Abstract Background The aim of this study was to compare long-term mortality, morbidity, and cumulative healthcare costs between antithyroid drugs, radioactive iodine, and surgical treatment for patients with persistent or relapsed Graves’ disease. Methods Data on patients with persistent or relapsed Graves’ disease between 2006 and 2018 were retrieved from the Hong Kong Hospital Authority. Hazard ratios (HRs) estimated by Cox proportional hazards regression models were used to compare the risks of all-cause mortality, cardiovascular disease, atrial fibrillation, psychological disease, Graves’ ophthalmopathy, and cancer across treatment groups. The 10-year healthcare cost and change in co-morbidity status were also estimated. Results Over a median follow-up of 79 months (22 636 person-years), a total of 3443 patients (antithyroid drug 2294, radioactive iodine 755, surgery 394) were analysed. Compared with antithyroid drug treatment, surgery was associated with significantly lower risks of all-cause mortality (HR 0.40, 95 per cent c.i. 0.36 to 0.45), cardiovascular disease (HR 0.54, 0.48 to 0.60), atrial fibrillation (HR 0.11, 0.09 to 0.14), psychological disease (HR 0.85, 0.79 to 0.92), Graves’ ophthalmopathy (HR 0.09, 0.08 to 0.10), and cancer (HR 0.56, 0.50 to 0.63). Patients who underwent surgery also had a lower risk of all outcome events than those in the radioactive iodine group. The 10-year direct cumulative healthcare cost was €14 754 for surgery compared with €17 390 for antithyroid drugs, and €17 918 for the radioactive iodine group. Conclusion Patients who underwent surgery for persistent or relapsed Graves’ disease had lower risks of all-cause mortality and analysed morbidities. The 10-year cumulative healthcare cost in the surgery group was lowest among the three treatment alternatives.