Optimistic interpretation of trial data by proponents of polypill
Over the past years, a consensus has emerged that treatment decisions
should be based, wherever possible, on empirical evidence. And in few
areas of medicine have the demands placed on that evidence been as
stringent as in the field of prevention of cardiovascular disease in
general and coronary heart disease in particular.
The paper by Wald and Law, however, seems to cast doubt on this
rigorous approach, suggesting that a cocktail of six medications in a
single pill (“polypill”) would lower the incidence of ischaemic heart
disease events by 88% and of stroke by 80%, and that it might therefore be
given with impunity to “everyone aged 55 and older and [to] everyone with
existing cardiovascular disease”.
In spite of the claim by the authors that widespread use of this
cocktail might have a “greater impact on the prevention of disease in the
Western world than any other single intervention”, we feel obliged to
introduce a note of caution into the argument before this approach is made
the basis of treatment recommendations.
It is generally accepted that treatment effects can only be
determined within the context of a randomised controlled clinical trial.
The results of the trial must take non-compliance and a range of dose-
responses into account. It is not permissible to extrapolate from a post-
hoc analysis of a trial what result might be expected with 100% compliance
and taking account only of those individuals who showed maximum benefit in
terms of risk-factor reduction. Thus the theoretical figure of an expected
61% reduction in ischaemic heart disease events from cholesterol-lowering
using statins is about twice the actual risk reduction seen in any statin
trial to date. The same reasoning can be applied to blood-pressure
lowering – here trials have shown a lowering of ischaemic heart disease
risk of about 20%, not the 46% suggested by Wald and Law. Trial evidence
for the benefits of anti-thrombotic therapy in primary pervention is not
as strong as that for cholesterol or blood pressure lowering, while trial
evidence of a benefit from folate use does not exist. Moreover,
combination therapies must be subjected to the same scrutiny as therapies
containing single ingredients before claims can be made for their
effectiveness, and, of course, no such trial evidence for the polypill
It would be nice if we lived in a polypill world. One for heart
disease, one for mood (perhaps the “happy pill” suggested by the Dr.
Richard Smith in the editorial accompanying the paper), and maybe even one
for finding the right partner. However this is not the case, and we are
forced to deal with each problem in turn, often compromising and accepting
a less than perfect result. There are no quick fixes, in life or in
medicine, and the polypill is not one for heart disease.
Competing interests: No competing interests