Abstract: Methods based on molecular recognition mechanisms for the clean-up of veterinary drugs and their residues, such as immuno-, receptor- and acceptor-affinity and molecularly imprinted polymers (MIPs), have been described as selective tools to improve the selectivity and the reliability of analytical results. In this work, we tested the extraction recovery performances of a MISPE column, designed for multi- residual clean-up of beta-agonists. For this purpose, 18 different samples of calf urine were spiked at 0.25, 0.50 and 1.00 ppb with pooled standard solutions of clenbuterol (Clen), tulobuterol (Tolu), isoxsuprine (Isox), brombuterol (Brom), mapenterol (Mape) and ractopamine (Racto) and analysed on two independent analytical sessions, on a LC-MS/MS ion trap detector. Averaged recoveries, constant for each molecule considered, were 64.6% for Racto, 63.0% for Salm, 59.9% for Form, 54.7% for Brom, 52.0% for Clen, 41.8% for Mape, 38.6% for Tolu and 34.5% for Isox, respectively. Reproducibility studies gave a CV < 11% at the 0.25 ppb level. The decision limit for the identification of the target drugs ranged from 0.01 ppb for mapenterol to 0.19 ppb for salmeterol, when considering one precursor, and two product ions as identification points. Such findings indicate that the choice of the appropriate molecule as template in the MIP preparation is the critical factor to guarantee a reliable analytical multi-residue approach for beta-agonists, despite the structural differences among molecules exploiting almost the same pharmacological effect. (C) 2004 Elsevier B.V. All rights reserved