- Design
- Retrospective multicentre cohort with multivariable Cox models
- Population
- 1,798 JAK inhibitor courses in 1,240 people with rheumatoid arthritis, Japan
- Primary outcome
- Discontinuation for ineffectiveness and for adverse events
- Effect
- Ineffectiveness vs tofacitinib: upadacitinib HR 0.33 (0.24–0.47); filgotinib HR 0.59 (0.42–0.85)
A retrospective multicentre analysis of the Japanese ANSWER cohort included 1,798 courses of JAK inhibitor treatment in 1,240 people with rheumatoid arthritis between 2013 and 2025, covering tofacitinib, baricitinib, peficitinib, upadacitinib and filgotinib. Discontinuation for ineffectiveness and for adverse events was compared, adjusting for patient, disease and prior treatment.
Compared with tofacitinib, discontinuation for ineffectiveness was lower with upadacitinib (HR 0.33, 95% CI 0.24–0.47) and filgotinib (HR 0.59, 0.42–0.85); baricitinib and peficitinib did not differ significantly. Upadacitinib had fewer stops for adverse events than peficitinib (HR 0.57). Prior IL-6 or JAK inhibitor use raised the chance of stopping for ineffectiveness with tofacitinib, baricitinib and peficitinib, but not clearly with upadacitinib or filgotinib.
This is observational, and prescribers may have chosen newer agents for different patients over a 12-year period in which practice changed. Treatment persistence is a useful real-world proxy but not the same as remission. It offers some guidance on which JAK inhibitor to try after a previous targeted therapy has failed.
- Consider upadacitinib or filgotinib when choosing a JAK inhibitor after failure of an IL-6 inhibitor or another JAK inhibitor.
- Apply the same cardiovascular, thrombotic and malignancy risk assessment before any JAK inhibitor.
- Treat persistence data as supportive, not as proof that one agent is more effective.
- Check local availability and cost; not every agent studied is available in India.
Why it matters
Head-to-head trials among JAK inhibitors are scarce, so large real-world comparisons fill a practical gap.
Don't overread it
Observational — differences in persistence may reflect prescribing patterns rather than drug effect.
The statistics, in plain English
A hazard ratio of 0.33 means the rate of stopping for lack of effect was about a third of that with tofacitinib. Because treatment was not randomised, part of this may reflect who received which drug and when.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for rheumatology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free