The concentration ladder
Tapering does not fail at the beginning. It fails at the bottom, in the last few milligrams, where the tablet runs out before the patient does. The ladder is how we get past that.

The last milligram is the hardest
A patient coming off 20 mg of diazepam usually does fine for the first several reductions. Twenty to fifteen to ten — those are comfortable steps. The trouble starts near the end, when the dose is small and the tablet cannot be divided any further.
The smallest diazepam tablet made in the United States is 2 mg. Split it in half and you are at 1 mg, and then the tablet runs out. What follows is the sentence that ends a great many tapers: "You are on such a small dose now, you can just stop."
That instruction fails for a reason that is pharmacological, not psychological.

Why equal steps are not equal steps
The relationship between the dose of these medications and their effect at the receptor is not a straight line. It is a curve — steep at the bottom, saturating at the top.
At high doses the receptor is progressively less responsive, so removing a few milligrams changes very little..
At low doses every milligram is doing a great deal of work. The same 1 mg reduction a patient did not notice at 20 mg can be, at 1 mg, the single largest change in receptor occupancy of the entire taper — delivered at the very end, when the nervous system has the least reserve left to absorb it.
The evidence behind that curve is strongest for the antidepressants, where human imaging has measured receptor occupancy directly across the dose range. For the benzodiazepines the curve is modelled from receptor theory and animal imaging rather than measured in people. The shape is well supported for both; the precise numbers are established only for the first.
This is why tapers that look sensible on paper get harder rather than easier as they go. "Reduce by 2 mg every two weeks" is a linear schedule laid over a non-linear pharmacology. Patients report the pattern consistently.

A dose reduction that is trivial at the start of a taper can be the largest change of the entire taper at the end.
 

What a gradual taper actually requires
If each step is going to feel the same to the patient, each step has to remove the same proportion of what remains — not the same number of milligrams. That produces a schedule in which the milligram steps get progressively smaller as the dose approaches zero.
The arithmetic is unforgiving. Take a patient at 20 mg of diazepam reducing by 10% of the current dose at each step, which sits inside the range current US clinical practice guidance describes. Reaching 0.1 mg takes roughly fifty steps, and the final steps are on the order of 0.01 mg. That is two hundred times smaller than the smallest tablet made. No tablet, no pill cutter, and no single-concentration liquid can deliver it. 

This is not only our reading of the problem. The 2025 joint clinical practice guideline on benzodiazepine tapering — developed by the American Society of Addiction Medicine with nine partner societies under an FDA grant — records the same constraint. It notes that once a patient reaches the lowest available strength, reducing by 5% or 10% becomes difficult, and states plainly that “the availability of a greater range of low-dose formulations for commonly prescribed BZDs would help facilitate BZD tapering.”



One liquid is not enough either
Liquid looks like the obvious answer, and for most of these drugs an oral solution or concentrate already exists. Diazepam has both — an approved oral solution at 1 mg per mL and a concentrate at 5 mg per mL. This only solves half the problem.

Escitalopram oral solution is 1 mg per mL. That is fine at 10 mg — ten millilitres, easy to handle. It is not fine at 0.05 mg, which is five hundredths of a millilitre: a drop sitting in the tip of a syringe, where the measurement error is larger than the dose itself.
Precision in liquid dosing is a function of volume, not of milligrams. To deliver a very small dose accurately, the dose has to arrive as a volume large enough to measure. One concentration cannot span a four-hundred-fold dose range and stay accurate at both ends of it.
And for some of the drugs that matter most, no approved liquid exists at all. There is no FDA-approved oral liquid formulation of venlafaxine in the United States — a drug repeatedly identified in the literature as among the hardest antidepressants to stop.

The ladder
So each drug is built as a set of concentrations — a ladder — designed so that every part of the taper is dosed at a concentration that keeps the delivered volume inside a comfortable, accurate range.

The patient stays on the same molecule the whole way down. No cross-taper to a different benzodiazepine, no switch to a longer half-life drug part-way through, per patient compounded preparation of uncertain potency. One drug, one platform, a continuous path from the therapeutic dose to zero.
 

What this means for the patient

• Steps sized to the pharmacology, not to the tablet. Sizing each step to the pharmacology rather than to what the tablet allows is designed to stop the patient hitting a cliff at the end of the taper.

• Withdrawal is easier to tell apart from relapse. Small, even steps are intended to make it easier to tell a withdrawal reaction from the return of the underlying condition — and that distinction decides whether a patient is put back on the medication for another year.

• The taper can be slowed, held, or reversed. A clinician who sees a patient struggling can hold the current dose or step back a fraction. With tablets, the nearest available option is often twice as far away as the one you want.

• Zero becomes reachable. The last stretch of the dose range is where tapers break down. That is precisely the range the ladder is built for.

• A manufactured product, not a preparation. The active ingredient is one the FDA has already approved and clinicians already know. What is new is the formulation and the concentration range, and that formulation is investigational. It is made under cGMP in a registered facility — not compounded per patient, not an unapproved import, not a tablet quartered on a kitchen counter.

Where we are
Extalvo is developing these formulations under the FDA’s 505(b)(2) pathway, which is built for new formulations of already-approved active ingredients. Twelve programs across five drug classes are in development, with IND-enabling work underway on the four flagship programs. The full pipeline and its stage gates are on the Pipeline page.

Extalvo’s products are investigational and have not been approved by the FDA. Nothing on this page is medical advice. Tapering should be undertaken only under the supervision of a prescribing clinician, and benzodiazepines and antidepressants should never be stopped abruptly.