Current information on Waldenström macroglobulinemia (WM) is based on retrospective or single-institution studies of patients requiring therapy. Between 1992 and 1998, 231 patients with WM were enrolled in a prospective observational multicenter clinical trial. Of these, 182 patients with symptomatic or progressive disease were treated with 4 to 8 cycles of therapy with a purine nucleoside analogue, fludarabine (FAMP; 30 mg/m(2) of body-surface area daily for 5 days every 28 days). A serum beta2-microglobulin (beta2M) level below 3 mg/L and a hemoglobin level of at least 120 g/L (12 g/dL) at presentation predicted a lower likelihood of requiring therapy. The overall rate of response to FAMP therapy was 36% (95% confidence interval, 29%-44%), with 2% complete remissions. Patients who were 70 years old or older had a substantially lower likelihood of response (odds ratio, 0.34; P =.004) than younger patients. On multivariate analysis, a serum beta2M level of 3 mg/L or higher, hemoglobin level below 120 g/L, and serum IgM level below 40 g/L [4 g/dL] were significant adverse prognostic factors for survival. We developed a simple staging system for WM by using these variables and identified 4 distinct subsets of patients with estimated 5-year overall survival rates of 87%, 64%, 53%, and 22%, and 5-year progression-free survival rates of 83%, 55%, 33%, and 12%. Prognosis in WM is highly variable and serum beta2M was the dominant predictor of a need for therapy and of survival. FAMP has activity against WM. Our staging system may provide guidance for a risk-based approach to the treatment of WM.