Robust predictive biomarkers of response to rheumatoid arthritis treatment remain elusive, researchers said in a review of the recent literature on this topic, which was published online Oct. 28 in Annals of the Rheumatic Diseases.
The ability to identify those rheumatoid arthritis (RA) therapies to which a patient is likely to respond is crucial in order to prevent unnecessary exposure to toxic adverse effects and allow physicians to "abandon the trial-and-error approach to treatment in favor of evidence-based guidance."
Dr. Paul Emery
It would also help control treatment costs, said Dr. Paul Emery, who is the arthritis research campaign professor of rheumatology and head of the academic section of musculoskeletal disease at the University of Leeds (England), and Dr. Thomas Dörner, who is professor of rheumatology at Charité Universitätsmedizin Berlin.
They assessed the recent literature for evidence supporting the use of various biomarkers in this regard. These fell into four categories: genetic, protein, and cellular biomarkers, as well as predictive patient factors.
Genetic markers show the most promise as indicators of RA treatment response, but studies have been limited to date, and better-designed research is needed to confirm their potential usefulness.
For example, several polymorphisms of the tumor necrosis factor (TNF)–alpha gene have been studied, but the results have been inconsistent. Some studies have suggested that certain polymorphisms predict the response to TNF inhibitors such as etanercept, adalimumab, and infliximab, but others have contradicted this finding.
"There may be some merit in further investigation of the [single nucleotide polymorphism] in the TNF-308 promoter region to confirm its potential" to predict response to TNF inhibitors, but so far "the data are not conclusive because studies ... have been limited by the small patient numbers, heterogeneity in the underlying genetic background, [and] differences in study design," Dr. Emery and Dr. Dörner said (Ann. Rheum. Dis. 2011 Oct. 28 [doi:10.1136/ard.2010.148015]).
Some investigators have suggested that some major histocompatibility complex alleles that confer susceptibility to RA, such as the shared epitope in the human leukocyte antigen region, may be potential markers of treatment response. But overall, findings from studies thus far "indicate that the [shared epitope] motif is not a robust genetic marker for predicting response to TNF inhibitors."
The findings have been similarly sparse for other gene polymorphisms such as the p38 network of mitogen-activated protein kinases and variations in the PTPN22, STAT4, TRAF1-C5, PADI4, and CTLA-4 genes. Current data are too limited to determine whether markers for these genes might predict treatment response.
Analyzing mRNA expression of thousands of genes simultaneously is emerging as a promising technique to identify unique sets of genes that are expressed differently between responders and nonresponders to a given therapy. However, there have been few studies of this method, and some of those few analyzed samples from the peripheral blood before it became known that samples from the synovial fluid are more useful.
On another front, several proteins involved in cartilage turnover, bone resorption, or the TNF-signaling pathway have been proposed as possible markers of treatment response, particularly to anti-TNF agents. But the research to date has involved very small numbers of subjects. "Consequently, no robust protein biomarkers have yet been confirmed as predicting response," the investigators said.
Similarly, it may be possible in the future to use the dominant type of interferon signature in a patient’s peripheral blood sample to predict response to rituximab in particular, but this hasn’t yet been confirmed in multiple studies.
And several autoantibodies present in blood samples from RA patients may eventually move to "the forefront of RA management, although we are not quite there," Dr. Emery and Dr. Dörner added.
With regard to cellular biomarkers, certain subsets of B cells have been noted during either depletion or repletion phases of treatment. "However, the utility of assessing B cells before and during treatment requires confirmation by independent controlled studies and rigorous validation before [it] can be recommended for use in clinical practice," the researchers said.
Finally, both clinical and demographic patient traits have been proposed as predictive of treatment response, including the baseline 28-joint disease activity score and C-reactive protein levels. But none has proved to be sufficiently predictive to guide therapy, they said.
The writing of this review was supported in part by Hoffmann-La Roche. Dr. Emery reported ties to Pfizer, Abbott, Bristol-Myers Squibb, Merck, Novartis, and Roche. Dr. Dörner reported ties to Chugai, Roche, and UCB.