Estrogens are important immunomodulators, exerting significant effects on cell proliferation, apoptosis, cytokine production and differentiation of hematopoietic cells. Estrogen receptors are expressed on normal B and T lymphocytes, bone marrow and in leukemia and lymphoma cell lines. Epidemiologic evidence for the association of menopausal hormone use with risk of non-Hodgkin's lymphoma (NHL) has been mixed; however, all of the investigations have been observational. We analyzed the data from Women's Health Initiative hormone therapy trials where conjugated equine estrogens (CEE; 0.625 mg/d) plus medroxyprogesterone acetate (MPA; 2.5 mg/d) (n = 16,654) or CEE alone (women with prior hysterectomy) (n = 10,685) were tested against placebos and the intervention lasted a median of 5.6 years in the CEE + MPA trial and 7.2 years in the CEE alone trial. During 13 years of follow-up through September 20, 2013 383 incident NHL cases were identified. We used the intent-to-treat approach to calculate incidence rates of NHL, hazards ratios (HR) and 95% confidence intervals (CI) by treatment group. Incidence of NHL was virtually the same in the treatment and placebo groups. The HR was 1.02 (95%CI 0.74-1.39) for CEE alone, 0.98 (95% CI 0.76-1.28) for CEE+MPA, and 1.00 (95% CI 0.82-1.22) for both combined. There were no specific NHL subtypes associated with either type of the treatment, except a marginally decreased risk of plasma cell neoplasms (HR= 0.53 95% CI 0.27-1.03) in the CEE-alone group. These results do not support a role of estrogen alone or combined with progestin in the development of NHL among postmenopausal women.