How to Taper Benzos Without Making Things Worse

Expert consensus: go slow, go gradual, and let the patient lead

STUDY: Brandt J et al, Ther Adv Psychopharmacol 2026
STUDY TYPE: Modified Delphi consensus study
FUNDING: Independent

Background

Standard taper protocols lower benzodiazepines by 25–50% every one to two weeks. They work for some patients but provoke severe withdrawal in others. This consensus study of 35 clinicians and patient advocates attempts to define best practices, focusing on patients who’ve been on them long-term or who failed standard tapers.

Results

The panel reached strong consensus on these core principles:

• Obtain informed consent before starting a taper, including education on withdrawal-induced akathisia and protracted withdrawal.
• Use gradual tapering as the foundation. For physically dependent patients, this can take months to years.
• Let the patient set the pace. The rate should flex with withdrawal symptoms, not a fixed calendar.
• Use hyperbolic dose reductions: cut by 10% or less of the previous dose each month, so reductions get smaller as the dose falls. Some patients may need final doses as low as 1% of the starting dose.
• Pause or revert to a prior dose if withdrawal becomes intolerable, then resume more slowly.
• Offer psychosocial interventions, ideally cognitive behavioral therapy, and peer support resources.

Switching to a longer-acting agent like diazepam reached only moderate consensus. Adding non-benzodiazepine medications reached weak consensus, with experts cautioning that few adjunctive agents have shown consistent benefit, and some may worsen symptoms in a sensitized nervous system.

The hyperbolic model is supported by pharmacologic theory, but lacks controlled trials. In practice, it often requires compounded liquid formulations; standard tablets don’t allow small enough cuts. Ask your compounding pharmacy about lorazepam 0.1 mg/mL or diazepam liquid. The Maudsley Deprescribing Guidelines and the Ashton Manual, both freely available, offer concrete taper schedules you can adapt.

Tapering in the Hospital

STUDY: Karavolis Z et al, Harvard Review of Psychiatry 2026 
STUDY TYPE: Narrative review
FUNDING: Independent

In a paper published the same week, researchers at Harvard reviewed the best course for inpatient benzodiazepine tapers.

They recommend converting to a long-acting benzodiazepine like chlordiazepoxide or diazepam to provide steady plasma levels and reduce rebound symptoms. However, some are unable to switch, particularly with alprazolam (which has unique receptor binding).

A sample inpatient schedule starts with chlordiazepoxide 50 mg every two hours as needed on day 1, then reduces the total daily dose by 50% each subsequent day.

Phenobarbital is a useful alternative, especially for patients with complex persistent benzodiazepine dependence, short-acting benzodiazepine use, or multiple prior taper failures. Its long half-life creates a natural self-tapering effect. They recommend weight-based loading (10 mg/kg of ideal body weight), given intramuscularly in divided doses or intravenously. Unlike long-acting benzodiazepines, phenobarbital doesn’t accumulate unpredictably and doesn’t require daily dose adjustments.

Adjunctive agents have supporting but limited evidence:
• Gabapentin and pregabalin reduce autonomic hyperactivity and may lower the total benzodiazepine dose needed
• Clonidine and dexmedetomidine ease autonomic symptoms but lack strong benzodiazepine-specific data
• Carbamazepine and valproate have stronger evidence for alcohol withdrawal than for benzodiazepines

Cognitive-behavioral therapy is the behavioral intervention with the strongest evidence; it outperforms supportive counseling when combined with a slow taper.

Protracted withdrawal — sleep disruption, fatigue, difficulty concentrating, nerve pain — can last months to years after acute detoxification.

Practice Implications

1. This consensus of experts with lived experience and clinicians provides a nice balance to the 2025 expert-led guidelines. Those guidelines tried to take an evidence-based approach, but the evidence in this scenario is sorely lacking.
2. Don’t push discontinuation as the only goal. Some patients stabilize at a low dose and can’t go further, and that’s a reasonable outcome.
3. Here’s a summary of guidelines on deprescribing stimulants and other psychotropics.