Non-fermented and fermented milk intake in relation to risk of ischemic heart disease and to circulating cardiometabolic proteins in swedish women and men: Two prospective longitudinal cohort studies with 100,775 participants

Abstract

Background: The effect of milk on the risk of ischemic heart disease (IHD) and acute myocardial infarction (MI) is unclear. We aimed to examine the association between non-fermented and fermented milk consumption on these endpoints and investigate the relationship between milk intake and cardiometabolic-related proteins in plasma.

Methods: Our study is based on two Swedish prospective cohort studies that included 59,998 women and 40,777 men without IHD or cancer at baseline who provided repeated measures of diet and lifestyle factors and plasma proteomics data in two subcohorts. Through registry linkage, 17,896 cases with IHD were documented during up to 33 years of follow-up, including 10,714 with MI. We used time-updated multivariable Cox regression analysis to examine non-fermented or fermented milk intake with time to IHD or MI. Using high-throughput multiplex immunoassays, 276 cardiometabolic plasma proteins were measured in two subcohorts. We applied multivariable-adjusted regression models using a discovery-replication design to examine protein associations with increasing consumption of non-fermented or fermented milk.

Results: The results for non-fermented milk differed by sex (p-value for interaction = 0.01). In women, we found a pattern of successively greater risk of IHD and MI at non-fermented milk intake levels higher than 1.5 glasses/day. Compared with an intake of 0.5 glass/day (100 mL/day), non-fermented milk intake of 2 glasses/day in women conferred a multivariable-adjusted hazard ratio (HR) of 1.05 (95% CI 1.01-1.08) for IHD, an intake of 3 glasses/day an HR of 1.12 (95% CI 1.06-1.19), and an intake of 4 glasses/day an HR of 1.21 (95% CI 1.10-1.32). Findings were similar for whole, medium-fat, and low-fat milk. We did not detect higher risks of IHD with increasing milk intakes in men. Fermented milk intake was unrelated to the risk of IHD or MI in either sex. Increasing non-fermented milk intake in women was robustly associated with a higher concentration of plasma ACE2 and a lower concentration of FGF21.

Conclusions: We show a positive association between high amounts of non-fermented milk intake and IHD in women but not men. We suggest metabolic pathways related to ACE2 and FGF21 potentially underlie the association.

Keywords: ACE2; Cohort; FGF21; Fermented; Ischemic heart disease; Milk; Myocardial infarction; Non-fermented.

Fig. 1

Sex-specific spline curves of the relation between non-fermented milk intake with time to ischemic heart disease. A (age-adjusted) and B (multivariable-adjusted) illustrate the pattern for women, and C (age-adjusted) and D (multivariable-adjusted) for men. Covariates were age, time-updated total energy intake, fermented milk intake, cheese intake, intake of fruit and vegetables, intake of red meat, intake of soft drinks and juice, intake of coffee, alcohol intake, total fat intake, saturated fat intake, vitamin- and mineral supplement use, body mass index, height, educational level, living alone, calcium supplementation, vitamin D supplementation, ever use of cortisone, leisure time exercise, walking/cycling, smoking status, baseline cardiovascular disease other than IHD, baseline diabetes mellitus, and baseline weighted Charlson’s comorbidity index. The bar plot shows the distribution of non-fermented milk intake. One glass of milk corresponds to 200 mL

Fig. 2

Sex-specific multivariable adjusted spline curves of the relation between non-fermented milk intake by fat content with time to ischemic heart disease. A–C illustrates the pattern for women, and D–F demonstrates the pattern for men. For each investigated fat % type of milk intake, individuals with milk intakes consisting of other fat concentrations of 200 mL or higher were censored [27]. Covariates were age, time-updated total energy intake, fermented milk intake, cheese intake, intake of fruit and vegetables, intake of red meat, intake of soft drinks and juice, intake of coffee, alcohol intake, total fat intake, saturated fat intake, vitamin- and mineral supplement use, body mass index, height, educational level, living alone, calcium supplementation, vitamin D supplementation, ever use of cortisone, leisure time exercise, walking/cycling, smoking status, baseline cardiovascular disease other than IHD, baseline diabetes mellitus, and baseline weighted Charlson’s comorbidity index. The bar plot shows the distribution of milk intake. One glass of milk corresponds to 200 mL

Fig. 3

Sex-specific spline curves of the relation between fermented milk intake with time to ischemic heart disease. A (age-adjusted) and B (multivariable-adjusted) illustrate the pattern for women, and C (age-adjusted) and D (multivariable-adjusted) for men. Covariates were age, time-updated total energy intake, non-fermented milk intake, cheese intake, intake of fruit and vegetables, intake of red meat, intake of soft drinks and juice, intake of coffee, alcohol intake, total fat intake, saturated fat intake, vitamin- and mineral supplement use, body mass index, height, educational level, living alone, calcium supplementation, vitamin D supplementation, ever use of cortisone, leisure time exercise, walking/cycling, smoking status, baseline cardiovascular disease other than IHD, baseline diabetes mellitus, and baseline weighted Charlson’s comorbidity index. The bar plot shows the distribution of fermented milk intake

Fig. 4

Baseline characteristic-stratified multivariable-adjusted hazard ratios of ischemic heart disease in women (A) and men (B) per time-updated serving (200 mL) of non-fermented milk or per serving (200 mL) of fermented milk intake (C for women and D for men). The model included age, time-updated total energy intake, non-fermented milk intake, cheese intake, intake of fruit and vegetables, intake of red meat, intake of soft drinks and juice, intake of coffee, alcohol intake, total fat intake, saturated fat intake, vitamin- and mineral supplement use, body mass index, height, educational level, living alone, calcium supplementation, vitamin D supplementation, ever use of cortisone, leisure time exercise, walking/cycling, smoking status, baseline cardiovascular disease other than IHD, baseline diabetes mellitus, and baseline weighted Charlson’s comorbidity index

Fig. 5

Multivariable-adjusted spline curves of the relation between non-fermented milk intake with the proteins ACE2 (A) and FGF-21 (B) in the discovery and the replication subcohort. Covariates were age, time-updated total energy intake, fermented milk intake, cheese intake, intake of fruit and vegetables, intake of red meat, intake of soft drinks and juice, intake of coffee, alcohol intake, total fat intake, saturated fat intake, vitamin- and mineral supplement use, body mass index, height, educational level, living alone, calcium supplementation, vitamin D supplementation, ever use of cortisone, leisure time exercise, walking/cycling, smoking status, baseline cardiovascular disease other than IHD, baseline diabetes mellitus, and baseline weighted Charlson’s comorbidity index