A recent prospective study from Switzerland followed 32 primary-care patients as they reduced antidepressants. The study adds useful evidence that symptom increases can cluster around dose changes. It also makes something else plain: measuring withdrawal is difficult precisely because many withdrawal symptoms overlap with depression, anxiety, ordinary stress, side effects, and other health problems.
That sounds straightforward. It isn’t.
When anxiety, insomnia, low mood, agitation, dizziness, poor concentration, or emotional volatility appears after a medication change, the quickest explanation is often, “The original condition is back.” Sometimes that will be the best explanation. Sometimes it will not.
The better question is not which explanation feels most familiar. It is whether the full timeline supports it.
A symptom can have more than one plausible cause
In medication decisions, labels often arrive before the investigation does.
A returning symptom may reflect relapse or recurrence of a prior condition. It may reflect withdrawal following physiological adaptation to a medication. It may be rebound, meaning a temporary intensification of a symptom the medication had been suppressing. It may be an adverse effect emerging during treatment, a new medical issue, a life circumstance, or some combination of these.
Those categories overlap. That is the hard part.
Withdrawal does not have to look unfamiliar. A person whose original reason for treatment involved panic, insomnia, depression, or anxiety can experience those same symptoms after a reduction. Similarity alone does not establish relapse. But timing alone does not establish withdrawal either.
I keep coming back to this: uncertainty is not a failure of clinical thinking. Pretending the uncertainty is not there is.
The timeline is evidence, not proof
A useful medication history does more than list names and doses. It reconstructs a sequence.
- What was happening before the medication was started?
- What problem was the medication expected to address?
- What changed after starting it, including benefits and unwanted effects?
- What changed after later increases, additions, or combinations?
- What symptoms appeared during long-term exposure?
- What changed after a reduction or discontinuation?
- How quickly did those changes appear, and how did they evolve?
This does not turn chronology into causation. It does something more modest and more useful: it prevents us from dismissing chronology because causation cannot be proven with certainty.
A symptom that appears tightly after a medication change deserves to be considered in relation to that change. A symptom that gradually returns months later, in a pattern closely resembling earlier episodes, may raise a different set of questions. Neither pattern supplies a verdict by itself.
The recent Swiss study is a good example of both progress and restraint. Researchers found that clinically meaningful symptom increases were more common in intervals following certain lower-dose reductions than in intervals without reductions. But it was a small observational cohort, not a definitive map of incidence for everyone taking an antidepressant. The symptom scale itself includes items that can reflect several causes. The researchers say as much.
That is how evidence should be handled. It can make one explanation more plausible without making every other explanation disappear.
Physical dependence is not addiction
This distinction matters because the language around medication can distort decisions before the actual facts are examined.
Long-term exposure to some psychiatric medications can lead to physiological adaptation. That can mean withdrawal symptoms occur following a reduction or discontinuation. Physical dependence is not the same thing as addiction, which involves compulsive use despite harm and other behavioral features.
Neither conclusion follows automatically from the other.
Withdrawal symptoms do not prove that a medication was inappropriate, harmful, or ineffective. They also do not prove that a person requires the medication indefinitely. They show that stopping or reducing a medication may be its own biological event, rather than a neutral test of whether someone is “still ill.”
That distinction is especially important in long-term treatment. Drug elimination from the bloodstream and the body’s adaptation after long exposure are not necessarily on the same clock. Persistent symptoms do not automatically establish protracted withdrawal, but symptoms lasting beyond a short expected window should not be ruled out as medication-related solely for that reason. Competing explanations still need to be examined.
What current guidance gets right, and what it cannot settle
In February, an American Society of Clinical Psychopharmacology task force published consensus recommendations on deprescribing psychotropic medications. Its central point is sensible: the value of an ongoing medication should be periodically reassessed, with benefits, adverse effects, interactions, function, and patient priorities all on the table.
That is not an argument for ending treatment. It is an argument against passive continuation.
The task force also emphasizes that medication changes can affect work, relationships, self-care, and other parts of daily life. This is a necessary correction to the narrow question of whether symptoms are technically present. A medication can have value while also carrying costs. A medication change can offer a possible benefit while also carrying real risk.
Still, consensus is not the same thing as a complete evidence base. The recommendations draw on expert agreement where direct comparative evidence is often limited. They cannot tell us, in advance, what any one person’s symptoms will mean after a change.
A 2026 randomized study of 103 people with remitted major depression illustrates the difficulty. It identified some symptoms reported only in the discontinuation group and found that withdrawal symptoms and early depressive symptoms were associated with later relapse risk. That association does not establish that withdrawal causes relapse. It may reflect overlapping processes, shared vulnerability, or limits in how cleanly the two can be separated.
Here’s the part worth looking at: a symptom can be real, consequential, and still not have one obvious cause.
The question worth carrying forward
Medication decisions become less reliable when every difficult symptom is forced into a single story: the medication is the problem, the diagnosis is the problem, the person is relapsing, or the person is withdrawing.
Sometimes one of those stories will fit. Often the honest answer is that more than one may fit, at least initially.
The stronger decision frame is to ask what the medication was meant to do, what it has actually done over time, what costs have emerged, what changed around any adjustment, and what evidence would genuinely distinguish the competing explanations.
That does not eliminate uncertainty. It gives uncertainty a proper place in the decision.
Written for Dr. Teralyn Sell, PhD
Psychology · Brain Health · Human Behavior
Sources worth knowing
- Incidence of antidepressant withdrawal reactions: A prospective longitudinal cohort study in primary care patients
- Recommendations for the Deprescribing of Psychotropic Medications: A Consensus Statement From the American Society of Clinical Psychopharmacology Task Force
- Do withdrawal symptoms predict depression relapse after antidepressant cessation?