To longitudinally assess serum concentrations of rituximab, it was administered intravenously to 25 children with opsoclonus-myoclonus syndrome at 375 mg/m2 on each of 4 consecutive weeks with (Group I and II) or without (Group III) conventional immunotherapy. Serum rituximab levels, drawn before and after each infusion and at later intervals, were analyzed by enzyme-linked immunosorbent assay. Rituximab concentration increased stepwise with each infusion, dropping by the next infusion, thereby forming 4 discrete peaks (Cmax) and troughs (Cmin). It then fell precipitously to trace levels at 4 months. However, Cmax and Cmin curves differed significantly between groups. Compared with the youngest children (Group I), the oldest (Group III) had a 34% lower rituximab concentration at the fourth infusion, 45% less IgM depletion 1 month later, and received 20% less rituximab when the dose was recalculated as mg/kg. Serum IgM and rituximab levels were negatively correlated. Peak rituximab concentration did not correlate with adrenocorticotropic hormone dose. These results indicate that the degree of serum IgM depletion is a useful indicator for rituximab dose equivalency in children of different ages. They also suggest that pediatric rituximab dosing should be based on body weight, not surface area. (ClinicalTrials.gov NCT00244361).