CoChat

Pharmacological principles for safe benzodiazepine and Z-drug dose reduction

Open access

The DOI matches its Crossref record; 93 of the 93 references it cites were found in a registry or an index; 49 of the 156 sources its claims cite were checked for support.

Abstract

Many guidelines recommend deprescribing benzodiazepines and z-drugs in most long-term users. Due to physical dependence in those using medication as prescribed (distinct from addiction), discontinuation can be difficult and is often unsuccessful, as withdrawal symptoms can be severe and long-lasting, especially after long-term use. There is a lack of clarity on how to taper patients in a tolerable manner that minimizes discomfort. Current guidelines vary with some suggesting linear reductions (e.g. 1 mg every 1-4 weeks) and some proportionate reductions (e.g. 5%-10% of the most recent dose per month, with smaller reductions as the dose decreases). The shape of the relationship between dose of benzodiazepine and activity at gamma-aminobutyric acid-A (GABA-A) receptors is hyperbolic: steeply inclining at low doses, flattening at higher doses. Consequently, linear dose reductions produce increasingly large changes in receptor occupancy, predicting escalating withdrawal effects, while hyperbolic or proportionate dose reductions produce linear reductions in pharmacological effect. Slower tapering (over months and years) may be more successful than tapering over days or weeks. In practice, rate of tapering should be adjusted according to severity of withdrawal effects. Final doses for some people will need to be very small (equivalent to as little as 0.2 mg of diazepam, or less) so that the last 'step down' to zero is not larger than prior tolerated reductions in terms of pharmacological effects. Liquids or compounded formulations of medication can be helpful at low doses to make small reductions. Guidelines should recommend hyperbolic tapering while awaiting randomized trials comparing linear with hyperbolic tapering.

Overview

CoChat Summary

TL;DR

This narrative review argues that benzodiazepine and z-drug tapering should follow hyperbolic/proportionate dose reductions because the dose–GABA-A occupancy relationship is hyperbolic, so linear reductions produce increasingly large pharmacological effects and likely worsening withdrawal. It recommends slow, symptom-guided tapering with very small final doses, while noting that randomized trials comparing linear and hyperbolic tapering are still lacking.

Key findings

  • Benzodiazepine dose and GABA-A occupancy are hyperbolically related: reducing diazepam by fixed 12.5 mg steps from 50 mg produces progressively larger occupancy drops—6.6 percentage points (50 to 37.5 mg), 10.0 p.p. (37.5 to 25 mg), 16.7 p.p. (25 to 12.5 mg), and 33.6 p.p. (12.5 to 0 mg). At low doses, 5 to 0 mg changes occupancy by 16.8 p.p., exceeding the 13.3 p.p. change from 100 to 50 mg.
  • Linear tapering is often unsuccessful: one trial found 90% of linearly tapered patients experienced withdrawal symptoms, with 32% on long half-life benzodiazepines and 42% on short half-life benzodiazepines unable to cease medication; only 36% successfully stopped after a 4-week taper at 15-month follow-up. A systematic review of 35 studies found best results with a 10-week taper, though 31% still could not stop.
  • Hyperbolic/proportionate tapering—approximately 5%–10% of the most recent dose every 1–2 weeks—produces roughly linear reductions in GABA-A occupancy (about 1–2 percentage points per step) and is recommended over linear tapering. Final doses may need to be very small, equivalent to as little as 0.2 mg diazepam or less, so the last step to zero is not larger than previously tolerated reductions.
  • Adding psychological support to tapering increased cessation rates (OR 1.82) with sustained benefit at follow-up (OR 1.88), whereas gradual tapering alone had a larger effect versus routine care (OR 6); no medication is approved for benzodiazepine withdrawal, and the ASAM panel found no single medication with enough data to recommend.

Methods

  • Narrative review of benzodiazepine withdrawal syndrome and commonly used tapering guidance from UK and US authorities, plus review of basic benzodiazepine pharmacology and withdrawal neurobiology.
  • Targeted literature searches on PubMed and Google Scholar for relevant domains; principles for tapering were developed from this understanding.
  • Focused on patients taking benzodiazepines as prescribed without benzodiazepine use disorder, stated to account for 98% of benzodiazepine users; management of addiction was outside scope.

Limitations

  • The paper states that PET imaging data used to model the dose–occupancy curve come from non-human primates rather than humans, represent single doses and population averages, and obscure inter-individual variability in receptor sensitivity, binding affinity, and pharmacodynamic response.
  • The paper states that no empirical studies have directly tested linear versus hyperbolic tapering; the recommendation remains a biologically plausible inference from receptor theory and clinical experience.
  • The paper states that modeling assumes steady-state pharmacokinetics and linear elimination, which may be disrupted by illness, polypharmacy, or metabolic variability; evident from the text, this is a narrative review without a randomized trial or systematic meta-analysis comparing tapering strategies.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1017/s0033291726104887

Cites

93 of 93 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. Different approaches to benzodiazepine tapering regimens recommended by NICE approximate hyperbolically reducing schedules. (a) A linear pattern of dose reduction is shown, involving a fixed percentage (in this case 25%) of the original dose reduced every week, as in many guidelines (McCormack et al., 2015; Permanente, 2019; Pottie et al., 2019). (b) The pattern of dose reduction recommended for tapering of diazepam from NICE and ASAM (for some patients) guidelines (of 5%–10% of the most recent dose each month) (Brunner et al., 2025; Nice, 2018). This pattern closely resembles a decreasing hyperbola. (c) cumulative reduction of diazepam per week as recommended by NICE guidance is represented (Nice, 2018). Note the similarity in the shape of this curve to the relationship between dose of benzodiazepines and both clinical and biological effects (other Figures). Figure 1. long description.
Figure 1. Different approaches to benzodiazepine tapering regimens recommended by NICE approximate hyperbolically reducing schedules. (a) A linear pattern of dose reduction is shown, involving a fixed percentage (in this case 25%) of the original dose reduced every week, as in many guidelines (McCormack et al., 2015; Permanente, 2019; Pottie et al., 2019). (b) The pattern of dose reduction recommended for tapering of diazepam from NICE and ASAM (for some patients) guidelines (of 5%–10% of the most recent dose each month) (Brunner et al., 2025; Nice, 2018). This pattern closely resembles a decreasing hyperbola. (c) cumulative reduction of diazepam per week as recommended by NICE guidance is represented (Nice, 2018). Note the similarity in the shape of this curve to the relationship between dose of benzodiazepines and both clinical and biological effects (other Figures). Figure 1. long description.CC BY · PubMed Central
Box 2. long description.
Box 2. long description.CC BY · PubMed Central
Figure 2. The differential effects of tapering benzodiazepines according to a linear or hyperbolic pattern of reduction. (a) The relationship between dose of diazepam and action at GABA-A receptors. Note the hyperbolic relationship between dose and estimated GABA-A occupancy. This graph was adapted from Brouillet et al. (1991), a study conducted in baboons, with dosages converted to human equivalents, as outlined in Nair and Jacob (2016). (b) Linear dose reductions of diazepam cause hyperbolically increasing reductions in effect at the GABA-A receptor. This may be associated with increasingly severe withdrawal effects. (c) Hyperbolically decreasing dose reductions produce linear reductions in effect at GABA-A receptors. This may be associated with more ‘evenly spread’ withdrawal effects. Figure 2. long description.
Figure 2. The differential effects of tapering benzodiazepines according to a linear or hyperbolic pattern of reduction. (a) The relationship between dose of diazepam and action at GABA-A receptors. Note the hyperbolic relationship between dose and estimated GABA-A occupancy. This graph was adapted from Brouillet et al. (1991), a study conducted in baboons, with dosages converted to human equivalents, as outlined in Nair and Jacob (2016). (b) Linear dose reductions of diazepam cause hyperbolically increasing reductions in effect at the GABA-A receptor. This may be associated with increasingly severe withdrawal effects. (c) Hyperbolically decreasing dose reductions produce linear reductions in effect at GABA-A receptors. This may be associated with more ‘evenly spread’ withdrawal effects. Figure 2. long description.CC BY · PubMed Central
Figure 3. The relationship between dose of benzodiazepine or z-drug and many biological and clinical effects is hyperbolic. (a) Diazepam produces a hyperbolic relationship between benzodiazepine-site concentration and GABA current in human recombinant GABA-A receptors. Adapted from Atack (2009). (b) Triazolam produces a hyperbolic relationship between dose and increase in seizure threshold in non-human primates. Adapted from Bottlaender et al. (1994). (c) Lorazepam demonstrates a hyperbolic relationship between dose and degree of anxiolysis. Anxiolysis was measured by mean percentage of open arm entries on the elevated plus-maze in mice. Adapted from Dubinsky et al. (2002). (d) Zolpidem produces a hyperbolic relationship between dose and total sleep time in human subjects. Adapted from Merlotti et al. (1989). Figure 3. long description.
Figure 3. The relationship between dose of benzodiazepine or z-drug and many biological and clinical effects is hyperbolic. (a) Diazepam produces a hyperbolic relationship between benzodiazepine-site concentration and GABA current in human recombinant GABA-A receptors. Adapted from Atack (2009). (b) Triazolam produces a hyperbolic relationship between dose and increase in seizure threshold in non-human primates. Adapted from Bottlaender et al. (1994). (c) Lorazepam demonstrates a hyperbolic relationship between dose and degree of anxiolysis. Anxiolysis was measured by mean percentage of open arm entries on the elevated plus-maze in mice. Adapted from Dubinsky et al. (2002). (d) Zolpidem produces a hyperbolic relationship between dose and total sleep time in human subjects. Adapted from Merlotti et al. (1989). Figure 3. long description.CC BY · PubMed Central

Verification

Claims the paper makes

35 Valid · 5 Not valid · 57 Not checked · 26 Own results and interpretations

77 more claims · 22 with their cited sources checked

Create a free account to see all 123 claims and to verify the 2 references still open.

See all claims — free account
  • It may state two separate claims as one. 33
  • Probably not a substantial claim. 9

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

4 of 4

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

1 of 1

Citations

Cites

1 retracted

Research works printed but not linked: 11

  1. White, 2015, Handbook of drug administration via enteral feeding tubes (third)
  2. Dhaliwal, 2025, Diazepam
  3. Parsons, 2012, Dependence on benzodiazepines or Z-drugs: Having that conversation, The Pharmaceutical Journal, 289, 399
  4. Dubinsky, 2002, Nonbenzodiazepine anxiolytic, Pharmacology, 303, 777
  5. 2018, MOC guidelines for tablet crushing in patients with swallowing difficulties
  6. Guy, 2019, Guidance for psychological therapists: Enabling conversations with clients taking or withdrawing from prescribed psychiatric drugs
  7. Shapiro, 2025, The psychopharmacology blue book
  8. 2004, Guidance on the use of zaleplon, zolpidem and zopiclone for the short-term management of insomnia
  9. Root, 2011, Pharmaceutical issues when crushing, opening or splitting Oral dosage forms, Royal Pharmaceutical Society, 1, 1
  10. Horowitz, 2024, The Maudsley deprescribing guidelines in psychiatry: Antidepressants, benzodiazepines, Gabapentinoids and Z-drugs
  11. Harriman, 2014, Deprescribing medication for frail elderly patients: Survey of Vancouver family physicians, Canadian Family Physician Medecin de famille canadien, 60, S16

Metadata

Authors
Horowitz, M., Lamberson, N., Brandt, J., Framer, A., Shapiro, B., Sørensen, A., Cadogan, C., Ritvo, A., & Taylor, D.
Affiliations
  • North East London NHS Foundation Trust (GB)
  • Benzodiazepine Information Coalition
  • University of Manitoba (CA)
  • Psychotropic Deprescribing Council
  • University of California, Irvine (US)
  • Rigshospitalet (DK)
  • Trinity College Dublin (IE)
  • University of Colorado Anschutz (US)
  • Columbia University (US)
Corresponding author
Horowitz, Mark
Venue
Psychological Medicine
Published
2026
Volume
56
Pages
e198-e198
ISSN
  • 0033-2917 (ISSN-L)
  • 1469-8978
OpenAlex
W7165192235
Cited by
0
References
81
FWCI
0.00
Type
journal article
Language
en
Open access
Locations
  • Psychological Medicine (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
Topics
  • Health Sciences › Medicine › Geriatrics and Gerontology › Pharmaceutical Practices and Patient Outcomes
  • Health Sciences › Medicine › Epidemiology › Substance Abuse Treatment and Outcomes
  • Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Opioid Use Disorder Treatment
Keywords
benzodiazepines, Z-drugs, deprescribing, dose reduction, tapering, withdrawal symptoms, physical dependence, chronic drug administration, compounded formulations
MeSH terms
Deprescriptions, Drug Tapering, Benzodiazepines, Dose-Response Relationship, Drug, Humans, Substance Withdrawal Syndrome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.