An interlaboratory study funded by the European Commission, Standards, Measurement and Testing Programme (4th Framework Programme) was performed to evaluate the effectiveness of an immunoaffinity column cleanup liquid chromatographic (LC) method for the determination of ochratoxin A in baby food at a possible future European regulatory limit (0.1 ng/g). The test portion is extracted in a blender with tert-butyl methyl ether (chosen to avoid use of chloroform but shown to give equivalent extraction efficiency) after addition of 0.5 mol/L phosphoric acid-2 mol/L sodium chloride solution. The extract is centrifuged and redissolved in a mixture of phosphate buffered saline solution and methanol. After removal of lypophilic substances with hexane, the extract is applied to an immunoaffinity column containing antibodies specific to ochratoxin A. The column is washed with water to remove the interfering compounds and the purified ochratoxin A is eluted with methanol. The separation and determination of ochratoxin A is performed by reversed-phase LC and detected by fluorescence after postcolumn derivatization (PCD) with ammonia. Test materials (baby food infant formulae), both spiked and naturally contaminated with ochratoxin A, were sent to 13 laboratories in 8 different European countries. Test portions were spiked at a level of 0.085 ng/g ochratoxin A. The average recovery for the spiked blank baby food was 108%. Based on results for spiked samples (blind pairs at 0.085 ng/g) as well as naturally contaminated samples (blind pairs at levels between 0.05 and 0.22 ng/g) the relative standard deviation for repeatability (RSDr) ranged from 18-36%. The relative standard deviation for reproducibility (RSDR) ranged from 29-63% and HORRAT values of between 0.4 and 0.9 were obtained.