|
Epi Wit & Wisdom Articles
Asthma Drug Controversy Climaxes
With Government Decision to
Restrict Use
Episode is Case Study of
Pressures in Epidemiology Today
[Editor’s note: The following
story has been pieced together from interviews and from a variety of
written materials including newspaper accounts, copies of
correspondence, letters to the editor, company statements, government
pronouncements, abstracts and original articles.]
A New Zealand Department of
Health decision to restrict the use of the asthma drug fenoterol has
brought to an end the latest chapter in a year long controversy over
the findings from two case-control studies. Over the course of the
battle, questions have been raised about the actions taken by the
Lancet in publishing the study, by Boehringer Ingelheim (BI), the
manufacturer of the drug, in defending its product, by the Department
of Health in regulating the drug, by the Asthma Task Force (ATF) and
the Medicines and Adverse Reactions Committee (MARC) in providing
advice to the government, by individual epidemiologists in reviewing
the study, and by the media in covering the story.
An understanding of the key
events and the issues raised is of interest to epidemiologists because
the profession is currently striving to obtain a grip on the ethical
issues it faces and to fashion useful guidelines. The fenoterol story
provides an example of some of the forces at work in epidemio]ogy
today and may be useful as a case study.
First Chapter
The first chapter in the story
began in December 1988 when researchers at the University of
Wellington in New Zealand prepared a report of a case control study
showing that the relative risk of asthma death in patients prescribed
fenoterol by metered dose inhaler was 1.55. Sub-group analyses showed
a relative risk as high as 13 for patients with the most severe asthma
(Lancet 04/29/89). The authors concluded that their findings were
consistent with the hypothesis that use of fenoterol increases the
risk of death in severe asthma.
Public Health Implications
Because of the potential public
health importance of these findings, the results were transmitted to
the Department of Health which asked for an independent review by a
panel headed by Australian Stephen Leeder. At the same time, the new
data were under review by the Asthma Task Force (ATF) which had been
constituted by the New Zealand Medical Research Council (MRC) to
investigate New Zealand’s asthma epidemic which began in the late
1970’s. According to the Christchurch Star, the ATF conducted the
largest study of asthma mortality ever undertaken between 1981 - 84,
and raw data from this project were obtained and analyzed separately
by Julian Crane and colleagues for their Lancet report.
The Leeder panel concluded that
“Assuming the basic data to be correct, these findings are sufficient
to justify public health action.” The ATF, which according to the
investigators had been critical of the Crane et al. study prior to the
review panel findings, seemed to accept the panel’s findings and the
results of a later validation exercise which showed that data had been
recorded properly. Nevertheless, the Director General of Health, who
also received advice from the Medicines Adverse Reactions Committee
(MARC), decided that “no pre-publication action is required by this
office to in any way influence the availability of use of fenoterol in
New Zealand.”
The reluctance of the Department
of Health to act at this time to restrict the use of fenoterol was
reportedly fueled by continuing criticisms of the study by members of
the ATF, by the advice of the MARC, and by criticisms from individual
epidemiologists hired by BI to evaluate the findings.
Next Chapter
The next chapter unfolds between
the time the manuscript was submitted for publication in early 1989
and the time it was actually published in April. To enable BI to make
comments on the paper to the MARC, the Department of Health, with the
permission of the authors, had sent a copy of the report to BI with
the understanding that the company would “do nothing to upset the
normal process of publication.”
BI commissioned reviews of the
study and convened a group including epidemiologists in New York to
discuss the fenoterol study. Some of the reviewers’ criticisms were
sent to the Lancet prior to publication. The review group expressed
the opinion that “...other explanations for the data were more
plausible than those reached by the authors... the study was found to
have serious flaws in design, execution and analysis which rendered
its results uninterpretable...the study provides no basis for either
physicians or patients to alter their use of fenoterol.”
After accepting the paper based
on suggestions from two reviewers for minor changes, the Lancet wrote
to Crane and colleagues to state that the new criticisms were causing
them “anxiety” and giving them “second thoughts.” Since the Lancet
felt it could not ignore the criticisms, it offered the authors the
option of withdrawing the paper, countering convincingly very critical
comments, or accepting publication in the second half of the journal
accompanied by a highly critical editorial. Disturbed that their
acceptance might be withdrawn following submissions made by the
company, the authors chose the option of seeking to counter the
criticisms in a lengthy 12 page reply to the Lancet.
Their successful reply resulted
in publication of the paper as the lead article in Lancet on April 29,
1989.
Clash Over Interpretation
The next chapter revolves around
the attempts of the Department of Health to inform physicians about
the significance of the Lancet report. It stated in a letter to
physicians that “fenoterol will not be withdrawn from the market, but
doctors should review and perhaps modify the treatment of severe
asthmatics. The use of inhaled fenoterol rather than alternative
medicines should be supported by cogent arguments.”
Anticipating the negative
effects that publication of the Lancet article and the Department
statement might have, the company had disagreed with the Department’s
statements and warned that “...we strongly protest against this and
highlight that the responsibility for legal and medical consequences
will rest with you.” BI had wanted the Department to state that
therapeutic conclusions could not be drawn from the Crane et al.
study.
Apparently acting out of this
concern, BI sent to physicians a copy of the statement by the review
panel disagreeing with the Crane et al. conclusion and a company
commentary “...to set this research in perspective to enable you to
have the fullest details about this research programme and its
interpretation...”
After the study was published,
it was further criticized by BI and the ATF. The MARC indicated it
would wait for the results from international debate before reviewing
its stance further. The New Zealand Medical Association called for an
independent review to resolve conflicting advice.
Second Study
The next set of developments
revolve around a second report on fenoterol by the Wellington group. A
major criticism of the first study was that drug data came from
different sources for cases and controls. A new case control design
was used to evaluate the same hypothesis with data from 1977 - 81
using the same source for drug information. A total of 57 cases and
227 controls were studied. The overall relative risk was 1.99 and
reached 9.82 in a subgroup with the most severe asthma. The authors
concluded that their second study added further support to the
hypothesis that inhaled fenoterol increases the risk of death in
patients with severe asthma.
Other Opinions
The results were presented at
the second annual clinical epidemiology meeting in Newcastle,
Australia in June 1989 and prominent epidemiologists were in
attendance, including Paul Stolley from the US who had done previous
work in this area. According to newspaper accounts, Dr. Stolley
concluded that evidence against the drug was too strong not to act. He
said there was only one important weakness in the design of the
research reported in the Lancet and that had been cleared up in the
new study. The Chairman of the United Kingdom’s Committee on Review of
Medicines, David Lawson, was reported to have said after hearing the
new data that if he were in New Zealand he would push to get fenoterol
off the market “tomorrow or the day after,” according to the Dominion
Sunday Times.
Replay
In an apparent replay of
previous developments, the Department of Health sought to obtain a
copy of the new study so that it could be reviewed by the MARC. The
researchers refused unless a “rock-solid” agreement could be reached
whereby the company would not receive a copy of the results prior to
publication. Under the Official Information Act, the Department
obtained the paper and passed it on to the company, according to the
Christchurch Star.
The MARC met again in July 1989
and this time concurred that the study implied long term use of
fenoterol was associated with a greater risk of death. However, the
MARC advised the Department that no further action on fenoterol was
required and suggested that animal studies be conducted to identify a
potential mechanism to explain the deaths. According to the
Christchurch Star, the researcher was to be a MARC member funded by
BI.
Public Health Reactions
Reacting to this decision, a
Public Health Association of New Zealand pharmaceutical task group
said it was “...mystified by this statement since looking for a
mechanism of death presupposes that the association is causal.
Furthermore, when there is strong evidence of an association between
fenoterol use and human death, it is a backward step to be doing
animal studies before taking further action.” The Public Health
Association wrote to the Minister of Health seeking “clarification of
the pharmaceutical company’s role in the process of policy making over
fenoterol” and recommended that all members of MARC should divulge to
the Minister any associations with the pharmaceutical companies.
In July 1989, BI convened a
second review group in Los Angeles to review the second study.
According to a newspaper account, questions were raised by the
director of the NZ Medical Research Council “...over the potential
conflict inherent in the company’s hosting such a scientific
review...,” and the Department of Health asked epidemiologist Marc
Elwood to prepare a report for them to be furnished to the MARC.
Apparently MARC has been composed primarily of clinicians and only
recently added an epidemiologist to its roster.
Government Decision
In December 1989, the Minister
of Health announced that the drug fenoterol would be restricted as of
August l, l990. The delay is to enable physicians time to assess the
need for fenoterol therapy in individual cases. The government
decision was reached following a recommendation of the MARC which had
received other reports from the company and from Marc Elwood. The
latter’s report apparently carried weight and stated that “Although it
still falls short of a totally acceptable standard of scientific
certainty, the evidence is strong enough to support clinical and
public decisions, and there do not appear to be strong
counterbalancing arguments for the benefits of this particular drug.”
According to the wording of the
government statement, the MARC agreed “that the balance of evidence
was now in favor of a causal association between fenoterol use and
asthma mortality.”
Issues Raised
Several issues are raised by the
events that have transpired in New Zealand, many of which cannot be
presented here. Those which seem of particular interest to
epidemiologic researchers include the following:
The Journal: What limits, if
any, should a journal set on the receipt and consideration of
unsolicited reviews about articles being evaluated for publication? Is
it permissible for journal editors to seek or consider reviews of
papers commissioned by a company whose interests may be affected by
the paper in question? The issue may be complicated even further if
reviews are considered after a paper has been accepted for publication
and the commissioned reviews are critical. The journal is placed in a
difficult position because it may feel that it cannot prudently ignore
the criticisms, yet they may not have been obtained during the normal
process of peer review. What are the appropriate limits on “peer
review?”
The Company: What limits, if
any, should be set on a company’s efforts to defend its interests in a
product when it believes the product is being wrongly implicated in
adverse reactions or deaths? Clearly, a company has a right to argue
against what it believes are weak data and an incorrect conclusion. It
does not serve either the company’s or society’s interest to withdraw
drugs prematurely. But what means has the company taken to reach its
conclusion about the data? What question should the company be asking
its consultants to answer (e.g., “Is there any chance the data are
wrong?” “Is there any chance the data are right?” Or should the
question be, “on balance, what conclusion is most likely true from the
data?”
The Epidemiologists: What
considerations should govern the formulation of opinions provided by
epidemiologists working for hire? Are these considerations any
different from those influencing the usual formulation of opinions
about the extent to which a hypothesis has been proven? Must a critic
consider both the strengths and weaknesses of the data, and both
directions of potential biases in rendering an opinion?
Epidemiologists are often defined as professional skeptics and
consider this one of their more distinctive and useful attributes.
What, if anything, is wrong with focusing on the reasons why some
hypothesis has not been convincingly established? Does this not best
serve the interest of science? Does it also best serve the interest of
public health?
Department of Health and
Advisory Groups: What safeguards should the government take to retain
useful drugs while at the same time not exposing the public to
unacceptable risks? Should a drug which becomes suspect for safety
reasons be considered innocent until proven guilty or guilty until
proven innocent? The normal safeguards for reaching correct
conclusions about the causal nature of an association are to require
repeat studies. However, when there are potentially serious public
health consequences to further delay, and when epidemiologic data such
as that from case control studies are not conclusive, decision makers
and advisory groups may not have the luxury of awaiting further data.
In that circumstance, what safeguards are appropriate? What
considerations should be paramount?
Published March 1990 v
|
