Folate deficiency is often discussed as a potential risk factor for CVD and some cancers. Reliable assessment of folate status in large-scale epidemiological studies is therefore of major importance. The present study assessed the value of plasma folate (PF) compared with erythrocyte folate (EF) as a marker of folate status in 363 participants in the European Investigation into Cancer and Nutrition (EPIC)-Potsdam cohort. EF and PF, total homocysteine (tHcy), pyridoxine, cobalamin, creatinine, total protein and packed cell volume were determined; glutamate carboxypeptidase (GCP) C1561T, reduced folate carrier (RFC) G80A and methylenetetrahydrofolate (MTHFR) C677T polymorphisms were analysed. Anthropometric measurements were taken and dietary intake was assessed with the EPIC-Potsdam food-frequency questionnaire. Comparison of EF and PF with factors that may modulate their concentrations was performed. Cross-classification of blood folates in quintile categories resulted in correct classification into the same or adjacent category of 75·5 % of all subjects. Age, BMI, pyridoxine and cobalamin, fruit and vegetable intake, and vitamin supplementation 24 h before blood draw were positively associated with EF and with PF. For tHcy an inverse association was found. Participants with the MTHFR 677TT genotype showed significantly elevated EF concentrations compared with those with 677CT genotype; EF and PF were more strongly correlated (r 0·78, P>0·0001) for participants with MTHFR 677TT genotype than for those with the 677CC or 677CT genotype. In summary, our present results indicate that plasma folate seems to be a suitable marker for assessment of folate status for use in large-scale epidemiological studies.