PSYREFLECT
RESEARCHJanuary 8, 20262 min read

Slow Tapering Plus Therapy Prevents Relapse as Well as Staying on Antidepressants

Key Findings
  • Network meta-analysis of 76 RCTs (17,379 participants, mean age 45.2) compared five antidepressant deprescribing strategies with and without adjunctive psychological support
  • Slow tapering (>4 weeks) combined with psychological support (MBCT, CBT) prevented relapse as effectively as continuing antidepressants at standard doses
  • The combined strategy could prevent relapse in 1 out of 5 patients compared to abrupt or rapid discontinuation
  • Abrupt discontinuation and fast tapering (≤4 weeks) showed significantly higher relapse rates regardless of psychological support

The question every prescribing clinician faces eventually: when and how to stop antidepressants in a remitted patient. This Lancet Psychiatry network meta-analysis of 76 trials provides the first head-to-head comparison of deprescribing strategies — and the answer is not simply "taper slowly." The answer is taper slowly and add therapy.

What the evidence shows

The analysis pooled data from 17,379 participants across five strategies: abrupt discontinuation, fast tapering (4 weeks or less), slow tapering (more than 4 weeks), reduced-dose continuation, and standard-dose continuation. Each strategy was evaluated with and without adjunctive psychological support — primarily mindfulness-based cognitive therapy (MBCT) and cognitive behavioral therapy (CBT).

The critical finding: slow tapering plus psychological support matched standard-dose continuation in relapse prevention over a mean follow-up of 46 weeks. Neither slow tapering alone nor psychological support alone achieved this equivalence. The combination works because tapering removes the pharmacological safety net while therapy builds the cognitive and behavioral skills to maintain remission without it.

Fast tapering and abrupt discontinuation performed significantly worse, even with psychological support. The withdrawal timeline matters — and four weeks is not enough.

What changes in your practice

If you are managing a patient in stable remission who wants to discontinue antidepressants, the protocol is now clear: extend the taper beyond four weeks and pair it with structured psychological support. MBCT appears particularly well-suited — it was the most frequently studied adjunct and directly targets the ruminative patterns that predict depressive relapse.

For psychologists and therapists: this positions you as an essential partner in deprescribing, not a downstream referral. The data says your involvement during the taper changes outcomes measurably.

Slow tapering combined with therapy prevents relapse as effectively as staying on antidepressants — making psychologists essential partners in deprescribing decisions.

Limitations

87.9% White participant pool limits generalisability. Most studies used MBCT; other therapy modalities less well-represented. Optimal taper duration beyond "more than 4 weeks" remains undefined.

Source
The Lancet Psychiatry
Comparison of antidepressant deprescribing strategies in individuals with clinically remitted depression: a systematic review and network meta-analysis
2026-01-01·View original
Tags
antidepressantsdeprescribingMBCTrelapse-preventionmeta-analysis
Related
Industry
NICE Now Recommends Mindfulness-Based Cognitive Therapy as First-Line for Depression — Not Just Relapse Prevention
NICE Clinical GuidelinesRead →
Research
Training the Addicted Brain to Self-Regulate: EEG Neurofeedback Shows g = 0.85 Across 17 RCTs
AddictionRead →
Research
Psilocybin for Depression: The First Meta-Analysis That Asks How, Not Just If
Neuroscience and Biobehavioral ReviewsRead →
PsyReflect · Free · Mon & Thu
Get analyses like this every Monday and Thursday.
Only what matters for practice. Curated by a clinical psychologist. 5 minutes instead of 4 hours of monitoring.
← Previous
AI Is Entering the Therapy Room — But Regulation Hasn't Arrived Yet
Next →
California's AI Chatbot Law Sets the First Mental Health Safety Standard for Consumer AI