228). There are no data regarding the combination of abatacept and rituximab. Patients with COPD treated with abatacept had more adverse events than patients treated with placebo; therefore use in RA patients with COPD should be undertaken with caution (Category D evidence149).

Based on theoretical concerns, live vaccines should not be given with abatacept or within 3 months of using abatacept (Category D evidence).

A numerical imbalance was reported in the number of lung cancers observed in the placebo-controlled trials (4 abatacept, 0 controls). An additional nine cases were reported during open-label extension studies (cumulatively 13 cases/4134 patient years). The overall incidence of lung cancer in abatacept-treated patients during these studies was not higher than expected, based on large RA cohorts, although it was higher than in the general population.

There has been one case of lymphoma occurring in double-blind period with abatacept versus 0 in the placebo group and four additional cases in the open-label extension (cumulatively 5/4134 patient years) (Category B evidence229). While this number is consistent with that expected from large RA cohorts, ongoing surveillance is necessary.

All patients in abatacept phase 3 trials were screened for TB and excluded if the screen was positive. The risk for activation of latent TB or for developing new TB when using abatacept is unknown. Until the risk is known, it is prudent to screen patients considered for abatacept therapy for TB according to local practice (Category D evidence).

Precautionary statements

The safety of abatacept is unknown or has not been established in the following situations:

  • viral infections, including HIV, hepatitis B and C

  • during pregnancy and/or lactation.

Other areas where knowledge is lacking are highlighted in the consensus group’s recommendations for areas most urgently requiring further research.


Among a number of potential areas requiring action and/or further research, the following projects or directions are important in each of three areas: registries, efficacy and toxicity:


  1. As per anti-TNF agents.


  1. What is the efficacy of abatacept in polyarticular juvenile arthritis, early arthritis, systemic lupus undifferentiated early arthritis and other rheumatic diseases?

  2. Rituximab versus abatacept in TNF failures.


  1. Can abatacept be used in patients who cannot be treated with TNF blocking agents because of CHF, demyelinating diseases, serious infections or tuberculosis?


Rituximab is a chimeric anti-CD20 monoclonal antibody, which was approved in 1997 for treatment of indolent CD20, B cell non-Hodgkin’s lymphoma. More than 1 000 000 patient exposures (usually four infusions per patient) have been documented over 9 years in post-marketing surveillance of these non-rheumatic diseases. Recently, a consensus statement on the use of rituximab in patients with RA (Category D evidence248) has been published.


Rituximab has been approved by the FDA in the USA for the treatment of moderate-to-severe RA in patients who have had an inadequate response to TNF inhibitors (Category A and D evidence;230232 FDA and European Medicines Agency label; Category C and D evidence233239). Patients should have at least moderate disease activity despite MTX therapy. Rituximab is administered intravenously as two 1 g rituximab infusions (given with 100 mg methylprednisolone or equivalent) separated by an interval of 2 weeks; two 500 mg doses can also be used with little decrease in efficacy (Category A evidence233240). It may also be used when TNF inhibitors are not suitable (Category D evidence232 241 242). In RA, it may be used alone or in combination with MTX (Category A and D evidence230232 241244). Appropriate supportive equipment should be available when rituximab is used in case of rare anaphylactoid reactions.

Current evidence on the efficacy of rituximab relates to rheumatoid factor (RF) positive patients (Category D evidence233 234). Divergent ACR responses were seen with rituximab in RF-negative patients, in TNF non-responders and in DMARD non-responders (Category D evidence233235). Patients who were both RF and anti-CCP negative had a reduced response in TNF non-responders (Category D evidence242 244). Since small numbers of RF-negative patients were evaluated, the role of RF and anti-CCP antibodies is not clear.

Clinical use

In clinical trials, rituximab results in significant improvement in signs and symptoms and/or laboratory measures by 8–16 weeks (Category D evidence230238 241244) in patients with an inadequate response to MTX who have failed conventional DMARDS or have one or more TNF inhibitors (Category A evidence240 245). Improvement has also been demonstrated in patient-related outcomes such as HAQ-DI, patient global VAS, fatigue, disability and quality of life. Evidence from randomised controlled trials show that the combination of rituximab with MTX yields superior clinical efficacy for RA when compared with monotherapy (Category A evidence233235 240 245). The optimal treatment schedule is currently under investigation (Category D evidence230 233235 242 244). Preliminary data have shown that repeat treatment courses are effective in previously responsive RA patients (Category D evidence;236 FDA). Most of the patients who received subsequent courses did so after 24 weeks after the previous course and none received repeated courses earlier than 16 weeks from the previous course (Category D evidence;236 FDA). No data are available on repeated treatment in patients who failed to respond to the initial course.

There are data indicating that rituximab can slow radiographic progression in patients who have had an inadequate response to one or more TNF inhibitors (Category D evidence234).

Rituximab has been used in primary Sjögren’s syndrome, systemic lupus erythematosus, Wegener’s granulomatosis, hepatitis C-associated cryoglobulinaemia, antineutrophilic cytoplasmic antibodies (ANCA), associated vasculitis disorders other than Wegener’s such as polyarteritis nodosa, dermatomyositis/polymyositis (Category C evidence), autoimmune haemolytic anaemia, idiopathic thrombocytopenia purpura, antiphospholipid syndrome and scleroderma (see Appendix 3).


The most frequent adverse events are infusion reactions, which are most common with the first infusion of each course (approximately 35%) and are reduced with the second infusion (approximately 10%). Intravenous corticosteroids were shown to reduce the incidence and severity of infusion reactions by about 30% without changing efficacy (Category A, C and D evidence230 231 233235 240242 244).

Recently, two cases of progressive multifocal leukoencephalopathy (PML) have been communicated in cases of patients with severe SLE who had received an extensive additional immunosuppressive therapy. Twenty-three cases of PML in SLE cases not using rituximab have also been documented. Twenty-seven cases of PML are known in the NHL literature. It is unclear whether this incidence reflects the overall PML occurrence in these heavily immunocompromised individuals. As of today, no case of PML has been reported in RA (FDA reports).

Serious infections, including bacterial infections, have been observed in patients receiving rituximab. Similar to the anti-TNF agents and the other biologicals, a small increase in serious infections (not intracellular infections) in patients receiving rituximab 2×1000 mg compared to placebo has occurred: 4.7/100 versus 3.2/100 patient years in the DANCER study and 5.2/100 versus 3.7/100 patient years in the REFLEX trial (Category D evidence230 233235 239 241).

Rituximab should not be given in the presence of serious or opportunistic infections (Category D evidence).

In general, patients who did not respond to TNF inhibitors will also have been pre-screened for the presence of active or latent TB. In the RA clinical trials of rituximab in TNF inhibitor non-responders, patients with active TB were excluded. Others were screened by chest x ray, but were not screened for latent TB by PPD testing. Moreover, there is no evidence of an increased frequency of TB in patients with lymphoma treated with rituximab (Category D evidence). Based on expert opinion, it is felt that at this time there is no evidence indicating the necessity to screen for TB prior to initiating rituximab in RA (Category D evidence).

Since decreased levels of immunoglobulin (Ig) M, A and G have been observed with rituximab, it may be useful to determine baseline Ig levels (Category D evidence,231 237 241 244). In clinical trials no increase in serious infections have been reported in the very few patients with reduced levels of IgM after rituximab therapy compared to their normal IgM levels previously (Category D evidence237 244). Further analyses are required to address this issue. B cell levels have been measured in clinical trials but their utilisation in routine practice has not been proven. Depletion of peripheral levels of the CD20+ B cell subpopulation was not found to be predictive of achieving or maintaining a clinical response in RA patients (Category D evidence238). This suggests that the timing of retreatment should be based on disease activity rather than repletion of peripheral B cell levels (Category D evidence).

The safety of TNF inhibitor use in patients (n = 78) who did not respond to rituximab and who had B cell levels below normal demonstrated a numerical increase in serious infections (Category D evidence239). Further data are needed to address this important issue.

Since rituximab causes B cell depletion, it is recommended that any vaccinations required by the patient, such as those to prevent pneumonia and influenza, should be given before commencing this agent. While on therapy, appropriate vaccination (such as against influenza) should be given when indicated, although the success may be suboptimal.

Until further data are available, the use of live attenuated vaccines should only be given prior to the use of rituximab (Category D evidence).

Patients should be screened for hepatitis B and C before starting rituximab as hepatitis B reactivation has been reported in oncological practice (Category D evidence246).

Until further data are available, rituximab should not be used during pregnancy.

There is no evidence that rituximab is associated with an increased incidence of solid tumours in RA. Nevertheless, vigilance regarding the occurrence of solid malignancies remains warranted during treatment with rituximab (Category D evidence).


The consensus group felt the following projects or directions were most important in each of three areas: registry, efficacy and toxicity.


  1. As per anti-TNF agents.


  1. What is the minimal effective dose of rituximab?

  2. What is the optimal strategy in using rituximab including timing, repeat dosing and combination with other drugs?

  3. Are there predictors of treatment response?

  4. What is the long-term efficacy of repeat dosing?

  5. What are the cost-effectiveness and ethical issues when treating patients with rituximab?

  6. What are the mechanisms underlying the efficacy of rituximab?

  7. Is induction therapy with rituximab possible with subsequent withdrawal or dose reduction?

  8. What is the role of rituximab in the treatment of other rheumatoid diseases (for example, systemic lupus erythematosus, scleroderma, myositis, Sjögren’s syndrome, vasculitis, psoriatic arthritis and AS)?


  1. What are the safety issues associated with long-term B cell depletion and lowered Ig levels?

  2. What is the safety profile with respect to immunoglobulin reduction vis-a-vis infection and response to vaccination?

  3. Does rituximab increase the frequency of TB and opportunistic infection?

  4. Can TNF blocking agents and other immunomodulators be used safely in patients with B cell depletion who are unresponsive to rituximab?

  5. What are the optimal strategies including vaccination to lower the risk of infection during treatment with rituximab?

  6. Can rituximab be used safely in pregnant and lactating women?

  7. Will long-term use of rituximab result in secondary loss of effect?

  8. Will long-term use of rituximab substantially reduce the incidence of B cell lymphoproliferative disorders?

  9. Can rituximab be used in patients with remote or current solid malignancies?

  10. Can rituximab be used safely in hepatitis C?

Appendices: categories of evidence

  • Category A evidence: based on evidence from at least one randomised controlled trial or meta-analyses of randomised controlled trials.

  • Category B evidence: based on evidence from at least one controlled trial without randomisation or at least one other type of experimental study, or on extrapolated recommendations from randomised controlled trials or meta-analyses.

  • Category C evidence: based on non-experimental descriptive studies such as comparative studies, correlational studies and case control studies which are extrapolated from randomised controlled trials, non-randomised controlled studies or other experimental studies.

  • Category D evidence: based on expert committee reports or opinions or clinical experience of respected authorities or both, or extrapolated recommendations from randomised controlled trials, meta-analyses, non-randomised controlled trials, experimental studies or non-experimental descriptive studies.

Note: abstracts have not been considered in the above evidence scheme, as they are not complete and may change by the time the data are published, or they may not be published as full papers at all. Evidence from abstracts alone, therefore, is considered as Category D evidence and noted as “abs” until those data are published as a complete, peer-reviewed paper.

Appendix 1: Anecdotal studies of anti-TNF agents

Appendix 2: Anecdotal studies of IL1ra (anakinra)

Appendix 3: Abstracts or anecdotal studies of rituximab