The aim of this work was to investigate the association between homocysteine levels and coffee consumption in a sample of cardiovascular disease-free men and women. From May 2001 to December 2002, we randomly enrolled 1514 men and 1528 women, stratified by age and gender, from the greater area of Athens. Blood samples were collected in the fasting state. Among other investigated factors, dietary habits (including coffee consumption in ml per day, adjusted for 28% caffeine containment) were evaluated using a validated food frequency questionnaire. Men consumed higher quantities of coffee compared with women (250 +/- 55 vs 150 +/- 60 ml/day, P = 0.001), while homocysteine values were also higher in men than in women (14.5 +/- 6 vs 10.8 +/- 3.5 micromol/l, P = 0.001). A dose-response relationship of homocysteine levels with coffee consumption was observed (r = 0.10, P = 0.034). In particular, we found that homocysteine levels were 11.2 +/- 5 micromol/l for no consumption, 11.7 +/- 7 micromol/l for <100 ml/day, 12.5 +/- 7 micromol/l for 200-400 ml/day, and 12.7 +/- 4 micromol/l for >500 ml/day consumption (P = 0.018). The observed trend remained significant even after controlling for the interactions between coffee consumption with gender, smoking habits, physical activity status, and eating habits. However, the sole effect of the consumption of filtered coffee on homocysteine levels was significant only in those who consumed more than 500 ml/day (P = 0.043). Although our findings cannot be evidence for causality, they can be the basis for hypotheses about the relation between homocysteine and coffee that can partially explain the mechanisms by which elevated homocysteine levels may influence coronary risk.