There are now some data to suggest that the immunopathogenesis of RA differs not only among different patients, but also at different stages of the disease in the same patient. We have numerous biomarkers that are used in the diagnosis and monitoring of patients with RA, but we really do not have the ability to select individual patients, in advance, for the most appropriate therapies. Thus, the therapeutic plan is often subject to course corrections, which manifest as trials of new therapies.

Wouldn’t it be nice to be able to predict which patient will do best on which drug? Perhaps, I’d add, at which point in the patient’s disease? That is, I’d really like it if, for my next patient with RA, I could do a test—perhaps a biomarker test—and predict that, for this individual patient, the disease is predominantly tumor necrosis factor (TNF)–driven. In this case, an anti-TNF agent would be most appropriate for him or her. Then, for the next patient, perhaps the disease is more interleukin-6 (IL-6)–driven, so, in that case, an agent that inhibits IL-6 would be preferred. Ideally, we might have a test to determine, for example, whether the patient’s disease is predominantly IL-6–driven or predominantly TNF-driven. Now, we obviously do not yet have the ability to select patients for individualized treatment, but it would be a wonderful thing.

Recent updates to the American College of Rheumatology (ACR) and the European League Against Rheumatism (EULAR) recommendations have recognized this need, suggesting that work be done to sort out which predictive biomarkers have shown promise and which have repeatedly failed in clinical trials.

Diagnostically, we have rheumatoid factor available, along with testing for anti-citrullinated protein antibodies, which has a higher sensitivity and specificity. Also available for diagnostic purposes is testing for 14-3-3 eta.

Disease activity can be measured using acute phase reactants, such as erythrocyte sedimentation rate and C-reactive protein (CRP), which are also measured as components of RA disease activity indices (eg, Disease Activity Score in 28 joints [DAS28]). Unfortunately, these tests are frequently normal in patients with active disease. We also have the multi-biomarker disease activity (MBDA) score, which measures serum levels of 12 proteins and was validated against the DAS28 CRP (DAS28CRP). The MBDA score includes measurement of serum levels of acute phase reactants, inflammatory cytokines, cell adhesion molecules, adipose tissue products, and matrix metalloproteinases.

However, to return to the overarching need for treatment predictive biomarkers, the MBDA does not deliver on this need, nor was it designed to do so. Baseline MBDA scores do not predict the clinical response to methotrexate monotherapy or to second-line therapies. Low and moderate MBDA do associate with a very low risk of subsequent radiographic joint damage. In addition, low levels of 14-3-3 eta have been shown to correlate with a low risk of radiographic progression. The reverse is also true.

Thus, although there has been a good deal of progress on diagnostic markers and on markers that correlate with the risk of progression, there continues to be a great need for new biomarkers that will help us to deliver on the promise of personalized medicine, whereby treatment is tailored to the individual patient—and his or her phenotype, disease stage, and predominant drivers.