To investigate the causal relationship between dietary habits and phenotypes of gastroesophageal reflux disease using Mendelian randomization.

A total of 24 dietary patterns, including cereal, meat (poultry, beef, pork, oily fish, non-oily fish, processed meat, and lamb/mutton intakes), vegetable (salad/raw vegetable and cooked vegetable intakes), fresh fruit, and beverage (alcohol intake frequency, alcoholic drinks per week, tea intake, and coffee, water, fizzy drink, and milk intakes) intakes, another food intake (cheese, starchy food, dark chocolate, and ice cream intakes), and type of diet (salt added to food, hot drink temperature, and low-calorie diet), were included from the Integrative Epidemiology Unit and UK Biobank. The GERD dataset was obtained from the FinnGen Consortium Release 9, and a Mendelian randomization approach was employed to analyze the causal association between these dietary factors and the risk of GERD. The main method of Mendelian randomization analysis was inverse variance weighting (IVW). The stability of the instrumental variables was tested by calculating the F-value statistic, and the heterogeneity was tested by Cochran’s Q statistics. In addition, weighted medians (WM). and MR-Egger regression and leave-one-out methods were used to assess sensitivity and pleiotropy. The Steiger test detects reverse causality.

IVW showed that salad/raw vegetable intake (OR=0.411, 95%CI: 0.189-0.894; P=0.025), and the IVW, WM and MR-Egger regression were in the same direction. Steiger test found no evidence of reverse causality (steiger pval=2.11×10^-50). The results of sensitivity analyses were robust, and there was no heterogeneity, pleiotropy, or reverse causality.

The results of Mendelian randomization analysis support a potential causal relationship between salad/raw vegetable intake and gastroesophageal reflux disease.