Are Psychiatric Medications Safe? The FDA's Answer May Surprise You

FDA logo

The Deputy Director for Safety at the FDA’s Division of Psychiatry Products answers our questions about how the risks of psychiatric drugs are evaluated – and explains what you need to know to better protect yourself.

Listen to the audio version of this story here:

Many of us assume that if a psychiatric drug has been approved for use in the United States, then the drug is reasonably safe. But is this assumption correct?

Certainly, a key role of the U.S. Food and Drug Administration (FDA) is to review findings from clinical drug trials that pharmaceutical companies provide to them, and then decide if a drug can be approved for sale. The FDA publicly releases records from these reviews and, in 2016-17, we mined them to help write Inner Compass Initiative's summaries about the safety and efficacy of psychiatric drugs. During this process, though, what kept striking me was how unsafe and outright dangerous many psychiatric medications are.

Benzodiazepines can cause intense physical dependence within weeks. Antipsychotics can cause diabetes, permanent motor dysfunction, catatonic states, and death. If blood levels are not maintained within a very narrow range, common doses of lithium can suddenly become lethal. Anticonvulsants can cause life-threatening rashes – and life-threatening seizures during withdrawal. And on and on the examples go. 

I started to wonder: How does the FDA actually decide if a psychiatric drug is "safe enough" or not for general use? Under what conditions is a psychiatric drug considered too dangerous to approve? I was also struck by how short most of the clinical trials submitted to the FDA were – usually only a matter of weeks or a few months. Does that provide enough information about the safety of a drug, especially when many people take these drugs for years?

Recently, I interviewed Marc Stone, a physician and the Deputy Director for Safety in the FDA’s Office of New Drugs, Division of Psychiatry Products.  (Some of Dr. Stone's own scientific papers can be found here.)


How dangerous is too dangerous?

I submitted some of my questions in advance, and Dr. Stone began by addressing my over-arching question about how the safety of drugs is evaluated by the FDA. He said the process is much less clear cut compared to how the effectiveness of a drug is measured.

"Efficacy is pretty easy to answer as a yes or no question," said Stone. "We can demonstrate that a drug isn't snake oil, that it has an effect that's different from placebo and a certain outcome that we think is meaningful. But safe is a relative concept. It has to do with risk and benefit. It's basically, 'We do think there are situations in which the benefits outweigh the risks.' And that's basically all we can say."

Of course, how "effective" a psychiatric drug is or is not, and in what ways, is worthy of its own discussion – but for this interview, I was most interested in how the safety of drugs is evaluated. (For information on effectiveness, see "How Outcomes are Measured in Psychiatric Research" or any of ICI's mini-booklets on the main classes of psychiatric drugs.)

"We can demonstrate that a drug isn't snake oil... But safe is a relative concept."

Dr. Stone pointed out that many psychiatric diagnostic categories often encompass very different kinds of experiences and wide ranges of intensity. For example, some people who call themselves depressed may be experiencing a troubling reduction in enjoyment, and others may be experiencing frequent suicidal feelings. Consequently, the degrees of risk or harm from an antidepressant that might be considered reasonable by FDA evaluators or anyone else varies equally widely. Because of these complexities, the FDA tends to leave key aspects of safety evaluating to physicians and patients to work out for themselves.

"[W]hen you work on efficacy you can generally demonstrate that the drug has an effect. A positive effect on some area of concern... some proportion of the time," said Stone. "Then you have to ask, 'Is it worth it?' And that's often a question that depends on the patient, depends on the patient's particular circumstances, the degree of disability, and the physician. Those are clinical questions... And it's often an individualized thing."

In some ways, the FDA's position is understandable: A drug aimed at a terminal illness, after all, might come with high levels of risk – but some people would consider the mere possibility of temporarily delaying death worth virtually any trade-off. On the other hand, though, when considering a psychiatric drug choice, many of us probably don't consider the fact that the FDA may have weighed the drug's risks against the premise that depression, anxiety, inattentiveness, vacillating moods, or inner voices might in extreme scenarios contribute towards some kind of fatal outcome.


The word "safe" is not regulated

I remarked to Stone on how often we hear that psychiatric drugs are "safe and effective". The FDA regulates whether a drug company is allowed to say that a drug is an "effective" treatment for a particular condition – and the U.S. Department of Justice sometimes sues companies for using the term "effective" inappropriately. But what about use of the term "safe"? 

Dr. Stone said no regulations exist around the word "safe." Regulations do exist, though, around a drug company saying that their drug is "safer" than another drug, or that their drug will not cause a specific type of harm. 

"It's generally not worth it for the companies to do [drug safety comparisons]. Often, it's better to deal with a certain degree of ambiguity about that."

"If they want to make a claim that their drug is more effective than another drug then they have to do head-to-head studies and they have to show a clear difference. And the same thing is true for any kind of safety outcome," said Stone. "That's why it's not commonly done. It's generally not worth it for the companies to do it. Often, it's better [from a drug company's perspective] to deal with a certain degree of ambiguity about that."

Essentially, drug companies tend to avoid the kinds of side-effect comparisons and specifics about safety that are regulated by the FDA, and instead simply declare outright that their drugs are categorically "safe".


Black box warnings don't highlight the most serious risks

Some of us see black box warnings prominently situated at the beginning of a drug label, and assume they warn us about the drug's most serious and likely risks. Once digesting these warnings, we may not read much further about the drug's potential adverse effects. But Dr. Stone said this is a mistake borne of not understanding the purpose of black box warnings.

Stone explained that black box warnings are issued in situations where the FDA thinks there's a high chance that people might miss or misinterpret a serious adverse effect – not that this possible adverse effect is necessarily worse or more common than any other ones listed many pages further into the drug label. He gave the example of the black box warning that SSRI antidepressants can increase the likelihood of suicidal feelings. 

"The main reason it's a boxed warning is because we want people to know about it, " said Stone. "The point is, that particularly if you have somebody who's depressed, and you start them on an antidepressant, and you start noticing that they're suicidal, the actual interpretation is, 'Well, it's just your depression worsening so, either the drug hasn't had time to work and we should wait a little bit longer, or we should increase the dose.' When, in fact, it may be the drug itself that's causing the suicidal tendencies. People really need to recognize that."


Psychiatric drugs are rarely considered too dangerous to approve

I asked Dr. Stone when a psychiatric drug might be considered "too dangerous" to approve. One situation, he replied, occurs when a new drug is no more effective than a drug already on the market, yet is clearly much less safe. He used as an example recent attempts by drug companies to get new ADHD drugs approved. 

"If you have an existing drug that treats the condition and the new drug doesn't look any more effective, but yet has some certain issues that might be worse and that the other drug doesn't have, then it's pretty clear to say, 'Well, even though the drug works, we probably shouldn't approve it.' And we run into that all the time," said Stone. "We've had drugs, for example, for ADHD, that are stimulants that seem to produce a higher rate of psychosis than other stimulants. So we don't think they're necessarily a good choice, given that risk."

"There are situations in which things just look so bad from a risk-benefit standpoint that we say, 'No, we can't in good conscience put this drug on the market.'...  But most of the time... it is going to depend a lot on the decision for the individual patient."

However, Stone clarified, it isn't always the case that riskier new drugs won't get approval. The FDA also always considers the possibility that different people may react differently to different drugs, and more drugs on the market means more choices for patients. In addition, if a new drug does seem in some way more effective than an old drug, then again, the FDA may accept a higher level of risk in the new drug.

In fact, Stone said, only very rarely has the FDA considered a psychiatric drug to be too dangerous. More commonly, the FDA simply leaves that question for individual patients and doctors to decide for themselves. "There are situations in which things just look so bad from a risk-benefit standpoint that we say, 'No, we can't in good conscience put this drug on the market.'...  But most of the time, if we can establish a drug is effective, then the question of whether the risk-benefit is worth it is going to depend a lot on the decision for the individual patient. Which is something that is the practice of medicine and not something that we try to determine at FDA."


Long-term effects are not well understood

Recent research has shown that more than 80% of people currently taking psychiatric medications have been taking them for more than three years. Yet drug companies typically submit to the FDA just two or three randomized, controlled clinical trials that last as short as one week, and generally less than twelve weeks. That's not enough time to detect slower-developing or longer-term adverse effects. Sometimes, these trials are followed up with monitoring of patients that might continue for six or twelve months. But, according to Dr. Stone, we should not rely much on these supplemental parts of the clinical trials, because they're uncontrolled and typically involve only the people who liked the drug enough to choose to continue taking it.

Consequently, Dr. Stone acknowledged that long-term effects are a major gap in the FDA's understanding of the safety of psychiatric drugs. "It's very, very difficult to identify long-term effects of the drug," he said. "To some degree you can do head-to-head comparisons between drugs and that may give you some outlook... We get post-marketing reports, but if something is subtle and slow to develop, then it's hard to distinguish that from some condition that may be completely unrelated to the drug."

"It's very, very difficult to identify long-term effects of the drug...There are a lot of epidemiological studies that get done, these have become very popular, but they're also in general very bad."

For the same reason, Stone said that "big data" and large population studies are not necessarily helping the situation, because it's too easy to find either good or bad patterns amid so much background noise. "There are a lot of epidemiological studies that get done, these have become very popular, but they're also in general very bad... If you've heard of Gresham's law, bad money driving out good, it's more like bad data driving out good. So they're often very, very difficult to interpret."


Dependence and withdrawal risks are real but often difficult to discern

Due to the poverty of our knowledge about long-term effects of psychiatric drugs, Dr. Stone felt the scientific research about long-term dependence and withdrawal symptoms is also underdeveloped. Nevertheless, Stone acknowledged growing scientific and public concern about withdrawal, and gave an example of a recently discovered issue with ADHD drugs that illustrates how dangerous withdrawal problems can be. 

Stone described children who were taking stimulants for ADHD, which increase blood pressure and heart-attack risks, and simultaneously taking another ADHD drug, guanfacine, which is sometimes also prescribed to lower blood pressure. The two drugs are intended to be taken together, Stone said, yet they can produce dangerous interactions during withdrawal – or even in between doses.

"Kids get stomach illnesses all the time, and they're vomiting, and so they can't keep their medication down," explained Stone. "So [interdose] withdrawal, even if you're not intending to withdraw, is a fairly common phenomenon with kids. And what we were seeing was that kids are on this blood pressure medication, their blood pressure would really shoot up once the medication was stopped – and particularly if they were also taking stimulants at the same time. The stimulants would amplify this counter-reaction to the withdrawal of the blood pressure drug. So we had to say, 'Well, you've got to be really careful about this and if you stop the one drug then you've got to stop the stimulants, too, to keep the blood pressure from shooting too high. And you also have to monitor them to make sure that the blood pressure doesn't get too high. And you may need to do an intervention to lower the blood pressure.' "

The FDA recently added a caution about this phenomenon of "rebound hypertension" to the drug label for Intuniv (guanfacine).


The FDA does not determine if drugs are "safe enough"

FDA WebsiteThroughout our discussion, Dr. Stone repeatedly emphasized that the FDA, contrary to a common misconception, does not evaluate if a psychiatric drug is "safe enough" for general use. Instead, the FDA assesses if a drug might provide some kind or level of benefit to some people under some conditions, and then attempts to simply identify the drug's main risks. Indeed, at times Dr. Stone expressed concern and frustration that many doctors and psychiatrists too often do not take sufficiently seriously their own responsibility for self-education and critical cost-benefit evaluations of drugs they're prescribing. (Though the FDA itself contributes to public confusion by, for example, describing its own mission prominently on its website as ensuring that all medications in the U.S. are "safe and effective.")

The real responsibility for determining if a drug is "safe enough" or "worth it", said Stone, lies with individual physicians and patients. Each of us, he suggested, must study and understand a drug's known effects and risks – and understand the limits on what is known – and then come to our own personal decisions as to whether the potential benefits seem to outweigh the dangers. 


Rob Wipond is a freelance writer and co-founder of Inner Compass Initiative.

Rob Wipond







gray dividing line

Learn more about the safety and effectiveness of psychiatric drugs in Learn/Unlearn.

Learn more about psychiatric drugs, dependence and withdrawal at The Withdrawal Project.