Dairy fat and cardiovascular disease: adjusting for potential confounders

Dear Editor:

Recent research commentaries have challenged the common belief that dairy fats and saturated fats generally are a significant risk factor for cardiovascular disease (CVD) (1). The recent study by Chen et al. (2) examined the relation between dairy fat intake and CVD in 3 prospective cohorts in the United States. They reported no significant relation between dairy fat and CVD. However, in an isocaloric substitution analysis, they noted that replacing “animal fats, including dairy fat, with vegetable sources of fats and PUFAs may reduce risk of CVD.” The dietary adjustors used for the substitution analyses were fruit, vegetables, coffee, and protein (the authors’ Figure 1, p. 1215).

A major problem with the findings was that the adjustment for potential confounders did not take into account the association of dairy fat intake with food patterns previously reported for the Health Professionals Follow-Up Study (HPFS) and the Nurses’ Health Study (NHS) prospective cohort samples used in the Chen et al. analyses. Dairy fat intake was correlated with other foods associated with an increased risk of CVD, such as sugar, refined carbohydrates, and red meat. Table 1 shows data from 2 published studies on food patterns in the NHS and HPFS samples. These data indicate that dairy fats, refined grains, sweet foods, and red meat are a central component of the “Western” pattern in both samples. The association is somewhat higher in the HPFS sample of men. In both the women’s (NHS) and men’s (HPFS) samples, the Western dietary pattern was associated with a significantly increased risk of coronary heart disease (3, 4). In the HPFS sample, the increased risk of coronary heart disease with the Western dietary pattern was RR = 1.64 (P < 0.001), not adjusted for dietary fats.

TABLE 1.

Dietary patterns: loadings on the “Western” factor for NHS and HPFS samples¹

Food or food group	Factor loadings
NHS cohort sample (women)
Refined grains	0.58*
Processed meats	0.56
Red meat	0.56
French fries	0.47
Condiments	0.44
Desserts and sweets	0.43*
Potatoes	0.41
High-fat dairy products	0.36*
Sugar beverages	0.32
HPFS sample (men)
Red meat	0.63
Processed meats	0.59
Refined grains	0.49*
Sweets and desserts	0.47*
French fries	0.46
High-fat dairy products	0.45*
Eggs	0.39
High-sugar drinks	0.38
Snacks	0.37
Condiments	0.36
Potatoes	0.33

¹The food labels shown in the table are those used in the published studies (3, 4). Factors loadings >0.30 are shown. *Dairy fats, refined grains, and sweets variables are indicated with an asterisk. HPFS, Health Professionals Follow-Up Study; NHS, Nurses’ Health Study.

In the Chen et al. (2) paper, their Table 1 showed the baseline characteristics, including selected dietary variables. Baseline rates for food group loadings in the “Western” factor, including refined starches and sugars, were not reported. Data from Table 1 indicate that high consumers of dairy fat in the HPFS and NHS samples had less-healthy lifestyles (higher alcohol consumption, less exercise, more frequent smoking), which were adjusted for, but they also consumed less cereal fiber, which is partially related to refined starches and sugars. The reported HRs for dairy fat consumption in both the direct and substitution analyses are very likely to be confounded by the unadjusted associations of dairy fat consumption with a Western dietary pattern, including refined carbohydrates and sugars. This interpretation is consistent with findings reported from the NHANES III Linked Mortality cohort, a prospective cohort (1988–2006) of a nationally representative sample of US adults. Yang et al. (5) reported a significant relation between added-sugar consumption and the risk of CVD mortality in the NHANES III sample.