Abstract: A new analytical method for the determination of the carcinogenic mycotoxin ochratoxin A (OTA) in red wines has been developed involving a two-dimensional solid-phase extraction (SPE) clean-up protocol on C18-silica and a target-selective molecularly imprinted polymer (MIP). Prior removal of the interfering acidic matrix compounds by C 18 solid-phase extraction was crucial for a successful clean-up as direct sample loading onto the MIP led to poor recoveries. The combined solid-phase extraction protocol afforded extracts suitable for sensitive ochratoxin A quantification by HPLC- fluorescence detection. Preliminary validation of the method performance with spiked (0.033-1.0 ng OTA/ml) and commercial red wines provided recoveries >90% and R.S.D-r < 10%, with limit of detection (LOD) and limit of quantification (LOQ) of 0.01 and 0.033 ng/ml. However, a similarly favorable performance characteristics was observed in control experiments in which the MIP was replaced by the corresponding non- imprinted polymer (NIP). These findings provide evidence that under the employed experimental conditions specific analyte binding to imprinted binding sites plays a minor role in selective OTA retention. In the framework of this study, other problems inherent to MIP-based solid-phase extraction have been addressed. These include the reproducible preparation of MIP materials with consistent molecular recognition characteristics, the potential for repeated use of MIP, unfavorable polymer swelling in application-relevant solvents, potential sample contamination by template bleeding, and slow analyte binding kinetics. (C) 2003 Elsevier B.V. All rights reserved