4.1 Analgesics

Analgesics in early RA are used as an adjunct to NSAID and DMARD therapy. There is evidence that paracetamol, coproxamol, nefopam, and codeine are effective in reducing pain in RA.^{51,52,53,54,55} Most of these trials were carried out more than 20 years ago and can be criticised for small patient numbers and short duration. Only a very small proportion of rheumatoid patients is likely to be controlled with analgesics alone.

Simple analgesics should be used in place of NSAIDs if possible and DMARDs should be introduced early to suppress disease activity.

4.2 Nonsteroidal anti-inflammatory drugs

There is abundant evidence that NSAIDs are effective and provide symptomatic relief of pain and stiffness without influencing the progression of disease.^56,57 The choice of short-, medium- or long-acting preparations can be tailored to fit a patient's particular lifestyle.

NSAIDs act by inhibiting cyclooxygenase (Cox) pathways. There are considered to be two isoforms of Cox. Cox1 produces prostaglandins which are cytoprotective and regulatory (GI mucosa, platelets and renal endothelium). Cox2 produces prostaglandins which mediate pain and inflammation and are the preferred targets in RA (see section 4.2.4).

Prescribers should be aware of the many potential drug interactions with NSAIDs and the side effect profiles of different drugs (see Annex 6 and British National Formulary).58

4.2.1 Adverse effects of NSAIDs

Toxicity is a major limiting factor and side effects are related to dose and duration of therapy.^59,60 Common side effects (especially in the elderly) are gastrointestinal toxicity, fluid retention and hypertension. Other less common but potentially serious side effects are renal disease and hypersensitivity (including asthma). Uncommon and not usually serious side effects are headaches, dizziness, tinnitus, rash (particularly with fenbufen) and abnormal LFTs (particularly with diclofenac). Evidence level 1+

4.2.2 Gastrointestinal toxicity

The use of NSAIDs is associated with gastrointestinal (GI) toxicity.^59,60 The following side effects occur to a varying extent with all preparations and all routes of administration:

• dyspepsia
• gastric erosions
• peptic ulceration
• small bowel inflammation and bleeding⁶¹
• perforation
• haematemesis or melaena
• occult GI blood loss and anaemia

According to the Arthritis, Rheumatism and Aging Medical Information System (ARAMIS), 13 of every 1000 RA patients who take NSAIDs for one year have a serious gastrointestinal complication. The annual relative risk of mortality attributed to NSAID-related GI adverse effects is four times that for those not using NSAIDs.⁵⁹

The rate of NSAID-related serious GI complications requiring hospitalisation has decreased in recent years. The reason for this is likely to be multifactorial. Intensive education programmes have alerted physicians and patients to the use of newer, less toxic NSAIDs and non-NSAID analgesics in populations at high risk. There has also been a much wider use of gastro-protective therapy.

Risk factors for NSAID-associated gastroduodenal ulcers are listed in Table 4.

Table 4

Risk factors for nsaid associated gastroduodenal ulcers

Some patients who have serious GI complications do not report antecedent dyspepsia. Thus every possible strategy should be employed to minimise risks of GI-related toxicity,⁵⁹ e.g. smoking cessation, and alcohol reduction. Eradication of Helicobacter pylori in NSAID-associated peptic ulcers has not been shown to be of value (see the SIGN guideline on Helicobacter pylori: eradication in dyspeptic disease, which is currently under review).⁶²

Surveillance and endoscopic studies have confirmed that the incidence of GI mucosal injury is reduced with nabumetone, etodolac and meloxicam. Mefenamic acid, azapropazone and piroxicam are considered unacceptably toxic for long-term use in RA.^{63, 64}

If NSAID use is unavoidable, gastroprotective agents may be used, as summarised in Table 5.

Table 5

Summary of gastroprotective agents in RA

It is important to note that the use of an enteric coated, parenteral or rectal NSAID preparation is not protective. The systemic effects of NSAIDs are the predominant cause of damage.

4.2.3 Renal toxicity

NSAID use is also associated with renal disease. Prostaglandins regulate and maintain intrarenal perfusion particularly under conditions where renal blood flow may be reduced (e.g. dehydration or blood loss, cardiac failure, chronic renal failure, diuretic use, or hypertension). By inhibiting prostaglandin synthesis under these conditions, NSAIDs may further impair intrarenal blood flow contributing to renal impairment (or overt renal failure), hyperkalaemia, oedema and hypertension. These problems are particularly likely in the elderly.

Interstitial nephritis is an uncommon, idiosyncratic side effect, unrelated to the above pharmacological action of NSAIDs.

No currently available NSAID has a completely safe renal profile. The effects of the new Cox2 agents (see section 4.2.4) on renal profile are as yet unknown. Preliminary work suggests that the effects of Cox2-selective NSAIDs on renal function are similar to those observed with non-selective NSAIDs.⁶⁷

4.2.4 Newer NSAIDs

Recent developments have concentrated on Cox pathways. Cox2-selective NSAIDs target the inhibition of inflammatory pathways while having less effect on cytoprotection and regulatory effects than Cox1. Endoscopic ulceration is no more than for placebo with celecoxib.⁶⁸ A recent comparative study of the Cox2-selective NSAID, celecoxib, compared with diclofenac in RA suggested equivalent efficacy with lower frequency of GI ulceration.⁶⁹ Similar findings with regard to the upper GI tract were noted with celecoxib in comparison with naproxen⁷⁰ and ibuprofen and diclofenac.⁷¹ The incidence of upper GI complication with rofecoxib (at present unlicensed for RA) has also been shown to be less than that with naproxen.⁷² In this study, both drugs exhibited similar efficacy.

As with all new drugs, the safety of Cox2 inhibitors remains under close review.^{73, 74}

4.2.5 Summary of strategies to minimise the risk of nsaid toxicity

The lowest NSAID dose compatible with symptom relief should be prescribed. NSAIDs should be reduced and if possible withdrawn when a good response to DMARDs is achieved.

Simple analgesics should be used in place of NSAIDs if possible and DMARDs should be introduced early to suppress disease activity.

Only one NSAID should be prescribed at a time.

Introduce gastro-protection in RA patients >65 years and in those with a past history of peptic ulcer.

Consider intra-articular corticosteroids, particularly when disease is localised (see section 4.5).

NSAIDs should be avoided in patients taking anticoagulants or corticosteroids.

4.3 Disease modifying anti-rheumatic drugs (DMARDs)

DMARDs are drugs which have a beneficial effect on the course of RA, as well as providing symptomatic benefit. Onset of benefit is slow (four to 16 weeks). The currently available agents, associated adverse effects, and their monitoring requirements are indicated in Table 6. Patients should be informed of the potential benefits, risks and monitoring requirements of these drugs.

4.3.1 Evaluation of DMARDs

Although the majority of the placebo-controlled studies on DMARDs were carried out in the 1960s, '70s and '80s with methodology which would not necessarily achieve the standards expected in the 1990s, studies have clearly shown benefit from existing DMARDs.^{75,76,77,78,79,80,81,82} The Felson meta-analysis⁷⁵ of 66 trials in 5,343 patients showed a significant improvement in articular index and ESR (see Annex 7). Evidence level 1+

Further supportive evidence for the use of DMARDs comes from cohort studies where increased disease modifying anti-rheumatic drug use was strongly correlated with better long-term disability index values. ⁸³

There is clear evidence from placebo controlled trials that DMARDs reduce symptoms in RA (as measured by joint pain, swelling, and tenderness, and duration and severity of morning stiffness). DMARDs also improve global wellbeing as assessed by both patient and physician.^{81,83,84,85,86,87,88}

Inflammatory markers such as ESR, CRP and elevated platelet count are reduced significantly by DMARDs (but not by NSAIDs) and this is associated with better long-term outcome.^{21,81,83,84,85,86,87} An improvement in anaemia of chronic disease is often observed. There is also consistent evidence that DMARDs have a beneficial effect on functional status, as measured, for example, by HAQ score.^83,85,86,89 Most DMARDs have been shown to have some effect in retarding radiological progression of disease.^85,90,91

4.3.2 Comparison of DMARDs

The best DMARDs for the treatment of RA are those that provide the most efficacy with the least toxicity over the long term.

Data are available from Fries et al and the large ARAMIS database, from the Felson meta-analyses and from direct comparative studies.^{60,75,78,81,83,85,87,91,92,93,94,95,96} The range of well-established DMARDs and a brief summary of the advantages and adverse events associated with each is illustrated in Table 6. Recommendations on the choice of DMARD are made in section 4.3.8. Evidence level 1++

4.3.3 Timing of dmard treatment in RA

It is becoming increasingly clear that DMARDs should be introduced as soon as possible. Four recent studies have highlighted the importance of early intervention with DMARDs.^81,86,87,88 Protracted benefit may be achieved in RA patients if appropriate DMARD therapy is introduced early. Evidence level1+

While longer disease duration prior to initiation of DMARDs does not influence the beneficial effect on symptoms or the acute phase response it does have an adverse effect on functional outcome. There is also evidence that delays in initiating DMARDs lead to long lasting negative effects on disease course. ^{81, 88} There is clear evidence that patients with early disease respond better to treatment.⁸¹

Early DMARD therapy in RA is important to maintain function and reduce later disability.

4.3.4 Sustained DMARD therapy

While early initiation of therapy is of importance, a sustained input is vital if disease suppression is to be maintained. Remission (see Annex 3) is the goal but is seldom achieved. Equally 'cure' is not attained, thus withdrawal of treatment is seldom appropriate.

Two randomised placebo controlled studies have demonstrated relapse on withdrawal of disease modifying agents.^97,98 In both these studies, disease modifying effect was unequivocal. These results confirm the efficacy of DMARDs in comparison with placebo, and demonstrate that sustained prescription of DMARDs is necessary to suppress disease activity. Serial use of DMARDs has been shown to be safe after 10-15 years.^37,99 Evidence level 1+

DMARD therapy should be sustained in inflammatory disease in order to maintain disease suppression.

4.3.5 Late harm

Uncontrolled disease activity can cause late harm to the patient and this must be weighed against concerns about cumulative or late toxicity when selecting the most appropriate DMARD.¹⁰⁰

Although evidence relating to late harm is patchy and incomplete, the development of malignancies in patients treated with immunosuppressive drugs has been studied.^101,102,103 In patients who had developed either neoplasms of the immune system or skin or bladder cancer, the adjusted incidence rate ratio for those in the highest cumulative exposure group was 3.7 compared with those in the lowest exposure group.¹⁰³

Table 6: DMARD profiles

4.3.6 Efficacy of DMARDs

Meta-analysis suggests similar efficacy of sulphasalazine, IM gold, penicillamine and methotrexate.⁷⁵ Double-blind RCTs show equal efficacy of sulphasalazine and methotrexate. Drugs of reduced efficacy from open RCTs include hydroxychloroquine and auranofin.^80,92,93

Recent studies suggest sulphasalazine,⁸⁵ methotrexate⁹⁶ and leflunomide have comparable efficacy (details of ACR 20 and ACR 50 are shown in Annex 7). There are no data from adequate studies with respect to azathioprine. Evidence level 1+

A beneficial effect on radiological progression of RA has been shown with all DMARDs except hydroxychloroquine (which has been shown to have less effect than sulphasalazine in at least one RCT).⁹³

DMARD choice should take into account patient preference and existing co-morbidity.

Sulphasalazine, methotrexate, IM gold, and pencillamine are equally effective DMARDs.

4.3.7 Toxicity of DMARDs

Toxicity assessment in the initial Felson meta-analysis came from 71 clinical trials that contained 129 treatment groups.⁷⁵ Over one year almost one third of the patients (30.3%) stopped therapy. Half of these did so because of drug toxicity. In this meta-analysis injectable gold had higher toxicity rates and higher total dropout rates than the other drugs. Antimalarials and auranofin had relatively low rates of toxicity. In a subsequent meta-analysis the same authors updated their previous meta-analyses by adding trials published through 1990 and trials of azathioprine.⁷⁶ Antimalarial drugs had the lowest toxicity rate of all those studied, but efficacy was only moderate. Methotrexate had the most favourable efficacy/toxicity trade off. Sulphasalazine scored close to methotrexate but in that meta-analysis was slightly more toxic.

A meta-analysis of RCTs of folic or folinic acid supplementation during low dose methotrexate therapy for RA showed that 5 mg folic acid weekly is useful in reducing mucosal and gastrointestinal side effects.^110,123

4.3.8 Choice of DMARD

Overall there is consistent evidence that hydroxychloroquine and auranofin^92,93,75,76 are relatively weak DMARDs with a slower onset of action, while intramuscular gold, penicillamine, sulphasalazine⁹² and methotrexate^75,76 have very comparable clinical effects on disease activity. More patients are likely to continue on sulphasalazine (and with better effect) compared to auranofin and on methotrexate compared with other DMARDs.

Sulphasalazine and methotrexate are the current DMARDs of choice due to their more favourable efficacy/toxicity profiles.

Successive DMARDs are required for most patients in the medium to long term.¹²⁴

4.3.9 Practical prescribing of DMARDs

Patients should be counselled about the benefits and risks of specific DMARDs, and should be provided with additional written information.

Good liaison between primary and secondary care is essential. Rheumatology nurse specialists have an important role in this aspect of care.

Monitor for continued efficacy (ESR/CRP, number of tender and swollen joints, duration of morning stiffness, function).

Monitor toxicity using British Society of Rheumatology/local guidelines or manufacturers' data sheet recommendations.

Clear advice about the monitoring of specific DMARDs should be available to the patient, GP and practice nurse.

4.3.10 Combination DMARD therapy

Combination DMARD therapy in RA is being increasingly used by rheumatologists, but evidence of benefit remains patchy.

A meta-analysis in 1994 concluded that combination therapy does not offer a substantial improvement in efficacy,125 whilst toxicity was increased. Evidence level 1+

Since the 1994 Felson meta-analysis¹²⁵ there have been six parallel (five blind,^{126,127,128,129, 130} one open¹³¹), five step-up^{132, 133, 134,135,136} (two using biological agents^128,135) and one step-down, combination DMARD137 studies. Overall the results have been disappointing.

The addition of cyclosporin to methotrexate in patients with established RA has been shown to be of benefit, but it is not known if the same effect could have been achieved by changing to cyclosporin alone.¹³³ Benefit of a combination of methotrexate, sulphasalazine and hydroxychloroquine has been shown but was not studied in early disease. ¹²⁸ A recent controlled study of iv infiximab in patients with partial response to methotrexate provided additional clinical benefit and prevented further radiological damage. However, the study was conducted in patients with longer disease duration than applies to the patients covered by this guideline.¹³⁶ Evidence level 1+

One study used very high doses of prednisolone (60mg daily initially) which could not be sustained for prolonged periods in clinical practice without unacceptable cumulative toxicity.¹³⁷

A recent open combination study using a variety of DMARDs showed significantly more patients in remission in the combination group and adverse events were of similar frequency. However, the mean improvement in symptoms, clinical signs and function were similar across the groups making this analysis difficult to interpret.¹³¹ Evidence level 1-

In the treatment of poor prognosis early RA, the combination of methotrexate, cyclosporin and targeted intra-articular corticosteroid did not have significant advantages over monotherapy with sulphasalazine in terms of joint damage and function after one year.¹³⁸

Benefit was seen with the addition of etanercept to methotrexate in RA patients with long disease duration and infliximab to methotrexate but at present there is no confirmatory evidence in patients with early disease. ^{134, 135}

At present the balance of evidence does not support the routine use of combination DMARD therapy in early RA.

4.4 Targeted immunotherapy

Tumour necrosis factor (TNF) is a product of macrophages which acts on the immune system to induce the production of powerful pro-inflammatory mediators. TNF is thus a potential molecular target for the treatment of RA. Two agents with anti-TNF activity have recently become available for RA patients:

1. Infliximab - a chimeric monoclonal antibody which is given as an intravenous infusion at 0, 2 and 6 weeks and then every 8 weeks. Side effects include mild infusion reactions and development of antibodies. Infliximab has been used with methotrexate in part to suppress antibodies against the drug. Results from a double-blind study as an addition to methotrexate are shown in Annex 7. There was also one death during the period of the study. ¹³⁴
   
   
2. Etanercept is a TNF receptor fusion protein designed to bind circulatory TNF. It can be administered alone or with methotrexate. It is given as twice weekly subcutaneous injection. Results from placebo-controlled studies are shown in Annex 7. Although no major complications were seen in clinical trials, serious and fatal infections have occurred post-marketing in the USA. A recent double-blind study comparing two doses of etanercept with methotrexate has shown significant advantage of 25 mg of etanercept twice weekly in terms of ACR 20, 50 and 70 outcomes and erosion scores. There were also fewer infections in the patients taking etanercept.¹³⁹

There are concerns that continued inhibition of pro-inflammatory molecules may increase the risk of infection and cancer, particularly lympho-proliferative malignancies. Evidence in this respect is still awaited. The exact role of TNF blockade in early disease has yet to be elucidated and their greater cost may preclude widespread early use.

4.5 Intra-articular corticosteroids

Intra-articular corticosteroid injections are widely used to provide rapid, and sometimes sustained, symptomatic relief in 'target' joints.

Intra-articular corticosteroid injections:

• allow local treatment of inflamed joints whilst minimising undesirable systemic effects
• provide symptomatic relief pending the onset of DMARD effect
• treat particularly troublesome joints where the overall disease control is good
• deal with mono/oligo arthritis in instances when DMARDs are deemed inappropriate.

However, there are few controlled trials in this area and no evidence on the long-term effect on disability or radiological progression. Experience from large cohorts suggests that complications such as joint sepsis are very rare.¹⁴⁰ Synovial fluid aspiration at time of joint injection has been shown to reduce relapse rate.¹⁴¹

Postinjection rest (24 hours) has shown enhanced improvement in symptomatic relief. Walking times were also improved by this approach.¹⁴²

4.5.1 Practical prescribing of intra-articular corticosteroids

Intra-articular injections can be used for rapid, and sometimes sustained, symptomatic relief in 'target' joints.

Intra-articular injections to any one joint should not be given more than three times in one year.

When administering intra-articular injections: use sterile technique advise patients how to seek help if joint fails to settle after injection always consider possible septic arthritis in differential diagnosis of mono/oligo flare in RA.

4.6 Systemic corticosteroids - oral and parenteral

4.6.1 Symptomatic benefit

The symptom relieving anti-inflammatory effects of corticosteroids are well established.¹⁴³ Recent randomised controlled studies have shown that this benefit is not sustained beyond 9 months when either continuous low-dose corticosteroids (7.5mg/day) ¹⁴⁴ or high-dose 'step-down' therapy,¹³⁷ is given as an adjunct to DMARDs or NSAIDs. Evidence level 2+

Bridge corticosteroids (usually IM) are an option to provide symptomatic relief until DMARDs become effective. These show benefit in some patients but their value is limited by possible 'rebound' flare of symptoms on discontinuation.¹⁴³ Evidence level 2+

4.6.2 Acute phase response

In many trials the effect of corticosteroids has been obscured by the concurrent use of other treatments known to affect the acute phase response. No additional benefit of corticosteroid compared to placebo was seen with low-dose corticosteroid study,¹⁴⁴ but a significantly more rapid fall in the ESR was apparent with the higher doses used by Boers.¹³⁷ The effect on acute phase response had disappeared by one year (by which time the corticosteroid had been discontinued).

4.6.3 Functional response

Although some benefit in function from corticosteroids was reported,¹⁴⁵ no objective long-term benefit was discerned. Recent studies using the well-validated health assessment questionnaire (HAQ)¹⁷ to measure function have shown an early advantage of high-dose 'step-down' corticosteroids.¹³⁷ However, the improved HAQ scores seen with adjunctive continuous low-dose corticosteroid were no longer significant by one year and disappeared beyond 15 months.¹⁴⁴ To date, no controlled study of systemic corticosteroid has truly addressed the important assessment of disability as a long-term outcome.

4.6.4 Radiological progression

A number of early studies suggested that oral corticosteroids may inhibit radiological damage and recent randomised controlled trials confirm this finding.^{137, 144}

The development of new erosions was reduced by 23% and the progression of existing damage was highly significantly retarded on hand x-rays of patients taking low-dose corticosteroid for two years.¹⁴⁴ On discontinuing corticosteroid therapy radiological progression was again seen at pre-treatment rate.¹⁴⁶ Evidence level 1+

A high dose step-down regimen of corticosteroid with combination DMARD also demonstrated inhibition of erosions but had no effect on joint narrowing.¹³⁷

It should be noted that a significant proportion of placebo treated patients in both studies did not develop erosions during the study.^{137, 144}

4.6.5 Cumulative toxicity

Osteoporosis was not formally measured in the Kirwan trial of continuous low-dose oral corticosteroid.¹⁴⁴ In a cohort study of 8,068 patients given a mean dose of 5mg of prednisolone per day, both reduced bone mineral density and increased fractures were demonstrated over 4 years. A relative fracture risk of 2:1 was derived for corticosteroid use after correction for multiple variables.¹⁴⁷ Evidence level 2+

Two case control studies show increased adverse events in corticosteroid treated rheumatoid arthritis patients, including cataracts, infections (the Committee on Safety of Medicines has drawn attention to the risks of chicken pox exposure in patients not previously infected¹⁴⁸), gastrointestinal bleeds, avascular necrosis and fractures.^149,150,151 Increased mortality has also been reported.¹⁵² There is the possibility that corticosteroid treated patients may have had more severe disease.

Both cumulative and average corticosteroid doses are independent, important adverse event predictors.¹⁵¹ Longer term studies are required to identify the cumulative long-term effect from continuous low-dose corticosteroid and from step-down or intermittent regimens.

Most clinicians withdraw oral corticosteroids slowly (e.g. 1 mg/month when below 15 mg daily) to avoid rebound flare of symptoms.

4.6.6 Practical prescribing of systemic corticosteroids

The short-term symptomatic benefit of systemic corticosteroids and the apparent prevention of radiological damage must be weighed against the risk of significant morbidity. Until additional, adequately powered and long-term studies are performed to address the benefits and risks the routine use of oral corticosteroids cannot be recommended. In specific situations where there are strong contraindications to NSAID prescription, or difficulties in using DMARDs, systemic corticosteroids may be acceptable.

Oral corticosteroids are not recommended for routine use, as there is no sustained clinical or functional benefit and there is high risk of toxicity with long-term use.

Inform patients of the risks of corticosteroids prior to prescription and issue steroid warning card.

Intramuscular corticosteroid allows control of dose and duration of therapy and may be preferable to oral therapy.

Oral corticosteroids should be withdrawn slowly to avoid rebound flare of symptoms.

The lowest possible dose of corticosteroid should be used for the shortest possible time.

Monitor patients closely for adverse corticosteroid effects. Be alert to the possibility of diabetes, cataract and infection. Inform patients not previously infected of the danger of chicken pox/shingles exposure.

Ensure adequate prophylaxis and treatment of osteoporosis in patients taking oral corticosteroids.

4.7 Complementary medicine

Few studies in this field relate specifically to rheumatoid arthritis and many studies have been excluded from the guideline on the grounds of small numbers and poor design.

There is unsatisfactory evidence of possible subjective benefit of homeopathy over placebo.^153,154,155 A proprietary remedy 'Rheumalex' appeared to help in pain relief, but the herbal substance, feverfew showed no evidence of benefit.^156,157 Acupuncture showed no benefit in one meta-analysis reviewed but the quality of this analysis was limited.^158,159 There is no evidence that Seatone or selenium produced any clinical benefit.^160,161,162

While no evidence of effectiveness is not the same as evidence of ineffectiveness the lack of adequate research studies precludes firm conclusions. Patients have a perception that because these treatments are 'natural' they are without side effects but this is not the case.¹⁶³

Further research is clearly needed and should include close monitoring of possible harm as well as potential benefit.

4.8 Other therapies

4.8.1 Hormone replacement therapy

Although RA patients on HRT have been shown to report a significant increase in general wellbeing compared with placebo, there is to date no evidence of alteration in indices of disease activity.^164,165 Hormone replacement therapy (HRT) is of undoubted benefit in improving bone mineral density in postmenopausal women with RA and associated osteoporosis.¹⁶⁶

4.8.2 Iron therapy

Anaemia is common in RA and is not always due to iron deficiency. Unnecessary iron supplements should be avoided. Patients who are truly iron deficient should be assessed and any dietary deficiency should be corrected. GI blood loss requires investigation and specific therapy e.g. proton pump inhibitor (see section 4.2.2). Ferritin acts as an acute phase reactant in RA, thus a 'normal' ferritin does not exclude iron deficiency.