Prediction Tool Targets Undifferentiated-arthritis Dilemma: To Treat or Not to Treat?

When it comes to deciding on a course of action for patients who present with early, undifferentiated arthritis (UA), physicians face a dilemma: 40% to 50% of such patients will spontaneously achieve remission without any intervention, but about a third of them will ultimately progress to rheumatoid arthritis (RA). A mounting body of evidence shows that for patients in the initial phase of rheumatoid arthritis, early and aggressive drug treatment within the first few months is crucial to stave off permanent joint and bone damage from the disease, and associated disability. Still, drugs for RA -- including the widely used methotrexate and other disease-modifying antirheumatic drugs (DMARDs) -- must be used carefully to avoid kidney, liver and bone marrow damage. So which patients with UA are likely to improve on their own -- and should simply be monitored -- and which ones are likely to develop rheumatoid arthritis, and should receive immediate treatment? Until now, physicians making this decision have had to rely solely on their clinical judgment, given the highly variable course of undifferentiated arthritis and the lack of any single, reliable test for rheumatoid arthritis. But a recently published prediction model and clinical tool from the Early Arthritis Clinic at Leiden University Medical Center (Netherlands), has been shown to accurately predict the disease course of patients with undifferentiated arthritis, and could help physicians and patients make treatment decisions. In undifferentiated arthritis, the patient's symptoms do not fit the American College of Rheumatology (ACR) diagnostic criteria for any specific type of arthritis. 9 key predictive variablesThe prediction model, published in the February 2007 issue of Arthritis and Rheumatism, consists of 9 clinical variables that are commonly evaluated in clinical practice and were determined to have independent predictive value for rheumatoid arthritis. By filling out a simple form (see p. 436 of the Arthritis and Rheumatism article), the physician enters the patient's data for the 9 variables and computes a score from 0 to 14 -- a process that takes just a few minutes. The higher the patient's score, the greater the likelihood that he or she will develop RA.The 9 key variables identified by the model are:

• patient gender
• patient age
• location of affected joints (small joints of the hands/feet; symmetric localization; both upper and lower extremities)
• degree of morning stiffness
• number of tender joints
• number of swollen joints
• C-reactive protein level
• rheumatoid factor positivity
• presence or absence of anti-cyclic citrullinated peptide antibodies (CCP)

Preventing undertreatment, overtreatment

The study authors, led by Tom W.J. Huizinga, MD, PhD, chairman of rheumatology at Leiden University Medical Center, say their prediction model can prevent undertreatment as well as overtreatment for undifferentiated arthritis. "Because the prediction rule is accurate and can be easily determined in daily clinical practice," their paper states, "the present model is an important step forward in achieving individualized treatment in patients with recent-onset UA."While a

decision model already existed

(also from the Leiden clinic, from 2002) to guide treatment decisions for patients with persistent, erosive arthritis, the latest model is the first one designed to evaluate patients with undifferentiated arthritis.

Not standard practice, but useful as a guide

U.S. rheumatologists specializing in early arthritis and RA say the new model won't likely become standard practice anytime soon. It still needs to be validated in other countries with larger, more heterogeneous populations of arthritis patients (research to do this is underway in Germany and the UK). And, more to the point, time-strapped physicians may not find it feasible or desirable to follow the model exactly."I'd be skeptical that doctors would take this model, follow every step and apply it to every patient," says Dennis W. Boulware, MD, professor of medicine in the division of rheumatology at the University of Alabama at Birmingham."Physicians are unlikely to sit there and do these calculations," echoes John J. Cush, MD, clinical professor of internal medicine at the University of Texas Southwestern Medical School. "They don't have the time and they don't want to follow a formula; they want to look at the patient and use their clinical judgment."Still, these U.S. researchers say the model is a well-done exercise and a helpful reminder of the key factors physicians should already be thinking of when evaluating UA patients. And they say such a tool, even if not followed exactly, could help primary-care physicians treat appropriate patients more aggressively -- a strategy rheumatologists have increasingly been advocating in recent years."This model can be useful as a guide, and it could help physicians feel more comfortable about the decision to start treatment," Boulware says, adding, "It's somewhat intuitive -- the more joints involved, the more symptoms, the more positive lab results, the more aggressive you should be."

Primary care is key

Primary-care management of all forms of arthritis will become increasingly important in the coming decades. According to the CDC's May 4, 2007

Morbidity and Mortality Weekly Report

, the number of U.S. adults diagnosed with arthritis is predicted to rise from about 46 million today to 67 million by 2030. Meanwhile, the increased demand for rheumatologists is already creating a shortage in some areas of the country, and it's expected to worsen. According to

"The United States Rheumatology Workforce: Supply and Demand, 2005-2025"

-- published in the March 2007 issue of

Arthritis & Rheumatism

-- "the consensus of practicing rheumatologists is that a shortage exists now," with average wait times for new patients at 38 days. Furthermore, according to the researchers' supply projection model, the demand for rheumatologists in 2025 is expected to exceed supply by 2,576. "The role of primary-care physicians for arthritis is incredibly important, since they're the first line of evaluation and therapy," Cush affirms.

Development of the model

Starting with clinical data for some 1,700 arthritis patients, the Leiden researchers identified 570 patients with recent-onset UA and monitored them for a year. At the culmination of follow-up, 177 of the original UA patients were determined to have RA and 150 had achieved remission; the remaining 94 had developed another rheumatologic disease. Through questionnaires, physical exams and bloodwork, the researchers identified the 9 (aforementioned) clinical variables with independent predictive value for rheumatoid arthritis. Based on their assessments, the researchers developed a 14-point score, with 0 representing the lowest risk and 14 the highest risk of developing RA. None of the study patients who scored 3 or less progressed to RA; all of those who scored 11 or higher did. The likelihood of developing RA increased with higher scores between 4 and 10.

Cutoff scores guide treatment decisions

Based on their findings, the researchers recommend using cutoff scores of 6 and 8 to guide treatment decisions, explains first author Annette H. M. van der Helm-van Mil, MD, PhD, of Leiden University Medical Center. In their study, 91% of the UA patients with a score of 6 or less did not develop RA; 84% of the patients with a score of 8 or greater did develop RA. Accordingly,

• Patients with a score of 8 or higher should begin treatment with an effective RA drug.
• Patients with a score of 6 or less should be monitored, but should not start on medications, except for anti-inflammatories.

For patients with a score of between 6 and 8, van der Helm says her group's research provides no guidance on how to proceed, since these patients have a 50% chance of progressing to RA and a 50% chance of not doing so.

Give earliest possible treatment

Given the serious consequences of delaying treatment for rheumatoid arthritis, however, Cush says most rheumatologists today would agree that even for patients at moderate risk, it's better to err on the side of earlier treatment."The evidence clearly shows that any delays in treatment, for even 3 months, can have devastating downstream effects" in terms of joint damage and disability, he says. "So the current thinking is, institute the earliest possible treatment." Cush cites the

latest results from the PROMPT study

(PRObable Rheumatoid Arthritis: Methotrexate versus Placebo Treatment Trial), a randomized multicenter trial which monitored for up to 30 months 110 patients with UA who fulfilled the American College of Rheumatology criteria for probable RA. The researchers found that the patients randomized to take methotrexate fulfilled the ACR criteria for rheumatoid arthritis at a later time point than in the placebo group, and fewer patients on methotrexate showed radiographic disease progression and joint damage over 18 months than the placebo group.For physicians who prescribe DMARDs, there's an important caveat: They must make sure to perform the recommended lab tests to monitor the drugs' effects on patients' kidneys, liver and bone marrow cells. A chart detailing the tests appears on pages 333-335 of the ACR's

Guidelines for the Management of Rheumatoid Arthritis

.

Related Links

• A prediction rule for disease outcome in patients with recent-onset undifferentiated arthritis: How to guide individual treatment decisions
  Arthritis & Rheumatism, February 2007

• American College of Rheumatology Classification Criteria for Rheumatic Diseases

• ACR Guidelines for the Management of Rheumatoid Arthritis: 2002 Update

• CDC Predicts Arthritis Boom
  WebMD News

• The United States Rheumatology Workforce: Supply and Demand, 2005-2025
  Arthritis & Rheumatism, March 2007

• Efficacy of methotrexate treatment in patients with probable rheumatoid arthritis: a double-blind, randomized, placebo-controlled trial
  Arthritis & Rheumatism, May 2007

Have comments or questions on this article? Please e-mail the author, Sara Selis, at sselis@cmp.com.