Modified Mediterranean diet and survival

BMJ 2005; 330 doi: http://dx.doi.org/10.1136/bmj.330.7503.1329-a (Published 02 June 2005) Cite this as: BMJ 2005;330:1329

Key confounder was missed

  1. Eddie Vos (vos{at}health-heart.org), maintains www.health-heart.org
  1. Sutton (Qc), Canada J0E 2K0

    EDITOR—The title “modified Mediterranean diet” for the study reported by Trichopoulou et al is unfortunate and vague,1 as would be a “modified Asian or American” diet. To many Westerners it might simply say: eat more pasta. The modification in the study diet variable was the addition of polyunsaturates to the numerator of the diet score. This, however, is problematic in not differentiating between the often excessive omega 6 linoleic acid2 and the mostly deficient family of the three main omega 3 fatty acids (one plant based and two fish based).3

    While justifying (on the basis of reference 12 in part 2 of the discussion section of the paper) adding polyunsaturates to the diet score numerator because of coronary heart disease benefits, the authors misrepresent the study. Plant based omega 6 had actually dropped (by 7% in plasma fatty acids, at one year) and by far the greatest change in fatty acids was a quantitatively small but proportionally large increase in plant based omega 3 α linolenic acid (+67% in plasma at one year).3 This study specifically tried to increase (successfully) the ratio of omega 3 to omega 6 with a margarine made from canola (rape, colza) oil that was supplied to the study, not to increase polyunsaturates. Total polyunsaturates were 5% lower in serum in study participants than in controls after one year, not higher (table 4 in De Lorgeril et al4).

    Part of the benefit of a traditional Cretan-Mediterranean diet was probably omega 3, as outlined by Leaf in an editorial.3 The study reported by Trichopoulou et al adds the omnipresent vegetable oil omega 6 linoleic (over 50% of the fatty acids in soy, corn, cottonseed, sunflower, and safflower oils, for example) into the equation, for which clinical benefits have never been shown.2


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.