Professor David Jefferys & Paul Walthall MICR
Session chaired by: Alastair Macdonald MICR
Reporter: Wendy Tomlinson MICR
Keywords
External validity, Global studies, Internal validity, Off-shoring, Masking of treatment effect
Clinical development needs
Professor
Jefferys explained that off-shoring has proven to be a popular solution
to increase the speed and predictability of patient recruitment, to
ensure increased number of patients required plus the increased
proportion of treatment-naïve patients; these are difficult to achieve
in the developed world. All of these activities have to be done in
parallel with the delivery of high quality data and at a reduced cost
per patient, with a view to presenting data that facilitates global
registration and influences key opinion leaders.
To date, a great deal of manufacturing, data
management, statistics, dossier product publishing and help desks have
already been off-shored to developing countries, with different levels
of success.
Off-shoring of clinical trials can’t be discussed as
a blanket topic: there are different situations when considering the
different phases. For example, how many companies would realistically
off-shore their first-in man studies? This is a stage which they are
more than likely to do in developed countries. Routine Phase I has
already been off-shored to a significant degree; however, trials
requiring thorough QT testing and microdosing continue to be maintained
in specialist centres and not off-shored. Similarly, with the advent of
innovative (eg, adaptive) later stage designs, such studies are likely
to be performed in the developed world whilst the larger-volume
‘simple’ studies continue to be off-shored.
Two documents have been recently issued related to
off-shoring, and Professor Jefferys was understandably keen to discuss
these further.
EMEA policy paper
The EMEA policy
paper ‘Acceptance of clinical trials conducted in third countries, for
evaluation in Marketing Authorisation Applications’
1
raises a number of issues. This paper discussed the additional resource
being put into third-country GCP and GMP inspections. The EMEA have
confirmed that they will contribute to building capacity in developing
countries to help these countries inspect more clinical studies to
satisfactory standards.
The paper also shows that of pivotal trials
submitted for EU Marketing Authorisation Applications over the past
three years >25% of patients were recruited from third countries
(eg, Latin America, Asia, CIS members and Africa). The EMEA now asks
industry to recognise that this high proportion could implications on
the acceptance of the data and in particular on the labelling of the
drug in developed countries. So, the pharmaceutical companies need to
consider if that cohort of patients is going to be acceptable: will it
result in the label as originally sought or will it lead to
restrictions?
Alongside this is consideration of the Health
Technology Assessment (HTA) implications. Obtaining approval for a drug
is important; however, so is the reimbursement. Therefore, if increased
numbers of off-shored studies reduce the acceptability after HTA,
getting those additional patients has actually undermined the initial
objective of producing a saleable product.
Professor Jefferys emphasised that these concerns
are not exclusive to the EMEA region: discussion is also raging in the
USA, with the Wall Street Journal in February using the headline ‘Most
Clinical Trials Done Abroad’
2 and with several debates having occurred in Congress.
Another challenge facing Europe is that current
clinical trial controls are still in place on a national basis.
Therefore, consideration is now being given to an optional centralised
system for clinical trials, which would help provide higher-level
overview of the clinical trial participant population; however, this
will require new regulation that will take 4 or 5 years to implement.
(This would also help solve one of the largest challenges we face in
Europe at the moment: that when a multi-centre, multi-national trial is
being run in Europe it requires individual country approval, with
different questions from different countries, and often within
different timelines.)
From another angle, the Declaration of Helsinki
article 29 has implications for some of the developing countries
regarding the ongoing provision of healthcare for subjects who are in
clinical trials. What does that mean? Does drug have to be supplied
until the drug is granted marketing approval in that country? Does it
mean that you are committed to applying for marketing approval in that
country in the longer term? Realistically the answer to this should be
‘yes’. Some countries are also looking for life long provision of
medical care to clinical trial participants; is that realistic? This is
another ongoing debate…
CHMP reflection paper
The second
paper Professor Jefferys mentioned was the Committee for Medicinal
Products for Human use (CHMP) have recently issued a reflection paper
3
on the extrapolation of results from clinical studies conducted outside
Europe to the EU-population. This is currently out for consultation
until the end of May, with industry comments being collated. Its focus,
leading on from the EMEA policy paper, is the extrapolation of results
from trials performed outside of Europe for their relevance to the
European population. The paper references the ICH E5
4
table laying out the intrinsic and extrinsic ethnic factors and raises
the question: is E5 being complied with or are pharma ignoring the
conditions they agreed to in 1998 in order to get the data quicker?
As laid out in the paper, “
Global drug development does not necessarily support the approval of unrestricted indications in an EU population.”
So, it’s important to consider what the implications are of the
extrinsic factors affecting non-EU populations when stipulating the
indications for which is it given approval. By off-shoring we could be
limiting the labelling to a certain sub-population in Europe. The CHMP
did also mention that populations within Europe could also ‘probably
sometimes’ be as different as across regions; however, this needs
clarifying. Clearly, this debate needs to occur as part of the broader
regulatory strategy in order to be transparent in the future.
Professor Jeffery’s then went on to mention the specific challenges for off-shoring:
Logistical requirements
These
include ethical compliance, the challenges and interpretation of
varying languages, varying review times for protocols and subsequent
amendments, the level of IT infrastructure and the cost of monitoring
and oversight in a wider geographically spread study
Scientific issues
There are
recognised differences in any given patient population’s
pharmaco-dymanic response, drug metabolism or in the patho-physiology
of the disease. Added to this is consideration of the presence of
genetic subtypes of a disease, the varied patient population’s diet,
the environmental differences and societal effects. Finally, what about
the differing availability of comparator therapies, and the safety
tolerance experienced by different populations? (Having considered
these points, Professor Jefferys was keen to acknowledge that there are
also differences within Europe on some of these points!)
Acceptability of data
How
representative is the trial data in relation to the population it is
going to be used in? Will the labelling be restricted if there is only
15% of the ‘relevant’ population included in the data, and what are the
implications for key opinion leaders or marketing messages?
In clinical research, the objective is to produce a
single global database of clinical studies supporting a simultaneous
global filing in all commercially appropriate markets. In order to
achieve this there are a number of steps:
- A comprehensive early drug development plan, before Phase I
- Early Phase I pharmacokinetic studies to prove the metabolism is the same in all the target populations
- Assess
the potential for regional sub-analyses including destructive analysis.
Professor Jefferys favours this approach, where regions are removed
from the developing world data to evaluate whether the trends and
outcomes remain the same. This means bringing in the statisticians
early.
Professor Jefferys mentioned the following points as future considerations:
- There has been an
initiative intended to evaluate if there is a pharmacokinetic
difference between the populations in Japan, China and Korea; the
consensus appears to be that this is no difference and if this is the
case then potentially there could be acceptance of data across the
territories. Subsequently, this could have an enormous impact upon the
status of clinical trials in the region.
- Incorporating HTA consideration into studies.
- How to incorporate adaptive seamless study designs when moving between regions.
Summary
In summary, Professor
Jefferys raised some valid points: has off-shoring gone as far as it
can taking into account the regulatory agencies increasing concern?
Also, from our regional point of view how can we improve the position
for EU and the US to make it more attractive to retain studies in this
regions (networks have had some success) and are there more initiatives
out there which need consideration?
Local validity of global data
Paul Wathall’s presentation provided examples as to concerns with the presentation of combined data.
Paul confirmed that the globalisation of clinical
trials is less than 30 years old and largely attributable to commercial
sponsors trying to contain costs in an environment where the cost of
drug development has increased whilst market values have declined.
Therefore, to facilitate speedy completion, maximise the period
available for on-patent sales, and minimise costs, increasing numbers
of clinical trials have included developing countries.
The advantages of off-shoring include the rapid
access to a large number of drug-naïve patients happy to participate in
clinical trials working with motivated, compliant investigators.
However, the logistical challenges faced along with the language,
custom requirements and legal practices when first working in a new
geography can be challenging.
Previous reviewers of the off-shoring process have
expressed concerns that trials which would have been rejected in
developed countries on ethical or safety grounds have been placed in
developing countries. Alongside this is the question as to whether it
is right to exploit patients from one region solely for the benefit of
patients in another? These drugs would often be too expensive for the
people in low/middle-income countries. Conversely, is off-shoring an
altruistic measure? Does it give patients in poorer countries access to
medicines that they wouldn’t otherwise have access to?
Paul asked “Are we achieving in our primary
objective of providing safe and effective drugs to the pharmaceutical
market, wherever that market may be?” Perhaps in a bid to reduce the
cost and increase efficiency, the process of drug manufacture has lost
focus on GCP compliance and possible ethical problems, as demonstrated
by previous cases of approved medicines prescribed to millions of
people being withdrawn due to safety concerns.
Comparing health risk factors
Paul’s
interest in data collected in developing regions and applied to the
west started whilst he worked in a cardiology research unit in York. He
reviewed an investigator newsletter and noticed that there was a
significant difference in the recruitment rate between eastern and
western countries, with the eastern countries having met their
recruitment targets when the western ones had barely started. Also, the
distribution of the recruitment targets was skewed: whilst only 12 of
the 33 countries were in Eastern Europe, these countries were tasked
with recruiting 60% of the study population. Therefore, Paul asked
whether the over-representation of any particular race or regional
sub-group represented a threat to the data.
Paul examined the data to consider if there was
likely to be issues due to generalising these results to a western
population. He compared the mortality and health status data related to
the involved countries using a World Health Organisation web site.
After performing normalising distributions, he applied independent
sample t-tests to compare summary statistics between eastern and
western countries. Paul highlighted one result from the analysis: that
in terms of life expectancy Eastern Europeans die on average 8 years
younger than those in the West. This was a highly statistically
significant finding. Also, Eastern Europeans appeared to experience 8
years fewer of healthy life.
Further analysis revealed many other highly
significant differences between Eastern and Western Europeans in a
number of other key areas including health expenditure per capita,
smoking prevalence and disability-adjusted life years lost to coronary
heart disease and stroke. From these results, Paul concluded that the
age and socioeconomic risk factors associated with an Eastern European
are not the same as those of a Western European of the same age.
Masking variation in treatment effect
With
this information in mind, Paul wondered if there were other differences
that could make extrapolation of data from Eastern Europeans to Western
Europeans inappropriate, exaggerating or underestimating the benefits
of certain treatments. For example, could an experimental drug appear
safer when administered to a drug-naïve group of patients than when
used in a population already with full access to many medications,
thereby masking potential for adverse drug reactions in the Western
population? Paul was concerned that drug-naïve Eastern European
patients with greater baseline risk might also gain greater benefit
from treatment than those in the West with a much lower baseline risk.
Paul then progressed his discussion to lay out that
it is also well known that treatment effects commonly vary among
patients with different levels of risk. Indeed there are many examples
where treatment is beneficial for patients with severe disease but
ineffective of even harmful otherwise. This may be particularly
important for two reasons: firstly, treatment complications are now a
leading cause of death in developed countries and secondly, baselines
risk frequently varies from one region to another. Paul illustrated
this by citing several examples that have been published in
peer-reviewed journals, including the following:
- A global study with over
60,000 patients enrolled with cardiovascular (CVD) disease or at least
3 risk factors for CVD. Eastern and Western sites were compared for
demographics, percentage prevalence of specific morbidities and
treatment. The Eastern European cohort were on average 6 years younger
at diagnosis. Although there were likely to have higher blood pressure
and smoke, they were very much less likely to be diabetic. Despite a
higher proportion of these having established coronary heart disease
and heart attacks, they were less likely to have heart surgery or
coronary angioplasty. Therefore, this suggests that at baseline the
Eastern European subjects carried more risk and received less treatment
than those in the West.
Paul was keen to emphasise why this information was
important by demonstrating the importance of regional effect
modification by showing a real example. In a simple randomised
controlled trial, over 3000 patients were randomised to carotid surgery
or medical management. At study completion the study showed a highly
statistically significant benefit for surgery amongst patients with
more severe disease which was subsequently reported in publications and
guidelines. However, sub-analysis showed that in fact surgery was only
truly beneficial for patients treated within 50 days of the onset of
symptoms. In the UK, treatment time was significantly longer then 50
days, suggesting that UK patients might not benefit from surgery, and
could in fact be exposed to more risk by being exposed to surgery so
long after symptom onset.
Next Paul discussed a mega-trial conducted entirely
in China exploring the benefit of adding clopidogrel to aspirin
following a heart attack. The results showed a modest but highly
significant effect of clopidogrel on both a composite outcome as well
as death from all causes at 28 days. As a result, the authors concluded
that whilst the study was completed entirely in China there was no good
reason to expect different results in other populations. However, the
Scottish Medicine Consortium did not agree: following review they
disagreed with the extrapolation of results and granted Clopidogrel
only a limited license in Scotland. As discussed earlier by Professor
Jefferys, this limited license demonstrated the concerns that there is
good evidence to suggest that Asian races have genetic differences that
can affect drug metabolism as well as how effectively a drug might work.
The final evidence provided by Paul supported the
concerns raised by Professor Jefferys that decision-makers are becoming
increasingly mindful of the possibility that off-shored clinical trial
data do not always transfer across regions. The FDA rejected a license
application having questioned the relevance of data presented for a
K-channel blocker for AF cardioversion due to concerns that all of the
data was obtained in Eastern Europe.
5 In fact, one panel member stated: “
the lack of a representative patient population to me is really troubling……we’re sitting here at the US FDA, not the UN FDA.”
Conclusions
Drawing to a close, Paul was keen to reassure that there is hope around
the corner: with clearer guidance being offered to sponsors for the
presentation of foreign data for licensing purposes. However, whilst
even the EMEA’s 1998 guidelines acknowledge that some studies
extrapolate data, they stress that others require additional research
in the form of bridging studies or in some cases the actual replication
of randomised controlled trials, for example where medical practice or
trial conduct differs significantly in between regions. The EMEA’s
recent strategy paper for the acceptance of trials conducted in third
countries has confirmed that they have been tracking the geographic
origins of patients included in pivotal trials, as discussed by
Professor Jefferys.
Paul’s conclusions were as follows:
- The impact of globalisation on external validity has largely been overlooked
- Extrapolation
of average effects from one region to another may be problematic in
some cases, leading to an under- or overestimation of benefits or harms
- Off-shoring may be a misguided false economy with the potential for harm
- External validity should be addressed in reporting guidelines and by regulatory authorities
- If
off-shoring were to be increased, tighter controls should be imposed on
numbers from developing regions enrolled to a single trial
- In the future, trials may be rejected if the data is not relevant to target populations
Discussion
During the Q&A
session, Mikhael Hounslow from Sanofi-Aventis raised the question of
whether there is a level of protectionism at work considering the
comments by the FDA’s representative especially when the USA has such a
large multi-ethnic, multi-racial population. Paul’s response was that
although there are variations within the population within the USA, it
is still a group of people living together in one socio-economic
environment (albeit with some variations) which needs to be reflected
along with the ethnicity of that population in the data. Professor
Jefferys re-iterated that the regulatory authorities have noticed that
new drugs just don’t seem to work as well in ‘real-life’ as they do in
clinical trials and that this is an issue which needs to be considered
and trials applied in true populations. Having said that, Professor
Jefferys did concede that some representatives on the committee may
have be investigators and yes, they might be looking for some of the
work to return.
David Newall asked what needs to be done to resolve
this issue. Professor Jefferys agreed that that many of the measures
put in place need updating and moving forward with the need for the
strategies to be agreed within the pharma companies early on. There may
be some therapies where off-shoring is fine as reactions across
populations are the same whilst there are others where it is not
acceptable.
Jenny Nutton from the Comprehensive Local Research
Network for Cheshire and Merseyside asked if the NHS objective of
doubling patient recruitment in UK clinical trials in the next 5 years
is achievable. Paul felt that, whilst the networks had done well
supporting academic studies, there is still some scope for development
to support commercial studies despite their promises. Professor
Jefferys agreed that this is a challenging target but could be possible
in innovative trials.
References
- EMEA Strategy Paper: Acceptance of clinical trials conducted in third countries, for evaluation in marketing Authorisation Applications. London: EMEA 2008
- Wang, S. Most Clinical Trials Done Abroad. The Wall Street Journal. 2009 February 19.
- Committee for Medicinal Products for Human use. Reflection Paper on the Extrapolation of Results from Clinical Studies Conducted outside Europe to the EU-population DRAFT. 2009 February 19
- E5 (R1) Ethnic Factors in the Acceptability of Foreign Clinical Data. EMEA September 1998
- Stiles.S FDA advisory panel rejects K-channel blocker for AF cardioversion The Heart.org 2007 December 13
Professor David Jefferys is Senior Vice President, Eisai Europe Ltd.
Paul Wathall is Senior Clinical Trials Manager at HCA International.
Wendy Tomlinson is Director, Clinical Operations with MDS Pharma Services.
