Why do some people experience severe withdrawal symptoms when stopping antidepressants, while others have few symptoms and some have none? Who are these “severe” people? Can we identify them before they start an antidepressant?
With so much debate and discussion about “how many” (including the previous two essays in this series), it’s surprising that so little has been written about “why some and not others?” Yes, being on an antidepressant for a long time increases the risk of severe withdrawal when stopping. That’s clearly one factor. Yet many people who’ve been on antidepressants for years can stop them with little difficulty. It appears clear that after long-term exposure to antidepressants, some sort of adaptation takes place in the central nervous system that takes a long while to reverse, for many people. Yet not everyone. What accounts for the difference?
Unfortunately, we don’t even understand how antidepressants really work. Yes, they affect serotonin and/or norepinephrine signaling. But how does that change mood? Yes, there are many known alterations that follow the change in serotonin and/or norepinephrine signaling. One common pathway for multiple antidepressant modalities, from pills to exercise to transcranial magnetic stimulation, is an increase in “neurotrophic factors”: cell fertilizers, if you will, that make neurons more likely to grow, to make connections with other neurons. In the face of chronic stress, neurotrophic factors can help prevent neuronal decline. But how does that change mood?
My point is: we just don’t know enough about the mechanisms of mood to know what the mechanisms of antidepressant withdrawal really are, in order to identify people at high risk for severe withdrawal before they start an antidepressant. Lacking an understanding of mechanism, is there any other way to identify the people who will have severe withdrawal symptoms when they try to stop an antidepressant?
Below, one possibility. But first, my apologies: this idea has been offensive to some people. I hope the following details may help make it less so. The idea comes from 20+ years of experience working with people with complex mood disorders.
The Basis for a Hypothesis
Most patients referred to me were already on an antidepressant, often their third or fourth or more. They usually had also tried many other approaches, but almost always those other treatments were done with an antidepressant. So, in search of something they hadn’t already done, a common alternative would be to try some of those previous treatments without an antidepressant—on the presumption that antidepressants might have interfered with response. Accordingly, the first step would begin with tapering their current antidepressant (while in most cases continuing whatever else they were on or doing). This strategy was so often helpful I used to hope that the next referred patient would be on an antidepressant, giving me at least one strategy to offer.
Here is the key observation: it seemed as though nearly all of these patients had withdrawal symptoms during their antidepressant taper. Despite taking small steps down, even with a month or more between reductions, they would still have a noticeable “wobble” right after the decrease. Their experience led me to the remarkable website, SurvivingAntidepressants.org, where similar experiences are described over and over again.
Now a second key observation: primary care colleagues, with whom I have worked closely, did not seem to be having the same experience. They asked for help with numerous issues, but not antidepressant withdrawal problems. They didn’t ask for help with tapering. So our research team looked at the numbers, finding that 90% of patients in our health care system who stopped their antidepressant did so without a prescription for the lowest available dose. Many of these patients may have had significant withdrawal symptoms, but not enough to go back to their prescriber and describe the problem and obtain a smaller pill to facilitate smaller steps down. In other words, a large majority of primary care patients seemed not to need the careful tapering that nearly all my patients did.
One difference could be trauma. In psychiatry, many of our patients have experienced it. Not quite so much so in primary care. But I’d see significant withdrawal problems even in patients without awful trauma histories (yes, I could have missed it).
So, here’s another possibility. This is a working guess, which could be mostly or completely wrong, but it is testable. I hope you’ll follow along as I explain my hunch. If I lose you, here’s the bottom line: it seems possible that some or much of the vulnerability to antidepressant withdrawal is related to “bipolarity”. Not bipolar disorder, but rather a mostly genetic element underlying bipolar disorders.
Bipolarity Is Not “Bipolar Disorder”
Bipolarity can be present in people who do not have bipolar disorder; i.e. their symptoms are not sufficient for that diagnosis. Really? Yes: most mood specialists around the world now accept that people with depression can have any number or intensity of manic symptoms, on a spectrum from “unipolar” (no manic symptoms, present or past) to “bipolar” (meeting official diagnostic criteria for bipolar disorder) with many patients in between these two poles. For example, in the most recent international guidelines, the mood spectrum is explicitly described (skeptics, see Figure 3 therein).
Instead, think of bipolarity as a kind of energy. Those who have a lot of bipolarity are likely to put a lot of energy into their activities: more drive, more passion, more certainty, more optimism, more likely to take the initiative. High bipolarity can be a stable trait with no depression problems or mood shifting or obvious cycling, so-called “hyperthymic temperament”. Any degree of bipolarity is possible, from none to hyperthymia to overt bipolar II or bipolar I.
My working guess: people with little or no bipolarity are less likely to have difficulty stopping their antidepressant, whereas people with a lot (but not necessarily meeting diagnostic criteria for “bipolar disorder”) are more likely to have severe withdrawal symptoms.
This hypothesis offends some people because the label “bipolar” carries more stigma than “depression”. I understand that. But my wish here is to go beyond the labels. Bipolarity—as opposed to bipolar disorder—is not an illness, it is a genetic trait. People can have none of it, or a little, or a lot.
Evidence in Support of the Hypothesis
In the next three subsections of this chapter, I’ll present three lines of evidence that support a connection between bipolarity and antidepressant withdrawal. First, high-energy symptoms during a depression (so-called “mixed state”) strongly resemble many of the symptoms of severe antidepressant withdrawal. Second, antidepressant dose reduction can induce hypomania and mania even in people who did not previously have recognized bipolar disorders. (Think about that for a minute.) And third, antidepressant discontinuation in patients with bipolar disorder is frequently extremely challenging, requiring small decrements and very long taper periods—much like the kind of tapering described on SurvivingAntidepressants.org.
1. Symptom similarity
Have a look at the figure below. Circle A presents physical symptoms of antidepressant withdrawal. Circle C presents symptoms and markers of bipolar disorder. Circle B presents the relatively more psychological symptoms of antidepressant withdrawal, drawn from two studies which quantify the frequency of withdrawal symptoms. (These lists are not exhaustive).
Circle B also represents symptoms of “mixed states”. Depressive mixed states are simply depression plus bipolarity: typical depressions mixed with symptoms associated with bipolar disorders. Debate continues about which high-energy symptoms qualify a depression for a mixed state diagnosis, but recently, several research teams have found four symptoms particularly prominent: anxiety, anger, agitation, and attention problems (dubbed “the Four A’s” by one specialist). As shown in the figure, all of these symptoms overlap with symptoms of antidepressant withdrawal. Is this just a coincidence?
Bipolarity admixed with depression (mixed state) is an extremely common variation of depression. For example, a recent study identified 29% of a sample of patients with major depression as having a mixed state, using empirically derived criteria. People with mixed depression do not necessarily have “bipolar disorder”, but they’re not “unipolar” either: most are in the broad middle between the two ends of the mood spectrum.
The point is: mixed states indicate bipolarity, and the symptoms of mixed states are remarkably like those of antidepressant withdrawal.
2. Induction of hypomania, mania, and mixed states
Ironically, while antidepressants can induce hypomania, mania, and mixed states when they are initiated, antidepressants can also induce these symptoms when rapidly discontinued. One way to understand this irony is to regard bipolarity as capable of creating unstable mood states. Pushing a person in either direction—toward mania, when antidepressants are started; or toward depression, when they are suddenly discontinued—can further destabilize some people’s depression into rapidly cycling mixtures of mood symptoms.
Antidepressant discontinuation-induced hypomania/mania has been observed even in people who have not previously been diagnosed with bipolar disorder. This suggests that the phenomenon is not simply a worsening of an underlying condition. It’s not limited to people who have bipolar disorder.
As a colleague who has reviewed this idea of mine pointed out, withdrawal from other psychiatric medications can induce hypomania, not just antidepressants. But this is far less frequent than has been described with antidepressant withdrawal (as follows, for the skeptics): hypomania during benzodiazepine withdrawal was described in a patient with severe intellectual disabilities; and during opiate withdrawal in a case series, where the authors noted that “features of hyperthymic temperament and family history of bipolarity were traced in most of the individuals”. Otherwise, in a review of psychotropic withdrawal syndromes, induction of hypomania/mania is not listed for alcohol, benzodiazepines, nicotine, opiates or stimulants.
When tapering my patients’ antidepressants, one goal was to avoid inducing the negative high-energy symptoms of mixed states. Or was I helping them avoid having antidepressant withdrawal? In patients with a diagnosis of bipolar disorder, I don’t think withdrawal and mixed states can be distinguished (except by additional physical symptoms such as nausea and brain zaps). If so, then are we seeing a connection between antidepressant withdrawal and bipolarity? Might this explain why primary care patients, with less complex depressions, appear to have much lower rates of antidepressant withdrawal than my bipolar-rich patient population?
Next, a third line of evidence supporting such a connection.
3. Similar taper rates
People with severe antidepressant withdrawal symptoms discover that to successfully discontinue the medication, they must lower the dose by extremely small steps, very slowly. For example, the website SurvivingAntidepressants.org recommends reducing one’s antidepressant dose monthly by 10% of the previous month’s dose. If the pre-taper dose is 100 mg/day of sertraline, the first step down would be to 90 mg/day, a month later to 81 mg/day, then 73, 66 and so on. Users may go faster, especially initially, but are warned to slow down if they are experiencing any increase in symptoms after a decrement.
In my clinical practice, a similar small-decrement/slow-taper process evolved as I focused increasingly on patients with bipolar disorders and sub-threshold bipolarity. In 2011, I suggested a four-month average taper based on cataplexy rates after antidepressant discontinuation.
Over time I received more and more referrals of people who had already tried four or more antidepressants. The probability that switching to a different antidepressant will alleviate your depression is very low once you’ve tried that many. But I slowly learned that some people can get better, at least a little better, by gradually stopping their antidepressant (while usually continuing everything else they’d ended up on by that time). A resident whom I supervised was so impressed with this approach, we published together a report of 12 cases in which we saw improvement in severe anxiety and suicidal ideation just by tapering patients’ antidepressants.
SurvivingAntidepressants.org suggests small steps down for people with antidepressant withdrawal. My target was a presumed underlying bipolarity. The taper programs are very similar. Of course, the necessity for very slow taper rates in both conditions does not by itself establish any direct relationship between these two groups. But this similarity in taper programs, together with the other two lines of evidence above suggests a possible connection.
At least this hypothesis leads to testable questions. If my hypothesis is correct, markers of bipolarity should be more common in people with severe antidepressant withdrawal symptoms than in people who can stop antidepressants with little or no difficulty. These markers include family history of bipolar disorder, early age of onset of depressions (around late teens or very early 20s), highly recurrent episodes of depression, postpartum depression, and strongly adverse reactions to antidepressants when or not long after they are started (although I reported one patient’s experience of having mixed state symptoms develop after seven years on sertraline that helped her “join the human race”).
For detecting bipolarity, a single simple two-page questionnaire combines three validated screening tools. Called MoodCheck, it consists of the Bipolar Spectrum Diagnostic Scale, a family history screener, and the Bipolarity Index. On this questionnaire, there is no “cut-off”, no threshold for a bipolar diagnosis. Rather, the point is to allow people to estimate their position on the spectrum of bipolarity, from none to a little to a lot.
If my hypothesis proves to be correct, more people with intermediate and high scores would have severe antidepressant withdrawal symptoms than people with low or zero scores. If so, a questionnaire like this could be used to identify patients who will need slow careful tapering; or better yet, patients who are at high risk for severe antidepressant withdrawal and who therefore should be even more cautious about starting one.
Again, my apologies to anyone who reads this essay and finds themselves angry or hurt. Bipolarity is not “bipolar disorder”, which though it is being actively destigmatized (e.g. BrainStorm the Film) is still more frightening to many than “depression”. If bipolarity really does correlate with severe antidepressant withdrawal, then identifying it could prevent a lot of suffering. If not, the question remains: why some people and not others—because the risk of antidepressant withdrawal does not seem to be similar for all users.
Mad in America hosts blogs by a diverse group of writers. These posts are designed to serve as a public forum for a discussion—broadly speaking—of psychiatry and its treatments. The opinions expressed are the writers’ own.
Mad in America has made some changes to the commenting process. You no longer need to login or create an account on our site to comment. The only information needed is your name, email and comment text. Comments made with an account prior to this change will remain visible on the site.