The association between presenilin 1 intronic polymorphism (rs165932) and late onset Alzheimer's disease (LOAD) has been a matter of controversy. Within China, varied results have been reported. Therefore, we collected a large sample from the North Chinese population to test the association of the PS1 polymorphism with LOAD. AD patients (467 total, mean age=75.3+/-7.3, age at onset=70.2+/-5.1) and age-matched normal elderly controls (480 total) were recruited. Genotypes of PS1 and apolipoprotein E (APOE) were determined by PCR and RFLP. The results showed that there were significant differences in the distributions of both alleles (chi(2)=45.305, P<10(-5)) and genotypes (chi(2)=53.055, P<10(-5)) of PS1 gene between the AD and control groups. The APOE epsilon4 allele was more prevalent in patients than in controls (chi(2)=46.389, P<10(-5)). It was significantly different when PS1 alleles and genotypes were compared between AD and controls with APOE epsilon4 negative. However, no significance was found when PS1 alleles or genotypes were compared between AD and controls with APOE epsilon4 positive. Furthermore, with PS1 2/2 genotype as a reference, the odds ratios (ORs) of LOAD with PS1 1/2, 1/1+1/2 and 1/1 genotypes gradually increased allele 1 copy number, suggesting that allele 1 is a crucial risk for LOAD. In summary, we found an association between presenilin 1 intronic polymorphism and LOAD, but no influence of APOE epsilon4 on the distribution of the PS1 intronic polymorphism. In addition, the larger sample size raises the possibility that ethnic and regional differences in China may explain the differences in reported results.