A recent front-page article in the New York Times noted that Many People Taking Antidepressants Discover They Cannot Quit. It pointed out that the SSRIs and SNRIs, the most commonly used antidepressants, often have a "discontinuation syndrome" where people feel sick after stopping them.  They describe symptoms like headaches, chills, mood swings, and tingling feelings. The article includes interviews with people who were disabled for months by the discontinuation syndrome, or who had to give up altogether on quitting the medications because of the syndrome.

The article's claims were concerning, but many of the nation's top psychiatrists wrote in to say they were exaggerated and irresponsible. A letter by 35 psychiatrists and professors of psychiatry affiliated with Columbia University stated that the Times' claims about discontinuation were "simply not the case" and that "the clinical consensus is [antidepressant withdrawal] is real, rare and always treatable.” Ronald Pies, a Tufts psychiatry professor, wrote that "based on my experience over 25 years of treating depressed patients, fewer than 10 percent will experience severe withdrawal symptoms when the antidepressant dose is tapered over three to six months." Another professor, Michael Serby, wrote that "withdrawal symptoms are readily avoided or minimized by careful tapering over a number of weeks."

The American Journal of Psychiatry published its own response by Dr. Roy Perlis, a leading psychiatrist. Dr. Perlis argued that psychiatrists have been aware of the discontinuation syndrome since these medications started being widely used, but that with a careful taper most patients will not encounter major problems. He also notes that many of the descriptions of "discontinuation syndrome" are unfair because they include anxiety, depression, and irritability as discontinuation symptoms. Since individuals that stop a working antidepressant are likely to relapse, this is not an avoidable problem or something that can be held against the medication. He accuses the Times of "refram[ing] a mental health success story into a conspiracy theory".

We have found that our clinical practice generally supports the psychiatric consensus, with a few caveats. The Times is right that many patients coming off antidepressants will have a few days or weeks of mildly unpleasant symptoms, and a few can have more severe symptoms that last months. But Perlis and the other experts are right when they say that if doctors and patients work together to stop an antidepressant in a gradual and well-considered way, most will not encounter anything like the Times' horror stories.

It's worth looking at the Times' evidence to see why they came to a conclusion so much different from Dr. Perlis' and our own. They present three studies which they purport show that "many" people who come off the drugs "often" encounter debilitating withdrawal symptoms.

First, they cite Ostrow et al. This is a study of 250 patients coming off various psychiatric medications. 54% described their withdrawal symptoms as "severe". However, the study did not give any information about whether any of the patients with severe withdrawal symptoms were on antidepressants. The New York Times tries to spin this away by saying the patients were "most commonly" on antidepressants, but this is meaningless given that many patients in the study were on multiple psychiatric medications. For example, 56% were on anxiolytics, likely benzodiazepines, which have a much more serious withdrawal syndrome than antidepressants. It's entirely possible that all 54% of people with severe withdrawal symptoms were benzodiazepine users, and that no patient rated their withdrawal from an antidepressant as severe. Using this study to estimate antidepressant withdrawal is inappropriate. Further, there is limited information about how the patients performed their taper, high risk of selection bias, and almost half the patients discontinued the medication without telling their doctor or against their doctor's advice.

The second study, Cartwright et al, looks at 180 current or former users of antidepressants. The Times says that "withdrawal symptoms were reported by over 130". But only 29 patients in the survey were off antidepressants; the large majority of people who said they had withdrawal symptoms were still on the drugs. While it is possible that some people's withdrawals were so bad that they restarted the medication, most likely this is describing people who forgot to take their medication for a few days, or who lost access to a prescription. This is different from a proper psychiatrist-assisted taper. And like the other study, this one has high risk of selection bias - people who have side effects are more likely to take an online survey on antidepressant side effects.

The third study, Rosenbaum et al 1998, avoided selection bias by being a randomized controlled trial, where researchers withdrew volunteers from their antidepressants under controlled conditions. The trial was placebo controlled, so patients did not know whether they were getting their real antidepressant or a placebo that would induce withdrawal. The study found no withdrawal symptoms from Prozac (ie no difference from placebo), but some withdrawal symptoms from Zoloft and Paxil. Once again, this was stopping the medication without any taper, which is not generally recommended.

What percent of people coming off antidepressants really get a discontinuation syndrome? This question is hard to answer accurately. Almost everyone will at least feel some mild change; what percent of people have a significant problem depends on where one puts the cutoff for "significant problem". But various researchers using standardized definitions of "discontinuation syndrome" have gotten a rough picture of rates for different antidepressants. The general consensus is that Prozac is about 10% (1), Lexapro is about 20% (1, 2, 3), and Paxil is 40% (1, 2). About 10% of people in placebo trials who think they are withdrawing from their antidepressant, but who are actually receiving it as usual, report a discontinuation syndrome.

Some of the responses to the Times claim that antidepressant discontinuation never happens with a proper taper. Unfortunately, that isn't entirely true. Studies are inconsistent in exactly how much a taper helps (1, 2), but it seems clear that some patients develop discontinuation even with the recommended few weeks of tapering. Tapers as long as Dr. Pies' recommended three to six months have rarely been fully investigated. There does seem to be some evidence that longer tapers are better than shorter ones, although even a short taper will be enough to prevent most patients from having serious problems.

If a discontinuation syndrome does develop, Fava et al report it as usually lasting only a few weeks. This doesn't mean that a few unlucky patients won't get a very serious discontinuation syndrome lasting months or more, but in our clinical experience this is quite rare.

The Times claims that "the medical profession has no good answer for people struggling to stop taking [antidepressant] drugs - no scientifically backed guidelines, no means to determine who's at highest risk, no way to tailor appropriate strategies to invididuals." This is false. There are plenty of good evidence-based guidelines, for example Guidance For The Discontinuation Or Switching Of Antidepressant Therapies In Adults and Antidepressant Discontinuation Syndrome: Consensus Panel Recommendations. There are plenty of treatments for discontinuation syndrome - the most common is to go back up to a comfortable dose that stops the symptoms, and then taper down more slowly. Other possibilities include converting the offending medication to Prozac, which very rarely causes discontinuation, and going off that one instead.

If you're worried about antidepressant discontinuation, please talk to your psychiatrist. They can guide you towards medications with low discontinuation risk, like Prozac, and avoid ones with higher discontinuation risk, like Paxil and Effexor, or help you investigate non-medical means for dealing with your depression.

When the Times cited the Cartwright survey of antidepressant users, they didn't mention one of its other findings - 85.5% of users said that antidepressants had improved their quality of life (54% "greatly" improved), and 89.4% said they reduced depression. Like the discontinuation numbers, these findings are certainly confounded by selection bias. But this underlines that the same people reporting withdrawal symptoms are also saying they are generally happy they went on the medication.

Antidepressants, like all drugs, are potentially dangerous chemicals with serious and real risks. But if you're depressed or anxious, there is a lot of evidence they are well worth taking.