MOG-DISCONTINUE
(2026)Objective
To investigate relapse outcomes and identify optimal maintenance immunomodulatory treatment (IT) durations following treatment discontinuation in patients with myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD).
Study Summary
• Relapsing disease course at discontinuation nearly doubled relapse risk (HR 1.95, 95% CI 1.25-3.06, P=0.003) in multivariable analysis.
• Positive/low-positive pre-discontinuation MOG IgG1 was more common in relapsers (89.8% vs 71.5%, P=0.005).
• Optimal IT duration estimated at 10-18 months for monophasic patients and 20-30 months for relapsing patients.
• 15-month treatment predicted 5-year relapse-free survival of 87.7% vs 47.9% for 3-month treatment (monophasic cohort).
Intervention
Retrospective observational analysis of maintenance immunomodulatory treatment (oral corticosteroids, azathioprine, mycophenolate, methotrexate, IVIG, or monoclonal antibody therapies) discontinuation in MOGAD patients.
Inclusion Criteria
Patients fulfilling 2023 MOGAD criteria seen at Oxford NMO Highly Specialised Service (Jan 2010-May 2025), ≥12 months follow-up, and who commenced and then discontinued eligible maintenance IT.
Study Design
Arms: Monophasic course at discontinuation (n=150 intervals) vs Relapsing course at discontinuation (n=86 intervals)
Patients per Arm: 150 monophasic vs 86 relapsing IT intervals (190 unique patients, 236 total intervals)
Outcome
• Relapsing course at discontinuation: HR 1.95 (95% CI 1.25-3.06, P=0.003) for relapse.
• Longer IT duration beyond 3 months progressively reduced relapse hazard.
• Monophasic patients: 10-18 months optimal IT duration (67.9-87.0% relapse risk reduction vs 3 months).
• Relapsing patients: 20-30 months optimal IT duration (63-68.4% relapse risk reduction vs 3 months).
• Negative MOG IgG1 associated with lower relapse in sensitivity analysis (HR 0.34, 95% CI 0.14-0.87, P=0.024).
Bottom Line
In MOGAD patients, maintenance immunomodulatory treatment should be continued for approximately 10-18 months after a single onset attack and 20-30 months in relapsing patients to substantially reduce post-discontinuation relapse risk, with a relapsing disease course at discontinuation nearly doubling relapse hazard.
Major Points
- Largest cohort study to date examining IT discontinuation outcomes in MOGAD (190 patients, 236 discontinued intervals).
- 39.0% of discontinuations were followed by relapse at median 5.4 months.
- Relapsing disease course at discontinuation was the strongest independent predictor of subsequent relapse (HR 1.95, P=0.003).
- Treatment durations beyond 3 months progressively reduced relapse hazard, with plateau at 10-18 months (monophasic) and 20-30 months (relapsing).
- 5-year relapse-free survival: 87.7% for 15-month treatment vs 47.9% for 3-month treatment in monophasic patients.
- Negative MOG IgG1 status prior to discontinuation was protective in sensitivity analysis (HR 0.34, P=0.024).
- Timing of onset attack acute treatment (early vs intermediate vs late) did not affect subsequent relapse risk.
- Cohort was predominantly corticosteroid-treated (84.7% of IT intervals used only corticosteroids).
Study Design
- Study Type
- Retrospective single-service observational cohort study with prospectively collected data
- Randomization
- No
- Blinding
- Not applicable (observational); relapses adjudicated by expert NMOSD consultants
- Sample Size
- 190
- Follow-up
- Median disease duration 4.9 years (IQR 2.5-8.2) at last follow-up; minimum ≥12 months
- Centers
- 1
- Countries
- United Kingdom
Primary Outcome
Definition: Time-to-relapse following IT discontinuation (relapse defined as new clinical attack >30 days after prior attack)
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| Monophasic course at discontinuation: 27.8% relapsed | Relapsing course at discontinuation: 50% relapsed | 1.95 (1.25-3.06) | 0.003 |
Limitations & Criticisms
- Retrospective single-center observational design limits causal inference
- Cohort predominantly treated with corticosteroids (84.7%), limiting generalizability to steroid-sparing agents
- 16.9% of IT intervals had missing pre-discontinuation MOG IgG1 status (addressed by multiple imputation)
- Optimal duration estimates derived from spline modeling rather than randomized comparison
- Wide confidence intervals in survival estimates at longer treatment durations reflect reduced precision
- Possible enrichment of relapsing patients in referral-based specialist cohort (partially mitigated by incident cohort analysis)
Citation
Yeh WZ, et al. Optimal strategies for treatment discontinuation in MOG antibody-associated disease. Brain. 2026. doi:10.1093/brain/awag006