MOGAD-Discontinuation
Optimal strategies for treatment discontinuation in MOG antibody-associated disease
Bottom Line
Prolonged maintenance IT (≥10-18 months for monophasic disease, 20-30 months for relapsing disease) significantly reduces post-discontinuation relapse risk in MOGAD; relapsing course at time of discontinuation nearly doubles subsequent relapse hazard.
Major Points
- 39.0% of 236 IT discontinuation intervals were followed by relapse at a median 5.4 months (IQR 1.4-20.1).
- Relapsing course at time of discontinuation independently doubled relapse risk (adjusted HR 1.95, 95% CI 1.25-3.06, P=0.003).
- Longer pre-discontinuation IT duration monotonically lowered relapse hazard, plateauing around 10-15 months (monophasic) and 20 months (relapsing).
- Estimated 5-year relapse-free survival for monophasic patients: 87.7% at 15 months IT vs 47.9% at 3 months.
- Estimated 5-year relapse-free survival for relapsing patients: 62.1% at 24 months IT vs 24.3% at 3 months.
- Negative MOG IgG1 within 12 months prior to discontinuation was protective in univariable analysis (HR 0.34, 95% CI 0.14-0.87, P=0.024) but not after multivariable adjustment.
- Timing of acute treatment for onset attack (early <5d, intermediate 5-14d, late/none) was not associated with post-discontinuation relapse risk.
Design
Study Type: Retrospective single-centre observational cohort study
Randomization:
Blinding: None (retrospective observational)
Enrollment Period: January 2010 - May 2025 (data extracted 2 May 2025)
Follow-up Duration: Median disease duration 4.9 years (IQR 2.5-8.2); median post-discontinuation follow-up 3.1 years (IQR 1.3-6.3)
Centers: 1
Countries: UK
Sample Size: 190
Analysis: Univariable and multivariable Cox regression with robust standard errors clustered by patient ID; IT duration modelled with restricted cubic splines; multiple imputation (100 datasets) for missing data
Inclusion Criteria
- Fulfilled 2023 International MOGAD Panel criteria
- Seen in the Oxford Neuromyelitis Optica Highly Specialised Service between January 2010 and May 2025
- Consented to Demyelinating Research Tissue Bank participation
- ≥12 months follow-up
- Commenced and then discontinued eligible maintenance immunomodulatory treatment (oral corticosteroids, azathioprine, mycophenolate, methotrexate, IVIG or monoclonal antibody therapy)
Exclusion Criteria
- Did not meet 2023 MOGAD criteria
- <12 months of follow-up
- Never commenced or never discontinued maintenance IT
- Ineligible maintenance IT regimen
Arms
| Field | Monophasic at discontinuation | Relapsing at discontinuation |
|---|---|---|
| Intervention | Maintenance IT (mostly oral corticosteroids) then discontinuation after a single attack, before any relapse | Maintenance IT then discontinuation after ≥1 relapse (median 1 relapse pre-discontinuation, IQR 1-2) |
| Duration | Median IT 9.6 months (IQR 2.9-15.6) | Median IT duration variable; 97% relapse-free on treatment at time of discontinuation |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| Time-to-relapse following IT discontinuation (relapse = new clinical attack >30 days after prior attack) | Primary | Not applicable (single-cohort observational) | Relapse after 92/236 (39.0%) discontinuations at median 5.4 months (IQR 1.4-20.1) | 1.95 | 0.003 |
| IT duration 6 months vs 3 months (overall cohort, multivariable) | Secondary | 0.65 | <0.001 | ||
| IT duration 12 months vs 3 months (overall cohort, multivariable) | Secondary | 0.36 | <0.001 | ||
| IT duration 18 months vs 3 months (overall cohort, multivariable) | Secondary | 0.3 | <0.001 | ||
| Monophasic subgroup: IT 12 months vs 3 months | Secondary | 0.25 | <0.001 | ||
| Monophasic subgroup: IT 18 months vs 3 months | Secondary | 0.13 | <0.001 | ||
| Relapsing subgroup: IT 24 months vs 3 months | Secondary | 0.34 | 0.02 | ||
| Relapsing subgroup: IT 30 months vs 3 months | Secondary | 0.32 | 0.024 | ||
| Negative MOG IgG1 pre-discontinuation (univariable) | Secondary | 0.43 | 0.045 | ||
| Negative MOG IgG1 within 12 months pre-discontinuation (univariable sensitivity) | Secondary | 0.34 | 0.024 | ||
| Predicted 5-yr relapse-free survival, monophasic, 15-month IT | Secondary | 87.7% (95% CI 77.0-100%) | |||
| Predicted 5-yr relapse-free survival, monophasic, 3-month IT | Secondary | 47.9% (95% CI 29.9-76.5%) | |||
| Predicted 5-yr relapse-free survival, relapsing, 24-month IT | Secondary | 62.1% (95% CI 38.6-99.9%) | |||
| Predicted 5-yr relapse-free survival, relapsing, 3-month IT | Secondary | 24.3% (95% CI 10.3-57.1%) | |||
| Discontinuations Due to Adverse Events | Adverse | 7/236 (3.0%) | |||
| Note | Adverse | AEs not systematically reported (observational cohort); the study rationale for time-limited IT was to balance relapse risk against long-term corticosteroid/immunosuppressant toxicity. | |||
Subgroup Analysis
Analyses stratified by disease course at discontinuation (monophasic vs relapsing). Sensitivity analysis restricted to 171 intervals with MOG IgG1 within 12 months of discontinuation confirmed protective effect of seroreversion. Incident cohort analysis (148 patients, 168 discontinuations) reproduced main findings: ≥10-month IT protective versus 3 months in monophasic patients.
Criticisms
- Retrospective observational single-centre design; susceptible to selection and confounding biases.
- Majority (84.7%) used corticosteroids only; unable to compare outcomes across specific steroid-sparing agents.
- Predominantly white cohort; race could not be examined as a predictor.
- Timing of pre-discontinuation MOG IgG1 testing was variable (reflecting real-world practice), possibly underestimating its predictive value.
- No MRI lesion data available to assess as a predictor.
- "Optimal" durations of 10-18 and 20-30 months are model-based estimates from spline curves rather than pre-specified thresholds, and should be individualised.
Funding
W.Z.Y. supported by an ECTRIMS Research Fellowship and Robert and Elizabeth Albert Travel Grant (RACP Foundation). Oxford Neuromyelitis Optica Service funded by NHS England Highly Specialised Services.
Based on: MOGAD-Discontinuation (Brain, 2026)
Authors: Yeh WZ, Francis A, Cooper S, ..., Palace J
Citation: Brain. 2026 (Advance article, published online). doi:10.1093/brain/awag006
Content summarized and formatted by NeuroTrials.ai.