A semi-automated procedure for the determination of caspofungin in human plasma using solid-phase extraction and HPLC with fluorescence detection using secondary ionic interactions to obtain a highly purified extract

A semi-automated assay for the determination of caspofungin in human plasma is presented. High assay throughput was achieved through the use of a robotic sample processor and 96 well format solid phase extraction (SPE). Drug and internal standard (an isostere) were extracted from plasma using a silica based, C-8 stationary phase. The extraction yielded a highly purified extract, as retention was mediated by a combination of reverse phase and secondary ionic interactions. Conditioned SPE plates (50 mg sorbent/well) were loaded with buffered (pH 4.9) plasma containing drug and internal standard. The wells were washed with water and neat methanol prior to elution with a reagent optimized for both recovery and selectivity (0.25 M ammonium hydroxide/0.05% trifluoroacetic acid in methanol). Excess residual water in the SPE wells during the methanol wash was found to cause variable drug recovery and was eliminated by centrifugation of the SPE plate. After evaporation of the SPE eluent, plasma extracts were dissolved in mobile phase and analyzed using a Keystone Betasil C-18 analytical column (4.6x50 mm, 3 mu m) with fluorescence detection (excitation 220 nm, emission 304 nm). The mobile phase was composed of a 3862 (vv) mixture of acetonitrile and 0.1% trifluoroacetic acid (adjusted to pH 3 with triethylamine) and was pumped at a flow rate of 1.5 mL/minute. Seven-point calibration curves over the concentration range 125-10,000 ng/mL yielded a linear response (drug concentration vs drug/internal standard peak height ratio) using a weighed (1/ x ) linear regression model. Based on the replicate analyses (n=5) of spiked plasma standards, intra-day assay precision was better than 5.7% coefficient of variation (CV) and intra-day accuracy was within 1.7% of nominal at all points of the standard curve. Inter-day precision, as assessed by daily analysis of high, mid, and low concentration quality control samples (n=6), was better than 5.3% CV. Inter-day accuracy was within 10.7% of nominal value.