Although numerous studies have dealt with changes in blood brain-derived neurotrophic factor (BDNF), methodological issues about BDNF measurements have only been incompletely resolved. We validated BDNF ELISA with respect to accuracy, reproducibility and the effect of storage and repeated freezing cycles on BDNF concentrations. Additionally, the effect of demographic characteristics in healthy subjects on BDNF was verified. Whole blood and serum was collected from 206 healthy subjects and a subgroup was genotyped for BDNF Val66Met polymorphism. The effect of age, gender, BDNF genotype and thrombocyte count on whole blood BDNF was assessed. The BDNF ELISA measurement was accurate, 91.6+/-3.0%, and showed high reproducibility, whereas inter-assay and intra-subject variations were modest, 8.4+/-5.2% and 17.5+/-14.1%, respectively. Storage of whole blood samples at 4 degrees C significantly decreased BDNF concentration, while repeated freezing cycles and storage at -20 degrees C was without any effect. Storage at -20 degrees C of serum, but not whole blood, was associated with a significant decrease in BDNF concentration. Women had significantly higher whole blood BDNF concentrations than men (18.6+/-1.3 ng/ml versus 16.5+/-1.4 ng/ml), and showed a right-skewed BDNF concentration distribution. No association between whole blood BDNF concentrations and thrombocyte count, age, or BDNF genotype was found. In conclusion, the BDNF ELISA assay determines whole blood BDNF accurately and with high reproducibility. Female gender is associated with higher whole blood BDNF concentrations whereas age, thrombocyte count and BDNF Val66Met polymorphism were un-associated.