BACKGROUND: Phthalates may pose a risk for perinatal developmental effects. An important question relates to the choice of suitable biological matrices for assessing exposure during this period. OBJECTIVES: This study was designed to measure the concentrations of phthalate diesters or their metabolites in breast milk, blood or serum, and urine and to evaluate their suitability for assessing perinatal exposure to phthalates. METHODS: In 2001, 2-3 weeks after delivery, 42 Swedish primipara provided breast milk, blood, and urine samples at home. Special care was taken to minimize contamination with phthalates (e.g., use of a special breast milk pump, heat treatment of glassware and needles, addition of phosphoric acid). RESULTS: Phthalate diesters and metabolites in milk and blood or serum, if detected, were present at concentrations close to the limit of detection. By contrast, most phthalate metabolites were detectable in urine at concentrations comparable to those from the general population in the United States and in Germany. No correlations existed between urine concentrations and those found in milk or blood/serum for single phthalate metabolites. Our data are at odds with a previous study documenting frequent detection and comparatively high concentrations of phthalate metabolites in Finnish and Danish mothers' milk. CONCLUSIONS: Concentrations of phthalate metabolites in urine are more informative than those in milk or serum. Furthermore, collection of milk or blood may be associated with discomfort and potential technical problems such as contamination (unless oxidative metabolites are measured). Although urine is a suitable matrix for health-related phthalate monitoring, urinary concentrations in nursing mothers cannot be used to estimate exposure to phthalates through milk ingestion by breast-fed infants.