Benzodiazepine Tapering Research

ASAM + 9 partner societies (published in J Gen Intern Med) · Feb 2025

The first US multi-society tapering guideline. Recommends starting with 5-10% reductions, generally not exceeding 25% every two weeks, slowing to 5-10% every 6-8 weeks for strongly dependent patients, and — critically — states clinicians should not abruptly discontinue a physically dependent patient. Conditionally supports diazepam crossover. The authors note most recommendations rest on expert consensus rather than trial evidence. The AAFP digest in American Family Physician (June 2026) is how most US family doctors will meet it.

Horowitz & Taylor, Wiley · 2024

The clinical reference for hyperbolic tapering, with drug-specific reduction schedules for every licensed benzodiazepine and Z-drug. The book behind the proportional-reduction approach our taper calculator implements.

Prof. C. Heather Ashton, Newcastle University · 2002

The document that established slow, patient-led tapering when almost nobody in medicine took benzodiazepine withdrawal seriously. Its diazepam-crossover schedules remain widely used; see our comparison of Ashton vs. hyperbolic tapering for where newer guidance departs from it.

Horowitz, Lamberson, Brandt, Framer, Shapiro, Sørensen, Cadogan, Ritvo, Taylor — Psychological Medicine · Jun 18, 2026

The first peer-reviewed, benzodiazepine-specific statement of hyperbolic tapering. Because dose-receptor occupancy is hyperbolic, cuts should be 5-10% of the CURRENT dose (so milligram steps shrink as you descend), final doses may need to be as small as 0.2 mg diazepam-equivalent, diazepam crossover is optional rather than default, and micro-tapering with liquids or compounded doses is explicitly endorsed. Co-authored by patient advocates alongside the Maudsley authors.

Humphreys et al. — JAMA Network Open · Jan 14, 2026

Randomized trial of 161 long-term users in primary care given a patient-directed electronic self-management program. Odds of full cessation at six months were 5.3x higher with the intervention, with no worsening of anxiety or sleep — evidence that informed, patient-led tapering works.

Bozinoff et al. — PLoS Medicine (1.8M-person Ontario cohort) · Jun 18, 2026

How the very first prescription is written shapes whether someone becomes a long-term user: initial supplies of a week or less, a single agent, and short-acting formulations all predicted higher likelihood of eventual discontinuation.

Wolitzky-Taylor et al. — Psychology of Addictive Behaviors · Jun 8, 2026

In patients co-prescribed opioids, about 60% achieved a 50%+ dose reduction whether or not CBT was added — but CBT produced a large improvement in anxiety during the taper (d = 0.86). Support helps you feel better while tapering, even when it doesn't change the dose curve.

Blazes et al. — Frontiers in Psychiatry (opinion) · May 26, 2026

A patient-advocacy-aligned critique arguing the ASAM guideline is ambiguous about when inpatient care is appropriate and risks people being sent to short detox stays without an outpatient taper plan — the exact failure mode the community has warned about for years.

Oldenhof et al. — Basic & Clinical Pharmacology & Toxicology · Mar 6, 2026

38 GPs ranked what they need to deprescribe safely: longer consultations, a written tapering protocol, and practice-level agreement. Useful evidence that many prescribers themselves feel under-equipped — worth remembering if yours seems unsure rather than unwilling.

Shade, Ritvo, Silvernail, Finlayson et al. — PLoS One · Aug 28, 2025

Across 46 studies with outcomes measured at least four weeks after cessation, roughly 61% contained signals of protracted symptoms — cognitive impairment, tinnitus, paresthesia — often recorded incidentally. The strongest literature-wide evidence yet that long symptoms after stopping are real and under-studied.

Ritvo et al. — PLoS One · 2023

The survey of 1,200+ current and former users that gave benzodiazepine-induced neurological dysfunction (BIND) its name, documenting symptoms persisting a year or more after cessation in a substantial minority.

Zeraatkar et al. — The BMJ (systematic review + meta-analysis, 49 RCTs) · Jun 17, 2025

The sobering one: across 39,000+ trial patients, patient education, medication review, and pharmacist-led education may increase discontinuation (all low-certainty), and no intervention had compelling evidence of improving patient outcomes. A reminder that this whole field runs ahead of its trial evidence — and that claims of a 'proven' method deserve skepticism.