Results
We first investigated the association between the FGF21 variant and macronutrient and alcohol intake. The G (major) allele of rs838133 was associated with lower intake of total sugars and alcohol, and higher intakes of protein and fat (Figure 1), consistent with the expected effect of an increase in FGF21 concentration.2 We scaled all results per additional G allele to mimic the effect of elevated FGF21 concentrations.
Higher genetically proxied FGF21 concentration was associated with greater body mass index, body fat percentage and waist and hip circumferences, but with lower waist-to-hip ratio (Figure 1). Additionally, higher genetically proxied FGF21 concentration was associated with lower concentrations of low-density lipoprotein cholesterol and triglycerides, lower systolic and diastolic blood pressure, and lower C-reactive protein concentrations, but was not associated with fasting glucose or fasting insulin (Figure 1). There was a positive association of higher genetically proxied FGF21 concentration with the liver enzyme alkaline phosphatase, but a negative association with aspartate aminotransferase, gamma glutamyltransferase, and direct and total bilirubin concentrations (Figure 1). There was suggestive evidence of a positive association between genetically proxied FGF21 concentration and lifespan (based on parental lifespan) (Figure 1).
In analyses of cardiovascular diseases, higher genetically predicted FGF21 concentration was strongly associated with a reduced risk of venous thromboembolism, and had suggestive inverse associations with coronary artery disease, heart failure, and ischemic stroke (Figure 2). There was a suggestive association of higher genetically proxied FGF21 concentration with reduced risk of Alzheimer’s disease (based on clinically diagnosed Alzheimer’s disease and Alzheimer’s disease by proxy cases and their corresponding controls), but no association with type 2 diabetes (Figure 2).