Endostatin, an endogenous anti-angiogenic protein, has been linked to reduced atherosclerosis in animal models. We conducted a nested case-control study to ascertain whether decreased circulating endostatin might be associated with increased odds of acute myocardial infarction (AMI) and whether this association varied by sex or race. Cases were 211 subjects who subsequently developed AMI, and controls were 173 subjects free of cardiovascular disease matched on age, sex, race and follow-up time. In conditional logistic regression adjusting for traditional risk factors, the odds ratio of AMI per 1 SD increment in endostatin was 0.85 (95% confidence interval, 0.73-1.00). This association varied by race (but not by sex) such that a statistically significant inverse relation was found among Asians and white individuals and a significant positive relation among black individuals. Further research is needed to replicate these findings and to elucidate potential mechanisms for these race/ethnic differences.