A Diagnosis Too New to Know if It Will Happen Again
His first attack is effectively over — near-complete visual recovery, three weeks out. What is unresolved is whether there will be a second, and no test can say. Treating now protects the children who would have relapsed and commits the rest to years of therapy they never needed.
J.K. missed his marching band's entire fall competition season. He is 15, a high-school sophomore, and woke three weeks ago with vision blurring in both eyes and a headache his parents put down to screens and stress. It was bilateral optic neuritis, confirmed by enhancement in both optic nerves on MRI and a positive MOG antibody on live cell-based assay — MOG antibody-associated disease, distinct from both MS and NMOSD, and formally defined by international criteria only since Banwell and colleagues published them in 2023. IV methylprednisolone worked; his vision is close to back.
Which is what makes the remaining question hard rather than urgent. This attack is finished. Whether it was the only one is not knowable today, and there is no marker that resolves it. The natural-history data behind the 2023 criteria put monophasic disease at roughly half of first presentations across all ages, with pediatric series running somewhere between a third and a half going on to relapse — most of those inside the first twelve months, so the window in which the question answers itself is short rather than open-ended. The European paediatric MOG consortium consensus sets maintenance therapy after a first relapse, not a first attack, reserving earlier treatment for children left with severe residual deficit. His deficit is nearly gone. What the group is weighing is whether that consensus describes a child whose recovery has been this complete, or whether the majority who never relapse is a poor reason to leave the minority unprotected for a year.
Three weeks after the first attack, reviewing what comes next
I'd start maintenance therapy now, after this first attack, rather than wait for a second. Real-world MOGAD cohorts show a meaningful fraction of patients relapse within the first year or two, and where children do relapse, most do it inside the first twelve months — so 'wait and see' is not a long, forgiving window. Every additional attack carries its own risk of incremental, potentially permanent deficit, which is worth avoiding in a fifteen-year-old with decades ahead of him.
If his vision recovery had been incomplete rather than near-total, I'd be even more insistent — but even with how well he's recovering from this attack, I don't think that changes the argument for preventing the next one, if there is one.
I'd observe closely without starting maintenance therapy. The natural-history data behind Banwell and colleagues' 2023 international criteria puts monophasic disease at roughly half of first attacks overall, and the pediatric estimates run more favorable still — series report somewhere between a third and a half of children going on to relapse, not more. The European paediatric MOG consortium consensus is explicit on the sequencing: maintenance treatment starts after a first relapse, not after a first attack, reserving earlier treatment for children left with severe residual deficit. He has near-complete visual recovery. He is not that child.
I understand the appeal of preventing even one more attack. But if more than half of children like him never relapse at all, treating now means committing the majority to years of maintenance therapy for a disease course they were never going to have — and you framed that as a coin flip, which I think overstates his risk rather than describing it.
There's a complication neither of you has weighed, and it cuts at the drug rather than the timing. You both said 'rituximab' as though it were the obvious agent. In MOGAD it isn't. Pooled pediatric data across several retrospective series put roughly 70 percent of children relapse-free on monthly IVIG against about 47 percent on rituximab, with mycophenolate and azathioprine in between; the large European pediatric cohort found IVIG cut the annualized relapse rate from 2.16 to 0.51, and it is the one maintenance agent associated with both fewer relapses and better disability outcomes. Anti-CD20 depletion works beautifully in AQP4-positive NMOSD and transfers poorly here.
This doesn't resolve the timing question, but it does change what's at stake in it. If we ever do treat him, the honest first-line here is monthly IVIG, not rituximab — which matters for today's decision because IVIG's burden is a recurring infusion commitment for a teenager, not years of B-cell depletion with the infection and hypogammaglobulinemia risk you were both weighing. The cost of treating turns out to be a different cost than the one this conversation has been pricing.
Agreed: observation without starting rituximab, with MRI and visual evoked potential follow-up at three months and a clear, written treatment plan already in place for a second attack if one occurs. J.K.'s family was walked through the actual probabilities directly, including the honest spread in them — somewhere between a third and a half of children in his position relapse, most within the first year — rather than given a falsely confident recommendation in either direction.
Not agreed: whether the relapse probability the Pediatric Neurologist is working from is the right one. He reads the pediatric data as closer to even odds and treats that as grounds to act now; the Neuroimmunologist reads it as a third to a half and treats the same range as grounds to wait. Neither conceded, and the group let the disagreement stand rather than average it into a number none of them believed. What did get settled is what happens next: if a second attack comes, the maintenance agent is monthly IVIG, not rituximab — written into the chart today precisely so that choice isn't made under pressure during an acute event.