Author: Dr. David Healy
Last updated: 2016
Other guides are available in Guides & Papers.
What are antipsychotics?
Antipsychotics are drugs licensed for the treatment of schizophrenia, psychotic episodes and mania. They were originally called neuroleptics in Europe and major tranquilizers in the USA.
They are divided by marketing into two groups:
- First generation antipsychotics
- Second generation (or atypical) antipsychotics
First generation drugs were marketed before 1990, and second generation since then. It was clear from the outset that there was unlikely to be much difference between these two groups of drugs but company marketing that was particularly successful in America made it seem like the second generation drugs were better. Second generation drugs except for clozapine can only be distinguished from first generation drugs in terms of much greater costs.
For new onset schizophrenia, clozapine is no better than first generation drugs. But for patients who are unable to tolerate other antipsychotics it is worth a trial as some patients benefit from it where they do not benefit from anything else.
Terms such as atypical antipsychotics or second generation antipsychotics are entirely marketing terms and essentially meaningless.1
For close to 50 years after their development, there was a general agreement that these drugs should be restricted to patients who had severe mental disorders. They were regarded as too toxic to take unless the person had a severe illness, such as schizophrenia or old style manic-depressive illness. Since then, this view has been relaxed in the interest of profit, with antipsychotics given to children as young as 1.
Antipsychotics should rarely, if ever, be given to anyone under the age of 20.
The benefits of antipsychotics
What do antipsychotics do?
Antipsychotics produce a tranquilizing effect.
This can be very helpful in the short term for states of acute agitation and confusion. It differs from sedation in that a good tranquilizer will reduce levels of agitation and leave the person able to function whereas a sedative puts you to sleep.
Antipsychotics act on the dopamine system, blocking D2 receptors. While this can produce a useful tranquilizing effect, it can also unhelpfully cause people to become immobilized and demotivated, particularly at higher doses. These drugs also block a variety of other receptors, and this can lead to varying degrees of weight gain, sedation, or neurological effects.
The term “antipsychotic” is misleading in that these drugs do not cure psychosis or stabilize moods (see Mood Stabilizers). They are more effective for the treatment of itch, nausea and vomiting, and for tics than for psychosis or mood stabilization. A range of them like metoclopramide and prochlorperazine are marketed for this purpose and when given for this reason can lead to all of the side effects of the antipsychotics.
Do antipsychotics work?
Unequivocally yes, if the question is do they tranquilize and can this with appropriate teamwork between doctor and patient be put to useful purposes. But possibly not if the question is what have controlled trials shown.
Other than for safety purposes, controlled trials aren’t needed when a treatment obviously does something either good — for instance using coal tar for strychnine poisoning or the right antibiotic for a life-threatening infection, or less good — for instance inhibiting sexual function with an SSRI. Trials are primarily done when there is a real doubt as to whether something either good or bad is happening.
In the case of the antipsychotics, trials aren’t needed to show they are tranquilizing or helpful for delirium or mania. But it is quite a different matter to establish that these drugs produce a benefit in psychoses in the longer run.
For most people saying a treatment “works” suggests it saves lives, or enables people to return to work, or makes the chances of an illness returning in the future less likely.
As of 2016, all the evidence points to the fact that antipsychotics increase rates of mortality when given over the longer term. Death rates in schizophrenia and other psychoses are — uniquely for any medical illness — higher than they were 100 years ago. These are acceptable risks to take in some cases of psychosis but much less so for manic-depressive illness, depression, anxiety or for children.
The evidence that antipsychotics get people back to work is limited — doctors believe it happens but clinical trials have not shown it. One of the better and more independent studies from 1988 showed that patients on antipsychotics were less likely to be readmitted to hospital if taking their medication continuously, but also less likely to get a job or get married.2
It has never seemed a good idea to companies to use Quality of Life rating scales in trials of the antipsychotics — because these drugs do not offer a feel-good factor.
The biggest problem with the antipsychotic studies is this: They clearly show that up to 33% of people with schizophrenia do not benefit or are made worse by treatment. But the trials appear to show the drugs “work”. This means that a doctor faced with a patient who is clearly doing worse on antipsychotics is likely to get the sack for not prescribing drugs that “work” to patients with such a terrible disease.
Another problem is that many people who do well in these trials have acute and transient psychoses — a condition that would clear up anyway even if not medicated. The problem for any patients with acute and transient conditions is that because of these trials, the system may force them to remain on treatment even after recovery, sometimes by means of forced treatment orders, when in fact these drugs can impair quality of life more than almost anything else short of chemotherapy.
What overall impact will antipsychotics have on how I function?
An antipsychotic may dramatically shorten a psychotic, delirious or manic episode, save your life and get you back to work, but when used chronically there can be risks.
If you have been ill for a relatively brief period of time and improve on treatment, your recovery may be helped by the acute tranquilizing effect of these drugs – your problem is then ensuring you can get off treatment. If you can identify a very clear benefit such as reduced anxiety, there is a chance the antipsychotic rather than placebo factors is making a difference, and the key thing is to make sure you are on the lowest effective dose for the shortest period of time.
If you are chronically psychotic (ie. over the longer term), the treatments are more likely to give you side effects than do anything useful. Unless you can identify a clear benefit, there may be a case for not being on drugs that do not work for you.
In terms of function, doctors should ask themselves this question — if I were on this cocktail of drugs in this dose, how would I function? If the answer is I probably wouldn’t be able to do much, the dose or number of drugs should be reduced.
Are the effects of antipsychotics all in the mind?
Absolutely not. Because the results in placebo controlled trials are not impressive, some may think the effects are all in the mind. But what the data from clinical trials shows is that a surprising number of people with psychosis can do better off medication or therapy than has often been thought.
If the question in these trials had been do antipsychotics tranquilize, the answer would have been clear-cut. They unequivocally do.
It may make sense to take the risks of treatment if you are getting a clear benefit from the drug you are on. But what the trials show is that many doctors, when they see a patient improve on treatment, assume that it is the treatment that has produced the benefit without asking the patient whether they can detect anything useful the treatment is doing. In contrast, if there are side effects they attribute these to the illness rather than the pill.
How do I know which drug to take?
Before starting treatment, unless you have been on something before, neither your doctor nor you can know whether you will respond to it. Lots of work could have been done to match people to treatments but this hasn’t been done — probably because research like this would have greatly reduced market share for blockbusters.
If you are put on an antipsychotic, you should clearly be able to identify a benefit in the first few hours or days. The drug should make you feel less bothered and if this feels good to you, it may be useful over the longer run. The first antipsychotic you are put on might feel extremely uncomfortable where another might feel quite good.
If you do not detect a benefit, and in particular if you feel worse in any way, you are likely on the wrong drug for you, or on too high a dose. The problem will rarely be because you are on too low a dose, and you should resist pressure to have the dose increased.
If severely ill what treatment should I take?
For severe psychoses or psychoses failing to respond to other drugs, the usual story is to turn to clozapine. This is not because clozapine is better than other drugs but because it is different and some who don’t respond to conventional treatments may respond to it — but overall it has a higher mortality profile than other drugs and therefore needs to be used with caution.
Many doctors have favorite drug combinations — adding different psychotropics to an antipsychotic. None of these combinations have ever been shown in controlled trials to work reliably.
If you do not show a rapid response to the addition of another drug, you should probably not remain on any combined treatment for long. It is more likely that something in the combination of treatments you are on is holding you back than it is that the combination will suddenly start helping.
Are there any problems with treatment combinations?
Yes. The reason combinations are prescribed is that each drug has supposedly been shown to work in clinical trials. But in fact, all these trials have shown is that it is not right to say the drugs do nothing.
If you aren’t getting better it might make sense to be on 4–5 drugs all of which work. But it makes little sense to be on 4–5 drugs about which we know very little other than that no-one has investigated the effect of combining them.
What will get me better?
In the case of psychosis, the first thing to establish is whether you have schizophrenia or an acute and transient psychosis (brief reactive psychosis). The latter, which make up roughly 15–20% of all psychoses, recover fully even without treatment, although antipsychotics may speed up this process.
If you have schizophrenia, it is more a matter of finding a treatment that suits reasonably well and makes you more functional, and then remaining on this in a dose that helps you function. It may be that a low dose of one antipsychotic will produce the best effects but that these effects don’t amount to a lot. In this case, the drug needs to be supplemented by an entirely different approach such as getting a job or engaging in group activities.
Even if chronically ill, someone needs to work with you to find out whether in fact you are better off drug free and psychotic, or drugged and psychotic.
How long do I have to be on treatment?
Half of those put on an antipsychotic stop within weeks — in great part because of side effects.
The most worrying problem at the moment is that through compulsory treatment orders or other means it is becoming more and more difficult for people put on antipsychotics to get off them. This makes it doubly important to have clinicians who listen to you.
What are the risks?
The standard list of side effects in the manufacturers information leaflets include motor problems such as Parkinson’s syndrome, a range of abnormal motor movements (dyskinesias), and painful muscle cramping (dystonias), some of which can be permanent, a state of intense inner restlessness (akathisia), marked weight gain, sexual dysfunction, sedation, constipation, fainting, palpitations, sweating, tremulousness, headache, blurred vision, rashes, and many more.
Depending on the drug, antipsychotics can also cause:
- Dependence leading to significant withdrawal problems — up to 50%.
- Anxiety or agitation — severe in 20% of cases. (See akathisia)
- Diabetes — all antipsychotics can cause diabetes but some like olanzapine are more likely to do so.
- Suicide — up to 1 in 100 will engage in a suicidal act, and a large number will succeed.
- Violence, aggression, and irritability — this affects an unknown number of people.
- Neuroleptic malignant syndrome — a potentially lethal complication of treatment.
- Tardive dyskinesia — probably as common in new as in old antipsychotics and likely to happen in children if they are put on antipsychotics.
- Demotivation — antipsychotics are profoundly demotivating. They take the colour out of life and leave someone much less likely to do anything. Doctors and other healthcare workers should ideally try these drugs themselves and find out how difficult it is to do much while on them.
For more details, see Side Effects of Antipsychotics.
Periods of risk
Just as for space shuttles, many of the dramatic problems antipsychotics cause cluster around take-off and landing (starting and stopping). The first few days of exposure to an antipsychotic is associated with agitation up to and including suicide. Problems also emerge when the dose is being reduced or shortly after the treatment is stopped. Difficulties also arise in the course of treatment when the dose is changed.
Aside from risk periods like these that are associated with agitation and dependence, there are a range of problems that may appear in the course of treatment and in some instances endure long after treatment stops — see list below.
- Altered bone densities
- Changes to blood clotting
- Cognitive problems
- Emotional numbness
- Gastrointestinal problems
- Sensory problems
- Sexual side effects
- Tardive dyskinesia
- Urinary difficulties
- Vision problems
Taking and stopping antipsychotics is not the same as never-taking. Antipsychotics come with significant withdrawal and legacy effects.
Depending on the antipsychotic, up to 50% of people may have withdrawal difficulties. Clozapine is probably the worst to stop. It may be possible to minimize withdrawal problems by tapering the drug very slowly using a liquid formulation. But this doesn’t work for everyone, and for many people withdrawal may be impossible.
The issue may be complicated by legacy effects of antipsychotics. Some people seem to develop a dysthymia that becomes manifest as they try to stop treatment. This does not seem to be withdrawal. Some features such as dysthymia, an intolerance for stress or memory difficulties can endure for years after treatment stops.
Antipsychotics can blunt normal emotions such as anger. Stopping them can lead to a resurgence of these emotions causing problems in anyone who is no longer used to dealing with them.
Myths about antipsychotics
1. Antipsychotics are neuroprotective.
This quite extraordinary claim is used to justify early intervention, even in children. The antipsychotics in particular are associated with much more obvious and visible brain damage than can be found with ECT. This is not a recipe for not using them but does suggest using just the same caution as you would take to having ECT. Children seem most at risk from the harmful neurological effects of antipsychotics used as mood-stabilizers.
2. Suicide on antipsychotics stems from patients not being on a high enough dose.
The leading cause of death in schizophrenia is suicide — over 50% of the deaths in the first 5 years of treatment. Despite the fact that antipsychotics cause severe akathisia which causes suicide, and controlled trials show an excess of suicides on active treatment, the field instead prefers to believe schizophrenia causes suicide (even though it didn’t in the 19th and early 20th century), or that the drugs restore insight and this leads to suicide, or that the drugs were not given in a high enough dose so that full insight would have been restored and the person would have just realized that in order to stay well all they needed to do was continue to take their antipsychotics.
The facts are that even healthy volunteers can become suicidal on antipsychotics, and suicide rates in schizophrenia are now 20 times higher than they were before the antipsychotics were introduced.
3. There is something wrong with the dopamine system in psychosis.
No one knows what is wrong in schizophrenia or other psychoses — but there is nothing wrong with the dopamine system. The antipsychotics would not produce any useful clinical effects if there were something wrong. When these drugs reduce vomiting we don’t say vomiting is caused by excess dopamine.
4. Antipsychotics do not cause addiction.
All antipsychotics cause physical dependence — it becomes difficult to stop because of how bad the person feels on stopping and the relief from restarting treatment. It is likely that some antipsychotics are worse than others (for instance clozapine) but because companies have denied there is any problem, it is difficult to know which drugs are the worst offenders and how to manage the problem.
Companies and their experts refer to discontinuation syndromes — another term for withdrawal or being hooked. It is not uncommon to hear people say it is harder to get off antipsychotics than off opiates, or benzodiazepines.
Publications and studies
What studies have been done on these drugs?
Almost all studies on the antipsychotics have been carried out or commissioned by the pharmaceutical companies that produce them. There are a small number of independent studies, which have in general shown that older drugs at least as good as and cheaper than newer ones. No government agency or independent authority runs studies.
What data underpin the use of these drugs?
The data from clinical trials of the antipsychotics is almost universally unavailable, even though no-one has given their consent to this.
Has anyone access to all the data?
No one has seen or has had access to all the data, not even regulators such as FDA.
What publications are there on the use of these drugs?
There have been hundreds of publications from hundreds of clinical trials involving antipsychotics.
Of the published studies, most (50–90%) appear likely to have been ghostwritten — that is written by a professional writer on behalf of a pharmaceutical company and published under the name of eminent physicians who may not have read the text.
For the most commonly prescribed antipsychotics more studies undertaken are published than is the case for the antidepressants because it’s close to impossible not to show some tranquilization with these drugs.3
But for many published studies there are multiple publications — the record appears to be 234 publications from 4 Zyprexa (olanzapine) studies, none of which contain a clear picture of the weight gain, raised lipid levels, or glucose levels or rates of suicide this drug can cause — in other words the benefits are published on multiple occasions but the problems are hidden.
In the 1980s, many of the studies that brought current antipsychotics on the market were run through a center where it is known that not all the patients actually existed. The drugs were still approved and non-existent patient data likely is included in some publications. More recently studies have moved to Eastern Europe and Asia and the ability of regulators to oversee what is happening is even more limited.
For example in a study looking at the long-term benefits of Abilify (aripiprazole) the drug failed to distinguish from placebo in over 30 centers in the US, but showed a big difference from placebo in two centers in Mexico. Adding in the data from the Mexican centers made Abilify look slightly better in terms of relapse overall. In none of the articles about this study or the benefits of Abilify is this aspect of the data revealed. FDA were aware of the issue but chose not to investigate.4
Many articles outlining the risks of antipsychotics have not been published owing to the concerns of medical journals that they will be sued by pharmaceutical companies, and a general bias against publishing even convincing case studies that outline the hazards of treatment.
How good are the studies?
Even if all studies were published by real authors, there would be a problem. Almost all the trials done have lasted only a few weeks. They do not look at outcomes that matter to patients such as whether I live or die, get back to work, or have a better quality of life. The rating scales used to decide if the drugs work in fact can show an improvement in your clinical state even if all that is happening is that you are suffering from side effects.
Some of the claims that antipsychotics work are based on recent studies that have stabilized patients on treatment and then randomized some to continue with treatment and others to placebo. Those remaining on the drugs do better. But all the drugs in this group are linked to dependence and withdrawal syndromes, and common sense suggests that what’s going on here is that those remaining on drug are not doing better — they just aren’t going into withdrawal.
The short duration of these studies mean that no-one can tell you whether it is safe or a good idea to take these drugs in the longer run.
Do the people in the studies resemble me?
Many of the people in antipsychotic trials were recruited by advert. Many were volunteered by their doctor rather than volunteered themselves. None gave informed consent. Some didn’t exist. There were very few average people in these trials. If you are being treated for another condition in addition to psychosis, there were few people like you in the trials.
Are there any problems if my doctor keeps to recognized guidelines?
It is customary to think that independent guidelines are superior to guidelines linked to pharmaceutical companies. But in fact the guidelines tend to be identical and independent guidelines may be more dangerous by virtue of their apparent independence.
When the Cochrane Center reviewed the antidepressants, they concluded that Sertraline (Zoloft) was among the most effective and safest, but taking unpublished data into account it ranks among the least effective.
When the Cochrane Center first reviewed the use of antidepressants for children, they concluded these drugs were safe and effective but when unpublished data became available it was clear they weren’t.
When the Cochrane Center first reviewed the evidence on Tamiflu they concluded it was beneficial but when unpublished data became available they had to revise their view.5
Commonly used antipsychotics
|Generic Name||Brand Name|
- Healy D (2002). The Creation of Psychopharmacology. Harvard University Press, Cambridge Ma.
- Johnstone EC, Crow TJ, Frith CD, et al. The Northwick Park ‘functional’ psychosis study: diagnosis and treatment. Lancet 1988; ii:119–125.
- Turner EH, Knoepflmacher D, Shapley L (2012). Publication bias in antipsychotic trials. PLoS Med 9, e1001189.
- Rosenlicht N, Tsai AC, Parry PI, Spielmans G, Jureidini J, Healy D (2012). Aripriprazole in the maintenance treatment of bipolar disorder: A critical review of the evidence and its dissemination into the scientific literature. PLoS Medicine, 8, e10000434.
- Jefferson R, Doshi P et al (2011). Ensuring safe and effective drugs: who can do what it takes? BMJ 342, 148–151.