Up to 50 % of women receiving aromatase inhibitor (AI) complain of AI-associated musculoskeletal symptoms (AIMSS) and 15 % discontinue treatment. We conducted a randomized, sham-controlled trial to evaluate whether acupuncture improves AIMSS and to explore potential mechanisms. Postmenopausal women with early stage breast cancer, experiencing AIMSS were randomized to eight weekly real or sham acupuncture sessions. We evaluated changes in the Health Assessment Questionnaire Disability Index (HAQ-DI) and pain visual analog scale (VAS) following the intervention compared to baseline. Serum estradiol, β-endorphin, and proinflammatory cytokine concentrations were measured pre and post-intervention. We enrolled 51 women of whom 47 were evaluable, including 23 randomized to real and 24 to sham acupuncture. Baseline characteristics were balanced between groups with the exception of a higher HAQ-DI score in the real acupuncture group (p = 0.047). We did not observe a statistically significant difference in reduction of HAQ-DI (p = 0.30) or VAS (p = 0.31) between the two groups. Following eight weekly treatments, we observed a statistically significant reduction of IL-17 (p ≤ 0.009) in both groups. No significant modulation was seen in estradiol, β-endorphin, or other proinflammatory cytokine concentrations in either group. We did not observe a significant difference in AIMSS changes between real and sham acupuncture. As sham acupuncture used in this study may not be equivalent to placebo, further studies with a non-acupuncture arm may be required to establish whether acupuncture is beneficial for the treatment of AIMSS.