Progress in the Treatment of Rheumatoid Arthritis

Rheumatoid arthritis (RA) is a chronic inflammatory arthritis involving 0.5% to 1% of the US population.¹ This disease affects women twice as often as men and its incidence rises with increasing age. In general, RA causes a symmetric polyarthritis affecting large and small joints in association with systemic manifestations, such as morning stiffness, fatigue, and weight loss. Although considered a joint disease, RA can have widespread effects on the entire body and can impair life expectancy by as much as 5 to 10 years. With progressive disease, patients with RA develop work disability, functional impairment, and radiographic evidence of joint damage.^{2 - 3}

While many conditions can produce polyarthritis, the diagnosis of RA can usually be established by the presence of persistent joint pain, swelling in a symmetric distribution, and prolonged morning stiffness. Supporting the diagnosis are laboratory findings, such as a positive test result for rheumatoid factor. The intensity of joint inflammation often correlates with the erythrocyte sedimentation rate or levels of serum C-reactive protein, but these abnormalities are nonspecific. In the management of RA, the primary care physician plays an important role by ensuring early diagnosis. If unsure about the presence or significance of synovitis, the primary care physician should consult a rheumatologist for this determination.

The management of RA has undergone a revolution in the past decade, reflecting a growing armamentarium of drugs and a shift in treatment strategies. While RA is a chronic condition, joint damage occurs at onset for many patients. Thus, in the current strategy, which has replaced the older, more gradualist treatment pyramid, therapy is begun immediately to control disease activity, reduce functional impairment, and prevent irreversible changes in cartilage and bone. Since prompt disease control is a major goal of therapy, the primary care physician should consider referring a patient with suspected RA to a rheumatologist for early treatment. A variety of surgical procedures is also available to reduce joint pain and improve function, but a discussion of these modalities is beyond the scope of this article.

Although designated an inflammatory arthritis, RA produces a more complicated lesion of the synovium composed of inflammation, proliferation, and destruction of cartilage and bone. By mechanisms that are as yet unknown, sustained inflammation stimulates synovial cell proliferation, forming pannus, an expansive, aggressive tissue that erodes cartilage and bone.⁴ Therapy in RA thus demands agents that block inflammation, retard synovial proliferation, and prevent joint erosion. Since proliferation is linked to inflammation, agents that limit immune responses may affect the entire process.

Drugs used to treat RA can be divided into 4 classes: nonsteroidal anti-inflammatory drugs (NSAIDs), corticosteroids, disease-modifying antirheumatic drugs (DMARDs), and analgesics. Drugs of the first 3 classes can reduce the number of painful and tender joints, duration of morning stiffness, and indices of inflammation, such as erythrocyte sedimentation rate and serum C-reactive protein. To be designated a DMARD, a drug must be shown to reduce joint destruction, such as radiographic joint erosions or joint space narrowing.

Nonsteroidal anti-inflammatory drugs are useful for the treatment of RA because of their anti-inflammatory and analgesic actions which, in some patients, may adequately suppress symptoms and obviate the need for DMARDs. In addition, NSAIDs may be used initially for patients with suspected RA to control symptoms temporarily until the diagnosis is established or the disease process warrants DMARD therapy.

Greater potential use of NSAIDs has come with the availability of cyclooxygenase 2 (COX-2) selective agents. These agents may have lower gastrointestinal toxicity than conventional NSAIDs, which inhibit both COX-1 and COX-2.⁵ Of the selective COX-2 inhibitors, celecoxib has been approved for use in RA and osteoarthritis, while rofecoxib has been approved for use in osteoarthritis and acute pain. Although selective COX-2 inhibitors have similar efficacy as conventional NSAIDs, their more favorable safety profile may lead to more frequent or sustained use in RA. As shown in recent studies, COX-2 selective inhibitors may cause fewer major gastrointestinal bleeding complications than conventional NSAIDs,^{6 - 7} though the extent and duration of this reduction may vary depending on the agent and comparator. Conventional NSAIDs nevertheless may be combined with gastroprotective agents, such as misoprostol, to lower toxicity.⁸

Despite their potential for adverse effects, low doses of oral corticosteroids (≤10 mg/d of prednisone, or its equivalent) can provide symptom relief for many patients and may serve as DMARDs. As shown in a randomized, double-blind, placebo-controlled trial, treatment with 7.5 mg/d of prednisolone can reduce the rate of radiographic progression of joint damage over 2 years.⁹ Low-dose corticosteroids can be useful in early disease for rapid control of disease activity or, during later stages, to reduce symptoms when DMARD treatment is inadequate or cannot be used because of toxicity concerns or multiple drug failures. In addition to systemic use, corticosteroids can be administered intra-articularly, especially in patients with isolated joint swelling.

Disease-modifying antirheumatic drugs are the cornerstone of aggressive treatment of RA with recent studies indicating that early DMARD therapy affords superior benefits compared with a more delayed approach.¹⁰ DMARD therapy for RA has been extensively studied, usually for a minimum of 6 months, with responses typically defined by criteria established by the American College of Rheumatology (ACR).¹¹ This definition calls for 20% improvement in disease measures that cross various health-related domains. To achieve an ACR20 response, a patient must show a 20% improvement in the tender and swollen joint counts and at least 20% improvement in 3 of 5 other measures (patient global assessment, physician global assessment, patient pain assessment, physical disability score, and serum levels of acute-phase reactants).

The DMARD class includes several agents (Table 1). Among these drugs, methotrexate, a folic acid antagonist, is the most frequently prescribed for the initial treatment of moderate to severe RA. Weekly methotrexate can produce ACR20 response rates of 50% to 60% and, as shown recently, can retard radiographic joint damage.¹² While once a major concern with methotrexate, hepatic toxicity (fibrosis and cirrhosis) can be limited by dose adjustment on the basis of periodic liver function testing.

In general, methotrexate produces more sustained responses over time than other DMARDs, such as sulfasalazine, parenteral gold, and hydroxychloroquine.¹³ These other drugs nevertheless can be used in certain clinical settings and as part of drug combinations. Sulfasalazine, for example, is a frequent choice for patients with mild or moderate disease activity and, like methotrexate, can reduce symptoms and slow radiographic progression.¹⁴ Cyclosporine, azathioprine, parenteral gold, and D-penicillamine all have demonstrated efficacy, but their use has been limited by toxicity and risk-benefit considerations. Hydroxychloroquine and minocycline can reduce the signs and symptoms of RA, although their effects on radiographic progression have not been established. These latter 2 drugs produce relatively few serious adverse effects and are primarily used for treatment of patients with mild disease.

Among the 3 new DMARDs approved for RA, leflunomide is a chemical compound while the tumor necrosis factor α (TNF-α) blockers are biologic agents based on naturally occurring proteins. Leflunomide inhibits pyrimidine synthesis and may affect T-cell activation in vivo. In a large clinical trial, Strand et al¹² reported ACR20 response rates of 52% for leflunomide compared with 46% and 26% for methotrexate and placebo, respectively, while Smolen et al¹⁴ showed that leflunomide was more effective than placebo and similar in efficacy to sulfasalazine. In these studies, leflunomide treatment also improved physical function and quality of life¹⁵ and slowed radiographic progression.¹⁶ Because of leflunomide's long half-life, treatment with cholestyramine may be required to speed its elimination from the body, if toxicity occurs.

Etanercept and infliximab act by inhibiting TNF-α, a key cytokine in joint inflammation. Etanercept is a genetically engineered, dimerized, fusion protein composed of a TNF receptor and an IgG1:Fc portion. Infliximab is a chimeric (mouse/human) anti–TNF-α monoclonal antibody. Both bind to TNF-α with high affinity, preventing the interaction of TNF-α with its receptor and the subsequent triggering of inflammation.

Clinical studies show that etanercept and infliximab are highly effective for treating patients with an inadequate response to other DMARDs. Etanercept (25 mg) given twice weekly produced an ACR20 response rate of 59% compared with an 11% rate for placebo.¹⁷ Another recent study has shown that etanercept compares favorably with methotrexate for treating early RA.¹⁸ In this 12-month study, the etanercept and methotrexate treatment groups had similar ACR20 response rates and changes in the primary radiographic end point. In a secondary analysis, however, etanercept was associated with less progression of joint erosions than methotrexate.

Infliximab has shown significant benefits in combination with methotrexate. In a 54-week study, infliximab, at 4 different doses when added to stable doses of methotrexate, produced ACR20 response rates of 42% to 59%, significantly higher than the 17% rate for placebo plus methotrexate.¹⁹ Infliximab also significantly retarded the development of radiographic joint damage in this study.

Etanercept and infliximab have been well tolerated in clinical trials. Etanercept can cause local injection site reactions, usually mild. Infliximab may cause infusion reactions, which occur in less than 5% of patients and are rarely severe. Because TNF-α functions in host defense, its blockade has been considered a possible risk factor for infection. In clinical trials, these agents did not increase the rates of serious bacterial infections. In subsequent practice, however, use of etanercept and infliximab has been associated with reports of serious infections. As a result, the US Food and Drug Administration has warned that these drugs may lead to infectious complications and should be discontinued in the event of a clinically important infection.

During postmarketing surveillance, the use of infliximab has been linked with the occurrence of tuberculosis, prompting the Food and Drug Administration to demand change of the package insert for this drug. Almost all of these cases occurred during the first several months of treatment, implying reactivation of latent tuberculosis, though this passive surveillance system does not allow determination of causation because of comorbid medical conditions and the use of other medications. Fewer cases have been reported with etanercept therapy. In view of these findings, patients should be evaluated for tuberculosis prior to TNF-α inhibitor therapy and, if indicated, receive antituberculosis prophylaxis for latent disease. Additional adverse effects reported with these agents include other opportunistic infections and neurologic events resembling multiple sclerosis. The long-term efficacy and safety of TNF-α inhibitors thus remain uncertain.

Since the response to single DMARD therapy is often incomplete, combinations of 2 or 3 DMARDs have been tested using various study designs to improve clinical outcomes. One design calls for the addition of a second DMARD (or placebo) to the regimen of patients with active disease despite methotrexate therapy. Table 2 summarizes trials using this design and the 2- and 3-DMARD combinations whose benefits exceeded those of methotrexate alone.^{19 - 22} This stepped up approach, in which DMARDs are added sequentially to achieve disease control, is commonly used in clinical practice.

Other trials of combination therapy have used a parallel-group design in which patients are randomly allocated to receive treatment with either a combination or a single agent. Table 3 provides examples of large trials comparing the value of combinations vs single drugs. These studies have primarily focused on patients with disease of relatively short duration,^{23 - 26} with the rationale that this group would most likely benefit from early intensive treatment. Thus far, the outcomes from these trials have produced variable results.^{23 - 27} Further studies are needed to determine the best initial DMARD therapy for RA.

In principle, maximal cost-benefits from early intervention would occur by reserving the most intensive treatment regimens for patients with the worse prognosis and using less expensive and potentially safer agents for those with more favorable projected outcomes. While genetic and other biomarkers with prognostic value have been sought, at present, treatment decisions remain empiric, with choices of agents based on clinical severity and individual patient preferences (eg, concern over adverse effects, tolerance for injections). In general, the level of disease activity guides treatment decisions, though the radiographic progression of joint damage may be a factor in decision making. The new DMARDs, especially the TNF-α blockers, should nevertheless improve outcomes and afford additional treatment options to reduce pain and disability in patients with this common arthritic disease.