Development of a surrogate matrix for cerebral spinal fluid for liquid chromatography/mass spectrometry based analytical methods.

Rationale: In recent years, several liquid chromatography/tandem mass spectrometry (LC/MS/MS) methods have been reported for the quantitative determination of drugs and metabolites in cerebral spinal fluid (CSF). Artificial CSF (aCSF) is often used as a surrogate for preparing calibration curves and quality control samples in these methods. However, aCSF does not accurately represent the composition of real CSF because it is missing all of the proteins and lipids, which may alter the electrospray ionization (ESI) response when performing LC/MS/MS analyses. In the current study we compared the mass spectral response of several compounds with a range of physiochemical properties in aCSF (essentially a mixture of salts and buffers), diluted plasma (ranging from 1:5 to 1:200) and real CSF to find the best surrogate for CSF in LC/MS/MS methods.

Methods: A number of analytes from polar to non-polar, high protein binding to low protein binding, employing different sample preparation methods, were prepared in diluted plasma, actual CSF or aCSF and tested using LC/MS/MS. The analytes included cotinine and its metabolites, quetiapine, norquetiapine, chlorpromazine, efavirenz and lamivudine. The similarity of MS responses from these compounds in aCSF and diluted plasma to CSF was assessed by comparing the slopes of the calibration curves generated from using linear regression modeling.

Results: For all compounds, the lowest percent difference in response ratio (0 to 17%) was observed from 1:200 diluted plasma. Our results indicated that, irrespective of the inherent physiochemical properties of the analytes or the method of sample preparation, 1:200 diluted plasma performed as the best surrogate for CSF in LC/MS/MS methods.

Conclusions: The percent difference in response ratio has been established to demonstrate how different compounds behave between CSF, aCSF and dilute plasma. Although among the compounds tested some of them showed a very similar MS response in actual and aCSF, there were analytes that demonstrated significant differences in ESI-MS signal when sprayed from these two matrices. However, even in such cases, 1:200 diluted plasma generated results with no significant difference from CSF. Therefore, we recommend that in order to develop robust and dependable bioanalytical LC/MS methods from CSF samples, it is more appropriate to prepare calibration curves and quality control samples in diluted plasma instead of aCSF. Copyright © 2016 John Wiley & Sons, Ltd.