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


Dr. Keith Jones, Chief Executive Officer
Medicines Control Agency
Market Towers, 1 Nine Elms Lane
London SW8 5NQ

12 February 2000

Dear Keith,

I have now heard from Mrs Thyer, in response (3 February) to my letter to you (1 December 1999) about the dosage of Prozac (fluoxetine). I'm writing back to you to emphasise concern about the message, not the messenger. I thought the Agency's response exquisitely unhelpful, a fine example of everything open government was never meant to be - even if that is what it seems destined to become.

The response to our enquiries conveyed studied indifference to an apparently significant development - the introduction in the USA of solid formulations of Prozac (fluoxetine) of one-half and one-quarter the strength of the only solid dosage form available in the UK. The manufacturers, Eli Lilly, have claimed specific benefits for some patients for the low dose version, and the FDA evidently accepted their arguments. It was natural and proper to ask if the MCA/CSM were satisfied that the manufacturer was justified in deciding not to supply Prozac in a scored 10mg tablet version in the UK. If you or anyone else your end feels like having another crack at it, I'd be happy to wait a bit. If not, I would want to request an internal review of the Agency's response, on these grounds:

1. In the circumstances, it was reasonable to ask and unreasonable not to properly respond to the request. Our enquiry raised important safety questions. Some were mentioned in the note "Prozac - dosage, akathisia and failures to warn" which accompanied the request; others come down to well-established principles of therapeutics:

[a] Effective dose titration is fundamental to good treatment; as a general principle, patients should be treated at the lowest effective dose. I question why fluoxetine should be an exception to this rule, taking into account scale of consumption, extensive use by elderly people and increasing use by children, variability of individual response, prevalence of interaction with other drugs, modest therapeutic margin, and lack of predictability of response. 

[b] Increasing use of Prozac by children is of particular concern. When Eli Lilly announced they were introducing a lower dose formulation in the US (see attached), they drew attention only to the benefits of Prozac for adult and elderly users - but they would, wouldn't they? Although "safety and effectiveness in pediatric patients have not been established", something like one million SSRI prescriptions have been written for US patients aged between 6-years and 18-years old. Presumably, many thousands of children in the UK would now be prescribed Prozac too.

[c] The evidence is not yet good enough, but it is still pretty clear that some adult patients respond well to dosages of fluoxetine below 20mg/day. A daily dosage of the order of 5mg/day has proved satisfactory for some - quite apart from the fact that a substantial part of the benefits attributed to the active drug are evidently due to placebo factors. 

[d] Higher than needed dosages unnecessarily increase risks for patients. Crude as they may be, data from the ADROIT adverse drug reaction reporting system suggest that fluoxetine is an unusually troublesome drug. Over 7,000 reports of some 12,500 adverse reactions have been reported to the CSM/MCA so far - i.e. Prozac accounts for something like 3% - 4% of reports received for all drugs. Serious adverse effects, including akathisia and risk of therapeutic dependence, may be involved.

[e] New drugs are commonly marketed at higher doses than some patients need, because of the importance manufacturers attach to achieving a therapeutic response in the majority. The UK regulators have been slow to recognise this and have made (but not acknowledged) some bad errors. [The requester has direct professional experience of two. The recommended 'usual' dose of the anti-arthritic drug, benoxaprofen, was identical to the specified maximum dose - extraordinary as it might have seemed to license a drug with a near-zero therapeutic margin for use in predominantly elderly population. In another case, the CSM permitted lorazepam to be marketed in the UK, for years, at dosages double those recommended in the US - greatly increasing the risk of therapeutic dependence, among other unwanted effects.]

2. The response to our enquiries fell far short of the standards we were entitled to expect, as set out in the paragraphs on Purpose and Aims of the Guidance on Interpretation of the Code of Practice on Access to Government Information. The MCA's response said, in effect: we see no evidence of harm and we're not curious enough, let alone concerned about any risks, to pursue the enquiries proposed. If the requester is bothered, he can do it himself. Such a response cannot be justified in the light of the professed Purpose (para 1) and Aims (para 2) of the Code. "The approach to release of information under the code should be positive". This response was not.  

3. I suggest that both the content and tone of the MCA's reply were unduly influenced by the Agency's attitude to the requester - though it is surely implicit that requesters have equal rights of access under the Code. I am inclined to think that, if the same concerns had prompted enquiries from someone not regarded as a nuisance and troublemaker, they would very probably have received a more appropriate and informative response. (The requester argues that this is unreasonable: he acknowledges provocation, but would argue that the Agency barely gets as good as it gives.)

4. The requester is "reminded" (as if he needed it) of an exemption clause in the Code that is nothing like as relevant as claimed, on which the Agency relies. It was clear from the enquiry that the MCA was not being asked to supply "information" obtained from the FDA, so much as being asked about its awareness of, and views on, the reasons for introducing a lower dosage formulation in the US, but not in the UK - and the implications thereof.

5. The reasons the MCA gives for not contacting the license holder seem extraordinary: "Nor is the Agency aware of any current safety issues with fluoxetine which relate specifically to the question of posology. This conveys that the Agency discounts evidence of the relationship between risk and dosage, outlined in 1 [a] to [e] above, also in the note attached to our original request, even before making simple enquiries. Is it not reasonable to ask the Agency to act in the face of reasonable evidence of avoidable risk - instead of waiting passively for hard evidence of harm? Given the volumes of evidence on our website and elsewhere of the abuse of secrecy by the MCA/CSM, it seemed impertinent rather than amusing to suggest I was free to contact the license holder myself.

6. The chronic inability of the MCA/CSM to give straight answers to straight questions - and its persistent abuse of the requester's rights to information under the Code - is influenced to an unacceptable degree by institutional conflicts of interest. This obsession with secrecy is overwhelmingly to do with hiding evidence of inaction and ineffectiveness, protecting reputations and vested interests, and the determination to avoid ever admitting that errors have been made. This demeans government and perverts honest science and inevitably contributes to failures in protecting the public from avoidable risk.

Thank you for your attention. I look forward to hearing from you.

Yours sincerely,
Charles Medawar


Contents page
List of correspondence with MCA/CSM

"posology" = the science of dosage