Standardization of Hormonal Assays for the 21st Century

Abstract

We write to reprise and expand the theme of a consensus statement on the subject of standardization of testosterone measurement published in the October issue of the JCEM (1) and encourage your review of that article. The statement from The Endocrine Society and the Centers for Disease Control and Prevention (CDC), with the endorsement of many other groups, addresses the impact of currently available, often inaccurate testosterone assays on clinical care, research, and health care expenditures. The concern is rooted in the fact that there is no recognized reference standard for the assay. Absent such a standard, trustworthy ageand sex-specific reference ranges for testosterone are not attainable. Inaccuracies in plasma testosterone measurement have been noted in our Journal for several years (2– 4) and apply to both total and free hormone measurements, particularly at the lower levels. The major problems with sensitivity and specificity of the assays have related to current immunoassay methodology, particularly those that are automated (platform-based), although variability has also been shown with analysis by mass spectrometry (5, 6). There are other problems that are unique to the validity of so-called free testosterone assays. For the most part, free testosterone “measurements” are rarely actually measured but reflect instead derived or calculated values by one or another method that has never been thoroughly validated or standardized. As a result, large errors in this calculation of 20–40% may exist (7–9), typically overestimating values obtained by equilibrium dialysis. The concept of measuring “bioavailable” testosterone is no less conceptually flawed and also provides values of questionable validity (10). Consequently, a calculated free testosterone measurement can in no way substitute for a total testosterone measured in a validated mass spectrometrybased method. An Endocrine Society Position Statement in 2007 warned of the significant limitations of testosterone assays (11). The lack of a “gold standard” test and the implications of inaccurate hormonal assays should be obvious and include mistaken diagnosis with subsequent inappropriate care, misleading research findings, and unnecessary health care costs, to name a few. Indeed, given that we strive to base our clinical practice guidelines (3, 12) upon the results of clinical investigation and clinical trials, such guidelines could be worthless to dangerous when rooted in inaccurate laboratory measurements. A recent study documented adverse cardiovascular events associated with testosterone administration; in that study, both initial participation and subsequent dosage adjustments of testosterone were based upon measurements of total and free testosterone (13). How reliable were those measurements, and how might they have influenced dosage changes directly linked to the observed adverse outcomes? The current consensus statement (1) reflects an ongoing initiative of the CDC (14, 15) in partnership with The Endocrine Society to illuminate this problem. The stated goal of the collaboration is “to improve the quality of research, patient care, and public health through broad implementation of standardized testosterone measurements that are accurate, reliable, and comparable over time.” Although the efforts cited above to standardize testosterone assays are both necessary and laudable, they represent an initiative based on only one hormone. At this time, the best results are achieved by methods based on tandem mass spectrometry. However, the issue is not the specific method, but rather the recognition that an assay’s accuracy be traceable to a scientifically arrived at, univer-