Design: Systematic review and meta-analysis of randomised trials
Population: 5,262 people with rheumatoid arthritis in sustained control, 27 publications
Primary outcome: Disease flare
Effect: Tapering RR 1.56 (95% CI 1.26 to 1.98); withdrawal RR 2.23 (95% CI 1.83 to 2.94)

This meta-analysis pooled 27 randomised trial publications, 5,262 participants, in which people with rheumatoid arthritis in sustained control had a DMARD tapered or withdrawn.

Tapering raised flare risk (RR 1.56, 95% CI 1.26 to 1.98), driven by biologics and seen mainly in the first nine months, not beyond. Withdrawal raised it more (RR 2.23, 95% CI 1.83 to 2.94), across all follow-up periods. Withdrawal was associated with radiographic progression on strong evidence, tapering on weak evidence. Neither reduced adverse events.

The usual reason to step down is to reduce harms and costs. These data show no fall in adverse events, a real flare cost, and a structural cost when drugs are stopped.

In practice, dose reduction or interval spacing of a biologic remains a reasonable discussion with a patient in stable remission, with a clear plan to restore the dose at the first sign of flare. Complete withdrawal should generally be avoided. Cost pressures in India make tapering attractive, but stopping altogether trades short-term savings for flares and joint damage.

Discuss tapering, not stopping, with patients in sustained remission
Expect most excess flares in the first nine months after tapering
Agree a rescue plan to restore the full dose quickly if a flare starts
Avoid complete DMARD withdrawal where possible
Do not expect fewer adverse events from de-escalation

Why it matters

It weighs the main promise of de-escalation, fewer harms, and finds it unmet.

The statistics, in plain English

A risk ratio of 2.23 means more than twice the risk of flare after withdrawal, with a confidence interval well clear of 1.0. For tapering, the raised risk faded after nine months, suggesting most patients who will flare do so early.