To the Editors:

We thank Drs. Costedoat-Chalumeau et al for their interest in our report on the validation of the MASRI when used in SLE. This questionnaire was validated to measure adherence to SLE medication in a clinical setting, using HCQ and prednisone as examples. These 2 medications were chosen for the validation process because many patients with SLE are treated with either or both of these medications. Our results suggest that the MASRI is a useful clinical measure of adherence to SLE medications using patient self-report.

Recently, advances have been made in relating exposure to rheumatic medications (rather than the daily dose of the medications), their pharmacodynamics, and even pharmacogenetics to medication efficacy and side effects. Such medications include, but are not limited to, rituximab (1, 2), HCQ (3, 4), mycophenolate mofetil (MMF) (5), and prednisone (6). Currently, however, exposure to rheumatic medications is rarely measured in clinical practice in North America for the purpose of tailoring medication regimens. The medication levels more commonly ordered by rheumatologists for patients with SLE may be those of MMF, for which, similar to what has been suggested by Costedoat-Chalumeau's group for HCQ, extremely low or undetectable levels can be considered a proxy measure of patient nonadherence.

Careful interpretation of such values must occur, however, especially in the cases of MMF and HCQ. With both medications, there can be remarkable between-patient and even within-patient variation in medication exposure. For example, transplant patients show such profound within-patient and between-patient variation in MMF exposure that patient-tailored MMF medication monitoring has been pursued in an attempt to improve graft survival (7). Similar to previous studies of patients with rheumatoid arthritis (8), the results of studies by Costedoat-Chalumeau et al support that patients with SLE who have low whole-blood HCQ levels have inferior disease control compared with those who achieve levels over 1,000 ng/ml, and that patients with SLE who have (almost) undetectable HCQ serum levels are very likely nonadherent (4, 9). However, given the large interpatient variation in HCQ pharmacokinetics (coefficient of variation 38–46%) (8, 10, 11), the interpretation of detectable HCQ levels remains difficult; lower-than-average HCQ levels can either be indicative of problems with adherence or the result of the large interpatient variation in HCQ pharmacokinetics. What might be helpful in distinguishing between both possibilities in this setting are serial HCQ whole-blood levels. Indeed, many large commercial laboratories in North America offer whole-blood HCQ measurement for about $150 per test.

As suggested by us and supported by Costedoat-Chalumeau et al (3), physicians are not well-versed in identifying medication nonadherence. Thus, more objective measures of patient adherence are needed for use in clinical practice. We consider the MASRI a useful tool to screen for SLE patient adherence to HCQ and other SLE medications. In the specific case of HCQ, however, and under the caveats mentioned above, HCQ whole-blood testing appears to be a valuable, although somewhat expensive, tool for screening for HCQ nonadherence.

Hermine Brunner MD, MSc*, Lena Das MD*, Murray Passo MD*, Sri Koneru MD†, Anne-Barbara Mongey MD‡, * Cincinnati Children's Hospital Medical Center, † Group-Health Associates, ‡ University of Cincinnati College of Medicine, Cincinnati, OH.