The CSM/MHRA expert working group on the safety of SSRIs was set up in May 2003. It took over from an earlier CSM/MCA enquiry, disbanded in April 2003, and was preceded by five earlier reviews.

The present working group originally expected to report its findings in September 2003. Then it uncovered data that showed that SSRIs should not be used to treat children. (An exception was made for fluoxetine: it is licensed for use in children in the US, but on very marginal grounds). That phase of the working group review was completed in December 2003 – at which point the final report was said to be due in the Spring, later amended to the Summer, of 2004.

By the time of the Panorama broadcast (Taken on Trust, BBC-TV, 3 October 2004), the deadline for publication of the working group report had become, "before the end of the year" – and now, perhaps, that moment has arrived.

What will the final report say? Clearly the main focus will be on drug-induced suicidal behaviour and on the problems of drug withdrawal problems and dependence. It remains to be seen how the expert group will tackle the many other matters arising, but this is how things seemed in December 2003:

"Many users had other problems, sometimes severe – especially mood changes, loss of libido and weight gain – and the neurological implications and clinical significance of the ‘electric head’ phenomenon had never been explained. Moreover, concerns about the risks of long-term use had barely been investigated, and long-term drug effectiveness was in question too. The risk of drug sensitisation needed investigation (Fava et al, 2003), as did the evidence of drug tolerance (‘poop out’) during treatment - and the tendency to increase dosage (Goldberg & Kocsis, 1996) or use ‘therapy augmentation’ to counteract it. In addition, basic problems to do with drug dosage and concentrations remained unaddressed, and little was known about tuning dosage to genetically determined differences and individual needs." (Medicines out of Control? pages 210-211)

In relation to withdrawal problems and the risk of dependence, the key issue will be the explanation given for the official view (1990 – 2002) that SSRI withdrawal symptoms were rare and mild – given the admission in June 2003 that they would affect one quarter of all paroxetine (Seroxat) users, and sometimes be severe. Were the original clinical trials really so ropey, or was the evidence suppressed - or both? And how would the regulators justify their own bizarre underestimates of the problem (Price et al, 1996) in the light of thousands of complaints about withdrawal problems from users?

Meanwhile, the Chairman of the MHRA (Medicines & Healthcare Products Regulatory Agency) seems to have been trailing the working group’s main conclusion on suicidality, ever since the Panorama broadcast (Taken on Trust, BBC-TV, 3 October 2004). On that programme, Professor Sir Alasdair Breckenridge said: "There is very good clinical trial evidence these drugs do not cause suicide; they do not cause suicidal thoughts in adults." He said this "with great firmness" – but how many viewers would have suspected this was the product of media training, rather than a critical scientific mind?

I heard Breckenridge pressing the same point at fringe meetings organised by MIND, at the New Labour and Conservative party conferences, at around the same time. This was just a few days before the US Food and Drug Administration announced that it was about to embark on "a huge undertaking" to examine the risks of drug-induced suicidality in adults? (Lenzer J, FDA to review risks of antidepressants in adults, BMJ, 9 October 2004, 329, 816). Why is the FDA not convinced by the analysis the MHRA has done?

Breckenridge is promoting what has been the official view for many years: yes, there is an increased risk of suicidal behaviour, soon after people start taking antidepressants - but it is not drug-induced. The official finger points at the drug user, but exonerates the drug: the increased risk of suicidality in the early phase of treatment is explained in terms of the delay before the drug ‘kicks in’.

Indeed, antidepressants have no measurable, beneficial effects early on. As Breckenridge says: "There is a period of time when the drug does not act. It takes a period of three or four weeks before it acts". This is not in dispute. The problem is with the denial that these drugs may exert nasty effects, before they kick in. It is plain they can.

On past regulatory performance, one might expect the working group’s report also to point to the Breckenridge solution – but perhaps alongside some magnanimous small-print admission that the possibly of drug-induced suicidality cannot be discounted. The regulators have traditionally relied on exclusion clauses of this kind, but this is hardly the moment for that.

The wider point is that the impact of this report will go far beyond the working group’s terms of reference – to look at "the safety of SSRIs". These terms relate to the safety of a class of drugs – when the central issue now is to do with regulatory competence in the management of drug risk. Why did it take the regulators some 15 years to properly investigate users’ concerns, and to recognise the scale and severity of the problems they described? The credibility of the whole regulatory system depends a lot on what the working group will and will not say. This was how it seemed when Medicines out of Control? went to press, this time last year:

"At the time of writing, there is even talk of the unthinkable - perhaps major prosecutions to come - but the terms of reference of the review seem to guarantee that its major output will be in the form of more prudent small-print warnings to promote safer drug use. It is already clear that the review will miss the wider point - the need for new management in medicines’ control, for a coherent policy on drug safety and for standards of public accountability that will sweep this dismal past away." (page 9)

Publicly, Professor Breckenridge now emphasises that there is no risk of antidepressant-induced suicide, even in a small percentage of adults. However, behind the scenes, the MHRA is still engaged in a long (if one-sided) correspondence with Professor David Healy about the basics (not detail) of interpretation of the relevant clinical trial results. Healy’s most recent communication – a 20-page letter sent to the MHRA on 5th November – spells out patterns of behaviour and evidence of drug risks that make the hair stand on end. It may take time but, honestly, there is no way his concerns can be dismissed.

When the working group has already established that these drugs can induce suicidal behaviour in children, and probably in young adults, how could it then credibly suggest that the same drug effect cannot exist in older people? That question seems all the more pertinent because most (but not all) adult patients, like children, respond to placebo as well as to any SSRI.

In a recent letter to Social Audit, the MHRA recalled some of the evidence that Andrew Herxheimer and I gave to the working group in Summer 2003. On the basis of an analysis of Yellow Cards, we advised that the risk of suicidality might exist not only when SSRIs were first started, but sometimes also when the dosage was either sharply increased or decreased – and the working group is now apparently intending to warn of this. Breckenridge wrote that he thought this risk to be "plausible" - notwithstanding the obvious conflict with his view that suicidality in adults indicated only that the drug had not kicked in.

For the regulators to deny the risk of suicidality in adults is also to disregard in its entirety the evidence of harm communicated in thousands of patient reports, yellow cards and clinical reports. The MHRA/CSM have been so distanced from public opinion, they have generally missed the political point: why should users trust the regulators, if the regulators don’t trust them?

Post Vioxx, the turbulence feels palpable, but also like some calm before a storm. The present tension seems only heightened by all the pumped up denial. Nothing much is going to change until regulators and companies alike begin to understand that reduced confidence in their standards of behaviour might well be justified, and reasonably and rationally explained.

Charles Medawar
3 November 2004

CM has been working since June 2004 as one of four specialist advisers to the UK Parliamentary Health Committee, on its inquiry into "the influence of the pharmaceutical industry". This will continue to be his main focus of work until the Committee reports, around Spring 2005.