University Centre for Pharmacy, Groningen Utrecht Institute of Drug Exploration, Ant. Deusinglaan 1, 9713 AV Groningen, The Netherlands.

To enhance the reproducibility of the data and to create a standardisable identification parameter, migration times were converted to effective mobilities, which were then corrected with 5 external reference barbiturates. The data were used to evaluate the Identification Power (IP) of the methods by calculating the Mean List Length, Discriminating Power and Discriminating Number.

The CE methods were very promising for the application in STA, providing fast analyses (within 5 min for CZE and 10 min for MEKC). The two methods showed low correlation and hence complement each other. Also, especially the CZE method showed low correlation with existing RP-HPLC and GC for STA. Reproducibilities were significantly improved using corrected effective mobilities instead of migration times. Both methods had good IP's but combinations of methods such as CZE plus MEKC, HPLC or GC and MEKC plus GC provided considerably better IP's than the individual methods.

Finally, CZE and MEKC appeared to be very favourable in STA when analysing a group of structurally similar compounds. In further studies, the two methods (CZE and MEKC) will be evaluated for compounds from different chemical classes.