KNOWLEDGE-BASED SYSTEM FOR THE AUTOMATED SOLID-PHASE EXTRACTION OF BASIC DRUGS FROM PLASMA COUPLED WITH THEIR LIQUID-CHROMATOGRAPHIC DETERMINATION - APPLICATION TO THE BIODETERMINATION OF BETA-RECEPTOR BLOCKING-AGENTS

Techniques for the preparation of biological samples are often based n owadays on solid-phase extraction (SPE). The different SPE steps can b e performed automatically on disposable extraction cartridges (DECs) b y means of a sample processor. A knowledge-based system was developed to facilitate the development of fully automated methods for the solid -phase extraction of relatively hydrophobic basic drugs from plasma, c oupled with their determination by high-performance liquid chromatogra phy (HPLC). The DEC filled with 50 mg of cyanopropyl-bonded silica pha se is first conditioned with methanol and buffer solution (pH 7.4). Af ter sample application, the DEC sorbent is washed with the same buffer . The analytes are then desorbed with an appropriate eluent and the el uate is finally diluted with the same buffer as used in the HPLC mobil e phase before injection. Under these conditions, only three variables are still to be optimized: the composition and volume of the elution solvent and the volume of buffer to be added to the eluate. On the bas is of this general strategy, a decision tree providing information abo ut suggested starting conditions and guidelines for the optimization o f the three variables was developed and implemented by use of a hyperm edia software. This didactic expert system was evaluated using several P-receptor blocking agents as model compounds and the operating condi tions obtained for the automated SPE of these compounds are presented. A method for the determination of propranolol in plasma using the SPE conditions deduced from the knowledge-based system was validated. The absolute recovery of propranolol is ca. 93% and the limit of detectio n is 1.3 ng ml(-1). The mean within-day and between-day reproducibilit ies are 2.3 and 3.6%, respectively.