A variety of evidence suggests brain-derived neurotrophic factor (BDNF) as a candidate gene for schizophrenia, and several genetic studies have shown a significant association between the disease and certain SNPs within BDNF (specifically, Val66Met and C270T). According to a recent study, the functional microsatellite marker BDNF-LCPR (BDNF-linked complex polymorphic region), which affects the expression level of BDNF, is associated with bipolar disorder. The goals of our current study were to 1) evaluate the quality of HapMap-based linkage disequilibrium (LD) tagging of BDNF-LCPR, 2) examine whether these tagging SNPs are associated with schizophrenia in a Japanese population, and 3) conduct a meta-analysis of the two most extensively studied polymorphisms: Val66Met and C270T. We genotyped eight tagging SNPs, including Val66Met and C270T. Our LD evaluation showed that BDNF-LCPR could be represented by these tagging SNPs in controls (with 73.5% allelic coverage). However, the functional A1 allele was not captured due to its low minor allele frequency (2.2%). In a case-control study (1117 schizophrenics and 1102 controls), no association was found in single-marker or multimarker analysis. Moreover, in a meta-analysis, the Val66Met polymorphism was not associated with schizophrenia, whereas C270T showed a trend for association in a fixed model (p=0.036), but not in a random model (p=0.053). From these findings, we conclude that if BDNF is indeed associated with schizophrenia, the A1 allele in BDNF-LCPR would be the most promising candidate. Further LD evaluation, as well as an association study in which BDNF-LCPR is genotyped directly, would be required for a more conclusive result.