Causal relationship between diet and knee osteoarthritis: A Mendelian randomization analysis

Lv X, Liang F, Liu S, Deng X, Lai R, Du J, Luo J · PLoS One

Genetic instruments for higher coffee intake were associated with higher knee OA risk (OR 2.04, 95% CI 1.61-2.58); instruments for dried fruit (OR 0.46), cereals (OR 0.63), cheese (OR 0.62) and oily fish (OR 0.67) with lower risk. Multivariable analysis suggested coffee and oily fish act through obesity, BMI, diabetes, hypertension and prolonged standing, and both lost significance once those were in the model. Neither cooked vegetables nor salad reached significance, a null worth stating alongside the observational vegetable associations.

Population

GWAS summary data for 18 dietary factors (n = 335,394-462,342) against knee osteoarthritis (n = 403,124)

Intervention

Genetically predicted intake of processed meat, poultry, beef, oily fish, non-oily fish, pork, lamb, alcohol, tea, coffee, dried fruit, cereals, cheese, bread, cooked vegetables, salad, and fresh fruit

Comparator

Lower genetically predicted intake

Limitations

Mendelian randomization inherits the weaknesses of the food-frequency phenotypes the GWAS were built on, and dietary instruments are notoriously entangled with each other and with socioeconomic position. Largely European-ancestry data. The multivariable finding that effects run through BMI means these are not independent dietary effects.