It is unclear whether running can affect iron stores. Results using the serum ferritin assay (SER FER) have been conflicting. Decreased red cell ferritin (RBC FER) values (< or = 4 ag/RBC) occur in iron depleted or inflammatory states. We compared the longitudinal changes of hemoglobin (Hb), SER FER, RBC FER, % saturation of total iron binding capacity (% sat TIBC), and daily dietary intake in 27 runners during a training program. These parameters were measured at days 0, 49 (range 48-52), and 115 (range 85-120). No significant changes occurred in the SER FER, % sat TIBC and Hb determinations throughout the study. Overall the RBC FER values trended down (mean values 11.7 ag/RBC to 7.7 ag/RBC; p = 0.06). Fifteen runners (56%) acquired RBC FER values in the iron deficient range (mean 6.8 ag/RBC to 2.4 ag/RBC; p < 0.05). These values differed significantly from the remaining 12 runners (mean 17.3 ag/RBC to 14.7 ag/RBC). The decline in RBC FER into the iron deficient range was primarily seen in a subset of runners who began with a RBC FER value < or = 10 ag/RBC (positive predictive value 0.79) and was independent of iron intake. We conclude that ferritin can be affected by running as recognized by the red cell ferritin assay. Moreover our results suggest that this decrease in red cell ferritin is likely a function of defective iron utilization rather than total body iron deficiency. A potential consideration is that this fall may occur as a result of repetitive running-associated injury and inflammation.