Dairy fat: does it increase or reduce the risk of cardiovascular disease?

Abstract

An important study published in the current issue of the Journal (1) reports on the association between dairy fat consumption and the risk of cardiovascular diseases (CVDs), 3 coronary heart disease (CHD), and stroke with the use of data from 3 large US cohorts (5,158,337 person-years of followup). In addition to analyzing the associations with dairy fat intake by using models containing typical confounding factors such as age, smoking, and BMI, they also modeled isoenergetic replacement of dairy fat with other fat sources and carbohydrate in relation to CVD risk. The authors' conclusion was that dairy fat intake was not significantly associated with the risk of total CVD, CHD, and stroke, but replacing energy from dairy fat with PUFA-rich fat and whole-grain carbohydrates (but not refined starch and sugar) was associated with a significantly lower risk of CVD, CHD, and stroke. The results for CHD risk are in accordance with previously published pooled US cohort analyses that replaced the saturated fat with PUFAs (2), but also with meta-analyses of cohort studies (3) and a meta-analysis of 8 randomized controlled trials (RCTs) with "hard" CHD endpoints (4). The replacement of dairy fat with PUFAs leading to lower stroke risk seems to deviate somewhat from the traditional diet-heart hypothesis, and it is notable that the replacement analyses for stroke are very similar to those for CHD. To our knowledge, this has not been previously reported. The current study found that replacing dairy fat with other animal (presumably meat) fat was associated with an increase in CVD risk, but unfortunately, the sources of this fat were not defined. This is important because other studies have shown red and processed meat to be associated with increased CVD risk, whereas the opposite was the case for poultry meat (5).A considerable limitation of the present study arises from the use of dairy fat without consideration of the dairy food that delivers it. It is complex to separate dairy fat from the dairy product in observational analyses with replacement models, because correlations between saturated fat from dairy products and the whole food itself will be high. A well-conducted RCT that exchanges dairy fat for fat from other sources in isoenergetic conditions is needed to find supporting evidence. As acknowledged by the authors, any impact of the food matrix could not be assessed but may well be important. A recent meta-analysis of RCTs compared blood lipid responses from butter and hard cheese and showed that cheese lowered