Elevated homocysteine increases the risk of vascular diseases but little information is available about this issue in the elderly. The aim of this cross-sectional study was to evaluate the relationships between homocysteinemia and gender, anthropometric, and life-style characteristics in a community-dwelling elderly population (65 men and 120 women; 67-78 years). Basal plasma homocysteine levels were determined by High Performance Liquid Chromatography (HPLC). Clinical records, and nutritional and anthropometric variables were collected in all subjects. Body composition was evaluated in all subjects by Dual energy X-ray Absorptiometry (DXA). Thirty-three percent of women and 66% of men had hyper-homocysteinemia. In women, a positive correlation was present between homocysteinemia, age, diastolic blood pressure and plasmatic creatinine, and a negative correlation between homocysteine, fiber intake and folates. In males, there was a positive correlation between plasma homocysteine, age, and body mass index. Multiple regression analysis showed that fat-free mass, cigarette smoking, fiber intake, vitamin B6 and total kcal intake accounted for 18% of homocysteine variance in males (R2 = 0.18, p<0.05). Significantly higher homocysteine values were found in women with a history of cardiovascular disease than in those without (16.6 +/- 9.4 vs 13.8 +/- 4.4 micromol/L, p<0.05). Homocysteinemia was significantly higher in elderly men compared to women (16.7 +/- 4.7 vs 15.3 +/- 7.6; p<0.05). Gender differences in homocysteine disappeared after adjusting for fat-free mass. This study confirms the age-related increase in plasma homocysteine. Life-style characteristics seem to influence significantly homocysteine levels in the elderly. Our study shows that gender effects on homocysteine may be attributed to differences in body composition.