Further letter from the Chairman, Medicines
and Healthcare Products Regulatory Agency to the Clerk of the
Committee (PI 124B)
Thank you for your letter of 21 February. I
am pleased to answer the questions you set in that letter.
Please will you send the Committee copies of all
papers relating to the EWG report tabled at the meeting of the
CSM held on 25 November and full minutes relating to this agenda
I am enclosing the papers which relate to the
EWG (Annex A) that were tabled at the meeting of 25 November and
the minutes of that meeting. Other papers were discussed under
this agenda item but were not tabled. We can provide these if
All data considered by the EWG was not included
in the final report . . . is the MHRA intending to publish or
make publicly available all the evidence on which the EWG based
The report of the Expert Working Group contains
the evidence which formed the basis for the conclusions of the
Group. In the 18 month period of the review the Expert Working
Group considered a large number of analyses and assessments of
data from different sources. It was not possible to include every
piece of data in the final report, which was designed to be a
readable and accessible document for both health professionals
and patients. Therefore the data provided in the report was that
on which the key decisions were based. The MHRA will, subject
to any relevant exemptions in the Freedom of Information Act,
make available on request the full data and assessments considered
by the EWG.
On the subject of efficacy, the EWG's remit
did not extend to re-evaluation of the efficacy of these products,
although the risks were considered in the general context of the
benefits. The EWG specifically looked at the balance of risks
and benefits with increasing dose of SSRIs and the efficacy and
safety data reviewed are included in the final report. One patient
group where the lack of evidence for efficacy had a fundamental
impact on the balance of risks and benefits of the products was
in paediatric use. Therefore the efficacy data for children is
included in the final report.
The Agency communications strategy
As requested, I am enclosing a copy of the MHRA
Communications Review conducted by Stonehenge, the public relations
firm appointed after a procurement exercise launched in late 2003.
The Report, entitled "A Case for Change" was presented
to the Agency in April 2004. The Report also contains the Patient
View Report at Annex C. I am also enclosing as requested the extracts
from the minutes of the Agency and Executive Boards where this
issue was discussed (Annex B). Should the Report be published
with the evidence of the Committee, I would be grateful for the
time to ensure that anyone named individually in the Report is
happy for their name to appear. I have also removed Appendix E
as this is a comparative study of the press functions of other
Agencies not within the remit of this Inquiry.
Stonehenge's report was comprehensive and extremely
detailed. It provided the basis for the development of short and
long term plans for our communications policy. Given the detail
the Report contained, and the relatively low base from which the
Agency started its work, much progress has been made.
Recruitment of a Director of Communications
started almost immediately. However, as with any senior appointment,
the process and the notice period which followed the successful
appointment, has meant that our new Director of Communications
was only able to take up post at the end of January 2005. The
new Communications Division, consisting of existing and newly
recruited staff under Simon Gregor's direction will come into
being on 1 April 2005. However, an internal working group set
up before Mr Gregor's arrival has progressed a number of issues,
including an audit of all internal and external publications,
the preparation for the launch of a procurement exercise seeking
advice on the "re-branding" of these documents using
an agreed Agency format and livery, the recruitment of a media
relations team and the development of the Agency website. The
results of these activities will be visible within the next few
months. For example, the Agency will be taking over responsibility
for its own press and media relations in mid March (work previously
carried out on our behalf by DH media centre). The new website,
with considerable additional functionality, will be launched in
the early Autumn of 2005. In the longer term, the Agency is planning
to work with key stakeholders in raising a better awareness of
risk literacy among patients and healthcare professionals.
Increasing the patient voice in regulation
The MHRA seeks to involve patients as stakeholders
in the regulatory environment in a number of ways, including formally,
through the Committee on Safety of Medicines and its sub-committees
and in less formal ways across the range of its work.
In 2004 you may be aware that the Agency consulted
on the reform of the Medicines Act Advisory Bodies. The Agency's
restructuring plans will include greater patient participation
in the regulatory process. There will be two lay members on the
on the Commission for Human Medicines and the Committees set up
under section 4 of the Medicines Act. The Agency is also establishing
Expert Advisory Groups to advise the Statutory Committees, and
each of these will have two patient members. The lay and patient
members will meet independently and regularly as a patient forum.
The new committee structure will be in place for the Autumn of
As you also will be aware, in 2003 Ministers
commissioned an independent review of access to the Yellow Card
Scheme Scheme, which looked in particular at transparency issues.
This publication is also available on the MHRAs website and gives
examples of some important early warnings of new ADRs identified
through the yellow card scheme. In May 2004, the Review concluded
that there should be greater access to Yellow Card data to ensure
the full potential of the data was realised, and proposed a system
for facilitating this. Of particular note, is the fact that recommendation
in the report on direct patient reporting to the Scheme was accepted
immediately and, after piloting several different options, launched
a direct patient reporting scheme in January this year.
The MHRA has been proactive in ensuring that
leaflets, labels and packaging meet patients' needs for clear
and authoritative information, especially in the recent Review
of EU legislation. Labelling of medicines will have to include
the name, strength and pharmaceutical form in Braille and changes
to the order of the information in the PIL will be required. The
Review importantly introduces a requirement for user testing which
should help to ensure improvements to the PIL to make them more
understandable for patients. Guidance on user testing is being
developed by the MHRA, supported by advice from a CSM Working
Group on Patient Information. The Group is also advising on possible
revisions to the European Guideline on Readability. This will
all involve patient consultation and be of direct benefit to patients.
In addition, the Agency is seeking to be more
transparent in its working ways of working. The Agency now publishes
details of all complaints received about medicines advertising
and the outcome of its investigations as well as being more open
about how such decisions are made. This transparency in decision
making has also extended to drug safety issues where the Agency
has released previously unpublished data. The MHRA has published
summaries of clinical trial data to support key communications
on drug safety when it has been in the public interest to do so.
Summaries of clinical trial data have been released in respect
of SSRIs used in children as well as for anti-psychotics (risperidone)
used in the treatment of dementia. A full and open consultation
process is also carried out before a decision to re-classify a
medicine (from Prescription Only to a Pharmacy or General Sales
List medicine). Patient groups and charities are included on this
Connections with Smithkline Beecham
From 1987 until the end of 1992, my advice was
sought at irregular intervals by Beecham Pharmaceuticals, later
to become Smith Kline Beecham, on general issues in the field
of clinical pharmacology. My advice was not sought on specific
drugs and did not involve discussion on Seroxat. From 1992 until
1997, I sat on a formal committee of Smith Kline Beecham which
met regularly to discuss broad scientific areas of drug development.
Specific products were not discussed. At the meeting of SCOP on
24 February 1998 there was a general discussion of SSRIs which
led to a decision to set up a review of the adverse effects of
the drug class of SSRIs and I felt it appropriate to chair that
meeting. At the meetings on 24 September and 22 October, there
was discussion about specific SSRIs, including Seroxat and it
was not appropriate to take part in these discussions, in spite
of the fact that I had left the Scientific Advisory Committee
of Smith Kline Beecham. My actions were taken in consultation
with the then Chairman of the Committee on Safety of Medicines
and were fully in line with the procedures in place at the time.
At Q797, you expressed the view that "when
a patient starts to take SSRIs, there is a period of time before
the benefit takes place and in that time . . . the patient is
at a great risk of suicide." . . . The Committee has not
been able to identify any such interpretation in the EWG final
My response to that question was not meant to
imply that there was no possibility of an adverse effect of treatment
increasing the risk of suicidal behaviour. The EWG considered
these issues very carefully and was unable to reach a firm conclusion
as to whether the increased risk of suicidal behaviour seen particularly
in early treatment was due to the antidepressant treatment, the
underlying disease, or part of the early recovery of depressive
Turning to Professor Healy's analysis of events
on withdrawal of paroxetine, the advice of the EWG was sought
on the calculations and conclusions drawn by Professor Healy from
the tables previously supplied to them. The Group considered this
an extremely important issue. Other analyses and data sets considered
by the EWG had not indicated an increase in suicidal events on
withdrawal of Seroxat in adults, although an increase in suicidal
thoughts and self-harm was seen in paediatric clinical trials
on withdrawal of Seroxat. The EWG advised that certain limitations
of the data, as presented by Professor Healy, had to be taken
into account. In particular, the analysis did not take into account
differences in the at-risk period and follow up between the populations
on Seroxat and placebo. This is important when focussing attention
on small periods of time at the end of clinical trials because
by this time follow up may differ substantially between groups.
In addition the denominator figures used in the calculations pertain
to those randomised to start treatment and were not necessarily
corrected for the immediate post-treatment phase. Furthermore,
patients are unblinded to treatment in the follow-up phase, which
may well affect the results.
The EWG considered that, although inconclusive,
the data supported the need for close monitoring of patients during
taper and withdrawal of Seroxat, which was a key recommendation
of the EWG. In order to fully clarify this issue the EWG advised
that a large placebo controlled double-blind randomised trial
of withdrawal events would be required.
You have asked for the best estimate of the
proportion of SSRIs prescribed for mild and mild-moderate depression.
As part of the EWG review of the safety of SSRIs, the MHRA commissioned
a study using the General Practice Research Database (GPRD). The
aim of the study was to investigate whether there was an association
between antidepressants and self harm in patients with first time
treated depression. As part of this study two psychiatrists carried
out a classification of the depression diagnoses entered by the
GPs into mild, moderate and severe depression. It should be noted
that patients with severe recurrent depression are likely to be
under-represented in this dataset as it considers first time treated
depression and is based in primary care. Furthermore, the classification
was carried out without seeing the patients and therefore relies
on the way that GPs use the coding system for entering diagnoses
into GPRD which is known to be variable.
In this study, approximately 95% of patients
with first time treated depression had diagnoses consistent with
mild or moderate depression. This proportion was the same across
SSRIs, TCAs and other antidepressants with no suggestion that
SSRIs were preferentially prescribed in mild depression. This
information is available on the website of the British Medical
Journal as supplementary material to the published GPRD study
(Martinez et al BMJ 2005;330:389).
At Q798, you were asked if the MHRA/CSM statement,
that SSRIs were effective medicines in the treatment of depression,
required some qualification, in view of the lack of evidence of
efficacy of the products in treating mild depression. Would you
address this point . . .? Could you specifically advise the Committee
why no restriction on the use of SSRIs was proposed in the Summary
of Product Characteristics?
SSRIs are indicated for treatment of depressive
illness. It is for the individual practitioner, informed by clinical
guidelines, to consider whether the risks and benefits of treatment
are appropriate for an individual. The recent NICE guidance on
treatment of depression concludes that antidepressants are not
recommended for the initial treatment of mild depression, because
the risk-benefit ratio is poor. However the guidance states that
the use of antidepressants should be considered in patients with
mild depression that is persisting after other interventions.
The currently licensed indications reflect the patient population
studied in clinical trials and allow for the individual prescribing
judgement necessary in each case.
In relation to paroxetine withdrawal reactions,
you wrote in your letter of 4 February that, "the MHRA has
never indicated that withdrawal was rare . . ." the Committee
would wish to know if the statement in your letter applies also
to the MCA. Specifically, the Committee invites you to comment
on the 1996 paper published by MCA/CSM staff that concluded "overall,
symptoms due to stopping an SSRI are rare . . .".
You ask whether the statement "the MHRA
has never indicated that withdrawal was rare" also applies
to the MCA in the light of the paper by Price et al published
in the British Journal of Clinical Pharmacology in 1986. I can
confirm that this is the case. The conclusions of this paper were
based purely on spontaneous reporting, a data source which has
limitations for the assessment of incidence of an adverse reaction
and particularly so for withdrawal reactions. Neither the MCA
nor the MHRA considered spontaneous reporting alone a sufficiently
robust data set to form regulatory decisions or prescribing advice
on the incidence of withdrawal reactions.
You have asked that I comment on your observations
relating to the incidence of withdrawal reactions. My comments
on each observation are provided below as numbered in your letter.
Lack of evidence for a taper regimen
(a) We fully accept that clinical trial
data for Seroxat do not provide direct evidence that gradual withdrawal
of an SSRI will reduce the incidence of withdrawal symptoms.
(b) The recommendation that the dose
should be tapered gradually at the end of treatment is based largely
on the knowledge of the pharmacology of these products and detailed
descriptions of tapering regimen described in individual case
reports from patients and prescribers. In addition, the clinical
trial data for venlafaxine and sertraline provide some evidence
to suggest that tapering of dose may be beneficial.
Failure to warn about incidence of withdrawal
(c) At the time of licensing, data
from 222 patients withdrawn abruptly from Seroxat did suggest
that some patients experienced events on withdrawal. This information
was included in the product information for Seroxat from the time
of licensing. The data sheet stated that "As with many psychoactive
substances it may be prudent to discontinue therapy gradually
because of the possibility of discontinuation symptoms such as
disturbed sleep, irritability and dizziness."
In responding to this point it may be helpful
if I provide a brief description of the paper published by Price
et al in the British Journal of Clinical Pharmacology in
1996. It was a comparison of spontaneous reporting data for four
SSRIs. The paper was not intended to be a review of all available
data sources and did not include any clinical trial data. It did
not reflect any agreed position of the MCA or CSM and was not
a document on which regulatory decisions were based.
The paper by Price et al clearly describes
the biases and confounding variables which influence the interpretation
of spontaneous reporting data. The authors highlighted data which
suggested that paroxetine was associated with a higher frequency
of reporting of withdrawal reactions than other SSRIs. Further
information on the nature of withdrawal reactions with paroxetine
was specifically sought through a retrospective survey of doctors
who had reported such reactions through the Yellow Card scheme.
In relation to severity, the paper includes the following information
"21% of reactions were said to be mild, 58% moderately severe,
and 21% severe."
The review which was considered by SCOP and
CSM in 1998 looked at all sources of data for all the SSRIs and
related antidepressants, including unpublished clinical trial
data. From review of all available data the CSM concluded that
"to date, studies had not been carried out of an appropriate
design to allow an estimation of frequency of withdrawal reactions".
As a result of this review the CSM advised that product information
for all SSRIs and related antidepressants should contain warnings
about withdrawal reactions. The CSM further advised that any reference
to withdrawal reactions being "rare" should be removed
from the product information for SSRIs (it was present in the
product information for fluvoxamine and sertraline only). This
review did not contradict the data on paroxetine which was available
at the time of licensingthe paper included these data and
resulted in strengthened warnings in product information on the
basis of accumulating post-marketing data.
The estimate of withdrawal reactions added to
paroxetine product information in 2003 was based on more recent
trials which included a taper phase during which events on withdrawal
were systematically recorded. These trials indicated that 25%
of patients experienced symptoms on stopping treatment and 15%
of those experienced symptoms described as severe. In the same
trials, 15% of patients experienced symptoms on stopping placebo
and 9% of those experienced symptoms described as severe. The
most common events reported in clinical trials as occurring on
stopping both paroxetine and placebo were similar and included
dizziness, nausea, insomnia, anxiety and headache.
Doubts about evidence for efficacy of paroxetine
in relapse prevention
(d) In responding to this point I would
like to draw your attention to a systematic review which looked
at evidence for efficacy of antidepressants in relapse prevention
in 31 randomised trials (including that by Montgomery and Dunbar)
and was published in The Lancet in 2003 (Geddes et al,
Lancet 2003; 361:653-51). The authors took into account
the possibility that withdrawal effects associated with stopping
active treatment could inflate the effectiveness of the active
treatment in prevention of relapse. However their review did not
identify an excess of cases of "relapse" within a month
after discontinuation, which argued against this possibility.
The recent NICE guidance on treatment of depression
concludes that responders to medication, who have had multiple
relapse, should stay on medication to avoid relapse, with appropriate
re-evaluation of treatment taking into account their individual
At Q800, you stated, in relation to paroxetine,
that, "There was additional, clear information for the patients".
The first reference to withdrawal symptoms appeared in the patient
information leaflet five years after the introduction of Seroxat.
The 1996-97 Seroxat patient information leaflet stated that withdrawal
symptoms were "unusual". Please could you clarify and
explain the basis on which withdrawal symptoms were described
There was no requirement for a medicine to have
a patient information leaflet when Seroxat was first licensed.
At that time the warning about withdrawal reactions was in the
information for prescribers.
The first patient information leaflet for Seroxat
after patient information leaflets became a requirement stated,
"Some people find that if they suddenly stop taking these
tablets, they feel dizzy, shaky, sick, anxious, confused or have
tingling sensations. They may also have difficulty sleeping and
vivid dreams when they do sleep. But these symptoms are unusual
and generally disappear after a few days. To avoid these symptoms,
your doctor may tell you to take smaller doses or to spread doses
further apart before you stop taking the tablets altogether".
This wording was considered generally to reflect
the wording in the summary of product characteristics at the time
which stated that "Symptoms including dizziness, sensory
disturbance (eg paraesthesia), anxiety, sleep disturbances (including
intense dreams), agitation, tremor, nausea, sweating and confusion
have been reported following abrupt discontinuation of `Seroxat'.
They are usually self-limiting and symptomatic treatment is seldom
warranted. No particular patient group appears to be at higher
risk of these symptoms; it is therefore recommended that when
antidepressive treatment is no longer required, gradual discontinuation
by dose-tapering or alternate day dosing be considered".
The word "unusual" has no specific regulatory meaning.
2 March 2005
118 Not printed. Back