Associations Between Physical Activity and the Risk of Hip Fracture Depending on Glycemic Status: A Nationwide Cohort Study

Abstract Context Although physical activity (PA) is recognized to reduce fracture risk, whether its benefits differ according to glycemic status remains unknown. Objective We investigated the effect of PA on incident hip fracture (HF) according to glycemic status. Methods We studied 3 723 097 patients older than 50 without type 1 diabetes mellitus (DM) or past fractures. HF risks were calculated using Cox proportional hazard regression. Participants were categorized by glycemic status into 5 groups: normal glucose tolerance, impaired fasting glucose, new-onset type 2 DM, type 2 DM less than 5 years, and type 2 DM of 5 years or greater. PA was evaluated using the Korean adaptation of the International Physical Activity Questionnaire Short Form. Results The highest HF risk were associated with the lowest PA level (<500 metabolic equivalent task [MET]-min/wk). While similar risks emerged across MET 500 to 1000, 1000 to 1500, and greater than 1500 categories, the relationship showed variations in different glycemic status groups. Exceptions were particularly noted in women with normoglycemia. However, a consistent inverse pattern, with few exceptions, was observed both in men and women with type 2 DM of 5 years or greater. Furthermore, the benefit of PA in the prevention of HFs was most evident in participants with type 2 DM of 5 years or greater. Compared to the reference group (lowest physical activity level <500 MET-min/wk within type 2 DM ≥5 years), the adjusted hazard ratios were 0.74 (0.62-0.88) in men and 0.74 (0.62-0.89) in women, suggesting a significant reduction in risk. Conclusion Higher PA levels are associated with a lower risk of HF. This protective effect of PA on fracture risk is greatest in patients with DM, particularly in those with DM of 5 years or greater.