Benzodiazepine Equivalency: A Clinician's Reference for Cross-Tapering

Benzodiazepine equivalency tables are routine reference tools when transitioning a patient from a short-acting agent to a long-acting one for the purpose of deprescribing. They are also among the most misapplied resources in psychopharmacology, because the published equivalencies were derived primarily from single-dose anxiolytic or hypnotic studies in benzodiazepine-naive volunteers, not from chronically tolerant patients undergoing withdrawal. The prescriber who treats these tables as a precise pharmacokinetic conversion — rather than a starting estimate that must be titrated against clinical response — risks producing iatrogenic over-sedation, breakthrough withdrawal, or a destabilised taper.

Why equivalency tables exist at all

The most widely cited benzodiazepine equivalency reference is the Ashton Manual (Heather Ashton, Benzodiazepines: How They Work and How to Withdraw, 2002, revised 2011). Its table is reproduced — with minor variations — in the Maudsley Prescribing Guidelines, the Maudsley Deprescribing Guidelines (Horowitz & Taylor, 2024), the BNF, and most institutional psychiatric formularies. The purpose of the table is narrow: it allows the prescriber to convert a patient's current benzodiazepine to an approximately equivalent dose of a long-half-life agent — almost always diazepam — so that gradual dose reduction can occur from a more pharmacokinetically forgiving substrate.

The clinical rationale is straightforward. Short-acting agents such as alprazolam (half-life 6–12 h), lorazepam (10–20 h), and oxazepam (4–15 h) produce repeated interdose troughs. In a tolerant patient, those troughs manifest as inter-dose withdrawal: rebound anxiety, autonomic arousal, insomnia, and sometimes perceptual disturbance several hours after each dose. Reducing the dose of a short-acting drug exaggerates the trough. Diazepam, by contrast, has a parent half-life of 20–50 h and an active metabolite (desmethyldiazepam, nordiazepam) with a half-life of 36–200 h. Plasma concentrations remain comparatively flat across the dosing interval, and small reductions produce a smoother decline in receptor occupancy.

The standard equivalency table

The figures below approximate diazepam 10 mg orally and are drawn primarily from the Ashton Manual, with cross-reference to the Maudsley Deprescribing Guidelines. They assume oral administration in a tolerant patient.

| Drug | Approximate dose equivalent to diazepam 10 mg | Elimination half-life (parent + active metabolites) | Common indication | ||-----------------------------------------------|------------------------------------------------------|-------------------| | Alprazolam | 0.5 mg | 6–12 h | Panic, anxiety | | Bromazepam | 5–6 mg | 10–20 h | Anxiety | | Chlordiazepoxide | 25 mg | 5–30 h (active metabolites 36–200 h) | Alcohol withdrawal, anxiety | | Clobazam | 20 mg | 12–60 h | Adjunctive anticonvulsant | | Clonazepam | 0.5 mg | 18–50 h | Panic, seizure | | Clorazepate | 15 mg | Active metabolites 36–200 h | Anxiety | | Diazepam | 10 mg | 20–50 h (metabolites 36–200 h) | Reference compound | | Estazolam | 1–2 mg | 10–24 h | Hypnotic | | Flunitrazepam | 1 mg | 18–26 h | Hypnotic (restricted) | | Flurazepam | 15–30 mg | Active metabolites 40–250 h | Hypnotic | | Lorazepam | 1 mg | 10–20 h | Anxiety, status epilepticus | | Lormetazepam | 1–2 mg | 10–12 h | Hypnotic | | Nitrazepam | 10 mg | 15–38 h | Hypnotic | | Oxazepam | 20 mg | 4–15 h | Anxiety, alcohol withdrawal | | Temazepam | 20 mg | 8–22 h | Hypnotic | | Triazolam | 0.5 mg | 2–5 h | Hypnotic | | Zolpidem (Z-drug) | 20 mg | 1.5–3 h | Hypnotic | | Zopiclone (Z-drug) | 15 mg | 5–6 h | Hypnotic | | Zaleplon (Z-drug) | 20 mg | 1 h | Hypnotic |

Several caveats apply to every row.

First, published equivalencies vary. Lorazepam, for example, is given as 1 mg ≈ diazepam 10 mg in the Ashton Manual but 0.5 mg ≈ diazepam 5 mg in some Maudsley references and 2 mg ≈ diazepam 10 mg in older anaesthetic equivalency literature. Differences of a factor of two between sources are common and reflect the underlying populations studied. Treat the table as a starting estimate, not as a chemical conversion factor.

Second, Z-drugs are included by convention but bind a subset of the GABA-A receptor pool (predominantly α1) rather than the broader α1/α2/α3/α5 binding of classical benzodiazepines. Cross-tolerance is incomplete. A patient on long-term zolpidem who is converted to diazepam may experience emergent symptoms not seen with benzodiazepine-to-benzodiazepine conversion, particularly daytime sedation from the longer half-life and re-emergence of anxiety symptoms that the α1-selective hypnotic was not suppressing.

Third, the equivalency assumes oral dosing. Sublingual lorazepam, intramuscular midazolam, and rectal diazepam have different bioavailabilities and onset profiles that make equivalency conversion clinically meaningless outside of acute care.

Why the numbers should never be applied mechanically

The equivalency figures derive from anxiolytic potency studies, not from receptor occupancy in tolerant patients. Several pharmacological realities distort the conversion in practice:

• Tolerance is not uniform across receptor effects. Tolerance to the sedative effect of benzodiazepines develops within days to weeks. Tolerance to the anxiolytic effect develops more slowly and incompletely. Tolerance to the anticonvulsant effect develops over months. A patient stabilised on clonazepam for panic disorder is not pharmacodynamically equivalent to a benzodiazepine-naive volunteer receiving an equipotent dose of diazepam.
• Receptor downregulation and subunit composition shift with chronic exposure. GABA-A receptor density and subunit expression change measurably after weeks of exposure (Vinkers & Olivier, Adv Pharmacol Sci, 2012). The clinical implication is that the substituted drug may produce a different subjective effect at the "equivalent" dose, even before any reduction is attempted.
• Lipophilicity governs onset and subjective reinforcement. Diazepam and alprazolam are both highly lipophilic and produce a recognisable peak after oral dosing. Oxazepam, lorazepam, and clonazepam are less lipophilic and produce a slower-onset, less reinforcing effect. A patient transitioned from alprazolam to an equipotent diazepam dose may report loss of effect not because the diazepam dose is too low, but because the perceived peak is absent.

For these reasons, the recommended starting practice when converting is to use the equivalency table as the upper bound of the diazepam dose, then titrate downward over the first one to two weeks based on symptom report and sedation score. Most patients stabilise on a diazepam dose somewhat lower than the table predicts.

Cross-tapering versus direct switching

There are two practical approaches to moving a patient from a short-acting agent to diazepam.

Direct substitution replaces the entire dose at once. This is appropriate only when the original drug has been used for a short period (weeks rather than months), when the equivalent diazepam dose is modest, and when the prescriber can review the patient within days. Direct substitution from high-dose alprazolam to diazepam in one step is rarely tolerated; abrupt removal of the alprazolam peaks produces interdose withdrawal even though total receptor occupancy is preserved.

Cross-tapering is the standard for chronic, higher-dose, or shorter-half-life regimens. A typical staged approach:

1. Calculate the total diazepam-equivalent dose using the table.
2. Substitute one dose of the original drug at a time (commonly the night-time dose first, since diazepam's longer half-life is well tolerated overnight) with the equivalent diazepam dose.
3. Allow several days at each step before substituting the next dose.
4. Continue until the patient is on diazepam monotherapy.
5. Only then begin gradual dose reduction.

Cross-tapering avoids two predictable failure modes: dumping the patient into withdrawal because the diazepam has not yet reached steady state, and oversedating the patient because both drugs are present at full dose during overlap.

The Maudsley Deprescribing Guidelines emphasise that cross-tapering and dose reduction should not occur simultaneously. Cross-tapering itself is a pharmacodynamic perturbation, and superimposing reduction obscures the source of any emergent symptom.

When diazepam is the wrong destination

Diazepam is the default substitute for sound pharmacokinetic reasons, but it is not universally appropriate.

• Hepatic impairment. Diazepam, chlordiazepoxide, and clorazepate undergo Phase I oxidative metabolism (CYP2C19, CYP3A4) and accumulate in cirrhosis or in patients on potent CYP3A4 inhibitors. Lorazepam, oxazepam, and temazepam undergo glucuronidation only ("LOT" mnemonic) and are preferred in hepatic impairment, even at the cost of shorter half-life.
• Older adults. Long half-life agents accumulate over days. The American Geriatrics Society Beers Criteria explicitly recommend against initiating long-half-life benzodiazepines in older adults due to fall and fracture risk. A taper from lorazepam in an octogenarian may be safer using lorazepam itself, with smaller decrements taken more slowly, than via conversion to diazepam.
• Patients who have failed prior diazepam substitution. Some patients describe diazepam as subjectively ineffective or dysphoric. In these cases, cross-tapering to clonazepam (half-life 18–50 h) is a reasonable alternative; clonazepam offers most of the pharmacokinetic smoothing without diazepam's accumulation profile.
• Z-drug discontinuation. Conversion of zolpidem or zopiclone to diazepam is controversial. Some clinicians taper the Z-drug directly using compounded reductions; others substitute partially to diazepam to facilitate slower decrements. Evidence is thin and individualised judgment dominates.

Special populations and confounders

Pregnancy. All benzodiazepines cross the placenta. Long-half-life agents accumulate in the fetus, and floppy-infant syndrome and neonatal withdrawal are documented after third-trimester exposure. Equivalency conversion in pregnancy is a specialist decision; the question is rarely "which benzodiazepine" but "is continuation justified at all."

Alcohol use disorder. Active drinking distorts both pharmacokinetics (CYP induction, hepatic dysfunction) and pharmacodynamics (GABAergic cross-tolerance). Equivalency tables should not be used to plan ambulatory taper in a patient still drinking; abstinence or supervised withdrawal must precede deprescribing.

Concurrent opioid use. The FDA boxed warning on concurrent benzodiazepine and opioid prescribing reflects substantial respiratory-depression mortality. Equivalency conversion does not change combined CNS depressant burden; if anything, the longer half-life of diazepam prolongs the window of risk. Prescribers should not interpret "switching to diazepam" as a harm-reduction step in a patient on chronic opioids.

Polypharmacy with CYP3A4 inhibitors. Diazepam, alprazolam, midazolam, and triazolam are CYP3A4 substrates. Initiation of fluconazole, clarithromycin, ritonavir, or grapefruit juice during a taper can produce levels well above the equivalency-predicted exposure. Lorazepam, oxazepam, and temazepam are not affected.

Documentation and consent

The equivalency conversion should be documented explicitly in the clinical record: original drug and dose, table or source consulted, calculated diazepam-equivalent dose, planned cross-taper schedule, and any deviation from the standard equivalency. Patients should be informed that the equivalency is an estimate, that breakthrough symptoms during the substitution phase do not necessarily indicate that the dose is wrong, and that adjustment in the first one to two weeks is expected.

Clinical pearls

• Treat the published equivalency as a starting estimate and an upper bound, not as a chemical conversion. Most stabilised patients require a diazepam dose somewhat below the table value.
• Cross-taper one dose at a time, beginning with the night dose, and allow several days between substitutions before progressing. Do not reduce the total benzodiazepine load during cross-tapering.
• In hepatic impairment, in frail older adults, or when long-half-life accumulation is undesirable, taper the original drug rather than convert to diazepam. The "LOT" benzodiazepines (lorazepam, oxazepam, temazepam) bypass oxidative metabolism.
• Z-drugs are not interchangeable with classical benzodiazepines on a milligram-equivalent basis. Cross-tolerance is incomplete due to α1-selective binding.
• Re-confirm the equivalency source when the original prescription is high-dose, parenteral, or sublingual — bioavailability and route of administration distort the tabled values.
• Document the equivalency source consulted, the calculated dose, and the planned schedule. Equivalency tables vary by a factor of two between published sources, and the chosen reference should be auditable.

---

For more clinician resources on safe deprescribing and tapering, visit tapermeds.com.