Morgan and colleagues [1] noted an association between elevations in the mean corpuscular volume and the development of toxicity to the folic acid antagonist methotrexate in patients with rheumatoid arthritis. Although such a correlation has been independently reported [2], there are potential pitfalls in relying on elevations in the mean corpuscular volume to indicate incipient methotrexate toxicity. Even in conditions typically characterized by macrocytosis, the mean corpuscular volume may be within the normal range in the presence of intercurrent conditions, such as iron deficiency, which predispose patients to microcytosis. Although determination of vitamin levels would be informative, a more cost-effective approach may be available.

Price-Jones [3] previously noted not only that the erythrocyte diameter increased in patients with pernicious anemia but also that the coefficient of variation in the erythrocyte size was twice that of normal erythrocytes. Currently, information on variation in erythrocyte volume is routinely provided in the automated complete blood count as the red cell distribution width (RDW). Combined analysis of the RDW (a measure of anisocytosis) and the mean corpuscular volume provide a useful approach to the differential diagnosis of anemia [4]. Moreover, elevations in the RDW often precede changes in the mean corpuscular volume or hemoglobin level [5]. An elevated RDW may be the only abnormality evident during the early stages of a nutritional deficiency and may be the most reliable indicator of mixed anemia. Thus, an elevated RDW might be a useful, early marker of the potential for impending methotrexate toxicity. Because the RDW is part of the routine complete blood count, such an approach would be cost-effective. We therefore suggest that the RDW be serially monitored in patients with rheumatoid arthritis who are receiving methotrexate. Elevations in the RDW may necessitate a change in the methotrexate or folate doses used, along with heightened vigilance for methotrexate toxicity.