The adjuvant use of aromatase inhibitors in breast cancer is associated with adverse effects on bone health. We previously reported a decline in bone mineral density (BMD) following the switch from tamoxifen to exemestane in the Intergroup Exemestane Study (IES). Here we report effects of endocrine treatment withdrawal on BMD, bone turnover markers (BTM) and fracture rates. 4,724 patients took part in IES, and 206 patients were included in a bone sub-study. BMD and BTM were assessed pre-randomization, during and after the end of treatment (EOT). To evaluate treatment withdrawal effects, 12- and 24-month post EOT BMD results are available for 122 and 126 patients, respectively. Similar patient numbers had BTM measured post EOT. Following treatment withdrawal, the differences in BMD observed between the two endocrine strategies were partially reversed. At 24 months from EOT, spine BMD increased by 1.53% (95%CI 0.63-2.43; p = 0.001) after stopping exemestane and fell by 1.93% (95%CI -2.91 to 0.95; p = 0.0002) following tamoxifen withdrawal. A similar pattern of changes was observed at the hip. At 2 years post EOT, BMD changes from baseline were similar with both treatment strategies. Corresponding inverse changes in BTM were seen, with an increase following tamoxifen withdrawal and a reduction after exemestane. A higher number of fractures occurred during exemestane treatment, but fracture rates were similar after treatment withdrawal. With the switch strategy used in IES, the on treatment adverse bone effects of exemestane are reversed. Ongoing monitoring of BMD is therefore not routinely required.