The pathogenesis of nephrotic syndrome is unclear. However, the efficacy of rituximab, a B cell-depleting antibody, in nephrotic syndrome suggests a pathogenic role of B cells. In this retrospective study, we determined by flow cytometry levels of B and T cell subpopulations before and after rituximab infusion in 28 pediatric patients with frequently relapsing or steroid-dependent nephrotic syndrome. At baseline, patients had lower median percentages of transitional and mature B cells than age-matched healthy controls (P,0.001). Rituximab induced full depletion of B cells (,1% of lymphocytes). At 1 year, most patients exhibited complete total and mature B cell recovery, whereas memory B cell subsets remained significantly depleted. Total T cell concentration did not change with rituximab, whereas the CD4 + /CD8 + T cell ratio tended to increase. Fourteen patients relapsed within 24 months, with a median follow-up of 11.2 months (interquartile range, 8-17.7 months). We observed no difference at baseline between nonrelapsing and relapsing patients in several clinical parameters and cell subset concentrations. Reconstitution of all memory B cell subpopulations, number of immunosuppressive drugs, and dose of tacrolimus during the last 4 months of follow-up were predictive of relapse in univariate Cox regression analysis. However, only delayed reconstitution of switched memory B cells, independent of immunosuppressive treatment, was protective against relapse in multivariate (P,0.01) and receiver operator characteristic (P,0.01 for percentage of lymphocytes; P=0.02 for absolute count) analyses. Evaluation of switched memory B cell recovery after rituximab may be useful for predicting relapse in patients with nephrotic syndrome.