A Second Biologic After Eight Good Months and a Slow Fade
Older guidance said the reason a first biologic stopped working should decide what comes next. A newer trial, built specifically to test that assumption, found switching classes was no better than cycling within one — right as this patient's own history sits ambiguously between the two categories the assumption depends on.
Desmond O. runs the kitchen at a busy downtown restaurant, a job that keeps him on his feet twelve hours at a stretch, which is exactly why the first eight months on etanercept felt like getting his life back — morning stiffness gone, able to bend into a walk-in freezer without wincing, ASDAS down from 3.6 to 1.8. Then, gradually, over about six weeks, the stiffness crept back in, first just at the end of a double shift, then most mornings, his ASDAS climbing back to 3.1 by the time he came in for this visit. He never stopped the drug, never missed a dose he can recall, and nothing about his life changed enough to explain it on its own.
Where exactly Desmond's history lands matters more than it might seem, because the two categories the field has historically used to plan a next step — primary non-response, meaning the drug never really worked, versus secondary loss of response, meaning it worked and then stopped — assume a cleaner split than his actual course shows. He had a real response, unambiguously, for eight months; the 2019 ACR/SAA/SPARTAN guideline would call this secondary failure and conditionally recommend cycling to a second TNF inhibitor rather than switching mechanism. But the more recent 2022 ASAS-EULAR update dropped that reason-for-failure distinction from its own recommendation entirely, proposing simply "an alternative bDMARD" without mandating a mechanism change either way — and the ROC-SpA trial, designed specifically to test which approach actually works better after a first TNFi failure, found switching to an IL-17A inhibitor was not superior to cycling to a second TNF inhibitor on its primary endpoint, undermining the very premise the older, reason-based guidance was built on. Desmond's own priority, stated plainly once the drug options were on the table, was less about the mechanism than the schedule — a twice-weekly injection has been genuinely hard to fit around split shifts that start before dawn and end past midnight. It is worth being honest about what that means for the decision rather than dressing it up afterward: with the guideline's reason-based rule withdrawn and the trial that replaced it returning a null, nothing in the evidence distinguishes his two options. A choice made on dosing schedule here is not a compromise of the clinical reasoning. It is what is left of it.
Cycle the class, or switch the mechanism
This reads as a classic secondary loss of response — eight months of real benefit, a gradual fade with no missed doses, rising CRP. My instinct is drug-specific failure rather than pathway failure — something about this molecule in this patient, which a structurally different TNF inhibitor might get around even though the mechanism class stays the same. I'd cycle to adalimumab before reaching for a mechanism switch he doesn't clearly need.
I'd be careful about which version of that story we tell, because the obvious one doesn't survive contact with etanercept specifically. Immunogenicity is the usual explanation for a fade like his, but etanercept is the wrong drug for it: it is the least immunogenic of the class, and where anti-drug antibodies against it are detected at all they behave as non-neutralizing — therapeutic-drug- monitoring series find no inverse relationship between anti-etanercept titer and etanercept level, the relationship that does hold for adalimumab and infliximab. So "antibodies ate the drug" is the one mechanism his case argues against. And we now have a trial that tested the actual policy built on reasoning like it. ROC-SpA randomized axSpA patients with a first TNFi failure specifically to compare switching to an IL-17A inhibitor against cycling a second TNFi, and found no superiority for the mechanism switch on the primary endpoint. That's the exact comparison we're having right now, tested directly rather than reasoned toward.
I'd push back gently on reaching for "secondary failure, therefore cycle" as settled logic — the 2022 ASAS-EULAR update deliberately dropped that reason-for-failure distinction from its own recommendation, and ROC-SpA is a real part of why: the field's confidence in that specific rule was higher than the evidence behind it turned out to support.
I don't think ROC-SpA settles this either way for Desmond specifically, and I'd rather say that plainly than pick a side the trial doesn't actually support. Its own exploratory analysis found non-significant trends favoring an IL-17A inhibitor after primary nonresponse, and favoring TNFi cycling when the first drug stopped for an adverse event. He's neither — a genuine secondary loss of response, no adverse event. Since neither favorable subgroup describes him, I'd let his other features decide: no psoriasis, no IBD, no uveitis pointing either class toward or away from him. Given that, and given his stated dislike of the twice-weekly etanercept schedule, I'd move to secukinumab — not because the evidence says it will work better, but because a monthly dosing interval is a real adherence advantage for a chef doing twelve-hour shifts, in a genuine toss-up.
Agreed: switch to secukinumab, with ASDAS and CRP rechecked at 12 and 24 weeks to confirm real response before assuming today's dosing-schedule tie-breaker paid off clinically.
Not agreed: whether the reason-for-failure distinction from the 2019 guideline still deserves independent clinical weight going forward, or should be treated as effectively retired by ROC-SpA's result. The first rheumatologist would still apply it as a tie-breaker in a future patient with a clearer immunogenicity picture; the clinical pharmacologist would treat the distinction as no longer load-bearing for any future decision, consistent with its removal from the 2022 ASAS-EULAR recommendation.