In the latest hit series Homeland Claire Danes plays Carrie Mathison a CIA agent with bipolar disorder taking Clozapine. She takes the drug to prevent herself tipping over into frank paranoia in a world where being paranoid is necessary for survival.
Anyone who knows anything about Clozapine knows Claire Danes is definitely not on it – she would not be as slim and svelte as she is if she were taking it. Weight gain is something Evident about Clozapine that stands in contrast to the Evidence showing no weight gain that companies have gone out of their way to produce for Clozapine and related drugs like Zyprexa and Seroquel (see False Friends).
Should Claire Danes get paranoid?
The question is what does Claire Danes know about Clozapine and should she get paranoid rather than just play the paranoid? As an actress is she killing people playing the part she plays? Is there anything else Evident about Clozapine being hidden by the Evidence?
Clozapine began life in 1958. It was given to the world’s leading psychopharmacologist Pierre Deniker to assess. At the time the neuroleptic/antipsychotic group of drugs was regarded as very safe. Several of Deniker’s patients died on Clozapine and startled by the number and range of deaths he said it was Evident that it should not be developed.
Deniker said Clozapine was too dangerous to be developed.
The company who made Clozapine (Wander) paid no heed to him; business and clinical evidence are two different things. Clozapine’s development continued even after Wander was taken over by Sandoz. Then in 1975 a series of deaths on Clozapine following drops in white blood cell counts happened in Finland. Clozapine was removed from markets in Europe and never made it to the US – Homeland Security (aka the FDA) intervened.
But it re-emerged in 1988 in the United States, in part because of efforts within Homeland Security. The history of Clozapine’s return has been spun and respun – see The Creation of Psychopharmacology – in the course of which a myth has been created that Clozapine is more effective than other antipsychotics (very important for someone on whose wits the fate of America depends) even though head to head trials in first episode psychoses show Clozapine to be no better than older drugs like chlorpromazine.
Recently the Lancet published a large study by Jouko Tiihonen and colleagues, ironically from Finland, which has thrown the cat among the safety pigeons. In line with expectations, Tiihonen showed that patients treated with antipsychotics had higher mortality rates than patients not on antipsychotics. Except for Clozapine. Clozapine had a lower mortality rate than all other antipsychotics, lower even than non-treatment.
Did Deniker get it badly wrong?
So did Deniker get it badly wrong? Have the Lancet redressed an historical error and are they helping Homeland Security by getting Claire Danes on Clozapine – or has the Lancet been turned and is it also now a threat to the American people if not to US business?
Having spent several posts pointing up the limitations of randomized controlled trials (RCTs), and before going on to labor them further, this Finnish study offers a wonderful example of where RCTs really are needed. The need for an RCT in this case should have been very apparent to the Lancet and its reviewers, and probably should have led to the Tiihonen paper being turned down – other than as a marvelous illustration of how badly wrong cross-sectional outcome studies can get things.
Here’s the problem. Outcome studies like the Finnish one are not randomized. The investigators in a study like this just look at what happens in patients taking older antipsychotics like haloperidol, newer drugs like Zyprexa (olanzapine) or Clozapine. Given that thousands of patients may be on each drug and they were followed up over longer periods of time than the usual clinical trial would follow them up, is randomization we might ask a big deal? Does the accumulation of lots of patients not ultimately manage the bias that randomization helps us manage? If Clozapine is dangerous could it conceivably show up in a study like this as the safest drug in town?
Outcome studies can get things this badly wrong. Here’s how it can happen. Patients never get Clozapine first line. They have to have taken several other antipsychotics first. This means they will never get Clozapine in their first year of treatment and perhaps not in the first three to five years. The significance of this is that by far the greatest cause of mortality in patients on antipsychotics, over 50% of the mortality, comes from suicide, primarily in the first year of treatment.
The next biggest source of mortality comes from heart attacks and strokes. But these happen for the most part in older patients given antipsychotics for acute and transient psychoses or delusional disorders. These older patients are at some risk of heart attacks or strokes even before treatment. Again these are patients who rarely if ever get Clozapine. There were deaths in the group of patients who were not on any antipsychotic and these most likely came primarily from older patients with acute and transient psychoses.
This is how the most lethal drug in town can appear the safest
Add these two exclusions together and it becomes clear why there could be fewer deaths on Clozapine in Finland but how at the same time it can be the most lethal antipsychotic. Clozapine causes problems of unparalleled severity in every body system with deaths from myocarditis, interstitial nephritis, cardiomyopathy, diabetes, neuroleptic malignant syndrome and a range of other causes, including suicide, much more than other antipsychotics, but it is not ordinarily been given to the patients at greatest risk.
Clozapine is a wonderful drug – to have in reserve. It is not the drug that is going to help Danes keep the Homeland more secure than others.
But when even the most lethal of the antipsychotics, with a range of cautions like no other, can be billed as the safest, it’s clear we have a system that can induce paranoia. Here’s a question for Danes – is the Lancet one of our heroes or has it in some way been turned? A series of Lancet articles touting the benefits of Agomelatin, and the work of Robert Gibbons, and a more general track-record embracing the latest treatments would make anyone wonder.
The reality is more likely to be as Homeland Security might say that the price of security is eternal vigilance. The issues are tricky and expertise is needed. Beyond expertise we also need to shift orientation back to the future – away from the current adoption of the latest chemical as a wonder fertilizer to be spread as widely as possible, and back to a traditional medical view of chemicals as poisons to be used judiciously.
Clozapine really does deserve a poison sign
We used to have a poison sign on most medicines (see We need to talk about doctors). Among antipsychotics Clozapine is the most deserving of such a symbol still.
Medicines are chemicals that used to come with information about their use in people. The drugs are still poisonous and the art of medicine should still involve balancing the risks of the poison against the risks of the illness, but the information that comes with the drugs that portrays them as free of risks is increasingly poisonous in its own right. Is it time to introduce a poison symbol with articles touting the benefits of pharmaceuticals? Pharmaceutical companies campaign actively against any references to their chemicals as poisons. Medical journals would most likely be equally resistant to the use of such symbols on their articles.
For some of us having a poison symbol on articles might feel like the best way to avoid tipping over into paranoia. But this was tried with quality marking controlled trials and when this was done pharmaceutical company articles garnered all the best quality marks. It’s likely in just the same way that company articles touting the benefits of treatment would end up festooned with symbols of purity while other studies pointing to the risks of drugs would have a poison symbol. Time to reach for the Clozapine.