Background: Standard treatment for thyroid eye disease is with systemic corticosteroids. We aimed to establish whether orbital radiotherapy or antiproliferative immunosuppression would confer any additional benefit. Methods: CIRTED was a multicentre, double-blind, randomised controlled trial with a 2 × 2 factorial design done at six centres in the UK. Adults with active moderate-to-severe thyroid eye disease associated with proptosis or ocular motility restriction were recruited to the trial. Patients all received a 24 week course of oral prednisolone (80 mg per day, reduced to 20 mg per day by 6 weeks, 10 mg per day by 15 weeks, and 5 mg per day by 21 weeks) and were randomly assigned via remote computerised randomisation to receive either radiotherapy or sham radiotherapy and azathioprine or placebo in a 2 × 2 factorial design. Randomisation included minimisation to reduce baseline disparities in potential confounding variables between trial interventions. Patients and data analysts were masked to assignment, whereas trial coordinators (who monitored blood results), pharmacists, and radiographers were not. The radiotherapy dose was 20 Gy administered to the retrobulbar orbit in ten to 12 fractions over 2 to 3 weeks. Azathioprine treatment was provided for 48 weeks at 100–200 mg per day (dispensed as 50 mg tablets), depending on bodyweight (100 mg for <50 kg, 150 mg 50–79 kg, 200 mg for ≥80 kg). The primary outcomes were a binary composite clinical outcome score and an ophthalmopathy index at 48 weeks, and a clinical activity score at 12 weeks. The primary analysis was based on the intention-to-treat allocation and safety was assessed in all participants. This study is registered with ISRCTN, number 22471573. Findings: Between Feb 15, 2006, and Oct 3, 2013, 126 patients were recruited and randomly assigned to groups: 31 patients to radiotherapy plus azathioprine, 31 to sham radiotherapy and azathioprine, 32 to radiotherapy and placebo, and 32 to sham radiotherapy and placebo. Outcome data were available for 103 patients (54 for sham radiotherapy vs 49 for radiotherapy and 53 for placebo vs 50 for azathioprine), of whom 84 completed their allocated treatment of radiotherapy or sham radiotherapy and 57 continued to take azathioprine or placebo up to 48 weeks. There was no interaction betweeen azathioprine and radiotherapy (pinteraction=0·86). The adjusted odds ratio (ORadj) for improvement in the binary clinical composite outcome measure was 2·56 (95% CI 0·98–6·66, p=0·054) for azathioprine and 0·89 (0·36–2·23, p=0·80) for radiotherapy. In a post-hoc analysis of patients who completed their allocated therapy the ORadj for improvement was 6·83 (1·66–28·1, p=0·008) for azathioprine and 1·32 (0·30–4·84, p=0·67) for radiotherapy. The ophthalmopathy index, clinical activity score, and numbers of adverse events (161 with azathioprine and 156 with radiotherapy) did not differ between treatment groups. In both groups, the most common adverse events were mild infections. No patients died during the study. Interpretation: In patients receiving oral prednisolone for 24 weeks, radiotherapy did not have added benefit. We also did not find added benefit for addition of azathioprine in the primary analysis; however, our conclusions are limited by the high number of patients who withdrew from treatment. Results of post-hoc analysis of those who completed the assigned treatment suggest improved clinical outcome at 48 weeks with azathioprine treatment. Funding: National Eye Research Centre, Above and Beyond, and Moorfields Eye Charity.