Background: Recent prospective studies reported that increased plasma homocysteine levels are an independent predictor of osteoporotic fracture in elderly persons. These studies, however, did not take into account folate and vitamin B12, which are the major nutritional determinants of homocysteinemia.
Methods: Incident osteoporotic fractures were assessed in 702 Italian participants aged 65-94 years with a mean follow-up of 4 years (1999/2000-2003/2004). A multivariable logistic regression model was used to study the relation of baseline plasma homocysteine, serum folate, and serum vitamin B12 with risk of fracture.
Results: After adjustment for sociodemographic and clinical confounders, the odds ratio (OR) for each increase of 1 standard deviation in log-transformed plasma homocysteine was 1.39 (95% confidence interval [CI], 1.01-1.91), but decreased to 1.22 (95% CI, 0.85-1.74) after further adjustment for serum folate and vitamin B12. The corresponding multivariable-adjusted OR for hyperhomocysteinemia (plasma total homocysteine [tHcy] > 15 micromoL) was 1.58 (95% CI, 0.71-3.53). Participants in the lowest serum folate quartile (< or =9.3 nmol/L) had an increased risk of fracture than did those in higher quartiles (multivariable-adjusted OR = 2.06; 95% CI. 1.02-4.18), but no dose-related protective effect for increasing serum folate levels was found (multivariable-adjusted OR = 0.84 for each increase of 1 standard deviation in log-transformed serum folate, 95% CI, 0.59-1.19). No independent association was found for serum vitamin B12.
Conclusions: Low serum folate is responsible for the association between homocysteine and risk of osteoporotic fracture in elderly persons.