| Summary: | Quantitative MS of human plasma lipids is a promising technology for translation into clinical applications. Current MS-based lipidomic methods rely on either direct infusion (DI) or chromatographic lipid separation methods (including reversed phase and hydrophilic interaction LC). However, the use of lipid markers in laboratory medicine is limited by the lack of reference values, largely because of considerable differences in the concentrations measured by different laboratories worldwide. These inconsistencies can be explained by the use of different sample preparation protocols, method-specific calibration procedures, and other experimental and data-reporting parameters, even when using identical starting materials. Here, we systematically investigated the roles of some of these variables in multiple approaches to lipid analysis of plasma samples from healthy adults by considering: 1) different sample introduction methods (separation vs. DI methods); 2) different MS instruments; and 3) between-laboratory differences in comparable analytical platforms. Each of these experimental variables resulted in different quantitative results, even with the inclusion of isotope-labeled internal standards for individual lipid classes. We demonstrated that appropriate normalization to commonly available reference samples (i.e., “shared references”) can largely correct for these systematic method-specific quantitative biases. Thus, to harmonize data in the field of lipidomics, in-house long-term references should be complemented by a commonly available shared reference sample, such as NIST SRM 1950, in the case of human plasma.
|
|---|
