In response to Thomas Kingston’s death, Katy Skerrett, the coroner at his inquest, wrote to the MHRA (Britain’s medicines regulator) and to NICE (Britain’s guideline body) suggesting that their communications around antidepressant hazards appeared to downplay the risks of suicidal reactions to SSRI antidepressants, perhaps contributing to his death. See Aunts, Ants and Regulators, and Kingston’s Rule.
Both bodies have responded more or less suggesting that, while over the course of the next year there would be further input on these matters, which might lead to a tweaking of things, they thought their communication was pretty good. MHRA Here – NICE Here.
Here is the key piece from the MHRA response:
The product information for all SSRIs warns that the risk of suicidal behaviour is considered to be greatest in the early stages of antidepressant treatment. This is likely to be related to antidepressants being effective only after a few weeks of taking the medicine and depression itself being associated with an increased risk of suicidal behaviour.
The Kingston family and others bereaved by treatment induced suicide will likely find this response jaw-dropping.
Some of the rest of us are better placed to be disbelieving rather than horrified.
Disbelieving for two reasons. One is that MHRA’s wording looks like it was dictated for them by GlaxoSmithKline (GSK) two decades ago and is unchanged since then. The other is the current wording goes to the heart of how to determine cause and effect and not just raises questions as to whether MHRA should have anything to do with this but positively indicates they shouldn’t.
Scripted in GSK
In 2002 and 2003 a series of BBC Panorama progams put GSK’s SSRI, paroxetine (Paxil/ Seroxat), in the firing line. Regulators like FDA, MHRA, and later EMA were faced with a crisis.
GSK – almost certainly having consulted with other companies – came to the rescue. A review of rates of suicidal events on paroxetine and other SSRIs, along with other problems like dependence on paroxetine, let GSK bundle up their clinical trial data and claim there were no problems. On average, paroxetine and other SSRIs did not cause suicidality or do anything for that matter other than correct a serotonin abnormality.
This claim involved Statistical Lap-Dancing. A not quite statistically significant G String got in between regulators and reality. For just this reason, medical statisticians had been saying for decades that statistical significance tests are completely irrelevant to a hazard like SSRI-induced suicidality.
The G String in this case was used not only to intimate there was no risk from treatment but that the substantial excess of suicidal events on SSRIs compared to placebo was a mirage – a veil between us and reality behind which some might mistakenly think or fantasize there was something.
We now know that in their trials companies had created non-existent placebo suicidal events, breaking regulations in the process and had done so fraudulently in some instances, GSK most notably. Companies do things like this if they expect regulators to turn a blind eye.
It says something about the state we are in that companies could expect to be able to pull a veil over the eyes of pretty well all of medicine.
Professional bodies like the American Psychiatric Association and Royal College of Psychiatrists duly trotted out messages (Suicide Notes) saying Antidepressants Save Lives which they repeat to this day. There is literally no evidence of the kind regulators, companies and APA claim is gold standard, to support this claim – See Professional Suicide and Royal College Suicide Note.
I think antidepressants can save lives but for entirely different reasons to those the Great and the Good are offering us.
GSK embodied one of the Lap-Dance veils in the following submission to MHRA:
It has been proposed that SSRIs, such as paroxetine, may induce agitation or akathisia like events soon after therapy is started and that the distress this causes in an already vulnerable group leads to suicidal behaviour [Healy, 1994]. However this suggestion is not supported by the findings from the adult clinical studies with paroxetine or the postmarketing adverse event data.
Hence, establishing whether paroxetine (or a concurrent medication or underlying psychiatric disorder) is the most likely cause of the onset of akathisia-like symptoms in any individual patient may not be straightforward. However, given that an association between the onset of akathisia or agitation and the development of suicidal behaviour is not supported by the data on paroxetine, there is not an absolute need for therapy to be immediately withdrawn. Rather, the physician will need to consider the range of management options available in the context of the severity of the akathisia symptoms and the potential benefit the patient may derive from continued therapy.
No additional guidance regarding the management of patients presenting with akathisia-like symptoms is considered appropriate, as the prescribing physician should consider each individual patient on a case-by-case basis.
This translates as:
You may be worse after starting treatment, even suicidal when you weren’t before, but in our clinical trials we damped down reactions like this by putting you on benzodiazepines and when we added the data from you and others who were worse right through to the end of the trial, we could still squeak out a teeny-weeny overall on-average benefit on the Ham-D rating scale.
This Ham-D score forced regulators to accept we (GSK) had met the standard for approval agreed with them which we spin as Our Drug Works – It Benefits You. You might have ended up dead but that was down to this terrible illness while our drug was en route to working.
Even if you and your doctor are convinced the drug is making you suicidal, there is no need for the label of the drug to tell either of you that your perceptions might be right. [We have carefully deflected regulatory attention from and told doctors nothing about our healthy volunteer trials where normal people become suicidal on treatment].
Trade-Offs
The wording of the statute governing drug approvals requires companies to demonstrate their drug is effective. Saving lives is the only unambiguous evidence for effectiveness. In the case of the SSRIs, related antidepressants, and many other drugs and vaccines, there is no evidence of lives saved. New treatments mostly show an excess of lives lost and other serious disabilities compared to placebo.
Decades ago, regulators defaulted into agreeing that instead of lives saved companies could demonstrate an effect – on the Hamilton rating scale for instance. An effect that suggests the drug might save some lives. All other rating scales in the trial might show negative results but if there is a bare minimum average drop in a Hamilton score in 2 of several trials done, regulators will license a company to claim their drug is antidepressant.
Not only is there no evidence for effectiveness, and in the case of the SSRIs absolutely minimal evidence for a supposedly beneficial effect, but in the case of sertraline (Zoloft) regulators demonstrated a willingness to license it without meeting the 2 trial criterion. Regulators were also willing to license antidepressant claims for these drugs in minors on the back of completely negative trials.
Companies, as it were, then place this Eucharistic benefit in the weighing scales and claim that even if you are gluten sensitive and killed by gluten in the Eucharist it will still have worked as the body of Xt. You may end up dead but look at it this way – you’ll go straight to Heaven.
Common Sense
A better explanation of what is going on is that SSRIs, created to produce a serenic effect, do so in some but not all of us as Carlsson, their creator, recognized. For some of us nicotine or a benzodiazepine is much better than an SSRI – see Restoring the Magic to Healthcare.
This serenic effect is apparent in roughly half of us taking an SSRI within a day or two of starting – just like the agitation or other effects we might experience. It is not the case that these drugs take 6 weeks to work. Recovery may take 6 weeks, but this is something different. The rates of recovery on placebo are almost indistinguishable from those on an SSRI at the 6 week point. The tiny benefit the entire SSRI group gets from some trial participants showing a serenic benefit, however – with some creative adjustments, was enough to get the entire SSRI arm of the trial onto the gravy train.
The MHRA response to the coroner features another Lap Dancing act. They intimate that children and young adults are particularly at risk of developing suicidality. This is just not true. There is no statistical basis for saying this.
The reason these drugs ended up with Black Box Warnings in the U.S. for children and young adults was because there was no demonstrable benefit on average in the trials against which the risks could be set.
This creates a conundrum. If half of us put on these drugs are not getting the desired benefit, and this is often obvious within a few days of starting, how legitimate is it to keep us on them for 6 weeks – on the basis that these drugs have been proven to produce a benefit by 6 weeks?
- A benefit on one rating scale that is not found on other scales, particularly scales filled by us (patients) such as the Beck scale or Quality of Life scales?
- A benefit that does not on average translate into lives saved, disability avoided or function restored?
Experts in What if Anything?
MHRA and FDA are not staffed by people who are treating us. If they have treated people like us in the past, they have ended up as regulators because they don’t like meeting us or aren’t any good at treating us. And they may be asked to deal with submissions on drugs they have no experience of giving for conditions they know nothing about.
Yet FDA and EMA claim they can decide these drugs cause cures and decide that they don’t cause harms or if they do it’s impossible for them, or anyone else, to decide if the drugs caused a harm in someone.
Faced with a coroner reporting that a doctor (me) and the family in the Kingston case believed his treatment had harmed Tom and stood no chance of curing him, the MHRA are refusing to face the obvious – that these drugs have the capacity to cause suicide – as evidenced by the healthy volunteers who’ve committed suicide on them. They pay no heed to company documents saying our drugs can do this.
This doesn’t mean that an SSRI has caused a suicide in someone who was on it and committed suicide. A doctor or family members who believe the drugs can cause suicide are much better placed than FDA, MHRA or EMA to say ‘we knew him and what else was going on and saw what he was like on the treatment, and we can say the drug did not cause him to commit suicide”.
Ditto for can save a life. If the doctor and family and patient know the desired effect is a serenic effect visible in 24-48 hours and they have clear evidence it was present and things work out well, it is not unreasonable then to say this drug likely saved or contributed to saving this life (if the depression was severe) and if it did so in this case, it has the capacity to do so in others.
If this isn’t the case medicine isn’t possible. If the combination of a doctor and the person on treatment is not mostly correct when we decide it is likely the drug is causing the problem and reducing the dose or stopping the treatment is the correct answer – medicine cannot be safe.
Regulators and Doctors
The chances of keeping you safe fall drastically if your doctor believes company and regulator endorsed propaganda that her view of what is happening you is misguided. If she has been seduced by Statistical Lap-Dances into viewing you as being over-emotional.
Regulators give the impression they are responsible for deciding on Cause and Effect. Ask them what doctors do in this respect and they fudge. They mumble that of course doctors have to treat you but regulators do not concede that in order to treat you safely doctors have to make cause and effect judgement calls and that their decisions on cause and effect in individual cases have or should have implications for what regulators do/think.
If Dr Morgan had turned up at Tom Kingston’s inquest, she could have been asked about all the professional literature, convincing cases, that SSRIs can make people like Tom suicidal. If she is quizzed about this and says she didn’t access it or have time to read it, she will be held liable for not being a responsible professional – she will be blamed.
The more successful company marketing, the more likely it is that, in addition to the drug label and NICE guidelines based on a ghostwritten literature, companies have shaped the culture so that people on Twitter or Tik-Tox will tell someone in Tom’s position, s/he needs to grit her/his teeth and persist until the drug starts working in perhaps 6 weeks time – see Restoring the Magic.
The more successful the company, the more at risk Dr Morgan and other doctors are, yet companies can still depend on professional medical organizations to trot out Antidepressants Save Lives rather than Doctors Save Lives messages.
Off-Piste Regulators
But here is the most extraordinary thing of all.
The regulations state companies should inform and warn when there are serious events that are life threatening, might land you hospital or cause other serious problems. Proven cause and effect is not needed for a clear warning in the case of a serious hazard. The regulations state that if there good grounds to think there might be a problem, there should be warnings.
Up to about 2000, knowing regulators were no good at, and neither in a position to nor trained to determine cause and effect, companies got their doctor to interview you if you or a family member had a serious problem. When they figured, as they often did, that there was no other way to explain what had happened other than their drug had caused the problem, they added this to the label of the drug.
But then, knowing regulators were no good at and not trained to do this, they began passing your reports or coroner reports to regulators and put an additional stumbling block in front of regulators. They also suggested to the regulators that it was not enough to decide there was a risk, the regulator really had to be sure that they wouldn’t kill people by warning. They might save some lives, but if suicide warnings deterred people from taking an antidepressant might the regulators be responsible for more deaths than they saved?
This company masterstroke depends on the fact that no studies have been done to explicitly see if antidepressants save lives. All the data points the opposite way. Even if there was pertinent data, there is no metric into which to put any data that might yield an answer that regulatory bureaucrats could use to create a paper trail in the event you or I were to question their judgement to warn or not warn.
Without a paper trail regulators are sunk. There is no place in their systems for a free standing act of judgement.
We make judgement calls when our lives are at stake. Doctors, like Dr Morgan, made the same calls once upon a time but now they have essentially become bureaucrats who stick to a paper trail.
Worse again, Tom Kingston was privileged in being able to access the same doctor every time. Most of us now get a different doctor every time and this makes it harder for the doctor to work with us when they can’t see that we clearly look and sound different to how we looked just a few days ago before we started the drug. A doctor who doesn’t know you is in a worse position to make a judgement call – and keeping to the paper trial becomes ever more appealing.
As a colleague looking at this put it recently – medicine has become a word-game rather than a place where lives are enhanced or saved. If we had a blood test that shows raised blood sugar levels, we get diagnosed with diabetes – adjudged to have diabetes – when we may not have diabetes at all even with a raised A1C level. This judgement call is almost impossible to eradicate from a record.
If we make a solid case we have SSRI induced suicidality, the record will read ‘patient thinks he has a treatment related problem’. This may even, ludicrously, become a bipolar diagnosis which has all sorts of implications for us.
The paper trail – the drug label or NICE guidance – doesn’t say that if the patient feels worse on treatment and like Tom Kingston stops a treatment he clearly believes is causing him to feel this way you should trust his judgment and not put him back on essentially the same drug.
The fact that the label doesn’t say this is a breach of regulations and at the very least breaking the spirit of the law. There is a regulation about warning if the hazard is serious, there is none creating a duty for regulators to consider what outcome a warning about a hazard will have on average.
As things stand, no-one, not even God, can make an on-average Benefit-Risk calculation. Weighing Benefits and Risks in the balance is an individual matter. It is a critically important duty for us and our doctor, more critically important than voting, paying taxes or almost anything else we do. It is up there with marrying and having children (along perhaps with choosing to die, our ‘rights’ to which are currently being debated in many countries).
Regulators have no place here other than to make all the information available. Legislators have not permitted regulators to be Censors. Companies have seduced them into this.
This post will be followed next week by Signal for the Goose is Signal for the Gander.
It maps onto recent RxISK posts that outline a rapidly deteriorating state of affairs.
Leave a Reply