BACKGROUND Patient specimens can be subject to subtle interference from cross contamination by liquid-based, potassium-containing EDTA anticoagulant, leading to misinterpretation of results. A rapid method for EDTA analysis to detect such contamination is described. METHODS An in-house EDTA assay on the Roche MODULAR analyser was assessed for accuracy and precision by comparison with an adjusted calcium difference measurement (atomic absorption and o-cresolphthalein complexone colorimetry). RESULTS EDTA method versus adjusted calcium difference showed: slope = 1.038 (95% confidence interval [CI] 0.949-1.131); intercept = 0.073 (95% CI 0.018-0.132) mmol/L; r = 0.914; n = 94. However, inter-assay precision of the calcium difference method was estimated to be poorer (coefficient of variation 24.8% versus 3.4% for the automated colorimetric method at an EDTA concentration of 0.25 mmol/L). Unequivocal contamination was observed at an EDTA concentration of > or =0.2 mmol/L. The automated method showed positive interference from haemolysis and negative interference from oxalate. The method was unaffected by lipaemia (triglycerides <20 mmol/L), icterus (bilirubin <500 micromol/L), glucose (<100 mmol/L), iron (<100 micromol/L), and citrate, phosphate or fluoride (all <2.5 mmol/L). CONCLUSIONS The automated colorimetric assay described is an accurate, precise and rapid (3 min) means of detecting EDTA contamination of unhaemolysed biochemistry specimens.