Requirement for Mass Spectrometry Sex Steroid Assays in the Journal of Clinical Endocrinology and Metabolism

Abstract

The accurate measurement of hormones is the pivot of modern endocrinology. Although tracing its roots to ancient writings, the modern scientific discipline came into existence after a series of mid-20th century Nobel Prizewinning achievements in identifying and measuring hormones. Arguably the most decisive step was the invention of immunoassay in 1959 (1), originally for peptide hormones but extended a decade later to nonimmunogenic steroids (2). This armed the nascent discipline with revolutionary impact through the ability to measure virtually every hormone at orders of magnitude lower than had ever been possible before with earlier cumbersome, insensitive, whole-animal bioassays. Yet the pioneers of steroid immunoassay understood its limitations as that decade’s delay was consumed in finding the additional workarounds required to create valid formats for immunoassay of small, subimmunogenic molecules like steroids. These essentials comprised solvent extraction, chromatography, and structurally authentic tracers, a triplet of validity criteria. Employing these required skills originally confined steroid measurement to manual methods in specialized laboratories; however, in the 1980–1990s, the steeply increasing demand for steroid measurement in the clinic and laboratory drove assay simplification to adapt steroid immunoassays into popular one-step kits and multiplex assay platform formats. This commodification resulted in “direct” steroid immunoassays, which bypassed all the original triplet of validity criteria (most fundamentally extraction), sacrificing accuracy and specificity for throughput speed and lower cost. Although problems with validity (analytical specificity, accuracy) of direct immunoassays were soon identified and were reported over 25 years ago (3, 4), the fundamental significance of violating these validity criteria resurfaced prominently in the last decade as the accuracy of direct hormone assays came under more stringent scrutiny—and as a solution, affordable and accessible mass spectrometry (MS) steroid assays came into view. The method-specific bias and analytical nonspecificity of direct T immunoassays are worst at the low circulating T levels in children, women, and men with pathological, functional, or experimentally induced T deficiency (5–7) where such direct immunoassays have been famously likened to random number generation (8). By modern standards, putative findings using such suboptimal methods would require verification by more specific MS-based assays. This renders direct T immunoassays methodologically inadequate for high-quality clinical reproductive medicine research. Not surprisingly, aiming to measure circulating (picomolar) estradiol concentrations 10to 100-fold lower than circulating (nanomolar) T levels that render direct T immunoassays unreliable is clearly problematic. Thus, direct estradiol immunoassays also lack accuracy in children (9), men (10, 11), and postmenopausal (3, 12–15) or aromatase inhibitor-treated (16) women who all share comparably low circulating estradiol concentrations. It is beyond doubt now that direct immunoassays of sex steroids are methodologically inadequate for high-quality clinical research. MS-based steroid measurements have been available for decades, indeed as long as immunoassays (17). Although they always remained the reference method for steroid specificity and structure, MS-based steroid assays were always too expensive and unavailable for routine use in the laboratory to support clinical research because they required highly skilled operators using complex, difficult-