Ultra-processed food consumption, genetic susceptibility, and the risk of hip/knee osteoarthritis

Wei Y, Zhang T, Liu Y, Liu H, Zhou Y, Su J, Chen L, Bai L, Xia Y · Clinical Nutrition

Prospective and large enough to see incidence rather than prevalence. The top UPF quartile had a 10% higher risk of knee OA (HR 1.10, 95% CI 1.03-1.18); hip OA showed no association. Substituting 20% of ultra-processed diet weight with unprocessed or minimally processed food was associated with 6% lower knee OA risk (HR 0.94, 95% CI 0.89-0.98). Genetic risk interacted with the association (p = 0.01) — the excess risk from ultra-processed food fell on those with *lower* genetic susceptibility.

Population

163,987 UK Biobank participants followed for 1,461,447 person-years; 11,540 developed osteoarthritis

Intervention

Highest quartile of ultra-processed food consumption from 24-hour dietary recall

Comparator

Lowest quartile

Limitations

Exposure comes from 24-hour dietary recalls, which capture a day rather than a habit. A 10% relative risk in an observational cohort is inside the range that residual confounding produces — ultra-processed intake tracks with income, education, smoking and activity. The substitution estimate is a statistical model, not an intervention.