While it is not practical to prepare all calibration standardand quality control samples from individual sources, someassessment of patient variability must be undertaken. Thefollowing approach has been used in the author’s laboratory.This protocol assesses the performance of a method usingblood from subjects who are the subject of the clinicalinvestigation being studied (i.e., transplant recipients, HIVpatients, uremic patients, etc). Three quality controlconcentrations, over the analytical range, in replicates offive are prepared using different individuals for each aliquot(i.e., each individual is used only once). The analyte ofinterest and internal standard are added pre- and postextractionand prepared in pure solution according to theprotocol of Buhrman et al. [27], previously described above.Using this protocol of 45 injections, matrix effects, absoluterecovery, process efficiency, and most importantly intersubjectvariability can be calculated. As an example, datafrom an HPLC–ESI–MS/MS method for the measurementof cyclosporin in whole blood that used solid phase samplepreparation are shown in Table 2 [40]. Three concentrationsof cyclosporin at 30, 400, and 1500 Ag/L were assessedusing blood obtained from transplant recipients not receivingthis drug. Matrix effects were minimal with less than 7%ion suppression. The intersubject variability, expressed ascoefficient of variation, was b3% for the three concentrations.These data compared favorably with the variabilityobtained from pooled blood of b5% coefficient of variation(data not shown). The results of this experiment suggest thatmatrix effects will have minimal influence on the results ofthis method. If the intersubject variability was greater than15%, changes to the extraction procedure and chromatographywould have to be undertaken. It is hoped that this typeof procedure will be common practice in the validation ofHPLC–ESI–MS/MS methods.