The epidemiology of selenium (Se) is mainly based on the determination of total serum selenium levels (TSe) which by many aspects is an inadequate marker of Se status. In this study we applied a recently developed LC-ICP-MS method, for the determination of the selenium content of the three main serum selenium-containing proteins, in a subcohort of the ATTICA study. This enables us to investigate whether the selenium distribution to selenoproteins may correlate with demographic (age, gender) and lifestyle variables (smoking, physical activity) that are crucial for the development of chronic diseases. A sub-sample from the ATTICA Study, consisted of 236 males (40 ± 11 years) and 163 females (38 ± 12 years), was selected. The selenium content of glutathione peroxidase (GPx-3), selenoprotein P (SelP) and selenoalbumin (SeAlb) was determined in serum by LC-ICP/MS method. We found that 26% of TSe is found in GPx-3, 61% in SelP and 13% in SeAlb. We have assessed the different ratios of selenoproteins' selenium content (Se-GPX-3/Se-SelP, Se-GPX-3/Se-SeAlb, Se-SelP/Se-SeAlb), showing that people with similar TSe may have different distribution of this selenium to selenoproteins. Total selenium levels and gender are the variables that mostly affect selenium distribution to selenoproteins while age, smoking, physical activity and BMI do not significantly influence selenium distribution. In conclusion, the simultaneous determination of the selenium content of serum selenium-containing selenoproteins is necessary for a thorough estimation of selenium status. The ratio of the Se content between selenoproteins may be proven a novel, valid marker of selenium status.