We read with great interest the article on cognitive impairment in rheumatoid arthritis (RA) published recently in Arthritis Care & Research (1). The authors reported on a comprehensive cognitive assessment of 115 patients with RA and noted cognitive impairment in 31% of the participants. Education level, income, glucocorticoid use, and cardiovascular disease were found to be the significant predictors of cognitive impairment. We wish to draw attention to 2 points that we believe are applicable to the interpretation of these results. The first point is about the potential effect of medications other than glucocorticoids, such as methotrexate (MTX), and the second point concerns the considerable challenges encountered in the assessment of cognitive function in RA.

Given that, in RA, pharmacologic treatment commonly consists of a combination of drugs (e.g., MTX, glucocorticoids, antiinflammatory agents, leflunomide, hydroxychloroquine, and biologic agents) rather than monotherapy, the singular effect of any one medication is difficult to isolate. In a recent study of 35 RA patients receiving long-term MTX treatment, most of whom were also taking glucocorticoids, cognitive function was tested before and after taking MTX (24 hours apart) ([2]). While no cognitive decline was detected, surprisingly little learning occurred given the practice effect opportunity. Participants taking a high dose of MTX performed somewhat poorer than those taking a low dose of MTX, suggesting some subtle and discrete cognitive change linked to medication.

The second point is in how to accurately assess cognitive function in RA. Studies using the same assessment tests ([1, 3]), but different dysfunction criteria (1 SD below the norm on 4 of 16 versus 5 of 16 tests), showed highly dissimilar cognitive impairment rates (31% versus 10%). Studies such as the one by Hanley and colleagues ([4]) that used other assessment tools (i.e., computerized tests) and dysfunction criteria (1.5 SDs below the norm on 4 of 10 tests with different adjustment for sensorimotor speed) reported even lower rates of cognitive impairment (9%). Known predictors of cognitive function, such as age and education level, certain symptoms (i.e., fatigue), and comorbidities (i.e., cardiovascular disease) need to be stringently controlled for before any direct effect may be attributed to RA.

In summary, we suggest a degree of reservation when interpreting results pertaining to cognitive dysfunction in RA until we identify RA-specific influences and reach a consensus on the cognitive assessment criteria for this population.