br Experimental protocols br Acknowledgements This research
Acknowledgements This research was funded by the Croatian Ministry of Science, Education and Sports (Programmes 098-0982915-2948 and 098-0982522-2525), the Austrian Science Fund FWF (DK-MCD W1226, SFB LIPOTOX F30, P22832), and the Croatian-Austrian Intergovernmental S&T Cooperation Programme (HR 08/12). The authors thank Dr. Biserka Kojić-Prodić, Dr. Marijeta Kralj, and Dr. Vesna Blažek (Ruđer Bošković Institute) for scientific discussions and Anton Ibovnik (Medical University of Graz) for technical assistance.
Introduction Serum or plasma concentrations of the cholesterol precursors lanosterol, lathosterol, and desmosterol are widely used as surrogate markers of endogenous cholesterol synthesis . The cholesterol metabolite 5α-cholestanol and the plant sterols campesterol and sitosterol are used as markers of cholesterol gaba receptors [2,3]. These non-cholesterol sterols (NCS) show even stronger correlations with cholesterol absorption and synthesis when expressed as ratios to total cholesterol, which standardizes for variations in sterol transport protein concentrations . Specifically, when reporting NCS as their ratios to cholesterol, the cholesterol measurement should ideally be performed from the same sample preparation as for the NCS analysis. The interest in surrogate markers of cholesterol metabolism is recently increasing and more laboratories develop new methods for the determination of cholesterol and NCS using different chromatographic separation and mass spectrometric detection methods. Conflicting absolute and cholesterol-corrected NCS concentrations are reported in the literature making it difficult or nearly impossible to define cut-off values [, , ] or to compare absolute or cholesterol corrected values from different studies in meta-analysis. For a better comparison of reported values and the identification of methodological sources of errors, we planned and performed the first international surveys for cholesterol and NCS. The first part of the cholesterol and NCS survey was initiated in the year 2013 under the expertise of the German Reference Institute for Bioanalytics (RfB) and the Laboratory for Specialized Lipid Diagnostic of the University Hospital, both located in Bonn, Germany. The aim of the first survey was to investigate the variations in cholesterol and NCS concentrations routinely determined by different laboratories in Europe, North and South America. Here, we reflect the comparability of cholesterol and NCS concentrations determined by different separation and detection methods and discuss the suitability of different compounds used as internal standards for sterol quantification. Twenty laboratories specialized in chromatographic lipid analysis, by either gas- or liquid-chromatography were enrolled to participate in this first survey part. The second part of the international cholesterol and neutral sterol survey took place one year later in 2014. Twenty-two laboratories participated in the second survey whereof five laboratories attended for the first time. Contrary to the first survey, the second was designed as an interventional trial. The focus was on the influence of the utilized calibration solutions and the participants were requested to use provided stock solutions for the quantification of the sample material.
Materials and methods The participants submitted results from six different methods for sterol determination: capillary gas chromatography-flame ionisation detection (GC-FID) with either 5α-cholestane (5α-chol) or epicoprostanol (epi) as internal standard (GC-FID-5α-chol and GC-FID-epi, respectively), capillary GC-mass selective detection (MSD) with either 5α-chol, epi or deuterium labelled sterols as internal standards (GC-MSD-5α-chol, GC-MSD-epi and GC-MSD-deuterium, respectively) and high performance-liquid chromatography (HPLC) with MSD and deuterated sterols (LC-MSD-deuterium) used as internal standard. Since each laboratory was requested to use its specific routine analytical method, the work-up procedure and determination setting are strikingly different for the individual participants. Table 1 shows an overview of the reported sample work-up conditions, chemicals, chromatographic columns, and instrumentation for GC analyses used in the surveys. Most of the laboratories used alkaline hydrolysis in order to deconjugate fatty acid esterified sterols and thus analysing total serum sterol concentrations after derivatisation of the free hydroxyl groups. One laboratory, using LC-MSD, quantified free and esterified fraction of the freeze-dried test-samples in order to calculate the total concentration. Some laboratories used fully automated peak integration and quantitation software belonging to the software package of the supplier. Others integrated the individual peaks half-automatically using self-created integration and evaluation macros with the option to supervise and correct integration of each peak. It should be emphasized that the individual sample work-up procedures or detection specifications cannot be reported within this article because of the anonymization of the participants. Since the freeze-dried serum samples were routinely used in round-robin tests of the RfB, it was possible to give a reference value (target concentration) only for cholesterol. The lyophilized serum samples were additionally used to measure other parameters in clinical routine biochemistry such as triglycerides, total protein or subspecies proteins such as albumin or gamma globuline in additional ring trials in Germany. All these parameters were within the normal range as characterized by reference values except for triglycerides in sample D of the second ring trial, where the median level was marginally increased (data not shown). Single values more than 50% different from the mean value of all submitted sterol concentrations have a marked effect on the mean values of the whole participant group. Therefore, individual values more than 50% different from the mean value were regarded as outliers and several of the above-mentioned parameters were additionally calculated without the outlier values and reported within the script. The guidelines of the German Medical Association for quality assurance of clinical laboratory investigations tolerates relative deviations for cholesterol within external proficiency testing of up to 13% . Due to the relatively small number of participants in our surveys we set a relative deviation, that can be considered as acceptable to ± 15% of the mean value of the included participants. The mean values, standard deviation, minimum and maximum values of all participants as well as the ratios of 5α-cholestanol, lathosterol, campesterol to cholesterol and the ratio lathosterol to campesterol are listed in Table 2, Table 3 and shown in Fig. 1, Fig. 2.