Skip to main content

Antipsychotic drugs and blood clots

Wednesday 22 September 2010

“Antipsychotic drugs taken by thousands in the UK raise the risk of dangerous blood clots,” BBC News reported.

Antipsychotic medicines are mainly used to treat psychiatric illnesses such as schizophrenia and bipolar disorder. The study behind this report compared their use in over 25,000 people with blood clots in their legs or lungs and in almost 90,000 people who did not have clots. It found that there was a 32% increase in risk of a blood clot in people currently using antipsychotics. However, the overall risk of having a blood clot was still very small, even among people taking antipsychotics. Overall, people in the study had about a 0.1% chance of having a blood clot each year.

The use of antipsychotic drugs is well established in treating conditions such as schizophrenia. If the small increase in risk of blood clots is confirmed by future research, this would have to be considered when weighing up the potential benefits and harms for each patient. People taking antipsychotics should not be concerned by this news and should not stop using their medication. Any concerns raised by this news can be discussed with a doctor.

Where did the story come from?

The study was carried out by researchers from Nottinghamshire County Teaching Primary Care Trust. It received no specific funding. The research was published in the peer-reviewed British Medical Journal.

BBC News and The Daily Telegraph reported this study. The BBC gave balanced coverage of the research. The Telegraph ’s headline mentioned the use of antipsychotics to treat nausea and vomiting, but this is not their most common use. Although a few of the drugs examined in this study (prochlorperazine, chlorpromazine and haloperidol) are used to treat nausea and vomiting, the primary use of antipsychotics is to treat psychosis, and their prescription for these very different conditions is not comparable.

There are also many different types of anti-sickness drugs (anti-emetics) with different uses depending on the type of sickness treated, and not all of them are used in the treatment of psychiatric illness. The antipsychotic drugs listed (prochlorperazine, chlorpromazine and haloperidol) are only a few of the anti-emetic drugs in common use. They are often used specifically in cancer care or when a person is also taking opiate-based painkillers.

What kind of research was this?

This nested case-control study looked at whether taking antipsychotic drugs increased the risk of blood clots. The researchers say previous studies have suggested that taking antipsychotics may increase the risk of blood clots, but that some uncertainty remains.

nested case-control study follows a group of people and identifies those who have experienced a particular outcome, in this case a blood clot. These people are the "cases". A group of control subjects is then selected from people who did not experience the outcome of interest. These controls are matched to the cases according to important factors such as age and gender.

A case-control study is a good way of looking at events which are rare, such as the potential harms of a drug. While drugs are normally tested with randomised controlled trials (RCTs), it is difficult to detect all possible side effects during these studies. They usually follow subjects for only a limited time and often include a relatively small number of people compared to the number who will eventually use the drug. This means that rarer harms may not be detected in RCTs.

As with all observational studies, the results may have been affected by differences between the groups in addition to the factors being compared. Ideally, the cases and controls should be as similar as possible and any important factors taken into account in the analyses.

What did the research involve?

The researchers used data from the UK QResearch primary care database, which holds anonymous medical records on over 11 million people who registered at one of 525 GP practices in the UK over the past 16 years. They extracted data on adults aged 16 to 100 years old who were registered with participating practices between 1996 and 2007. The researchers identified people who were recorded as having a first blood clot (venous thromboembolism) between 1996 and 2007 (the cases), and selected four matched controls for each of these cases. They then compared past use of antipsychotics between cases and controls.

In total, 25,532 eligible cases were identified and 89,491 matched controls selected from the database. Cases had either a deep vein thrombosis (15,975 people) or a blood clot in their lungs (pulmonary embolism, 9,557 people). Control subjects were matched to cases by age, gender and the GP practice at which they were registered. The controls were alive and registered with the GP at the date at which their matched case had a blood clot (the index date). The controls had not had a blood clot themselves up to this time.

People were not eligible for inclusion if they had less than two years of data available before the index date. Controls who were prescribed warfarin (an anti-clotting agent), cases prescribed warfarin more than six weeks before their clot, cases for whom controls could not be found or people with missing data were not included.

Based on their prescriptions, people were classed as:

  • current users of antipsychotics (one or more prescriptions for antipsychotics in the three months before the index date)
  • recent users of antipsychotics (one or more prescriptions for antipsychotics between 4 and 12 months before the index date)
  • past users of antipsychotics (one or more prescriptions for antipsychotics between 13 and 24 months before the index date)
  • non-users of antipsychotics (no prescriptions for antipsychotics in the 24 months before the index date)

Users in each of the different categories were compared with non-users. Researchers also looked at the effects of the specific drug prescribed, the dose and the class of antipsychotic used (newer “atypical” antipsychotics or older “conventional” antipsychotics).

The analyses took into account factors which could affect results, such as number of months of data available, any mental health diagnoses, socioeconomic status, co-existing medical conditions or prescriptions that could affect the risk of clots. Body mass index (BMI) and smoking were also taken into account in a separate analysis. Some data on these measures were missing, so the researchers estimated the missing values based on available data.

What were the basic results?

There were 118 blood clots for every 100,000 person-years of data collected (person years are a way to measure the total amount of follow-up data collected, calculated by adding up the length of follow up for each person in the study). The risk of having a blood clot increased with age. Compared to the controls, cases were more likely to have higher BMI, to live in deprived areas and to have risk factors for clotting (although some of these differences were small).

In the past two years, 8.3% of cases and 5.3% of controls had taken antipsychotics. After taking into account other possible risk factors for clots, people who had been prescribed antipsychotics in the previous two years had a 32% greater risk of having a blood clot than non-users of antipsychotics (odds ratio [OR] 1.32, 95% confidence interval [CI] 1.23 to 1.42).

People who had last used antipsychotics between 13 and 24 months ago were not at increased risk of blood clots compared to non-users. People who had started taking a new antipsychotic in the past three months had about twice the risk of non-users (OR 1.97, 95% CI 1.66 to 2.33).

The increase in clot risk was greater for people prescribed the group of antipsychotics classed as atypical than for those prescribed conventional antipsychotics (atypical antipsychotics: OR 1.73, 95% CI 1.37 to 2.17; conventional antipsychotics: OR 1.28, 95% CI 1.18 to 1.38).

Smoking and BMI did not have a large effect on results.

Based on their results, the researchers estimated that for every 10,000 patients aged 16 years and over treated with antipsychotics over one year, there would be four more cases of blood clots than in people not taking antipsychotics. If they looked only at people aged 65 and over treated with antipsychotics over one year, the risk was greater, with ten extra blood clots for each 10,000 patients compared to non-users.

This means that 2,640 patients of all ages or 1,044 patients aged 65 and over would need to be treated with antipsychotics to result in one extra blood clot.

How did the researchers interpret the results?

The researchers concluded that there is an association between use of antipsychotic drugs and risk of blood clots in primary care. They say that the increase in risk was greater among new users and those prescribed atypical antipsychotic drugs.


This study has found an increase in the risk of blood clots in people taking antipsychotics. It has a number of strengths. For example, cases and controls were identified from a large pool of people visiting their GPs, which should be representative of people in primary care in the UK.

Other strengths include the use of detailed recorded prescriptions rather than relying on people to estimate their past medication use, and the ability to adjust for a number of factors which could affect results. Other points to note include:

  • In a study like this, it is difficult to untangle the effects of use of medication from the effects of the conditions the medication is intended to treat. The researchers say that when they excluded people with diagnoses of schizophrenia and manic depression, the pattern of increased risk remained, suggesting that the effect is the same across the different conditions for which antipsychotics might be used. This finding fits with the possibility that the drug may be responsible for the increased risk observed.
  • The researchers relied on data from the QResearch database. There may be some inaccuracies or missing information in this data. However, the researchers say that the completeness of recording of diagnoses in this type of database had been shown to be good, and that data correspond well with other similar sources of this information.
  • The analyses were based on prescriptions. It is not possible to say for certain how many individuals took their medication as prescribed.
  • The absolute risk of a blood clot was very small, even among those taking antipsychotics. If 100,000 people aged over 16 years and over were followed for a year, only 118 would have a blood clot, and 2,640 patients of all ages would need to be treated with antipsychotics to result in one extra blood clot a year.
  • The increase in risk did not appear to be present in people who had stopped taking antipsychotics over a year ago.
  • The data available did not indicate the specific reason for the prescription of antipsychotics in most patients.
  • The Daily Telegraph raised the point that these drugs are used to treat nausea and vomiting. Although certain antipsychotic drugs examined (prochlorperazine, chlorpromazine and haloperidol) are used to treat nausea and vomiting, this is not the main use for antipsychotics, and their prescription to treat psychiatric illness should not be considered comparable. There are many different types of anti-sickness drug, which have different reasons for being used depending on the type of sickness, and not all of them are used in the treatment of psychiatric illness. The particular antipsychotic drugs listed (prochlorperazine, chlorpromazine and haloperidol) are only a few of the anti-emetic drugs in common use, and they are often used specifically in cancer care or when a person is also taking opioid painkillers. Due to the adverse effects associated with their regular use, they are only generally prescribed for psychiatric illness when there are specific reasons to do so. It is also worth noting that the study found a stronger association between blood clots and the use of newer ‘atypical’ antipsychotics. No atypical antipsychotic drugs are used in the treatment of nausea and vomiting.

This study adds to a body of evidence about the risks of blood clots in people taking antipsychotics. A systematic review would now be the best way to look at all of this evidence and come to conclusions based on the findings. The authors themselves say that their findings “would need to be replicated on another database before changes in clinical practice can be recommended, and larger numbers would be required to estimate the risks associated with individual antipsychotics”.

People on antipsychotic drugs should not be concerned by these findings and should not stop taking their medication. If they have any concerns, they should talk to their doctor for further advice.

Analysis by Bazian
Edited by NHS Website

Links to the headlines

Drugs to treat nausea and vomiting 'increase risk of clots'

The Daily Telegraph, 22 September 2010

Antipsychotic drugs clot 'link'

BBC News, 22 September 2010

Links to the science

Parker C, Coupland C, Hippisley-Cox J.

Antipsychotic drugs and risk of venous thromboembolism: nested case-control study

British Medical Journal 2010; 341:c4245