Vegetarian dietsBMJ 2009; 339 doi: https://doi.org/10.1136/bmj.b2507 (Published 08 July 2009) Cite this as: BMJ 2009;339:b2507
All rapid responses
The word limit for editorials may account for some of Dr Byatt’s
concerns. It was certainly not my intention to be dismissive, patronising
or condescending about a vegetarian diet.
The vegetarian dietary pattern has been associated with several
positive health outcomes including reduced rates of obesity and the fact
that relatively few vegetarians are smokers must contribute to the overall
lower rates of smoking related diseases amongst those who choose this
However, there is ample evidence that dietary practices independently
of smoking can indeed influence important indicators including blood
pressure, measures of body fatness and cholesterol levels. The fact that
we cannot disentangle which component of the vegetarian dietary pattern
confers health benefits does not in any way detract from its positive
I am very happy to acknowledge animal welfare as an additional
justification for considering vegetarianism. I was not implying any
particularly novel dietary pattern when referring to a “modified
conventional western dietary pattern”. Many diets consumed in western
countries are relatively high in saturated fat, sugars and energy dense
foods rich in these nutrients. I was merely attempting to indicate that
when the totality of evidence is considered, it appears that by modifying
what is typically eaten in such countries, one may be able to achieve the
health outcomes which have been attributed to traditional Asian,
Mediterranean, African and vegetarian diets.
Competing interests: No competing interests
I would like to thank Jim Mann for his very nearly excellent leading
article. It is extremely informative, positive and encouraging in places.
However, having read it several times, I am still trying to work out what
his message actually is here!
In paragraphs 2 to 5 (of 10) Mann carefully spells out the
statistically significant massive health benefits of a 'vegetarian diet'
(incidentally, without defining what he means by this term). He points
out that, for the population included in the substantial trials he quotes,
the overall all-cause SMR is 50% (although he doesn't explicitly specify
which of the trials yield this finding), and the cancer SMR is about 70%
(whether meat-eating or not) - in other words this is an extremely healthy
trial group overall, whether 'vegetarian' or meat-eating. Despite this
very much above average survival for both groups, he shows that within the
study population there is still a distinct survival benefit from having a
vegetarian diet (including a reduced IHD mortality of ~24%, a relative
risk of 0.39 for type 2 diabetes and =<0.45 for hypertension;
additionally a World Cancer Research Fund/American Institute for Cancer
Research report suggested meat as 'the only "convincing" dietary
association with colon cancer').
In paragraph 7 he sets out the possible health risks of a opting for
a 'vegetarian diet' and in paragraph 8 acknowledges that 'vegetarianism
covers a wide spectrum' of sub-groups with consequently differing
potential risk types and magnitudes.
In paragraph 9 Mann states that 'vegetarians who participate in long
term prospective studies have appreciably lower mortality than the general
population' (omitting to mention that meat eaters do, too [as shown in
para 6], albeit less so), but there is little evidence to show it is the
'meat avoidance' per se that achieves this. He also implies that much of
the benefit is because vegetarians tend to be non-smokers.
In his summary in paragraph 10, Mann states that vegetarianism is
therefore overall an 'acceptable' dietary practice, but not better than
'Mediterranean, Asian or "modified conventional Western" dietary
patterns.' - not defining any of these terms (especially the latter, which
means nothing to me, and a Google search for "modified conventional
Western diet" returns 0 hits).
His conclusion from all this expertly summarized mass of heavyweight
data is that "Claims for unique benefits from a vegetarian lifestyle may
be more appropriately made on ecological grounds." I don't argue with
this assertion, but I have four concerns:
1. Mann implies that vegetarians' (non-)smoking habits might account
for their lower mortality ratios. If that aspect of the work was done in
the original trials, then please summarize it explicitly. If not, how can
the study be regarded as meaningful? Surely it would be unethical not to
control for smoking in any study purporting to examine aspects of
cardiovascular epidemiology (did none of the quoted studies of diet on IHD
outcomes consider this factor?).
2. The Department of Health states on its website that the 'target on
obesity is the Government's first, high level response to the major public
health problem posed by the continuing rise in obesity.' Mann's
analysis mentions 'the low rates of obesity.. among vegetarians' although
not quantifying this. Why not investigate a possibly useful prevention
approach as suggested by these data, rather than suggesting an alternative
reason for its 'unique benefits?'
3. Why pick only the ecological argument for considering
vegetarianism to have an unique benefit? Animal welfare may be considered
as compelling a reason to consider vegetarianism (or veganism) as a
lifestyle choice. Why ignore this alternative consideration?
4. The tenor of the discussion in the last two paragraphs is almost
dismissive. I can accept the science shows that we do not yet know for
certain which aspect of being a 'vegetarian' confers the benefit.
However, as an analogy, NICE report that UKPDS showed an overall benefit
of metformin of 0.68 on all-cause mortality in overweight participants.
Do we know exactly which of metformin's biochemical actions achieves this
effect? We still use it, though.
I am left feeling that Mann might be being patronising about
vegetarianism and thus tacitly inviting his audience to condescend to
those who choose (for whatever reason) this lifestyle. Whether he does
this unintentionally or not is impossible to discern. The unqualified use
of the word 'acceptable' in his concluding paragraph can be taken to imply
'tolerable' or 'just good enough', although the same word can also be used
to mean 'absolutely fine'. This potentially confusing quasi-homonymy is
perhaps analogous to the current use of the word 'competent' in medical
education. It can be taken as meaning 'skilled' or 'expert' on the one
hand, or 'barely achieving an acceptable level' on the other, depending on
how it is said or qualified. This causes much unease amongst assessors
and trainees alike as to what was actually meant when it was deployed! I
realize that the constraints of a word-limited editorial result in much
qualifying text having to be redacted, but I hope in his response Jim Mann
will answer the numbered points above and also confirm which usage for the
word 'acceptable' he intended.
On the positive side, however, there are some very encouraging data
summarised here suggesting that at least some of the lifestyle choices of
vegetarians are evidence based!
(accessed 5 September 2009)
2. Humane Slaughter in British Abattoirs a 'Sham'
http://animalaid.org.uk/h/n/AA/HOME/ (accessed 5 September 2009)
3. http://guidance.nice.org.uk/CG66/Guidance/pdf/English (accessed 5
I am a vegetarian for health, ecological and animal welfare reasons
Competing interests: No competing interests