- Design
- Systematic review and meta-analysis of case series and comparative studies, PROSPERO CRD420251071702, GRADE assessed
- Population
- 466 patients with non-infectious uveitis refractory to anti-TNFα therapy, from 89 studies contributing to meta-analysis
- Primary outcome
- Pooled clinical response rate after biologic switching, intraclass versus interclass
- Effect
- Overall 83% (95% CI 76-90%); intraclass 77% (65-87%) versus interclass 89% (80-96%), P=0.0995; sensitivity analysis 71% versus 85%, P=0.044; low certainty
Up to a third of patients with non-infectious uveitis respond inadequately to a first-line anti-TNF agent, and the next move has been a matter of preference: try a second anti-TNF, or move to a different class.
This systematic review and meta-analysis, registered on PROSPERO, searched PubMed, EMBASE and Cochrane to January 2025 and included 136 studies, of which 89 contributed to the meta-analysis. The evidence base is mostly case reports and series — 129 of the 136 — with seven comparative studies, and risk of bias was assessed with the JBI checklists and the Newcastle-Ottawa Scale.
The overall pooled response rate after any switch was 83% (95% CI 76-90%, I² = 30.2%, 466 patients). Intraclass switching — anti-TNF to anti-TNF — pooled at 77% (95% CI 65-87%) and interclass switching at 89% (95% CI 80-96%), a difference that did not reach significance in the main analysis (P = 0.0995). In sensitivity analyses the gap held and did reach significance: 71% (59-82%) intraclass against 85% (77-93%) interclass (P = 0.044). Certainty of evidence for the overall rate was graded low.
The practical message is a nudge, not a rule. Both approaches work in most patients, and 83% overall is a genuinely encouraging number to give someone who has just failed adalimumab. But where the choice is open, this is evidence pointing towards a different class rather than a second agent in the same one — offered at low certainty, from an evidence base made mostly of case series.
- Both strategies work in most patients — 83% pooled response is the number to counsel with
- Where the choice is free, this points towards a different class rather than a second anti-TNF
- Immunogenicity to the first agent is a reason to prefer interclass; secondary loss of response is not always that
- The evidence is 129 case reports and series against 7 comparative studies — certainty is low
- Access and cost will decide many of these choices regardless of which has the better pooled rate
Why it matters
It gives a direction for the decision that follows a failed anti-TNF, which has been made on preference until now.
Don't overread it
The difference between strategies reached significance only in sensitivity analysis, on an evidence base dominated by case reports.
The statistics, in plain English
The headline comparison was not significant (P = 0.0995) and only became so in a sensitivity analysis (P = 0.044) — which is the reverse of the usual order and a reason for caution, not confidence. Pooling response rates from case series inflates success, because series of failures are less often published. GRADE low certainty means further research is likely to change the estimate. The I² values (18-54%) show the studies were reasonably consistent with each other, which is the one reassuring part.
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 ophthalmology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free