Dairy Product Consumption and Cardiovascular Health: a Systematic Review and Meta-Analysis of Prospective Cohort Studies.

The association between dairy product consumption and cardiovascular health remains highly debated. We quantitatively synthesized prospective cohort evidence on the associations between dairy consumption and risk of hypertension (HTN), coronary heart disease (CHD) and stroke. We systematically searched PubMed, Embase, and Web of Science through August 1st, 2020 to retrieve prospective cohort studies that reported on dairy consumption and risk of HTN, CHD or stroke. We used random-effects models to calculate the pooled relative risk (RR) and 95% confidence interval (CI) for the highest vs the lowest category of intake and for 1 serving/day increase in consumption. We rated the quality of evidence using NutriGrade. Fifty-five studies were included. Total dairy consumption was associated with a lower risk of HTN (RR for highest vs lowest level of intake: 0.91, 95% CI: 0.86-0.95, I2 = 73.5%; RR for 1 serving/day increase: 0.96, 95% CI: 0.94-0.97, I2 = 66.5%), CHD (highest vs lowest level of intake: 0.96, 95% CI: 0.92-1.00, I2 = 46.6%; 1 serving/day increase: 0.98, 95% CI: 0.95-1.00, I2 = 56.7%), and stroke (highest vs lowest level of intake: 0.90, 95% CI: 0.85-0.96, I2 = 60.8%; 1 serving/day increase: 0.96, 95% CI: 0.93-0.99, I2 = 74.7%). Despite moderate to considerable heterogeneity, these associations remained consistent across multiple subgroups. Evidence on the relationship between total dairy and risk of HTN and CHD were of moderate quality and of low quality for stroke. Low-fat dairy consumption was associated with lower risk of HTN and stroke, and high-fat dairy with a lower risk of stroke. Milk, cheese, or yogurt consumption showed inconsistent associations with the cardiovascular outcomes in high vs. low intake and dose-response meta-analyses. Total dairy consumption was associated with a modestly lower risk of hypertension, CHD and stroke. Moderate to considerable heterogeneity was observed in the estimates and the overall quality of the evidence was low to moderate.