Dietary determinants of post-menopausal bone loss at the lumbar spine: a possible beneficial effect of iron.

INTRODUCTION: Previous studies suggesting different effects of diet on post-menopausal bone loss may have given conflicting results because they sometimes failed to exclude confounding conditions or used imprecise methodology. DESIGN: To identify dietary determinants of bone loss from the lumbar spine after menopause in women not taking hormone replacement who developed no evidence of spondylotic or sclerotic degenerative disease, forty-three women were followed with repeated (mean = 12) measurements of bone mineral density (BMD) at L2-4 for 11-14 years. Eleven developed evidence suggestive of degenerative disease and were excluded. Diet was assessed at the beginning of the study and 2.5 years later using 3-day and 7-day periods of weighed intakes. Nutrients estimated were: carbohydrate, fat, protein, fibre, calcium, magnesium, iron, phosphorus, copper, zinc and six vitamins. We tested the ability of diet to predict post-menopausal bone loss using stepwise regression. RESULTS: Each woman's BMD change was described by a single coefficient after log transformation of the BMD data. The best model for BMD loss including dietary factors alone had two significant determinants: daily energy or protein (p=0.0003) intake was adverse, while dietary iron (p=0.002) was predictive of bone maintenance, an effect that persisted if iron was expressed as a ratio to energy intake. Adding body mass index to the model increased the goodness of fit (R (2)adj rose from 0.33 to 0.42) without affecting the statistical significance of the dietary determinants. CONCLUSIONS: Diet may influence bone loss after menopause, with dietary iron (or an associated factor) possibly having a protective effect on bone at the spine