Homocysteine has been identified to be associated with Alzheimer disease (AD) and methionine synthase (MS) is one of the enzymes involved in homocysteine metabolism. Confused data were reported on the association between the MS 2756 A>G polymorphism and AD. To determine if this polymorphism could affect the occurrence of AD, we investigated the association between the MS 2756 A>G polymorphism and AD risk in 353 sporadic AD patients and 346 controls in a Chinese Han population. No significant differences of allele and genotype distributions between the AD cases and the controls were observed in the total samples, neither when the samples were stratified by age/age at onset and gender. When the samples were stratified by APOE epsilon4 status, a trend of A allele and AA genotype over-representation in the AD patients in comparison with the controls was observed, but it was not statistically significant (for the alleles, A versus G OR=1.549, 95% CI 0.920-2.609, p=0.098; for the genotypes, AA versus AG+GG OR=1.485, 95% CI 0.861-2.560, p=0.153). Similar trend was observed in the APOE epsilon4 non-carrier samples of the >or=65 year subgroups and it was not statistically significant too (for the alleles, A versus G OR=1.682, 95% CI 0.901-3.140, p=0.099, for the genotypes, AA versus AG+GG OR=1.690, 95% CI 0.884-3.232, p=0.110). Our data did not reveal significant association between the MS 2756 A>G polymorphism and AD development. However, a weak effect of the A allele on developing AD could not be completely excluded.