Background: Physical activity during growth increases peak bone mass (BMD). BMD is also closed associated with fractures as it has been suggested that 1 standard deviation (SD) increased BMD reduces the fracture risk by 50%. But to recommend physical activity during youth as a strategy to reduce prevalence of osteoporosis, and fragility fractures, exercise-induced benefits must be retained at advanced ages. Purpose: This study aimed primarily to prospectively evaluate BMD changes in male athletes, from activity into long-term retirement. Secondary aims were to evaluate other bone traits and fracture incidence during the follow-up period. Methods: By single photon absorptiometry (SPA), this study evaluated whether exercise-associated high BMD in 46 active male athletes with a mean age of 22 years (range 15-40), that we have reported in the past, was retained a mean 39 years (range 38-40) later when the former athletes had been retired from sports for a mean 29 years (range 10-58). 24 non-athletic males of similar ages served as controls. BMD and bone structure were at follow-up also measured by dual energy X-ray absorptiometry (DXA), peripheral computed tomography (pQCT) and quantitative ultrasound (QUS). Both cohorts were normally distributed (Shapiro Wilk’s test). Data are presented as means with 95% confidence intervals (95% CI). Group differences were evaluated by Student’s t-test between means and chi-square test. Z scores, i.e. the number of SD above or below the age-predicted mean, were derived by linear regression using data from the controls. Results: There was no group difference in anthropometry either at baseline or at follow-up. The active athletes (baseline) had BMD Z score of 1.0 (0.7, 1.4) in the femoral condyles (SPA). The former athletes (follow-up) had BMD Z score of 0.7 (0.2, 1.1) in the distal radius (SPA) and 1.2 (0.8, 1.7) in the legs (DXA) Figure 1). There were no changes in BMD Z scores during the follow-up period, neither when estimated by the same SPA apparatus but in different skeletal regions (delta Z score -0.3 (-0.8, 0.2)) nor when estimated in the same extremity but with SPA at baseline and DXA at follow-up (delta Z score 0.0 (-0.4, 0.4)) (Figure 1). Furthermore, tibial cortical area was larger in the former athletes (Z score 0.8 (0.5, 1.2)) as was tibial strength index (Z score 0.7 (0.4, 1.0)), and 10.9% of the former athletes had sustained fractures in comparison with 20.8% of the controls (no statistics done due to low power). Conclusions: Exercise-associated BMD benefits seem to be retained even three decades after cessation from active training. In a wider perspective, this indicates that physical activity could be recommended in younger years as a feasible strategy to reduce osteoporosis in older ages.