A Reply Letter Questions How Far a Danish Psoriasis Registry Study Can Really Reach

A reply letter posted October 6, 2026 on the Journal of the American Academy of Dermatology's site, under the heading "Response to Schwarz C.W., et al.," pushes back on a registry study the journal ran online six months earlier. The exchange is narrow on its face, a comment on one drug-survival paper, but it maps onto a wider argument about how much a single national registry can tell a prescriber about which psoriasis biologic to try second.

The paper under dispute, led by Amélie Spangenberg and John Y.M. Koo according to the response's author listing, examined Danish DERMBIO registry patients who moved to a second biologic after adalimumab stopped working. Schwarz et al. reported in JAAD02509-0/fulltext) a standardized 2-year drug survival of about 0.71 for bimekizumab against about 0.45 for brodalumab, built from a Cox model adjusted for age, sex, weight, smoking and psoriatic arthritis.

The arithmetic behind that 0.71 figure deserves a second look before anyone treats it as settled. Only 53 of the bimekizumab patients had actually reached the full two years of follow-up when the study closed. The rest of that arm's estimate comes from statistical extrapolation beyond the data the registry had on hand, so the confidence interval on bimekizumab's side, 0.62 to 0.81, is wider than the headline number suggests.

Sample size cuts the same way. The study put 197 bimekizumab starts against 63 brodalumab starts, an unequal split that a registry analysis can adjust for but can't erase. A quasi-randomization scheme, assigning patients by the calendar period in which each drug became available rather than by clinician choice, approximates a trial design without being one.

Brodalumab's own history in other registries undercuts the idea that 0.45 is a stable number for the drug anywhere. Germany's PsoBest registry put 1-year survival near 76 percent. A separate Danish single-center study found 65.7 percent at one year, a figure its own authors flagged as lower than other published cohorts. Three registries, three different answers, before anyone gets to the two-year mark the JAAD paper reports.

One confound sits outside the adjustment set entirely. The FDA attached a boxed warning and a REMS monitoring program to brodalumab in 2017 after trial data showed an imbalance in suicidal ideation and behavior relative to other psoriasis biologics. A later postmarketing analysis out of UCSF found brodalumab's suicide risk comparable to other agents in the class once surveillance intensity was accounted for. The REMS paperwork itself, the extra visits and documentation it requires, gives prescribers and patients a reason to drop the drug that has nothing to do with how well it controls psoriasis, and the Cox model in the JAAD paper doesn't include a term for it.

Context from the same Danish dataset complicates the picture further. Schwarz and colleagues published a larger DERMBIO analysis in JAMA Dermatology in January 2026, covering 376 bimekizumab starts and 192 brodalumab starts across bio-naive and bio-experienced patients going back to 2007. In that wider cohort, ustekinumab outperformed newer agents among bio-naive patients, while bimekizumab, guselkumab and risankizumab all beat ustekinumab among bio-experienced ones. The narrower JAAD paper the October letter responds to draws on an overlapping slice of that same registry population, which raises a question about how independent the two findings really are.

A shorter-horizon Italian comparison points to a different metric entirely. Potestio and colleagues tracked bimekizumab and brodalumab over 36 weeks in roughly 125 patients and found bimekizumab ahead on PASI100 response at week 4, 41.5 percent against 23.6 percent, with the gap persisting at week 16. They found no significant difference in safety outcomes or treatment failure between the two drugs across that stretch. Drug survival and PASI response are measuring different things, staying on a drug versus clearing skin, and a study built for one doesn't automatically speak to the other.

What the October 6 letter is asking, in effect, is whether a single-country registry comparison of 260 total patients, with a REMS-burdened drug on one side and an extrapolated two-year tail on the other, can support a claim about which biologic to prescribe second, anywhere. The JAAD exchange is still open. Schwarz and co-authors have not yet posted a reply to the reply, the format the journal typically uses to close out this kind of correspondence.