Abstract
BACKGROUND: Mental illness has been presenting serious public health concerns. Studies have linked mental disorders to poor dietary patterns. The study aimed to estimate the causal relationship between dietary macronutrient intake and mental disorders using a bidirectional two-sample Mendelian randomization (MR) analysis. METHODS: The causal relationships between dietary macronutrient intake (i.e., carbohydrate, fat, protein and sugar) and mental disorders (i.e., anxiety, bipolar disorder, depression, neuroticism and schizophrenia) were explored on the basis of UK Biobank and 5 datasets. Inverse-variance weighted (IVW) and weighted median (WM) methods were applied to calculate the potential causation between macronutrient and mental disorders. MR-Egger's intercept and MR-PRESSO approaches were used to verify the MR assumptions. Outcomes were reported as odds ratios (OR) and regression coefficients (β), which were adjusted by False Discovery Rate (FDR) corrections. RESULTS: Forward MR analyses revealed that genetically predicted high-fat and high-protein intakes were related to an increased risk of anxiety [fat: OR = 3.103, 95% confidence interval (CI): 1.015 to 9.486, P(IVW) = 0.047, FDR = 0.209; protein: OR = 3.364, 95% CI: 1.071 to 10.567, P(IVW) = 0.038, FDR = 0.189]. High-carbohydrate intake was associated with BD (OR = 1.468, 95% CI: 1.052 to 2.048, P(IVW) = 0.024, FDR = 0.160), and high-protein intake was also linked to depression (OR = 1.443, 95% CI: 1.227 to 1.696, P(IVW) < 0.001, FDR < 0.001). Reverse MR analyses indicated that neuroticism resulted in higher sugar intake (β = 7.556 g, 95% CI: 0.610 to 14.502 g, P(IVW) = 0.033, FDR = 0.189). Additionally, schizophrenia was associated with lower intakes of carbohydrate, fat, protein and sugar. CONCLUSIONS: Our findings demonstrated that although the effects of psychiatric diseases on diet patterns have been well understood, the potential impacts of dietary intake on mental disorders are needed to be highlighted. CLINICAL TRIAL NUMBER: Not applicable.