Social Audit Ltd
P O Box 111 London NW1 8XG
Telephone/Fax 44 (0)171 586 7771


Heather Simmonds, Director
Prescription Medicines Code of Practice Authority
12 Whitehall
London SW1A 2DY 1 August 2002

Dear Mrs Simmonds,

Case AUTH/1318/5/02 Safety of Seroxat (paroxetine)

Thank you for your letter of 22 July, advising us of the outcome of the Panel’s consideration of our complaint. This letter gives notice that we have decided to appeal and outlines the basis on which we shall do so. We rely on all evidence previously offered as well as one or two new pieces, together with the following observations and clarifications.

This will be our second appearance before the Appeal Board in less than a year; we should therefore explain why we are pursuing this complaint. It is not because of the wording of one clause or another, nor to do with the niggling detail of who said what and how. The point is, there is a problem with paroxetine and similar drugs and that both the cause and remedy have everything to do with honest communication of precisely the kind the Code of Practice says it requires.

It is now clear that SSRI and related antidepressants, and paroxetine in particular, induce withdrawal symptoms of sufficient severity and frequency to cause unprecedented levels of adverse reports from doctors and patients. We offer as new evidence (See Appendix) data from the Medicines Control Agency, showing the number of Yellow Cards sent in (to July 2002) for the top 20 drugs suspected of causing withdrawal reactions. The absolute numbers are almost meaningless, but the ranking order clearly indicates a problem: Paroxetine is top of the list, in a league of its own, and five of the top six drugs are SSRI/SRNIs.

Our original complaint included a list of comments from 40-odd paroxetine users. Their meaning should be interpreted in the light of these many Yellow Card reports. We made it clear that these comments were like many thousands of others, all saying the same kind of thing: many people who want to stop taking paroxetine, find they can’t – they feel addicted, well and truly hooked. Of course these comments cannot compare with good, strong "scientific" evidence - but we honestly cannot conceive of anyone reading them through and not concluding that some sort of dependence-related problem does exist, and that for some it is severe.

We contend that reports from users, collectively, are of real value in helping to understanding the aetiology and nature of this problem, also in pointing towards solutions. The people who wrote to us, and many like them, will not be happy to learn that the company proposes to bury their problems in some database, loftily declining to offer any comment on their overall meaning, murmuring about bias, invalidity, lack of reliability and the impossibilities of verification. Here we are in this brave new world of transparency, "concordance" and "The Expert Patient" (ABPI, 2000), teetering on the brink of trusting companies enough not to fear Direct-To-Consumer Advertising - and there is the biggest pharmaceutical company in Europe, simply saying it doesn’t want to know. We ask the Appeal Board to rule that users’ comments may be counted as admissible evidence under the Code, and to request GSK to give a proper account of its assessment of their meaning.

We believe these comments point to a deep and dangerous misunderstanding that now exists and think it reasonable to request the Company takes steps to remedy the persisting failures of communication that will only make it worse. The nub of the problem is that the company is relying on a definition of "dependence" that is profoundly confusing to users and probably many doctors as well.

That is not surprising because the word "dependence" was recently radically redefined. Until the 1990s, health professionals understood that withdrawal symptoms on their own signalled "dependence" – just as users knew that not being able to stop a drug, especially after repeated and determined attempts, signalled that the drug was habit forming and that the user might feel addicted or hooked. Then the definition was radically changed: suddenly "dependence" meant there had to be other features of drug abuse, as explained in our original complaint. It is clear that there is great scope for confusion and we contend that the company’s response is seriously deficient in not addressing such an obvious and important point. GSK surely realises that the definition they rely on is widely misunderstood, but they do not acknowledge the possibility at all.

The company has simply not responded to the WHO statement (1998) cited in our complaint, that the SSRIs may indeed be drugs of dependence even within the formal definitions that now apply. Does GSK accept this; if not why not?

"There is obviously some confusion about the concept of dependence ... The simplest definition of drug dependence given by WHO is 'a need for repeated doses of the drug to feel good or to avoid feeling bad' (WHO, Lexicon of alcohol and drug terms, 1994). When the patient needs to take repeated doses of the drug to avoid bad feelings caused by withdrawal reactions, the person is dependent on the drug. Those who have difficulty coming off the drug even with the help of tapered discontinuation should be regarded as dependent, unless a relapse into depression is the reason for their inability to stop the antidepressant medication.

… In general, all unpleasant withdrawal reactions have a certain potential to induce dependence and this risk may vary from person to person. Dependence will not occur if the withdrawal symptoms are so mild that all patients can easily tolerate them. With increasing severity, the likelihood of withdrawal reactions leading to dependence also increases …" (WHO Drug Information, 1998)

We invite members of the Appeal Board to take account of what members of the public might infer, when assured that a drug is not habit forming, addictive or liable to cause dependence. We also offer as evidence of the existence of some sort of "dependence" problem with paroxetine – and confusion over meanings - the definition relied on by DAWN. This is the nationwide Drug Abuse Warning Network operated by the Substance Abuse and Mental Health Services Administration of the US Department of Health & Human Services:

"Dependence: A physiological or psychological condition characterized by a compulsion to take the drug on a continuous or periodic basis in order to experience its effects or to avoid the discomfort of its absence (e.g., had to take, had to have, needed a fix)"

If GSK completely rejects this interpretation – which seems to us to pretty much reflect public understanding - it should explain why. Why does the company see no need to need to clarify its meaning, and to reduce the high likelihood of misunderstandings, given the relevant requirements of the Code? Why does it rely instead on a narrow interpretation by the two authors of a BMJ article (especially when they failed to disclose that they were echoing the views of a seven strong "Consensus Panel" convened by Eli Lilly & Co, the manufacturers of Prozac)?

It seems unacceptable that resolution of this issue through self-regulation should stand or fall on the question of exactly what Mr. Chandler said. On the balance of probabilities, he clearly conveyed to the journalist in question that what some paroxetine users felt to be a problem was barely a problem at all. Assuming he followed the brief in the GSK "Reactive Key Messages and Issues Document" (19 December 2001), that is certainly what he would have wanted to convey.

The issue here is not simply whether this brief complies with statutory standards but whether, in the light of the above, it satisfies the requirements to provide clear, reliable and balanced information, as specified below. We contend that this company brief falls short of Code requirements in failing to recognise and address the widespread and evident controversy and confusion over definitions and meanings, as explained above. Relevant code provisions include the following:

"…. All promotion-making claims concerning medicinal drugs should be reliable, accurate, truthful, informative, balanced, up-to-date, capable of substantiation and in good taste. They should not contain misleading or unverifiable statements or omissions likely to induce medically unjustifiable drug use or to give rise to undue risks …." (Article 7: World Health Organisation, Ethical criteria for medicinal drug promotion, 1998).

"Information must be provided with objectivity, truthfulness and in good taste accurate, fair and objective and presented in such a way as to conform … to high ethical standards" (I.2) … based on an up-to-date evaluation of evidence that is scientifically valid and should not give an incorrect or misleading impression (I.3) …"… in "Communications to the Public" … all information "should be accurate, fair and not misleading", and companies "should adhere to the highest standards of accuracy". (I.7). IFPMA Code

"Information about medicinal products must be accurate, balanced, fair, objective and sufficiently complete to enable the recipient to form his or her own opinion of the therapeutic value of the medicinal product concerned. It should be based on an up-to-date evaluation of scientific evidence and reflect that evidence clearly. It must not mislead by distortion, undue emphasis, omission or in any other way." (EFPIA Code, Article 3)

"Information, claims and comparisons must be accurate, balanced, fair, objective and unambiguous and must be based on an up-to-date evaluation of all evidence and reflect that evidence clearly. They must not mislead either directly or by implication." (ABPI Code, clause 7.2)

"Information about medicines made available to the public .. must be … presented in a balanced way … and must not … be misleading with respect to the safety of the product" (ABPI Code, clause 20.2)

In our original complaint (relating to Clause 7.2), we conceded that Mr Chandler’s statement that, "there is no scientific evidence that Seroxat leads to addiction and dependency" was not inconsistent with the SPC – though we thought it unfair, ambiguous and misleading, especially when directed to a lay readership. We still do, contending that Clause 3.2 of the ABPI does not override the above requirements – i.e. that concordance with the Summary of Product Characteristics is necessary, but not in itself sufficient for compliance with the ABPI Code. ("The promotion of a medicine must be in accordance with the terms of its marketing authorization and must not be inconsistent with the particulars listed in its summary of product characteristics"). This distinction appears to have been lost on the company: "the information in our briefing document is consistent with our Seroxat SmPC and published data, being a factual and balanced document, which underwent the required internal approval process." We invite the Appeal Board to rule that this clause is intended to mean that product information must comply with statutory requirements, not that compliance with this provision satisfies the requirements of the Code provisions cited above.

Interpreting the Code in this spirit, we contend that the brief on which Mr. Chandler would have relied ("Reactive Key Messages and Issues Document" - 19 December 2001) is itself in breach of Code requirements, in several respects: 

"Seroxat is not addictive. There are well-defined international criteria for drug dependency and addiction and Seroxat is clearly shown as being neither addictive nor causing dependence."

Comment: The categorical statement, "Seroxat is not addictive", in the absence of any qualification, is highly misleading. If requested, complaint would produce compelling evidence of high levels of "health illiteracy", and much more. We believe the assertion, "Seroxat is clearly shown as being neither addictive nor causing dependence" is also highly misleading and inconsistent with the SPC. No such thing has been "clearly shown": the lack of available evidence clearly precludes any such finding (EMEA, 2000). This is tantamount to claiming a drug is "Safe", without qualification. We invite the Appeal Board to contrast this categorical denial, directed at the lay public, with the much more tentative wording in the SPC directed at health professionals:

Relevant findings and recommendations by EMEA/CPMP (2000): "The available clinical evidence does not suggest that the SSRIs cause dependence. However the lack of evidence does not prove the absence of a problem…"; "For the majority of compounds, evidence from well-designed preclinical studies with respect to dependency and withdrawal was incomplete …"; "The available preclinical and clinical evidence does not suggest that SSRIs cause dependence."

Statements directed to the lay public: "Seroxat is clearly shown as being neither addictive nor causing dependence"; "Seroxat is not addictive".

  • "Discontinuation symptoms are completely different to addiction or dependence …"

Comment: This is highly misleading for reasons stated; it also demonstrates the blurring of meaning achieved with the word, "discontinuation". Withdrawal symptoms are a prime indicator of dependence (WHO, 1998) and, in the formal definition used until 1990, were in themselves sufficient for a formal diagnosis of ‘dependence’.

The company increases the risk of misleading drug users and prescribers by persisting in use of the term "discontinuation". This is not consistent with the SPC (Clause 3.2). The EMEA (2000) evaluation on which GSK otherwise relies states: "The term ‘withdrawal reactions’ should be used, not ‘discontinuation reactions’, as has been proposed by some marketing authorisation holders." Similarly, the CSM/MCA (1998) concluded, "that it would be inappropriate to change medical terminology in this way." In effect, the Company is claiming: "Withdrawal symptoms are completely different to addiction or dependence". This is clearly not true.

"The European Regulatory Body the CPMP (Committee for Proprietary Medicinal Products) have recently completed (April 2000) a thorough review of safety data collected following the discontinuation of all SSRIs and other newer serotonergic antidepressant medications. The MCA (Medicine Control Agency) and CPMP have concluded that SSRIs do not cause dependency/addiction."

Comment: again this statement is misleading, in the absence of any qualification. For lack of evidence, both the CSM and the EMEA were a great deal more tentative in their conclusions than the company. GSK has wrongly concluded that lack of clear evidence of dependence is equivalent to clear evidence of lack of dependence.


• "There has been no reliable scientific evidence from either preclinical studies, long term clinical trials or clinical experience, to suggest that ‘Seroxat’ is addictive, shows dependence or is a drug of abuse."

Comment: GSK again relies on lack of evidence to make exaggerated assertions. It is also inaccurate. There is, for example, reliable evidence of abuse from the afore-mentioned Drug Abuse Warning Network. DAWN has recently published tabulations of reports received from a sample of hospitals operating 24-hour Emergency Departments, in 21 US metropolitan areas, of episodes involving deliberate use of prescribed/diverted pharmaceutical products. (Excluded are accidental overdose or adverse reactions, unless these occurred in combination with an illicit drug). Benzodiazepines accounted for 8% of mentions; antidepressants for 6% of mentions:

"In 2000, the most frequently mentioned SSRIs (Table 2.4.0) were:

-Citalopram (3,458 mentions), which more than doubled from 1999 to 2000,

-Fluoxetine (7,939 mentions), which decreased 19 percent from 1998 to 2000,

-Paroxetine (8,020 mentions), which rose 105 percent from 1994 to 2000, and

-Sertraline (6,670 mentions in 2000), which was unchanged from the previous 2 years."


  • "As recommended by the British National Formulary (BNF) and the European Medicines Evaluation Agency (EMEA), the likelihood of discontinuation symptoms is minimised by gradually tapering the daily dose."

Comment: Neither the BNF nor the EMEA statement uses the word ‘minimise’. It is inappropriate because it adds to the impression that withdrawal reactions may readily be controlled although, "as yet there is no controlled data to recommend its effectiveness, the length of time over which it should occur or the minimum dose that one should taper to" (Haddad, 2001). The BNF proposes a six months taper in patients ending long-term treatment: that makes it quite clear that it can be very hard to stop, even if many people manage well.

In short, we are suggesting that whatever Mr Chandler said, it would have reflected some of the lack of balance, misleading information and partial interpretation in the company’s briefing document, "Reactive Key Messages and Issues Document" (19 December 2001).

Finally, we contend that the finding in favour of the company, on the grounds that the Panel was unable to determine precisely what had been said, is unjustified for the following reasons:

1. The Panel gives no evidence that it made contact with the Editor of Mental Health Today to establish whether or not she stood by the remarks published. Assuming no such checks were made, we believe it was quite inappropriate to speculate that Mr. Chandler might have been misquoted (albeit on two different occasions) and to find in the Company’s favour on the grounds of uncertainty about what Mr Chandler actually said.

2. Mr. Chandler was given every opportunity to deny that he could or would have made the remarks quoted, but he did not do so. As the remarks published in Mental Health Today were evidently inappropriate, one would have expected Mr Chandler to make a shocked and categorical denial, rather than some belated and hedged response. The inference would be that Mr. Chandler did say something very much along the lines of the remarks attributed to him, and/or that he did not appreciate how inappropriate the remarks attributed to him actually were. If he had, he would surely have made some effort to persuade a critical complainant that this certainly wasn’t the impression he intended to give.

3. The Supplementary Information in the Code, relating to Clause 20.2, clearly suggests that companies should be in a position to provide copies of information supplied: "particular care must be taken in responding to approaches from the media to ensure the provisions of the this clause are upheld. In the event of a complaint which relates to the provisions of this clause, companies will be asked to provide copies of any information supplied, including copies of any relevant press releases and the like. This information will be assessed to determine whether it fulfils the requirement of this clause." (Emphasis added). We note the company has no transcript. But did it not make a tape recording, or keep any other record of what was said? This is surely a question the Appeal Board should pursue.

4. It would not reflect the spirit of the Code; it might undermine confidence in self-regulation.

Returning to the point we made at the outset, we emphasise that we would readily drop this appeal if GlaxoSmithKline accepted the need to properly address the problems that now exist and agreed to take prompt and effective steps to deal with them. Failing this, we would wish to pursue our complaint, notably under Clause 2 of the Code, relating to promotional activities that "bring discredit upon, or reduce confidence in, the pharmaceutical industry." Now having seen the promotional materials ("Reactive Key Messages and Issues Document" of 19 December 2001), which guided Mr. Chandler, and on which the Company generally relies, we are more than ever convinced of an established pattern of unacceptable behaviour.

I appreciate that the appeal process follows a strict time schedule, but please note that I shall be out of circulation until 2nd September.

Yours sincerely,
Charles Medawar



The Top 20 medicines associated with reports of suspected withdrawal reactions on the UK ADROIT database (Attachment to letter from Medicines Control Agency (Dr June Raine) to Social Audit, dated 3 July 2002)


Drug substance

Number of UK reports of
withdrawal reactions













































Pages 43-44, Trends in other substances of abuse, from the DAWN report referenced 1, below


ADROIT: Medicines Control Agency drug analysis print July 3rd 2002.

EMEA/CPMP: Position Paper on Selective Serotonin Uptake Inhibitors (SSRIs) and Dependency/Withdrawal Reactions, EMEA/CPMP/2775/99, (London: European Medicines Evaluation Agency, Committee on Proprietary Medicinal Products, 12 April, 2000).

P. Haddad, Antidepressant discontinuation syndromes, Drug Safety, 2001, 24(3), 183-197.

International Federation of Pharmaceutical Manufacturers Associations, Self-Regulation of Marketing Practices (Geneva: IFPMA, 1997), Code of Pharmaceutical Marketing Practices (1982-2001); .

World Health Organisation: Ethical Criteria for medicinal drug promotion (Geneva, WHO, 1988)

World Health Organisation: WHO Drug Information, 1998, 12, 3, 136-138.

Yours sincerely,
Charles Medawar


Contents page
What's New?