Intended for healthcare professionals

Rapid response to:

Research

Anatomy of health effects of Mediterranean diet: Greek EPIC prospective cohort study

BMJ 2009; 338 doi: https://doi.org/10.1136/bmj.b2337 (Published 24 June 2009) Cite this as: BMJ 2009;338:b2337

Rapid Response:

Relative impact of dairy product intake in the Mediterranean diet remains to be determined; differences in the contribution of various dietary components in this study were not statistically significant

Trichopoulou et al in the British Medical Journal (2009)(1) analyze
the relative contribution of the individual dietary components of the
Mediterranean diet on mortality in the Greek cohort of the EPIC study
(European Prospective Investigation in Cancer and Nutrition). The abstract
of the article states that the Mediterranean diet was associated with a
significant reduction of mortality and that the contribution of the legume
dietary component to this finding was 9.7%, while the contribution of
dairy products was “minimal”. However, the findings of the paper suggest
that dairy product intake and legume intake had a similar impact on
mortality in this cohort and that the only dietary component with a
statistically significantly impact on mortality was alcohol intake.

In the study, there is a table labeled “mutually adjusted mortality
ratios associated with intake of components of Mediterranean diet”. In
that table, the mortality ratio associated with a high intake of legumes
is 0.942 (0.825-1.076, p=0.379) and for a high intake of dairy products is
1.069 (0.931-1.227, p=0.345). This indicates the negative impact of a high
dairy product intake on mortality is similar (insignificantly greater)
compared to the contribution of high legume intake on mortality when the
impact of the individual components of the diet are mutually adjusted.

However, when the authors analyze the same data in this paper using a
Mediterranean diet score, the results suggest that the relative
contribution to the mortality benefits of the Mediterranean diet for
legume intake is 9.7% and for dairy intake is 4.5%.
The disparity of relative impact of dairy products and legumes on
mortality between the mutually adjusted analysis and the analysis using
the Mediterranean diet score is a function of the limitations of using the
Mediterranean diet score in assessing the impact of the individual dietary
components. With the Mediterranean diet score analysis, the components are
not mutually adjusted for the effects of all the other dietary components.

The Mediterranean diet score has been quite valuable in prior
publications in analyzing the effect of the adherence to a Mediterranean
diet on mortality.(2) However, an analysis to determine the impact of
individual dietary components on mortality where a mutual adjustment of
all dietary components is made is preferable to an analysis using the
Mediterranean diet score in which the mutual adjustment of all dietary
components is not made. This is analogous to a multivariate regression
analysis of individual factors being more reliable than a univariate
analysis of those same factors when trying to assess the impact of the
individual factors on mortality.

Furthermore, it should be noted that the differences in the relative
contributions of the individual components in regards to overall mortality
using an analysis of the Mediterranean diet score in this article are not
shown to be statistically significant. When a low consumption of meat is
said to make a 16.6% relative contribution, fruit consumption 11.2%,
legume 9.7%, and high dairy consumption 4.5%, there is no demonstration or
specific claim made that these differences in the relative contribution of
the components are statistically significant. Using the Mediterranean diet
score analysis, there were trends for the differences in relative impact
of the individual dietary components without statistical significance
being present. When the data was analyzed by the authors using the
mutually adjusted mortality ratio approach without using the Mediterranean
diet score, only the alcohol component was statistically significant.

Hence, the statistically significant relative impact of the various
dietary components of a Mediterranean diet on mortality, including a low
dairy product intake, is an issue that remains to be defined.

1. Trichopoulou A, Bamia C, Trichopoulos D. Anatomy of health effects
of Mediterranean diet: Greek EPIC prospective cohort study. BMJ 2009;
338:b2337.

2. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to
a Mediterranean diet and survival in a Greek population. NEJM 2003; 348:
2599-2608.

Competing interests:
None declared

Competing interests: No competing interests

18 December 2009
Eric Roehm
cardiologist
Austin, Texas 78731