One hundred years ago patients with psychosis were 4 times more likely than the rest of their contemporaries to be dead at the end of their first 5 years of treatment. The main cause of death was tuberculosis. The asylum was a place where if you had the wrong genetic makeup you were at great risk of catching tuberculosis, particularly if you were a young woman.
In 1954, chlorpromazine, the first of the antipsychotics, was introduced. It seemed extremely safe. In the early days chlorpromazine was given in doses of 50 mg three times a day. Twenty years later antipsychotics in some cases were being given in doses equivalent to 50,000 mg of chlorpromazine per day and the fact that patients still walked around fairly normally seemed to confirm the essential safety of the drugs.
Even if there was evidence the drugs shortened lives, in the 1960s a few years of life lost seemed a reasonable trade to many for the chance to get out of the asylums. But you can’t depend on people to be grateful for ever – and while losing a year or two of life might be a reasonable trade off, losing 10 or maybe 20 years is another matter.
Shockingly, over the last 5 years, a series of large studies, some looking at national databases, have shown that patients with psychosis are 2 to 3 times more likely to die in any one year than the rest of us. Death is primarily by heart attack or stroke. Being two or three times more likely to be dead may not sound much but other studies point to 15-20 years of lost life. The results have been consistent and have raised concerns about the contribution from the antipsychotics especially the second generation drugs which have bad cardiac profiles.
Even more shockingly, almost no-one knows what happens to patients with psychosis 5 and 10 years after they are first admitted – they do not know in New York, London, Berlin or Paris. No one in North Wales knew until this year. The studies that have already been done survey cross sections of patients and don’t answer this question. No doctor can in fact tell patients or their families how likely they are to be alive or dead 1, 5 or 10 years after first admission. If there are risks they cannot tell them what the risks are. This is important because if we know what the risks are we know what to look out for in order to minimize those risks.
What does happen in the first five years? Patients with psychosis, just as they were 100 years ago, are now 4 times more likely to be dead after 5 years of treatment than the rest of us. Patients with schizophrenia are 11 times more likely to be dead – this is much worse than 100 years ago.
Patients with schizophrenia are 10 times more likely to be dead at the end of the first year of treatment than they were 100 years ago. There is no other illness in medicine where such a statement could be made.
Death in the early years of schizophrenia does not come from heart attacks or strokes – it comes from suicide. In their first year of treatment, patients with schizophrenia are over a hundred times more likely to commit suicide than the rest of us.
Heart attacks and strokes do happen but they happen in patients over the age of 65 with delusional disorders or acute and transient psychoses.
See Healy et al (2012) Mortality in Schizophrenia.
Some of the reviewers of this article went into orbit – they did not want to see it published. Among the points raised were that of course we know schizophrenia causes suicide – if you didn’t find it one hundred years ago it’s because the hospital hid it well or patients were straitjacketed – anything except accept that we are doing worse.
Does schizophrenia cause suicide?
No, it doesn’t. The historical data show that the suicides do not come from the illness. Patients 100 years ago did not commit suicide – there does not seem to be a significant risk inherent in the illness.
Did patients 100 years ago have an opportunity to commit suicide?
Yes they did. They spent 99% of their time working on farms or in kitchens or sewing rooms.
Did the hospital hide the suicides?
No. The staff were under a legal obligation to report suicides to the authorities. When patients with mood disorders committed suicide, the records make it very clear what happened and that these reports were submitted.
Does de-institionalization cause these suicides?
No it doesn’t. If patients today were dying 5-10 years into their illness, when they have lost their social networks, jobs, and hopes, the idea that deinstitutionalization might be contributing would seem more likely but in fact they are dying before they lose their families, networks and hopes. One possibility though is that the institution was more of a protective factor for patients 100 years ago than we have appreciated. However it did not protect patients 100 years ago with mood disorders who went on to commit suicide in hospital.
So what causes the suicides? The evidence points to the antipsychotics. In placebo controlled double blind trials these drugs show an excess of suicides and suicidal acts with drugs like Zyprexa having the highest suicide and suicidal act rate in clinical trial history.
This is good news because if most deaths in young people with schizophrenia come from suicide and the antipsychotics make a contribution to this, there is an opportunity to correct the problem. The problem almost certainly stems from drug induced dysphoria. Patients are not on the right drug for them. A simple question – do you like the effect of this drug and do you find this effect useful, along with a willingness to switch from a drug that isn’t suiting someone to one that does, could eliminate the problem.
This is worth trying to make part of the culture of clinical practice because if we can eliminate suicides in year one of treatment we could go very close to restoring the life expectancy of patients with schizophrenia or other young people with psychosis to normal.
There may be no other measure in medicine that for such little cost could make such a difference to life expectancy.
There may be nothing else quite as good as this that psychiatrists could do to save their own skins. They have lost antidepressant prescribing to family doctors, clinical psychologists, nurses or pharmacists. But they cannot blame anyone else if things go wrong in the schizophrenia and psychosis domain. If they cannot ensure their patients safety, what brand value is there left in psychiatry?
There is good news on the heart attack front also. The increased risk of heart attacks and strokes lies in people over the age of 65 with acute and transient psychoses or delusional disorders. These patients are at prior risk of heart attacks or strokes, and in many cases we can detect this. The antipsychotics work on dopamine which has significant effects on the cardiovascular system, and in this way the drugs may tip these patients over the edge.
The antipsychotics already come with black box warnings for patients over the age of 65 – who have dementia. These warnings need to be extended to anyone over 65.
These are the patients who need to be screened before being put on treatment and monitored after they start. But it is a much smaller and more manageable group than all patients with psychosis. The acute and transient patients are also a group who do not need in any case to remain on antipsychotics in the longer term. This puts a premium on stopping treatment where possible.
Owing to the hostility of reviewers in order to get published we “toned down” the findings in an article just out in the BMJ Open – Healy et al 2012, so that the abstract of the article points only to the relatively benign tip of an iceberg that lies beneath. The fact that we toned things down may have meant BMJ didn’t press release the article in the way they have done recently for articles pointing to minor elevations in risk of Alzheimer’s in patients taking benzodiazepines. Getting journalists interested has proven difficult. Schizophrenia is not sexy. Young people dying unnecessarily lights no-one’s fire it seems – at least not these young people.
But that’s not all there is as the next posts La Reine Margot and the St Bartholomew’s Day Massacre may bring out.