TaperCommunity
For decades, psychiatric drug withdrawal was treated as a quick afterthought. Doctors halved doses, then halved again, then stopped within a few weeks. Patients who reported brutal symptoms were told it was their underlying illness returning. Then in 2019, two clinicians, Mark Horowitz and David Taylor, published a paper in The Lancet Psychiatry that quietly upended the field. Their horowitz taylor hyperbolic tapering research showed that the linear tapers everyone was using had no pharmacological logic. The relationship between drug dose and brain receptor occupancy is curved, not straight, which means equal milligram drops produce wildly unequal effects on the brain. This is now the single most important body of work in modern deprescribing.
This article walks through what the research actually says, why it matters, and how it informs the Maudsley Deprescribing Guidelines. You do not need a pharmacology degree to follow it. You do need to understand it if you are tapering an SSRI, an SNRI, or a benzodiazepine, because almost every doctor still operating from a 2010 mental model will get this wrong.
By the late 2010s, evidence was piling up that antidepressant withdrawal was more common, more severe, and longer lasting than the official guidance acknowledged. A 2019 systematic review by James Davies and John Read found that 56% of people experienced withdrawal effects when stopping antidepressants and nearly half rated those effects as severe. National guidance in the UK and US still described withdrawal as mild and self-limiting within one to two weeks.
The mismatch was not just academic. People were being misdiagnosed with relapse, restarted on medication they had been trying to leave, or told their symptoms were psychological. Online communities were full of patients describing waves of dizziness, electric shock sensations, emotional lability, and insomnia lasting months after a textbook taper.
Horowitz and Taylor approached this from a pharmacological angle. Mark Horowitz had personal experience tapering an SSRI and had found published guidance useless. David Taylor was lead author of the Maudsley Prescribing Guidelines, the most widely used psychiatric prescribing reference in the English-speaking world. They asked a basic question: if antidepressants work by occupying serotonin transporters in the brain, what does the dose-occupancy curve actually look like? The answer, drawn from existing PET imaging studies, was that it was sharply hyperbolic. This single observation explained why standard tapers were failing so many people.
A hyperbolic curve climbs steeply at low doses, then flattens. For most SSRIs, 80% of serotonin transporter occupancy is reached at roughly the minimum therapeutic dose. Doubling or tripling the dose beyond that adds only a few percentage points of occupancy. This is why higher doses often do not feel proportionally stronger.
Run that curve backwards and the implication is stark. For a drug like Lexapro, a cut from the top of the dosing range to half that dose may only nudge occupancy down a few points. The next halving moves occupancy a little more. But the final step from the smallest available pill down to zero strips off a huge share of receptor binding in one motion. In linear terms it looks like a tiny reduction. In neurobiological terms it is the largest change of the entire taper.
This is why so many people sail through the early reductions and crash at the end. The brain has been adapting to high receptor occupancy for years. Removing the last few milligrams strips off the largest functional dose the person has experienced in the taper. Horowitz and Taylor argued the only pharmacologically rational approach is to make reductions that produce approximately equal percentage drops in occupancy, which means progressively smaller milligram reductions as the dose gets lower.
The seminal paper, titled "Tapering of SSRI treatment to mitigate withdrawal symptoms," appeared in March 2019. It synthesized PET occupancy data for citalopram, escitalopram, sertraline, paroxetine, and fluoxetine and translated those curves into proposed taper schedules.
The reductions implied by the occupancy curves were dramatic compared to conventional practice. Final doses before stopping ended up being a small fraction of the lowest commercially available pill for drugs like citalopram and Zoloft. The implied tapers stretched over months for most people and over a year or more for those with long medication histories or prior failed tapers. Specific schedules belong to a prescriber working with the individual, but the broad picture is that the endpoint of a taper is much further down the dose curve than older protocols ever reached.
The paper did not invent hyperbolic tapering. Clinicians like Heather Ashton had been using non-linear reductions for benzodiazepines for decades, and the Ashton Manual remains a touchstone for benzo withdrawal. What Horowitz and Taylor did was apply the same pharmacological logic systematically to antidepressants, publish it in a high-impact peer-reviewed journal, and force the conversation into mainstream psychiatry. The paper became one of the most cited deprescribing references of the decade.
In 2024, Horowitz and Taylor published the Maudsley Deprescribing Guidelines, the first comprehensive deprescribing manual in the Maudsley series. It covers antidepressants, benzodiazepines, z-drugs, antipsychotics, and gabapentinoids, all built on the same hyperbolic framework.
The guidelines describe proportional reductions of the most recent dose as a reasonable starting framework for many people, slowing further if symptoms emerge. This is not a rigid prescription. It is a default that respects the curve. Some people taper faster, some need much slower reductions, and a substantial minority cannot complete a taper at all within standard timeframes. The guidelines make clear that flexibility and patient feedback drive the schedule, and that the specific numbers belong to a prescriber working with the individual.
Critically, the Maudsley framework treats withdrawal symptoms as pharmacological signals, not relapse and not psychological weakness. When symptoms appear, the protocol is to hold the dose, allow stabilization, and then resume reductions at a smaller increment. This is a sharp departure from older guidance that conflated withdrawal with returning illness.
The guidelines also describe practical issues like obtaining compounded liquid formulations, using compounding pharmacies to make sub-therapeutic doses, and how to read pill cutters and weighing scales accurately. The science would be useless without the logistical roadmap, and the book treats both as inseparable.
A second strand of Horowitz and Taylor's work concerns the nature and duration of withdrawal. A 2023 paper in Addictive Behaviors, co-authored with Davies and others, argued that protracted withdrawal lasting more than six weeks is plausible from a neuroadaptation standpoint and consistent with thousands of patient reports.
Their argument rests on a simple analogy with other drug classes. Long-term exposure to opioids, benzodiazepines, and alcohol all produce neuroadaptations that take months to reverse. There is no biological reason to expect serotonergic drugs to be different. The receptor systems are different but the principle of homeostatic downregulation is shared.
This view is still contested. Some psychiatrists argue that long-lasting symptoms after antidepressant cessation represent relapse, anxiety about withdrawal, or coincidental new illness. Horowitz and Taylor's counterargument is that the symptom profile, including dizziness, paresthesias, electric shocks, and sensory hypersensitivity, does not match major depressive disorder or generalized anxiety and is not what relapsing patients describe. The debate is active and the evidence base is still building, but the framework that withdrawal can be protracted has moved from fringe to mainstream within five years.
The practical implications are straightforward. Linear tapers, in which you halve the dose and then halve again over a few weeks, are pharmacologically irrational for most psychiatric drugs and predict the worst kind of late-taper crash. Reductions should get smaller in absolute terms as the dose gets lower. Final doses before stopping are often a fraction of the smallest commercially available pill.
This usually requires compounded liquid formulations, custom-made tapering strips, or careful weighing of pill fragments. None of these are easy to access through standard pharmacies, and many prescribers are still unfamiliar with the practical mechanics of micro-dosing reductions. The Maudsley Deprescribing Guidelines now serve as a clinical reference that patients can show their doctor.
Equally important is the time horizon. A taper informed by the hyperbolic curve is measured in months for a short medication history and often a year or more for long-term users. This is not slow for slowness's sake. It reflects the underlying biology of receptor adaptation and reverse-tolerance.
If a prescriber tells you to cut your dose in half tomorrow and stop two weeks later, the Horowitz and Taylor framework would describe that as a setup for severe withdrawal in the vast majority of long-term users. You are entitled to ask for a different plan, and the published research now backs that request.
Linear tapering makes equal milligram cuts at each step, such as 20, 15, 10, 5, 0. Hyperbolic tapering makes cuts that produce equal percentage drops in receptor occupancy, which means progressively smaller milligram reductions as the dose gets lower. The hyperbolic approach reflects the actual pharmacology of how psychiatric drugs bind to brain receptors.
No. Heather Ashton and others used non-linear tapers for benzodiazepines for decades. What Horowitz and Taylor did was apply the framework systematically to SSRIs and SNRIs using PET imaging data, publish it in The Lancet Psychiatry in 2019, and later codify it in the Maudsley Deprescribing Guidelines in 2024.
Acceptance is growing rapidly but is not universal. The Royal College of Psychiatrists updated its guidance in 2019 to acknowledge that antidepressant withdrawal can be severe and prolonged. National guidance in several countries now references hyperbolic principles. Many individual prescribers still use older linear protocols, so patient advocacy often matters.
There is no universal answer because it depends on the drug, the duration of use, prior tapering experience, and individual sensitivity. The Maudsley Deprescribing Guidelines describe proportional reductions of the most recent dose as a starting framework, with adjustments based on symptoms. For long-term users, total taper duration is often a year or more. Specific schedules should come from a prescriber.
The 2019 Lancet Psychiatry paper "Tapering of SSRI treatment to mitigate withdrawal symptoms" by Horowitz and Taylor is the foundational reference. The Maudsley Deprescribing Guidelines (Horowitz, Taylor, et al., 2024) is the practical clinical manual. Both are widely cited and accessible through medical libraries and academic search engines.
The Horowitz and Taylor research did not invent slow tapering, but it gave the practice a peer-reviewed pharmacological foundation and a clinical manual the rest of the field could no longer ignore. If you are tapering, you are part of a moment where the official guidance and the lived experience of patients are finally converging.
The community at taper.community exists in part because of this shift. Sharing real schedules, real symptoms, and real timelines is how the next decade of refinement happens. Come join the conversation, ask questions, and contribute what you are learning.
Medical disclaimer: This article is for educational purposes only and is not medical advice. Tapering psychiatric medications should be done with the support of a qualified prescriber. Do not adjust your medication based on anything you read online.