Twenty patients with cyclical breast pain were enrolled in a double-blind cross-over trial in which either a soy protein drink or a flavoured cow's milk was taken orally each day for 3 months before crossing over to the alternate drink for a further 3 months. Records of pain scores were taken throughout the study. Blood was also taken before and after 3 and 6 months for the measurement of phytoestrogents to assess compliance. Two women withdrew from the study at the outset leaving 18 evaluable patients who completed the study. Of these 10 (56%) felt that soy protein improved breast pain (two of whom received soy as first treatment) and two (11%) felt that cow's milk alleviated symptoms (one receiving this as first preparation) and the remaining six (33%) experienced no relief of pain with either dietary preparation. Blood levels of diadzein and genistein were elevated after the ingestion of soy protein in only 13 patients (seven of whom felt that soy improved their breast pain); in the remaining five patients (three of whom suggested that soy protein improved breast pain) phytoestrogen levels were no higher than pretreatment values. Although the ingestion of soy protein may be associated with relief of breast pain, these results illustrate the problem of non-specific effects in studies of mastalgia in that 1) cow's milk also relieved breast pain in some patients and 2) that the benefits of soy protein were not always associated with evidence of elevated circulating levels of phyto-estrogens, indicating the difficulty of compliance in dietary intervention studies using soy foods.