Folic acid, homocysteine, and cardiovascular disease: judging causality in the face of inconclusive trial evidence

Sir,

When, in future, Wald et al. turn their attention to marketing their
polypill, they will presumably rely heavily on the results of randomised
trials to support the inclusion of statins, beta-blockers, diuretics, ACE
inhibitors and aspirin. They might, however, reflect on how this will
square with their readiness to dismiss the clinical trials of folic acid
in the prevention of vascular events. [1]

In order to promote the folic acid component of their polypill, the
authors emphasize the data from observational studies which correlate
raised homocysteine levels with an increased risk of vascular disease.
Although a causal link has been disputed, [2] they argue that such a
conclusion is unwarranted. However, even if they were correct, it would
not follow that folic acid reduces vascular events. Thus, contrary to
their assertion, clinical trials were still required. Of seven randomised
trials, none showed any statistically significant difference in favour of
folic acid reducing ischaemic heart disease while only one reported
benefit in reducing stroke. Moreover, meta-analysis showed no efficacy in
either disease. Despite this evidence, Wald et al. maintain the contrary
view that folic acid reduces vascular events. But they do so only by
cherry picking, as well as twisting and distorting the data, to satisfy
their vested interests.

This is not to argue the case for the primacy of randomised trials or
meta-analyses over observational studies - indeed, there are sound reasons
for rejecting both methodologies as providing nothing but entertainment
for statisticians. Rather, it is to draw attention to the poverty of the
data emanating from both randomised trials and observational studies that
allows so much interpretation and manipulation. [3]

References

[1] Wald DS, Morris JK, Law M, Wald NJ. Folic acid, homocyteine, and
cardiovascular disease: judging causality in the face of inconclusive
trial evidence. BMJ 2006;333;1114-7.

[2] Lewis SJ, Ebrahim S, Davey-Smith G. Meta-analysis of the MTHFR C
to T polymorphism and coronary heart disease: does the totality of
evidence support a causal role for homocysteine and the preventive
potential of folate? BMJ 2005;331;1053-6.

[3] Penston J. An enlightening analogy. Rapid response letter. BMJ
22nd November 2006.

Competing interests:
None declared