Grass pollen immunotherapy is the only treatment for hayfever that is both effective and confers long-term benefit. Immunotherapy may act by altering the local nasal mucosal T helper type 2 (Th2) to type 1 (Th1) cytokine balance either by down-regulation and/or immune deviation of T-lymphocyte responses. There is controversy as to whether these changes are detectable in peripheral blood. We therefore examined both local nasal and peripheral T-cell responses to allergen exposure in the same subjects before and after immunotherapy. In a double-blind trial of grass pollen immunotherapy, nasal biopsies were obtained at baseline and during the peak pollen season following 2 years of immunotherapy. Placebo-treated patients showed a seasonal increase in CD3+ T cells (P = 0·02) and in interleukin-5 (IL-5) mRNA+ cells (P = 0·03) and no change in interferon-γ (IFN-γ ) mRNA+ cells (P = 0·2) in the nasal mucosa. In contrast, in the immunotherapy-treated group, there were no changes in the number of CD3+ T cells (P = 0·3) and IL-5 mRNA+ cells (P = 0·2) but a significant increase in the number of IFN-γ mRNA+ cells (P = 0·03). Furthermore, clinical improvement in the immunotherapy-treated group was accompanied by a seasonal increase in the ratio of IFN-γ to IL-5 mRNA+ cells in the nasal mucosa (P = 0·03). In contrast, there were no significant changes in peripheral T-cell proliferative responses or cytokine production for IFN-γ or IL-5 in response to grass pollen either within or between the two treatment groups. We conclude that successful grass pollen immunotherapy was associated with an increase in the ratio of IFN-γ to IL-5 mRNA+ cells in the nasal mucosa, whereas these changes were not reflected by alterations in peripheral blood T-cell proliferative responses or cytokine production before/after treatment.