← Back
NeuroTrials.ai
Neurology Clinical Trial Database

GEMINI 1 and GEMINI 2

Tolebrutinib versus Teriflunomide in Relapsing Multiple Sclerosis

Year of Publication: 2025

Authors: Jiwon Oh, Douglas L. Arnold, Bruce A.C. Cree, ..., Heinz Wiendl

Journal: New England Journal of Medicine

Citation: N Engl J Med 2025; DOI: 10.1056/NEJMoa2415985


Clinical Question

Is tolebrutinib, a brain-penetrant BTK inhibitor, superior to teriflunomide in reducing annualized relapse rates and preventing disability worsening in patients with relapsing multiple sclerosis?

Bottom Line

Tolebrutinib was not superior to teriflunomide in reducing annualized relapse rates in patients with relapsing multiple sclerosis. While disability worsening appeared numerically lower with tolebrutinib, formal statistical testing could not be performed due to the failed primary endpoint.

Major Points

  • Two identical phase 3, double-blind, double-dummy, event-driven trials (GEMINI 1: n=974, GEMINI 2: n=899)
  • Primary endpoint failed: annualized relapse rates were similar between groups (rate ratio 1.06 in GEMINI 1, 1.00 in GEMINI 2)
  • Disability worsening ≥6 months numerically favored tolebrutinib (8.3% vs 11.3%, HR 0.71, 95% CI 0.53-0.95) but no formal testing
  • Tolebrutinib is brain-penetrant and targets CNS-resident microglia and B cells, unlike peripherally-acting therapies
  • Safety profiles were generally similar, though tolebrutinib had higher rates of minor bleeding (petechiae 4.5% vs 0.3%)
  • Liver enzyme elevations occurred in both groups (5.6% tolebrutinib vs 6.3% teriflunomide with ALT >3x ULN)

Design

Study Type: Two multicenter, double-blind, double-dummy, event-driven, active-controlled, randomized phase 3 trials

Randomization: 1

Blinding: Double-blind, double-dummy design with matching placebos

Enrollment Period: June 25, 2020 to August 8, 2022

Follow-up Duration: Median 139 weeks (event-driven design)

Centers: GEMINI 1: 162 sites in 24 countries, GEMINI 2: 155 sites in 25 countries

Countries: 24-25 countries including US, Europe, and others

Sample Size: 1873

Analysis: Intention-to-treat analysis using negative binomial regression for primary endpoint, Cox proportional hazards for time-to-event analyses, hierarchical testing procedure


Inclusion Criteria

  • Age 18-55 years
  • Relapsing multiple sclerosis per 2017 McDonald criteria
  • EDSS score ≤5.5
  • At least one relapse within previous year, OR at least two relapses within previous 2 years, OR at least one gadolinium-enhancing brain lesion on T1-weighted MRI within previous year
  • Written informed consent

Exclusion Criteria

  • Primary progressive multiple sclerosis per 2017 McDonald criteria
  • Nonrelapsing secondary progressive multiple sclerosis
  • Other exclusion criteria detailed in protocol and Supplementary Appendix

Arms

FieldTolebrutinibControl
InterventionOral tolebrutinib 60 mg once daily with meal plus matching placebo for teriflunomideOral teriflunomide 14 mg once daily with meal plus matching placebo for tolebrutinib
DurationEvent-driven design, median 139 weeksEvent-driven design, median 139 weeks

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Annualized relapse rate (confirmed relapses per participant-year)PrimaryGEMINI 1: 0.12, GEMINI 2: 0.11GEMINI 1: 0.13, GEMINI 2: 0.11P=0.67 GEMINI 1, P=0.98 GEMINI 2
Confirmed disability worsening ≥6 months (pooled)Secondary11.3% (106 events)8.3% (78 events)HR 0.7195% CI 0.53-0.95 (no formal testing due to hierarchical procedure)
Confirmed disability worsening ≥3 months (pooled)Secondary15.3% (144 events)11.7% (109 events)HR 0.7395% CI 0.57-0.94
Confirmed disability improvement ≥6 months (pooled)Secondary10.4% (98 events)12.6% (118 events)HR 1.2295% CI 0.94-1.60
New gadolinium-enhancing lesionsSecondaryGEMINI 1: 0.29, GEMINI 2: 0.22GEMINI 1: 0.53, GEMINI 2: 0.46Rate ratio 1.86 (1.36-2.55) GEMINI 1, 2.12 (1.50-2.99) GEMINI 2Favored teriflunomide
Any adverse eventAdverse86.3% (810/939)84.9% (792/933)Not specified
Serious adverse eventsAdverse8.2% (77/939)9.8% (91/933)Not specified
PetechiaeAdverse0.3% (3/939)4.5% (42/933)Not specified
Heavy mensesAdverse1.0% (9/939)2.6% (24/933)Not specified

Subgroup Analysis

Results were consistent across both individual trials. Hierarchical testing procedure prevented formal statistical testing of secondary endpoints due to nonsignificant primary endpoint results.


Criticisms

  • Primary endpoint was not met, limiting conclusions about efficacy
  • Hierarchical testing procedure prevented formal statistical testing of potentially important disability outcomes
  • No direct comparison with other modern high-efficacy DMTs like anti-CD20 therapies
  • Event-driven design may have been underpowered for the observed low relapse rates in both groups
  • Tolebrutinib showed worse performance on MRI inflammatory markers compared to teriflunomide
  • Study population was relatively young (18-55 years) and may not represent full MS population
  • 84-86% completion rates suggest some loss to follow-up that could affect results

Funding

Supported by Sanofi

Based on: GEMINI 1 and GEMINI 2 (New England Journal of Medicine, 2025)

Authors: Jiwon Oh, Douglas L. Arnold, Bruce A.C. Cree, ..., Heinz Wiendl

Citation: N Engl J Med 2025; DOI: 10.1056/NEJMoa2415985

Content summarized and formatted by NeuroTrials.ai.