<jats:title>Abstract</jats:title><jats:p>Quantification of F<jats:sub>2</jats:sub>‐isoprostanes is considered a reliable index of the oxidative stress status <jats:italic>in vivo</jats:italic>. Several immunoassays and chromatography/mass spectrometry‐based assays are available for 15‐F<jats:sub>2t</jats:sub>‐isoprostane quantification. However, it remains unclear if results of immunoassays using different assays can be compared with those of liquid chromatography/mass spectrometry (LC/MS) assays. Previous studies comparing enzyme‐linked immunosorbent assay (ELISA) and more specific gas chromatography/mass spectrometry assays have already indicated that ELISAs may overestimate 15‐F<jats:sub>2t</jats:sub>‐isoprostane concentrations in human plasma.</jats:p><jats:p>Concentrations of 15‐F<jats:sub>2t</jats:sub>‐isoprostane in 25 human plasma and urine samples were measured by three commercially available ELISA assays (Assay Designs, Cayman Chemical and Oxford Biomedical Research) and compared with the concentrations measured with a validated, semi‐automated high‐throughput HPLC tandem mass spectrometry assay (LC/LC‐MS/MS).</jats:p><jats:p>All three ELISAs measured substantially higher 15‐F<jats:sub>2t</jats:sub>‐isoprostane concentrations (2.1–182.2‐fold higher in plasma; 0.4–61.9‐fold higher in urine) than LC/LC‐MS/MS. Utilization of solid‐phase extraction (SPE) columns, especially isoprostane affinity purification columns, brought ELISA isoprostane urine concentrations closer to the LC/LC‐MS/MS results. However, SPE did not have much of an effect on ELISA plasma concentrations which remained significantly higher than corresponding LC/LC‐MS/MS results. A poor correlation not only between LC/LC‐MS/MS and immunoassay results, but also among the immunoassays was found.</jats:p><jats:p>Especially in plasma, ELISAs grossly overestimate 15‐F<jats:sub>2t</jats:sub>‐isoprostane concentrations and are not comparable with each other or with LC/LC‐MS/MS. It is most disturbing that a sample with relatively high concentrations measured with one ELISA may show low concentrations with another ELISA, and vice versa, potentially affecting the conclusions drawn from such data. The use of specific mass spectrometry‐based assays seems advisable. Copyright © 2011 John Wiley & Sons, Ltd.</jats:p>