Mass spectrometry for steroids

This issue of the Annals includes several articles reporting analysis of steroids using mass spectrometry (MS).^1–3 Many papers on the use of MS for steroid measurement and assay interference have been published recently in the Annals, reflecting a growing appreciation of the power of this physicochemical technology.^4–9 Formerly, MS was primarily used for steroids in clinical research and reference centres. Combining MS with gas chromatography (GC–MS) is still a highly valued technique for profiling the range of urinary steroid metabolites^10,11 in the diagnosis of adrenal and some gonadal disorders, and for the most accurate of steroid quantitative assays (e.g. reference/gold standard methods ¹² ). The significant developments in MS for clinical laboratories in the last 15 years have come through the combination of liquid chromatography (LC) with tandem MS (LC–MS–MS) adopting new techniques of forming ions at their interface.

Samples for LC–MS/MS ideally need minimal preparation before analysis. Protein precipitation, liquid–liquid extraction, ³ solid phase extraction ¹ and supported liquid extraction^2,8 are all used. In some cases, sample preparation can be automated^1,2 or performed in 96-well plates. ¹ A precolumn may be incorporated before the LC process to improve sensitivity by holding back potential interferents and hence reduce background signal. Phenyl, ¹ C8³ and C18² high-performance LC columns have undergone refinement with increasingly smaller particles being used to enhance resolution. Mobile phase constituents include water or aqueous buffers (ammonium acetate and formic acid) with organic modifiers. Ammonium fluoride is added in oestrogen analysis to improve sensitivity. ² The electrospray source is most widely used for steroids. Instruments vary in design to focus the sample beam, which is important when measuring low concentrations such as testosterone in saliva. ³ The aim is to deliver the analyte to the mass spectrometer over a narrow time frame; positive ions are usually detected for steroids but for oestrogens ² negative ion analysis is preferred. Selected reaction ion monitoring is a two-step process to focus ions in the first sector (MS1) and then generate fragments in a collision cell, which are examined by the second sector (MS2).

Steroid assays are usually stable isotope dilution methods with variable levels of deuterium incorporation in the internal standard. This increases the likelihood that the internal standard will behave in exactly the same way as the analyte throughout the procedure and therefore compensate for any effects on signal intensity through ion suppression or enhancement.^13–17 The substitution of three atoms of deuterium is common but when more than three stable isotope atoms are added, the labelled steroid elutes ahead of the natural steroid thus risking independent ion suppression. The sites of incorporation may not always be appropriate for MS–MS methods depending on the fragmentation path and stabilities of charged ions. The final procedure must undergo validation for ion suppression, accuracy, recovery, imprecision, linearity, limit of quantitation and detection, specificity and stability. Some useful guidance is available to support assay development.^18–20 Specificity for steroids requires knowledge of LC retention times, molecular weights and fragmentations. Owen et al., ² for example, tested 15 naturally occurring steroids and 12 synthetic steroids in their validation of serum concentrations of oestradiol and oestrone.

LC–MS/MS methods in clinical laboratories have now been reported for nearly all steroids, and in some cases reference ranges have been generated. ⁶ Many of these methods are classified as laboratory-developed tests and have no regulatory approval (e.g. Conformité Européenne [CE] marking). Commercial kits are available with reagents permitting multiple steroid analysis in a single LC run, although not all of the steroid results generated are needed for clinical decision making. A number of factors can affect clinical results from LC–MS/MS. Blood samples vary considerably in nature and composition (e.g. variability in drug and metabolite concentrations and interferents including plasticisers from blood collection tubes) that may interfere by suppressing or enhancing ion signal intensity. Several steroids encountered in blood have identical molecular weights (isobaric compounds, e.g. 17-hydroxyprogesterone and 11-deoxycorticosterone). Steroids in blood are bound to albumin and specific steroid-binding proteins. The use of oral contraceptives affects binding protein concentrations and hence total steroid concentrations. Isomers are often encountered ¹ but are recognized and quantified when separated by the chromatographic step.

Two quantitative methods^2,3 and one study where MS was used to substantiate interferences from epi-isomers in 25-hydroxyvitamin D determination ¹ are reported in this issue. Keevil et al. ³ suggested differences in salivary gland testosterone metabolism between men and women from measurements of salivary testosterone and serum-free testosterone. In a previous issue of the Annals, norethisterone was shown not to interfere in the LC–MS/MS analysis of testosterone. ⁴ Sensitivity down to 5 or 10 pmol/L has been demonstrated for serum oestradiol/oestrone ² and salivary testosterone ³ using on-line sample preparation and a research level mass spectrometer to optimize signal generation: few hospitals would have equipment of this specification limiting the sensitivity that can be achieved.

The adoption of MS has been driven by (i) recognition of its superior specificity and sensitivity, (ii) reduction in sample preparation time and (iii) cost. If the purchase of a mass spectrometer is spread over several years or the equipment is leased, then above these hardware costs the running costs (mainly the solvents and columns) compare favourably with immunoassay. For steroid analysis, immunoassay may become a technique of the past. Developing a new method, in-house, clearly requires much laboratory skill. Validated methods can be adapted, although the MS/MS equipment does not perform identically. Setting up and establishing a new service based on LC–MS/MS is demanding for laboratory personnel, requiring the acquisition of new skills but the product will ultimately benefit patients and lead to harmony of results.