Key Takeaways
- Some withdrawal symptoms after a dose cut are expected. A cut that leaves you unable to sleep or think in a straight line tells you the rate was too fast, and it says nothing about you.
- There are three different responses: a hold keeps the dose where it is, a step back returns to the last dose you tolerated, and full reinstatement means going back to the original dose or restarting a benzodiazepine that was stopped.
- The Ashton Manual says never go backwards: you can stand still at a difficult stage for a few weeks, but try to avoid ever increasing the dose again.
- A hold is the first response and often the only one needed, because it gives up no ground and lets the symptoms from the last cut build and then ease.
- When a hold does not restore a livable baseline, some prescribers step back a single step to the last stable dose, and that decision belongs to you and your prescriber. The 2020 benzodiazepine label update from the Food and Drug Administration and the Maudsley Deprescribing Guidelines both describe that sequence.
- Restarting a benzodiazepine weeks after stopping is unpredictable, because a nervous system that has been through withdrawal and re-exposure may react more strongly, so it is a decision for a physician who knows your history.
- Once you are steady again, the next cut should be smaller than the one that caused the trouble and the hold after it longer: the Maudsley Deprescribing Guidelines say to resume at half the previous rate or less.
- A seizure, severe confusion, or thoughts of suicide are emergencies: call 911 or your local emergency number, and in the United States the 988 Suicide and Crisis Lifeline answers calls and texts at any hour.
Physical dependence on a prescribed benzodiazepine is not addiction; it is the expected result of regular use, and a slow taper exists to unwind it.
What happens when a taper cut goes badly?
A dose reduction that brings ordinary withdrawal symptoms is expected. A reduction that brings a person to a standstill, unable to sleep or to think in a straight line, is a different event, and it raises the question this article answers: should the dose go back up?
There are three ways to respond, and they are not the same thing. A hold keeps the dose where it is until the nervous system catches up. A step back returns to the last dose the person tolerated. Full reinstatement means returning to the original dose, or restarting a benzodiazepine that was stopped weeks ago.
A person who took a benzodiazepine as prescribed and now cannot reduce it without symptoms is physically dependent, not addicted, and going back up is a question about a nervous system that has adjusted to a drug, not about willpower. So, what does the Ashton Manual say about going back up? It says do not.
The Ashton Manual says stand still, never go backwards.
Professor C. Heather Ashton put her position in plain words in Chapter II, in the section on designing the withdrawal schedule: “As far as possible, never go backwards. You can stand still at a certain stage in your schedule and have a vacation from further withdrawal for a few weeks if circumstances change (if for instance there is a family crisis), but try to avoid ever increasing the dosage again.” The same chapter says it “is important in withdrawal always to go forwards,” and that at a difficult point “you can stop there for a few weeks if necessary.”
Every reduction asks the nervous system to adjust, and going back up hands that adjustment back and lengthens the road. The Manual is not harsh about the person who did go back: Chapter III notes that people who have gone back on benzodiazepines after a failed first attempt “can be just as successful at tapering as first-timers.”
The Manual is this article’s primary reference, and it argues against what the title describes. What follows on stepping back is Ashton Manual Guide’s own account of later practice, not something Professor Ashton said.
A hold is the first response, and often the only one needed.
A hold means keeping the dose exactly where it is and cancelling the next reduction until the person is steady again. In fact, it is the Manual’s own remedy for a hard stretch, and it costs nothing, because no ground is given up. Symptoms after a cut tend to build over the first days and then ease, and a hold gives that curve room to finish.
How long is a hold? As long as it takes. Professor Ashton wrote of “a few weeks,” and the useful test is function rather than comfort: sleeping some of the night, eating, getting through a day. This site’s article on holding periods between cuts covers the stabilization a hold is meant to produce.
Some prescribers step back to the last tolerated dose. That is a later development.
When a hold does not bring a person back to a livable baseline, some prescribers return the dose to the last level at which the person was stable. The situation they have in mind is not discomfort but a person who has not slept in a week, cannot keep food down, or cannot safely be left alone, after a cut larger than the nervous system could absorb.
Of course, the step has limits. It is one step, to a dose already known to work, not a percentage worked out at the kitchen table. Whether to take it, and when, are decisions for you and your prescriber.
The 2020 United States Food and Drug Administration (FDA) benzodiazepine label update points the same direction. It required a gradual taper, warned that stopping abruptly can bring life-threatening seizures, and the labeling revised under it tells prescribers that when withdrawal reactions develop, the taper may be paused or the dose returned to the previous step before reductions resume more slowly. The Maudsley Deprescribing Guidelines describe the same sequence, stabilizing at the last tolerated dose and then proceeding by smaller steps.
Ashton Manual Guide reads this as consistent with Professor Ashton’s principles. Her method rests on a pace chosen by the person, “you must be in control and must proceed at the pace that is comfortable for you,” as Chapter II puts it, and a single step back to a tolerated dose, followed by smaller cuts, serves that principle when a hold alone has not.
Full reinstatement is rarely wise, and restarting after weeks off is unpredictable.
There is a difference between a step and a retreat. While a step back to a tolerated dose gives up little, going back to the starting dose surrenders every adjustment the nervous system has made over months of tapering, and it rarely buys more relief than the smaller step would. Only after a very fast taper, such as one measured in days, is the last tolerated dose likely to be near the original, and even then it is a step back with a slower plan behind it.
Reinstatement in the strict sense, restarting a benzodiazepine that has been stopped completely, is a different decision. Some people who restart within days feel relief. Unfortunately, the longer the gap, the less reliable the response becomes, and some find that the drug no longer settles the symptoms it once did, or brings new ones.
The pattern behind that has a name, kindling, and this site’s article on benzodiazepine kindling explains it: a nervous system that has been through withdrawal and re-exposure may react more strongly to the next reduction. Kindling is not in the Manual, whose encouraging view of the second attempt still stands. Both can be true, which is why restarting after weeks off belongs with a physician who knows the history.
After stabilizing, the next cuts are smaller. That is the hyperbolic shape.
Once a person is steady again, whether after a hold or a step back, the next reduction should be smaller than the one that caused the trouble, and the hold after it longer. And, the Manual already leans this way. Chapter II suggests reducing “by up to one tenth at each decrement,” and its own schedules shrink the step as the dose falls, from 2 mg steps at 40 mg of diazepam to 1 mg steps from 20 mg.
Hyperbolic tapering, the shape the Maudsley Deprescribing Guidelines describe, takes that instinct to its conclusion. Each cut is a fraction of the current dose rather than of the starting dose, so the cuts shrink as the dose does, because the low end of the dose range does a disproportionate share of the work at the receptor.
Think of walking down a hill that gets steeper near the bottom: the steps have to get shorter as the slope increases, or the walker ends up running. The site’s benzodiazepine taper calculator draws that shape for a given starting dose, as a planning picture to bring to a prescriber, never as a schedule to follow alone.
If you have already gone back up.
Many people arrive at this page after the fact. They raised the dose during a hard week, felt better, and now regret it, or wonder whether the taper is ruined. It is not ruined, and the regret is common.
The Manual comes first here too: Professor Ashton would have preferred that the dose stand still, and that is the direction for the next difficult stretch. Yet, the dose has already gone up, and the useful question is how to move forward from here: hold at the new dose until symptoms are stable, then resume with smaller steps and longer holds than before.
One option to raise with the prescriber is a partial return, not all the way back down to the dose that failed, but somewhere between it and the one that worked. Whether that fits depends on how far the dose went up and how the person is doing now. It is a conversation, not a rule.
Which benzodiazepine is involved changes the arithmetic.
A step back is usually taken in the same drug the person is tapering, because switching drugs during an unstable week adds a variable. The exception is a short-acting benzodiazepine producing interdose withdrawal, meaning symptoms between doses as each dose wears off, where a prescriber may prefer a crossover taper to diazepam, the switch Chapter II of the Manual advises for drugs such as alprazolam and lorazepam.
Table 1 of the Manual, in Chapter I, gives the equivalences a prescriber would use. Diazepam 10 mg is listed as roughly equivalent to clonazepam 0.5 mg, alprazolam 0.5 mg, and lorazepam 1 mg. The same table gives half-lives of 20 to 100 hours for diazepam (36 to 200 hours counting its active metabolite), 18 to 50 hours for clonazepam, 10 to 20 hours for lorazepam, and 6 to 12 hours for alprazolam.
Professor Ashton’s note under the table matters as much as the numbers: the equivalents “are firmly based on clinical experience but may vary between individuals.” A conversion from a table is a starting point for a prescriber, not a dose for a reader.
When this is an emergency.
Some symptoms get listed as reasons to updose, and they are not. A seizure, sudden confusion or severe confusion, or thoughts of suicide are medical emergencies: call 911 or your local emergency number. In the United States, the 988 Suicide & Crisis Lifeline answers calls and texts at any hour.
The Manual is reassuring on the rarest of these. Chapter III notes that withdrawal fits are “extremely rare with slowly eliminated benzodiazepines (e.g. diazepam) or with slow dosage tapering,” which is one more reason never to stop a benzodiazepine abruptly.
Frequently asked questions
Does going back up mean starting the taper over?
No. A step back to the last tolerated dose means the taper resumes from a dose already below the starting point, with smaller cuts from there. What is spent is time, and covering the same ground twice is what Professor Ashton’s warning is about.
Can a benzodiazepine be restarted after being off it for weeks?
Sometimes, and the response is less predictable the longer the gap has been. Some people get relief, some get little, and some react more strongly to the next reduction, so it is a decision for a physician who knows the history.
How can withdrawal be told apart from the original anxiety?
Withdrawal tends to arrive in waves, to bring physical symptoms such as muscle tension, tremor, and sensory disturbance, and to include things the person never felt before the drug. The original condition usually looks like it did before, and an experienced physician can help sort one from the other.
Hold first, one step back at most, then smaller cuts.
The Manual stops at the first step, and the 2020 FDA label update and the Maudsley Deprescribing Guidelines add the second. A cut that goes badly is information about the rate, not a verdict on the person.
Whether to taper, how fast, and whether to step back are decisions for you and your prescriber. Fortunately, the nervous system that protested the last cut is the same one that will adjust to the next, smaller one, given the time it asks for.
This article is educational. It is not medical advice, and reading it does not create a doctor-patient relationship. Decisions about starting, continuing, or tapering any medication should be made with your own physician.
