Determinants of total body, total femur, and lumbar spine BMD (g/cm2) from backward stepwise multiple regression models in the athletes (n = 186)
| Site | MD | EDs | Body fat (%) | Sport groups | Weight (kg) | R2* |
|---|---|---|---|---|---|---|
| Adjusted R square (R2) for the final model for each measurement site is presented. Significance levels are shown, followed on the line beneath by unstandardised β coefficients (β) and standard error (SE). | ||||||
| *Age of sport specialisation (when they started competing in their sport), smoking, height, oral contraceptive use, use of medications that can affect bone metabolism, and menarche were not significantly associated with BMD and are therefore not included in the table or in the R2 value. | ||||||
| EDs, eating disorders; MD, menstrual dysfunction; NS, not significant. Sport groups are classified into low impact, medium impact and high impact sports. | ||||||
| Total body | NS | p<0.05 | p<0.01 | p<0.001 | p<0.001 | 0.48 |
| β (SE) | – | −0.023 (0.010) | −0.003 (0.001) | 0.043 (0.007) | 0.007 (0.001) | |
| Total femur | NS | NS | NS | p<0.001 | p<0.01 | 0.27 |
| β (SE) | – | – | – | 0.086 (0.012) | 0.003 (0.001) | |
| Lumbar spine | p<0.05 | NS | p<0.01 | p<0.001 | p<0.001 | 0.37 |
| β (SE) | −0.043 (0.021) | – | −0.004 (0.002) | 0.077 (0.013) | 0.009 (0.001) | |