Background: Relapses of multiple sclerosis decrease during pregnancy, when the hormone estriol is increased. Estriol treatment is anti-inflammatory and neuroprotective in preclinical studies. In a small single-arm study of people with multiple sclerosis estriol reduced gadolinium-enhancing lesions and was favourably immunomodulatory. We assessed whether estriol treatment reduces multiple sclerosis relapses in women. Methods: We did a randomised, double-blind, placebo-controlled phase 2 trial at 16 academic neurology centres in the USA, between June 28, 2007, and Jan 9, 2014. Women aged 18-50 years with relapsing-remitting multiple sclerosis were randomly assigned (1:1) with a random permuted block design to either daily oral estriol (8 mg) or placebo, each in combination with injectable glatiramer acetate 20 mg daily. Patients and all study personnel, except for pharmacists and statisticians, were masked to treatment assignment. The primary endpoint was annualised relapse rate after 24 months, with a significance level of p=0.10. Relapses were confirmed by an increase in Expanded Disability Status Scale score assessed by an independent physician. Analysis was by intention to treat. The trial is registered with ClinicalTrials.gov, number NCT00451204. Findings: We enrolled 164 patients: 83 were allocated to the estriol group and 81 were allocated to the placebo group. The annualised confirmed relapse rate was 0.25 relapses per year (95% CI 0.17-0.37) in the estriol group versus 0.37 relapses per year (0.25-0.53) in the placebo group (adjusted rate ratio 0.63, 95% CI 0.37-1.05; p=0.077). The proportion of patients with serious adverse events did not differ substantially between the estriol group and the placebo group (eight [10%] of 82 patients vs ten [13%] of 76 patients). Irregular menses were more common in the estriol group than in the placebo group (19 [23%] vs three [4%], p=0.0005), but vaginal infections were less common (one [1%] vs eight [11%], p=0.0117). There were no differences in breast fibrocystic disease, uterine fibroids, or endometrial lining thickness as assessed by clinical examination, mammogram, uterine ultrasound, or endometrial lining biopsy. Interpretation: Estriol plus glatiramer acetate met our criteria for reducing relapse rates, and treatment was well tolerated over 24 months. These results warrant further investigation in a phase 3 trial. Funding: National Institutes of Health, National Multiple Sclerosis Society, Conrad N Hilton Foundation, Jack H Skirball Foundation, Sherak Family Foundation, and the California Community Foundation.