A Flare on the Drug That Was Supposed to Prevent It: Switching or Adding After Fourteen Months
A single patient who has relapsed clinically while actively on tocilizumab. Neither trial that could guide the next step was built for a patient whose disease broke through on the drug itself — so the disagreement is over which kind of leap is smaller.
Eleanor P., 77, ran the reference desk at the county library for thirty-two years and still volunteers there twice a week, reshelving in the aisles she says she could navigate blindfolded. Her GCA was diagnosed fourteen months ago; tocilizumab and the standard prednisone taper handled it cleanly enough that she finished the taper six months ago and has been on tocilizumab alone since, symptom-free, her inflammatory markers holding at a stable ESR of 8 to 12. This past week the jaw claudication is back — real difficulty chewing, the same sensation that first sent her to a rheumatologist over a year ago — and her ESR and CRP have both climbed sharply, to 61 and 34. There is no vision change, no new headache beyond the jaw symptom itself, but this is, unambiguously, a clinical relapse. What makes it different from an ordinary flare is where it's happening: on active, ongoing weekly tocilizumab, not during a taper or after stopping it. Her disease has broken through the exact drug that has otherwise controlled it for over a year.
Neither of the two obvious next moves rests on evidence built for this situation. Adding methotrexate on top of ongoing tocilizumab has real precedent as a glucocorticoid-sparing combination in GCA, but the supporting data for that specific pairing are modest, and none of it was generated in patients whose disease had already broken through IL-6 blockade rather than in patients starting both together.
Switching to upadacitinib reaches for a genuinely different mechanism — JAK inhibition rather than IL-6 receptor blockade — which has real biological appeal after an apparent escape from one pathway. But SELECT-GCA, the phase 3 trial that won it FDA approval for GCA in April 2025 on a sustained-remission rate of 46.4% against 29.0% for placebo, enrolled patients with new-onset or relapsing GCA; it was not built or powered around patients whose disease had broken through an IL-6 inhibitor, and Eleanor's situation sits outside the question it was designed to answer. Choosing between augmenting a drug that mostly worked for fourteen months and switching to one whose own evidence doesn't cleanly cover her situation either means accepting that whichever path is chosen is itself an extrapolation, not a documented answer.
Add, switch, or bridge first
A breakthrough flare on an IL-6 inhibitor is a real signal that pathway may not be sufficient for her disease anymore. I'd switch to upadacitinib — a genuinely different mechanism makes more biological sense than adding a second agent to a pathway that's already shown it can be escaped.
I'd be more conservative before abandoning a drug that controlled her disease cleanly for fourteen months. Adding methotrexate is a smaller step than switching entirely, and its steroid-sparing role alongside biologic therapy in GCA, while modest, is established enough to be a reasonable next move before concluding the whole pathway has failed her rather than just needing reinforcement.
I take the mechanistic argument for switching seriously, but one breakthrough flare after fourteen months of control isn't the same as the drug never having worked — and SELECT-GCA wasn't built around patients failing tocilizumab either, so the switch is its own extrapolation, not a documented answer. I'd also want it said plainly that this is a switch and not an addition: upadacitinib's label specifically advises against combining it with a biologic DMARD, so tocilizumab stops if we go that way.
Whatever gets decided about the biologic question, she needs glucocorticoid control now — any new regimen, added or switched, takes weeks to take full effect, and jaw claudication this active shouldn't sit untreated in the meantime. I'd also flag that at seventy-seven, cumulative treatment burden matters on its own terms, not just as a tiebreaker between two mechanisms — whichever path is chosen should come with a clear plan for what would count as it also failing, so she isn't left escalating indefinitely without a defined off-ramp.
Agreed immediately: re-escalate prednisone now for symptom control, regardless of which biologic path is ultimately chosen.
Not agreed: whether to add methotrexate to ongoing tocilizumab or switch to upadacitinib. The geriatrician's request — a defined off-ramp if the chosen path also fails — was accepted by both other voices, but the underlying add-versus-switch disagreement was not resolved at this visit and will be revisited once glucocorticoid control is re-established.