Switching biologic in anti-TNF-refractory uveitis: most patients responded, with higher response when the class changed

Design

Systematic review and meta-analysis of 136 studies (129 case reports or series, 7 comparative; PROSPERO registered)

Population

466 patients with anti-TNF-refractory non-infectious uveitis in the pooled analysis

Primary outcome

Pooled response rate after switching biologic, intraclass vs interclass

Effect

83% overall (95% CI 76% to 90%); 77% intraclass vs 89% interclass (P = 0.0995)

This systematic review and meta-analysis looked at what happens when non-infectious uveitis fails an anti-TNF biologic. It included 136 studies, 129 of which were case reports or series and 7 comparative, with 89 in the meta-analysis (466 patients).

The pooled response rate after switching was 83% (95% CI 76% to 90%). Switching within the anti-TNF class gave 77% (95% CI 65% to 87%) and switching to a different class 89% (95% CI 80% to 96%), a difference that was not significant in the main analysis (P = 0.0995) but was in sensitivity analysis (71% vs 85%, P = 0.044). Certainty of evidence was low.

The data are mostly uncontrolled case series, and response definitions vary. They support switching rather than abandoning biologic therapy, and hint that a change of class may work better. They cannot decide which agent to choose.

• Consider switching biologic, rather than stopping, when anti-TNF therapy fails in non-infectious uveitis.
• Weigh a change of class, which had a higher pooled response, against the low certainty of the evidence.
• Review adherence, drug levels or antibodies and the diagnosis before declaring failure.
• Agree the response definition with the patient and rheumatology or uveitis team.
• Remember that publication bias may inflate response rates from case series.

Why it matters

Up to a third of patients do not respond to first-line anti-TNF therapy, and there has been little guidance on what next.

Don't overread it

Mostly case reports and series, with low-certainty evidence; it does not show that one class is better.

The statistics, in plain English

A pooled response rate of 83% from case series is probably optimistic, because failures are less likely to be reported. The 89% versus 77% gap was borderline in the main analysis and became significant only when the analysis was restricted, so it is a signal rather than proof.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.