Effect of caffeine intake on self-reported and genetic prediction of osteoarthritis: an epidemiological study and Mendelian randomization analysis

Ji Z, Shen Y, Chen D, Liu Z, Dai J, Dai B, Deng W · Frontiers in Nutrition

Runs the observational and the genetic analysis side by side in one paper, and here they agree. Fully adjusted odds of osteoarthritis were 1.19 (95% CI 1.01-1.40) at 95-206 mg of caffeine a day and 1.30 (1.10-1.52) at 206 mg or more, with a spline inflection near 95 mg; genetically predicted coffee intake gave an odds ratio of 1.94 (1.47-2.52) for knee OA across 22 variants, consistent across methods. Running both in one paper matters because observational and genetic coffee analyses have often disagreed.

Population

14,580 adults aged 20 and over in NHANES 2009-2016, 1,753 of them self-reporting a doctor's diagnosis of osteoarthritis, plus a two-sample Mendelian randomization of coffee intake (UK Biobank, 428,860) against knee OA (24,955)

Intervention

Caffeine intake from a 24-hour dietary recall in four bands, and genetically predicted coffee intake

Comparator

Under 11 mg of caffeine a day; lower genetically predicted intake

Limitations

Self-reported diagnosis and a single day's recall in the observational arm, and the usual instrument limitations in the genetic arm, whose outcome GWAS is small. Funded by a Nanjing Medical University college research fund; no conflicts.