OBJECTIVE To present and discuss three cases of apparent reactivation of Graves' orbitopathy (GO) after orbital decompression and to evaluate the incidence of this phenomenon. METHODS Observational case series and retrospective follow-up study. METHODS A few weeks after surgery 2 patients with GO (patients 1 and 2), treated at our institution with rehabilitative bony orbital decompression during the static phase of the disease showed clinical and radiologic evidence of reactivated orbitopathy. After this observation, a sample of 249 patients who had consecutively undergone the same treatment for the same reason before the second of the 2 observed patients was selected for this study. METHODS The records of the selected patients were retrospectively reviewed searching for cases presenting with clinical and radiologic evidence of GO reactivated as a consequence of any type of bony orbital decompression. Patients treated with perioperative systemic glucocorticoids or who had concurrent periorbital diseases, injuries, or surgeries, or who had immunocompromised conditions or a follow-up of < or =2 months, were excluded. METHODS Incidence of reactivation. Clinical history, clinical and radiologic characteristics, treatment modalities, and time course of the reactivation in patients presenting with this phenomenon. RESULTS Decompression surgery took place between 1994 and 2000. Eleven patients were excluded for having been treated with perioperative glucocorticoids. Only 1 patient (patient 3) presented with reactivation. The incidence of the phenomenon that we regard as reactivation of GO after rehabilitative bony orbital decompression was therefore 1.3% (3/239). In all 3 patients, the reactivation took place a few weeks after surgery, after an early normal convalescence period and could be controlled with systemic immunosuppression or orbital radiotherapy. None of the patients we report developed further episodes of reactivation during the follow-up period (mean, 7.5 years). CONCLUSIONS Based on its clinical characteristics, we suggest naming our observation delayed decompression-related reactivation and we propose using its acronym DDRR when referring to it. Although DDRR appears to be a rare event, it is important for physicians and patients to be aware of its possible occurrence with rehabilitative decompression surgery.