Does smoking protect against osteoarthritis? Meta-analysis of observational studies

Hui M, Doherty M, Zhang W · Annals of the Rheumatic Diseases

Pooled analysis found a modest inverse (apparently 'protective') association between smoking and OA (OR 0.87, 95% CI 0.80–0.94) — but the effect was confined to hospital-based case-control studies and the authors conclude it is likely false, driven by selection bias.

Population

537,730 participants across 48 observational studies of smoking and OA

Intervention

Smoking (observed)

Comparator

Non-smokers

Limitations

The headline inverse association disappears in community cohorts; classic index-event/selection bias example; OA diagnosis definitions heterogeneous.