Aromatase inhibitors are more effective than is tamoxifen in prevention of breast-cancer recurrence, but at the expense of increased musculoskeletal side-effects, such as carpal tunnel syndrome. The aim of this study was to assess risk factors and the prognostic value of musculoskeletal symptoms during treatment with the steroidal aromatase inhibitor exemestane or with tamoxifen after 2-3 years of tamoxifen. In the Intergroup Exemestane Study, postmenopausal women treated for early invasive breast cancer who remained disease free and on treatment after 2-3 years of tamoxifen were randomised to switch to exemestane or to continue tamoxifen for the remainder of the 5-year period of endocrine treatment. The primary endpoint for this retrospective analysis was occurrence of carpal tunnel syndrome and any musculoskeletal events, analysed in the safety population, which consisted of all patients who had received any trial treatment. As well as case-report forms, questionnaires were distributed retrospectively to gain more details of cases of carpal tunnel syndrome. The relation between musculoskeletal symptoms reported by 6 months from randomisation and survival from 9 months onwards was assessed by Cox proportional hazards models. The trial is registered, number ISRCTN11883920. It has completed accrual and follow-up is continuing for enrolled participants. After a median follow-up of 91·0 months (IQR 83·0-99·2), carpal tunnel syndrome had been reported for 66 (2·8%) of 2319 patients in the exemestane group compared with 13 (0·6%) of 2338 in the tamoxifen group (odds ratio [OR] 5·23, 99% CI 2·39-11·49; p<0·0001). More events occurred during treatment in the exemestane group than in the tamoxifen group (66 [2·8%] vs seven [0·3%], adjusted OR 9·90, 99% CI 3·52-27·82; p<0·0001). There was no significant difference between groups in events in the post-treatment period (ten with exemestane [0·4%] vs seven with tamoxifen [0·3%]; p=0·46). More patients in the exemestane group (1082 of 2319 patients, 46·7%) had musculoskeletal symptoms than in the tamoxifen group (901 of 2338, 38·5%; OR 1·48, 99% CI 1·32-1·67, p<0·0001). More events occurred during treatment in the exemestane group than in the tamoxifen group (984 [42·4%] vs 776 [33·2%], adjusted OR 1·59, 99% CI 1·32-1·91; p<0·0001), with this difference persisting to some extent in the post-treatment period (449 [19·4%] vs 390 [16·7%]; p=0·017). Of 73 on-treatment cases of carpal tunnel syndrome, 58 (79·5%) completed questionnaires were available. 27 patients (46·6%) had bilateral carpal tunnel syndrome and 31 (53·4%) had unilateral disease; 40 (69·0%) underwent surgical release. The disorder greatly affected daily-life activities in 21 (36·2%) cases. Occurrence of musculoskeletal symptoms, including carpal tunnel syndrome, was associated with improved disease-free survival in unadjusted analysis (p=0·023), but not with overall survival (p=0·36). However, after adjustment for possible confounding factors, musculoskeletal symptoms were not associated with disease-free survival (hazard ratio [HR] 0·96, 95% CI 0·82-1·14, p=0·67) or overall survival (HR 1·02, 95% CI 0·84-1·25, p=0·82). Occurrence of carpal tunnel syndrome is higher in patients with breast cancer given exemestane than in those treated with tamoxifen, and surgical release might be necessary in most cases. Development of musculoskeletal symptoms in the first 6 months of treatment is not an independent biomarker of improved disease outcome. Further investigation is warranted into the relation between treatment-emergent musculoskeletal symptoms and clinical outcome in patients with breast cancer receiving hormonal therapy. Pfizer.