Subjects included 389 community-dwelling women aged 69 y and older from the Muramatsu cohort initiated in 2003; follow-up ended in 2009. We obtained data on physical characteristics, osteoporosis treatment (with bisphosphonates or selective estrogen receptor modulators), physical activity, calcium intake, serum 25-hydroxyvitamin D, undercarboxylated osteocalcin, serum albumin, and bone turnover markers as predictors. The outcome was a 6-y change in forearm BMD (螖BMD). Osteoporosis treatment was coded as 0 for none, 1 for sometimes, and 2 for always during the follow-up period. Stepwise multiple linear regression analysis was used to identify independent predictors of 螖BMD.
Mean age of the subjects was 73.3 y. Mean values of 螖BMD and 螖weight were 鈭?.019 g/cm2 (鈭?.8%) and 鈭?.2 kg, respectively. Stepwise multiple linear regression analysis revealed baseline BMD (尾 = 鈭?.137, P < 0.0001), osteoporosis treatment (尾 = 0.0068, P = 0.0105), serum albumin levels (尾 = 0.0122, P = 0.0319), and 螖weight (尾 = 0.0015, P = 0.0009) as significant independent predictors of 螖BMD. However, none of the other nutritional or biochemical indices were found to be significant predictors of 螖BMD.
Our findings indicate that adequate general nutrition and appropriate osteoporosis medication, rather than specific nutritional regimens, may be effective in preventing bone loss in elderly women.