Comment on: Standardization of steroid tests and implications for the endocrine community

I was rather astonished to read Dr Honour’s recent editorial in which he appears to be blaming external quality assurance (EQA) programmes for the failure by assay manufacturers to make more accurate and specific methods for steroid hormones. ¹ This was especially surprising in that Dr Honour was a special advisor to the UK NEQAS for steroid hormones, of which I was Organizer from 1987 to 2010.

From the start of my EQA career, I was communicating with assay manufacturers about the poor analytical quality of their steroid assays. I continued the excellent work of Dr Graham Groom, my predecessor, by regularly distributing special serum pools with values assigned by the reference method, isotope dilution gas chromatography mass spectrometry (ID-GCMS), which highlighted poor calibration and lack of analytical selectivity. For many years, I used the services of the world-renowned reference laboratories of Professor Lothar Siekmann and Professor Linda Thienpont to provide target values for these materials.

I regularly presented the data from these studies and from recovery, linearity and interference exercises, at national and international conferences, major EQAS participants’ meetings, papers in this journal and summarized my views in a major textbook. ² I wrote a 100-page annual review, distributed to more than 300 participants and manufacturers all over the world, where these data and the problems they demonstrated were described. I also attended diagnostic company user group meetings and made robust criticisms of their methods, explaining how their characteristics could be improved using the well-established, fully traceable, steroid hormone reference measurement system. I contributed to major publications of the European EQA working groups and sat on the International Federation of Clinical Chemistry Joint Committee for Traceability in Laboratory Medicine (JCTLM) for four years.

With respect to cortisol, following the recommendations of a major review on the standardization of cortisol assays, ³ Jane French and I made a reference panel of 34 single donations of serum with endogenous cortisol concentrations spanning the clinically important range and had these targeted by ID-GCMS at Siekmann’s and Thienpont’s laboratories, prior to lodging them at the Institute for Reference Materials and Measurements as the certified reference panel mentioned in Dr Honour’s editorial. We designed this panel specifically for diagnostic manufacturers to use in ‘split sample’ exercises to assess the calibration and selectivity of their methods.

At the time I retired as Organizer, routine use of liquid chromatography tandem mass spectrometry (LC-MS/MS) techniques for steroid hormones was still relatively new, but I had already planned to use the trimmed method mean of the mass spectrometry group as the target value for several analytes in the UK NEQAS service when it was robust enough, as an alternative to using pools targeted by a JCTLM-listed reference laboratory, as the cost of the latter would not have been supportable. I am pleased to say that my successor, Finlay MacKenzie, has recently changed the UK NEQAS target values for cortisol, testosterone, androstenedione, dehydroepiandrosterone sulphate and aldosterone in this way.

I think this demonstrates my commitment to improving steroid hormone assays, so why did my efforts apparently have no effect? The answer is straightforward: manufacturers will not change unless there is strong pressure from purchasers. EQA programmes can only assess and advise as part of their service to participants and post-market vigilance role under the EU (European Union) IVDD (In Vitro Diagnostics Directive). They cannot directly pressure manufacturers to make improvements; they can only do so indirectly through their reports and advice to participants and advisory bodies.

Methods for steroid hormones have radically changed over the years. Laboratories have followed technological advances in assay system design and responded to the challenges of reduced staffing and greatly increased work volumes, by employing highly automated modular immunoassay systems, often within managed service contracts with a single company, which has meant analytical compromises. However, the introduction of LC-MS/MS for therapeutic drugs has enabled laboratories to break free from these analytical constraints and persuade their hospital trusts and commissioners that extending this technology to steroids will actually save money and be beneficial to clinicians and patients. I believe that through the excellent work of, for example, Professor Brian Keevil and colleagues^4,5 in enabling and promoting the use of this technology for steroid hormones, it will develop rapidly, become part of automated modular systems and eventually replace immunoassay.

In the meantime, laboratories should use the validation and verification requirements of ISO 15189 accreditation to require manufacturers to make available details of their traceability chains and the results of split sample studies with certified reference panels, with the explicit requirement that patients’ results are metrologically traceable to the highest level of the reference measurement system.