An Accountant Without a Headache: How Long to Treat the GCA That Never Announced Itself
A single patient with large-vessel GCA and no cranial symptoms at all. The disagreement is over whether the imaging that found his disease should also be the thing that decides when his treatment ends.
Marcus L., 63, wound down his accounting practice to three clients last year and spends most of his newly free time on a restored sailboat he keeps calling 'still a work in progress,' the same phrase, he's noted, that's applied to his own diagnosis. Four months ago it was fevers and unexplained weight loss that brought him in, not a headache — he has never had one, never had scalp tenderness, never had jaw claudication, the classic cranial trio absent from start to finish. A temporal artery biopsy came back negative, which would have ended the workup for a more typical presentation; instead, a PET-CT ordered to chase the fevers found FDG-avid inflammation running through his aorta and both subclavian and axillary arteries — large-vessel-predominant giant cell arteritis, a real and distinct phenotype from the cranial disease the diagnostic criteria were originally built around. He started tocilizumab and prednisone, and by every clinical measure he has answered well: fevers gone, weight back, prednisone down to 5mg, feeling, in his own words, better than he has in a year. A repeat PET-CT is planned, and that scan is what the team is actually arguing about, because the disease it's designed to find behaves differently from the disease his imaging phenotype excludes.
GiACTA, the trial tocilizumab's entire treatment-duration evidence rests on, enrolled overwhelmingly cranial-predominant GCA — patients whose disease was found by temporal artery biopsy the way most GCA still is. Large-vessel-predominant disease like Marcus's carries its own documented risks that don't map cleanly onto that trial's framework: higher relapse rates in some cohorts, and a real, if late, risk of aortic aneurysm or dissection that cranial-predominant disease carries much less of. Complicating any imaging-based answer further is a separate, well-described phenomenon reported across the large-vessel GCA imaging literature: FDG uptake can persist on PET-CT for many months in patients who are clinically doing exactly what Marcus is doing — no symptoms, normal inflammatory markers, feeling well — because vessel-wall healing lags behind clinical remission in ways the scan can't distinguish from ongoing disease activity. Treating his upcoming scan as a simple yes-or-no answer about whether to keep going or stop risks reading residual, expected healing as active disease that isn't there, or the reverse.
Reading a scan that outlasts the disease
His phenotype is exactly the kind GiACTA under-enrolled, and large-vessel disease carries real long-term risk — aortic aneurysm chief among them — that cranial-predominant disease doesn't carry to the same degree. I'd plan to continue tocilizumab well beyond the standard trial-modeled course and let his imaging, not a calendar built for a different phenotype, drive how long he stays on it.
I'd be careful about letting the scan itself set the stopping point. Residual FDG uptake in large-vessel GCA is a well-described phenomenon in patients with excellent clinical response — vessel-wall healing genuinely lags behind clinical remission, sometimes for well over a year. If we chase the scan to zero before stopping treatment, we risk prolonging immunosuppression in a man who, by every clinical measure available, is well.
I don't disagree that his phenotype carries real long-term structural risk — I'd just separate that risk from the treatment-duration question. The aneurysm risk is a reason for long-term vascular surveillance, not necessarily a reason to keep treating an active-looking scan that may not reflect active disease.
I think you're both right about different questions, and the disagreement is partly an artifact of treating them as one decision. How long to continue tocilizumab should be driven by his clinical course — symptoms, inflammatory markers, his own trajectory — with imaging as supporting rather than determining evidence, given how unreliable a persistently avid scan is as a stopping signal in this phenotype specifically. Separately, and regardless of when tocilizumab eventually stops, he needs a long-term vascular imaging surveillance plan for the aneurysm risk his phenotype carries — a decision that shouldn't be tied to the biologic decision at all, since aneurysm risk doesn't end when treatment does.
Agreed: continue tocilizumab based on his clinical course, with the upcoming PET-CT read as supporting rather than determining evidence — a persistently avid scan in a clinically well patient will not, by itself, trigger escalation or an extended course beyond what his symptoms and labs already justify.
Also agreed, and treated explicitly as a separate track: a long-term vascular imaging surveillance plan for aortic aneurysm risk will be set up independent of the tocilizumab-duration decision, since that structural risk doesn't end whenever treatment eventually does.