Longitudinal Associations of High-Volume and Vigorous-Intensity Exercise With Hip Fracture Risk in Men

Abstract

Maintenance of vigorous exercise habits from young to old age is considered protective against hip fractures, but data on fracture risk in lifelong vigorous exercisers are lacking. This longitudinal cohort study examined the hazard of hip fractures in 1844 male former athletes and 1216 population controls and in relation to exercise volume and intensity in later years. Incident hip fractures after age 50 years were identified from hospital discharge register from 1972 to 2015. Exercise and covariate information was obtained from questionnaires administered in 1985, 1995, 2001, and 2008. Analyses were conducted using extended proportional hazards regression model for time-dependent exposures and effects. During the mean ± SD follow-up of 21.6 ± 10.3 years, 62 (3.4%) athletes and 38 (3.1%) controls sustained a hip fracture. Adjusted hazard ratio (HR) indicated no statistically significant difference between athletes and controls (0.84; 95% confidence interval [CI], 0.55-1.29). In subgroup analyses, adjusted HRs for athletes with recent high (≥15 metabolic equivalent hours [MET-h]/week) and low (<15 MET-h/week) exercise volume were 0.83 (95% CI, 0.46-1.48) and 1.04 (95% CI, 0.57-1.87), respectively, compared with controls. The adjusted HR was not statistically significant between athletes with low-intensity exercise (<6 METs) and controls (1.08; 95% CI, 0.62-1.85). Athletes engaging in vigorous-intensity exercise (≥6 METs at least 75 minutes/week) had initially 77% lower hazard rate (adjusted HR 0.23; 95% CI, 0.06-0.86) than controls. However, the HR was time-dependent (adjusted HR 1.04; 95% CI, 1.01-1.07); by age 75 years the HRs for the athletes with vigorous-intensity exercise reached the level of the controls, but after 85 years the HRs for these athletes increased approximately 1.3-fold annually relative to the controls. In conclusion, these data suggest that continuation of vigorous-intensity exercise is associated with lower HR of hip fracture up to old age. © 2022 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).

Keywords: AGING; EXERCISE; FRACTURE PREVENTION; LONGITUDINAL STUDIES; OSTEOPOROSIS.

Fig. 1

Nelson‐Aalen cumulative fracture hazard curves of hip fractures for controls (Ctrl) and all former athletes (Athlete). HR adjusted for occupation and its 95% CI are displayed at the top. Hip fractures were followed from January 1, 1972 until time of death, diagnosis of hip fracture, or end of follow‐up on December 31, 2015. HR = hazard ratio; CI = confidence interval.

Fig. 2

Nelson‐Aalen cumulative fracture hazard curves of hip fractures for controls (Ctrl), and athletes divided into two subgroups according to total exercise volume (A) and average exercise intensity (B). HRs and their 95% CIs are shown in the inset tables. The models were controlled for body height, body weight, living situation, occupational class, alcohol use, and smoking status. Nelson‐Aalen curves of hip fracture risks for former athletes were adjusted by possible changes in exercise level during follow‐up. Time‐dependent effect observed for exercise intensity (B) is partitioned into two components: initial HR is the hazard ratio at the start of the follow‐up period and time HR refers to risk modification over time. Hip fractures were followed starting from the first questionnaire participation (athletes) or from 1985 (controls) until time of death, diagnosis of hip fracture, or end of follow‐up on December 31, 2015. Exercise and other characteristics were obtained from questionnaire studies in 1985, 1995, 2001, and 2008. See Table 1 footnote for description of calculation of exercise volume and intensity with MET values. HR = hazard ratio; CI = confidence interval; MET = metabolic equivalent.