Research article
Meat consumption and mortality - results from the European Prospective Investigation into Cancer and Nutrition
BMC Medicine 2013, 11:63 doi:10.1186/1741-7015-11-63
This paper has misrepresented the data:
Zoe Harcombe (2013-03-19 10:28) University of West of Scotland
1) All of the hazard ratios [HR (b) HR (c) and HR (d)] in Table 2 have used the second
lowest intake of red meat, processed meat and poultry as the baseline and not the
lowest intake.
Had the baseline been set at the lowest red meat intake, HR (c) would have been 1.00/0.93/0.94/0.93/0.96/1.07
in order of lowest to highest red meat consumption i.e. consumption of between 10
g/day and 160 g/day having a lower HR than either extreme. The absence of association
with red meat was noted by the researchers, but the selection of the second lowest
intake group flatters the non-association figures.
Had the baseline been set at the lowest processed meat intake, HR (c) would have been
1.00/0.96/0.99/1.05/1.16/1.38 in order of lowest to highest processed meat consumption
and the 1.44 headline number from the article would be lower.
2) Studies normally allocate food consumption to quartiles or quintiles. This study
has allocated the results such that 98.63% of all deaths occur in the categories below
the highest for red meat and 99.15% of all deaths occur in the categories below the
highest for processed meat.
The headline should have been "99% of deaths in a population of 448,568 people over
a 12.7 year follow up were not associated with meat", but then that wouldn't have
made the headlines.
Competing interests
None declared
Its the tobacco and not the meat
Peter Hill (2013-04-02 12:33) Rhodes University, South Africa
There is no doubt that the title of the EPIC meat consumption and mortality study ( BMC Medicine 2013,11:63) is eye-catching and it may be possible that the authors conclusions will find their way into the media, albeit couched as a dire warning to the consumers of processed meats.
Unfortunately, the authors have not reflected on some key data contained the Results section nor on important comments made in the Discussion section in crafting their Conclusions :
�� Results (paragraph 1, page 7 ) ��There was also a statistically significant interaction between smoking and processed meat consumption (p-interaction 0.01), with mortality being significantly increased among former (HR=1.68, 95% CI 1.29 to 2.18) and current smokers (HR=1.47, 95% CI 1.18 to 1.83) but there was no association among never smokers (HR=1.24, CI 95% 0.89 to 1.72). However, the small number of deaths among never smokers has to be taken into account (n=44; 72 among former and 107 among current smokers in the top consumption category)��.
�� Discussion (paragraph 6, page 10) ��We cannot exclude residual confounding, in particular due to incomplete adjustment for active and passive smoking. The sub-group analysis for processed meat showed heterogeneity according to smoking, with significant associations only in former and current smokers and no associations in never smokers, which is compatible with residual confounding by smoking.��
If , as the authors found, there was no significant association between processed meat consumption and increased mortality in the never smokers cohort but that there was a significant association between smokers, current and former and increased mortality, then surely it is reasonable to conclude that it is tobacco use and not processed meat consumption that points to an association with ��increased risk of early death.��
Perhaps Conclusions along the lines of the following may more accurately reflect the study findings: �� The results of our analyses suggest that men and women with a high consumption of processed meat may not be (my inclusion) at increased risk of early death, in particular due to cardiovascular diseases but also due to cancer, unless they are former or current smokers (my inclusion)��.
The use of the words ��are at increased risk�� in the Conclusions section implies causality, whereas the study design is able, at best, to suggest association.
PW Hill, PhD
Competing interests
None
top