Suspected PML in a Natalizumab Patient: Whether to Accelerate Clearance Before Confirmation
A natalizumab-treated MS patient's new, subacute neurologic decline looks more like PML than a relapse — and plasma exchange clears the drug faster, but nobody is certain that doing so actually helps rather than just moving the danger earlier.
C.V., a 44-year-old woman, has worked as a hospital pharmacy technician for the past nine years — a detail she brings up herself, half-joking that she “already knows exactly what everyone in this room is about to argue about.” She was diagnosed with relapsing-remitting MS eleven years ago and has been on natalizumab infusions every four weeks for just over three years, after two other disease-modifying therapies failed to control her relapse rate. She is JC virus antibody positive, checked at the start of natalizumab and confirmed again eighteen months ago, a fact her neurologist has revisited with her at each infusion as her risk has changed with duration of therapy. Over the past three to four weeks, her coworkers and her closest friend from nursing school have separately noticed something different from her usual relapses: gradually worsening clumsiness in her right hand, word-finding trouble that comes and goes but is generally worsening, and — the detail that finally brought her in — an uncharacteristic irritability and flatness her friend says “isn't her.” Unlike her prior relapses, which came on over a day or two and partly resolved with steroids, this has been slow, weeks-long, and steadily progressive rather than stepwise.
Her anti-JC virus antibody index was 2.4 at the last check. Plavina and colleagues (2014) found eighty-four percent of natalizumab-associated PML cases arose in patients with an index above 1.5, and she has sat above it through her whole third year on the drug. She is not, though, in the highest cell of Bloomgren's 2012 stratification: that band requires prior immunosuppressant exposure, and interferon beta and glatiramer acetate don't count. The imaging is what turned a routine infusion week into this visit — a new, poorly marginated area of subcortical white matter signal change, without the ovoid shape, sharp margins, or contrast enhancement typical of an active demyelinating plaque, which in a patient with her serology, her duration of therapy, and this kind of subacute, non-relapsing decline is specifically the picture PML surveillance is built to catch rather than one MS lesion among others. CSF has gone for JC virus PCR, highly specific when it is positive and unreliable when it is negative this early. The same alpha-4 integrin blockade that has held her MS in check is what let JC virus go unwatched. Whatever happens next, everyone agrees the drug itself has to stop today. What nobody here can tell her is whether clearing it faster helps: that question has been put to registry data rather than to a trial, and the answer that came back was not encouraging.
In clinic, the week the MRI changed the conversation
I want to start plasma exchange this week, not wait on the CSF PCR. The MRI pattern — poorly marginated, non-enhancing, subcortical, arriving as a slow multi-week decline rather than a relapse — is close to the textbook picture PML surveillance is built to catch in a patient on natalizumab this long. CSF JC virus PCR is highly specific when it's positive, but it can be falsely negative early in a genuinely evolving case, so a negative result this week wouldn't actually rule anything out. Natalizumab blocks alpha-4 integrin, which is exactly what's kept lymphocyte surveillance out of her brain for three years — as long as meaningful drug is still on board, that suppression continues. Plasma exchange is the only lever we have on that: Khatri and colleagues (2009) showed three exchanges over five to eight days clear roughly seventy to eighty percent of circulating natalizumab and restore leukocyte function, which is faster than her own kidneys and liver will manage it.
I don't disagree the picture is worrying, and stopping the drug today isn't in question. What I'd push back on is treating PLEX itself as low-cost: central line placement, real hypotension and citrate-related electrolyte risk, and — the bigger issue — the doubt about whether accelerating clearance changes survival is not vague, it has names on it. Landi and the Italian PML group (2017) titled their paper “No evidence of beneficial effects of plasmapheresis in natalizumab-associated PML,” and the manufacturer's own pharmacovigilance analysis of seven hundred and twenty-three confirmed cases — eighty-five percent of whom had plasma exchange — found two-year survival numerically worse in the exchanged group at every viral load tertile. Not significantly worse. But not better, either, and natalizumab's own half-life means most of it leaves over the coming weeks regardless.
What PLEX reliably does speed up is exactly the return of immune cells into a brain that's had almost none for three years — that's the known trigger for immune reconstitution inflammatory syndrome, and in some published cases the IRIS that followed rapid clearance caused more acute damage than the slower course would have on its own.
Then the honest framing isn't PLEX versus waiting — it's controlled clearance versus uncontrolled clearance, because she's clearing the drug either way over roughly the same several weeks. What actually differs is whether that happens with corticosteroids and close monitoring already staged for IRIS, or unpredictably while everyone's still waiting on a PCR result that might not even resolve the question. I'd send the confirmatory testing urgently, with a repeat sample built in if the first comes back negative, and start planning IRIS monitoring now regardless of whether we accelerate clearance with PLEX this week or let it happen on its own timeline — either path needs the same readiness, and neither of you is actually disagreeing about that part.
Agreed: natalizumab stopped today; urgent CSF JC virus PCR sent with a planned repeat sample if the first result is negative; corticosteroids and IRIS-focused monitoring staged and ready regardless of which clearance path is ultimately chosen.
Not agreed, and carried forward explicitly rather than resolved: whether to proceed with plasma exchange this week or await confirmatory — possibly repeat — CSF testing first. The neurologist and the neurointensivist were left at genuinely different comfort levels about the coming days, neither treating the other's position as wrong. The infectious disease physician's monitoring-and-readiness framing was accepted by both as the right immediate plan, without resolving which of their two underlying positions on PLEX timing is actually correct.