The old paradigm of therapy was called the "pyramid approach." The bottom of the pyramid included anti-inflammatory drugs, such as aspirin and nonsteroidal anti-inflammatory drugs, that were prescribed along with a basic program of exercise, rest, and education. Second-line therapy was initiated only after several years of treatment with antiinflammatory drugs and generally after radiographic evidence of joint damage. In fact, many rheumatologists waited for such evidence, considering it to be the reason to start second-line therapy. Part of the justification for delaying the initiation of second-line therapy was the perceptions that rheumatoid arthritis was a benign disease and that second-line therapies were very toxic. Over the past decade, we have come to realize that rheumatoid arthritis is not benign, that second-line therapies are not as toxic as initially thought, that first-line therapies have substantial toxicities of their own, and that the pyramid approach does not affect functional, clinical, or radiographic progression. This has generated an important dialogue among rheumatologists about modifying that traditional treatment approach [1-4].

Patients with rheumatoid arthritis have increased mortality compared with the general population. Factors predictive of death include (in addition to age and comorbid illnesses) baseline disease severity as measured by joint counts, functional status, and level of formal education [5]. A significant decline in functional status over time has also been seen with this disease [6]. Patients with rheumatoid arthritis have an increase in work disability; 50% are work disabled within 10 years of disease onset. Disability from rheumatoid arthritis and other types of symmetric polyarthritis costs more than 6.5 billion dollars per year (in 1986 U.S. dollars) in lost earnings [7]. A threefold increase in medical costs for patients with this disease compared with an age-and sex-matched population was reported in 1989 [8]. These are costs for a disease that affects approximately 2.5 million persons in the United States and for which no definite cure exists. In our current, controlled, and perhaps restricted health care environment, the implications of the cost (both direct and indirect) are clearly evident and of great concern.

One important finding is that radiographic damage occurs early in the course of rheumatoid arthritis [9, 10]. In fact, van der Heidje and coworkers [10] reported that 70% of patients in their study developed radiographic damage within 3 years of onset [10]. The rate of radiographic progression was higher in the first than in the second or third year of disease activity. This suggests that the opportunity to intervene and possibly modify the course of disease before structural damage occurs may appear earlier rather than later. Our conventional second-line therapies do not generally alleviate or heal radiographic damage. Even though second-line therapies may not uniformly halt radiographic progression, several of them have been shown to slow the rate of progression and reduce the formation of new erosions [11-14]. If we wait for the development of erosions and joint space narrowing before starting second-line therapy, we may be too late to affect disease activity and functional status. This will be even more critical as more potent and possibly more effective therapies are developed.

Second-line therapies are now started much sooner in the course of disease. In this issue, van der Heide and colleagues [15] report that patients who had disease activity for less than 1 year and were randomly assigned to receive various second-line therapies (with or without concurrent nonsteroidal anti-inflammatory drugs) had a greater reduction in disability and pain and a greater improvement in joint counts than patients randomly assigned to receive nonsteroidal anti-inflammatory drugs alone. However, a subset of patients who had received only nonsteroidal anti-inflammatory drugs had a beneficial response. This finding is difficult to interpret, because patients with rheumatoid arthritis have a 10% remission rate in the first year of disease, independent of treatment, and we cannot identify those patients a priori. In this study, no difference was seen in radiographic progression, but the short duration of the study and low radiographic activity scores in the study groups may have limited the power of the study to detect such outcomes. Other investigators have also reported that earlier intervention improves clinical and functional outcomes in patients with rheumatoid arthritis [16, 17].

The decision about which second-line therapy to use is still arbitrary at this point. Specific drugs (hydroxychloroquine, sulfasalazine, methotrexate, gold salts) are recommended on the basis of an individual patient's clinical disease activity, functional status (including the ability to perform activities of daily living), work status, age, underlying medical status, and lifestyle (including plans for conception and alcohol use). These decisions are made on a case by case basis; it is difficult to develop a specific, widely applicable drug algorithm for the treatment of rheumatoid arthritis. Consultation with the rheumatologist is helpful in assessing the patient and advising the patient and primary care provider about which specific second-line therapy is recommended. Recommendations about second-line therapy also require in-depth knowledge about the duration of onset of each second-line drug, drug interactions, monitoring guidelines, and—most importantly—drug toxicity. Proper management involves not just pharmacologic interventions but may require consultation with physical and occupational therapists, podiatrists, practitioners of orthotics, nursing specialists, vocational rehabilitation workers, and reconstructive surgeons. In one study [18], patients treated by a rheumatologist had a lower rate of functional disability than patients managed exclusively by a nonspecialist. This difference in the disability rate was associated with more intensive use of second-line therapies and joint surgery among patients treated by the rheumatologist.

A second and even more fundamental problem with rheumatoid arthritis concerns substantial delays in diagnosis. In one study at a health maintenance organization [19], a median lag time of 36 weeks occurred before the diagnosis was established. Cooperative education programs and greater interaction between the primary care physician and the rheumatologist are required so that arthritis can be diagnosed sooner and second-line therapies can be started earlier. Unfortunately, despite this considerable progress in our understanding, we still cannot tell whether any one second-line therapy is more effective than the others in improving outcomes when started early in the course of disease. Further study of this complex question is badly needed. What is clear, however, is that the evidence now supports greater short-term effectiveness in improving functional status for so-called second-line therapies when used early, and with acceptable levels of toxicity, relative to nonsteroidal anti-inflammatory drugs alone. We need to modify our old paradigm of treatment and adopt the new concept of "treat now, not later!"

Michael E. Weinblatt, MD

Brigham and Women's Hospital

Harvard Medical School

Boston, MA 02115