A Call for Better Estradiol Measurement

Research over the decades has demonstrated that the sex steroid hormones, primarily estradiol and testosterone, not only exert control over the development and maintenance of reproductive capacity but also have a guiding hand on a good many other biological processes. The evolving understanding of these latter functions has placed new measurement demands on both estradiol and testosterone, particularly at low levels that reflect non-reproductive physiology. In the case of testosterone, a concerted effort during the past 5 years by a broad public-private consortium has made great strides in achieving more robust and standardized measurements. Now, an Endocrine Society panel has turned its sights on estradiol measurement, calling for a universally recognized traceable standard, age- sex- and biologically-specific reference ranges, and more.

"We have the same problem with estradiol as we've had with testosterone, which is that the current assays, particularly direct immunoassays, are inadequate for many of the purposes we need estrogen measurements for. That includes in men, post-menopausal women, and in adolescents and children," said the panel's chair, William Rosner, MD, a professor of medicine at the Columbia University College of Physicians and Surgeons in New York City. "This is a very ambitious agenda, and we don't think it's going to happen tomorrow. But at the end of the day, how can we as physicians practice medicine without trusting the numbers we're getting from these assays?"

New Uses Outpace Methods

Panelist Patrick Sluss, PhD, recalled that the uses of and demands on estradiol assays have changed considerably since the immunoassays to measure this hormone were first developed. "Back in the 1970s, we used estradiol testing primarily in younger women for fertility management where measuring relatively high levels was the target," said Sluss. "Since that time our picture of both estradiol and testosterone as being gender-specific hormones has changed. However, whether you're looking at estradiol in men or testosterone in women, you need very low levels of detection. Whereas the assays were designed to measure testosterone in men and estrogen in women, which, in both cases, are at high levels." Sluss is director of special chemistry at Massachusetts General Hospital and associate professor of pathology at Harvard Medical School in Boston.

Levels of estradiol in women of childbearing age usually are >60 pg/mL, depending on the day of the menstrual cycle. Women undergoing induction ovulation for infertility treatment might have considerably higher levels, 250–2,000 pg/mL. In contrast, levels in post-menopausal women, men, and pre-pubertal children and adolescents typically are <20 pg/mL, and even lower, <5 pg/mL, in elderly men and women. The lowest levels, <1 pg/mL, probably can be found in women with breast cancer receiving aromatase inhibitor therapy to suppress estradiol levels. To meet the measurement needs of these disparate populations, the panel suggested that estradiol methods must be capable of yielding results over an impressive range, by a factor of 10,000 (J Clin Endocrinol Metab 2013; 98:1353–60).

That tall marching order poses challenges for just about all methods available to measure estradiol. The lower limit of quantitation of first-generation direct immunoassays is 30–100 pg/mL, which, even with a coefficient of variation of 20%, is adequate for use in premenopausal women and those undergoing fertility treatments. Liquid chromatography tandem mass spectrometry (LC-MS/MS) and second-generation direct immunoassays have lower limits of detection, about 10 pg/mL, but are not in all cases robust enough to capture the normal distribution of values for populations known to have low estradiol concentrations. High-performance (HP) LC-MS/MS potentially has adequate sensitivity and specificity for low estradiol concentrations, but still is not without potential analytical pitfalls. Similarly, the gold standard method for measuring estradiol, isotope dilution gas chromatography-MS, is too complex and time-consuming for routine clinical use (See Figure, below).

Rosner explained that each successive assay method has made it easier to meet the growing need for estradiol testing, but all also have their own challenges. "The early radioimmunoassays involved extracting the biological fluid, followed by a chromatography step, and then the assay itself. They had good sensitivity and specificity at that time," he said. "To get more efficient, direct immunoassays were developed which didn't require extraction and chromatography steps. That led to the ability to automate, which drove the cost down considerably. However, the price of that automation was very often inaccuracy. Direct immunoassays work fine in some cases, but not very well in others. Estradiol and testosterone are examples of that."

Arguably one of the most significant challenges for direct immunoassays is that antibodies in these tests cross-react with multiple forms of the hormone. According to the panel, estradiol circulates in blood as more than 100 conjugated and unconjugated metabolites. Furthermore, patients may have circulating estrogens from exogenous sources such as conjugated equine estrogens and nutritional supplements. Some of these forms circulate in high enough concentrations that even small cross-reactivities can cause "profoundly deranged results," the panel wrote.

The Limits of Mass Spectrometry

Even as the authors cited problems with direct immunoassays, they were quick to point out that MS methods have shortcomings too. "Although these technologies have a number of theoretical advantages over immunoassays, they too may show considerable variability, as has been shown for other analytes such as testosterone," the panel wrote. "As is true for all methods, measurements based on mass spectrometry depend upon the accuracy of calibration, freedom from interference, and appropriateness of any analytical corrections. Thus, it is important to realize that it is not solely the method per se that determines validity and utility."

Panelist Hubert Vesper, PhD, commented on the pros and cons of MS-based methods for measuring estradiol. "Mass spectrometry does have some advantages in terms of specificity of the measurement and in some cases also sensitivity of the measurement. Mass spectrometry can provide very accurate measurements if it's performed correctly, but it's a very complex technology and some of the problems we see in it are the same ones we see with immunoassays," he said. "This means we need to make sure that mass spectrometry assays, as well as immunoassays, need to be calibrated properly." Vesper is director of clinical standardization programs and chief of the protein biomarker laboratory in the Centers for Disease Control and Prevention's (CDC) Division of Laboratory Sciences.

Rosner elaborated that while MS has come along over the years in terms of simpler instrumentation and faster through-put, it also is not a panacea for measuring estradiol, a point echoed by panel member Margaret Wierman, MD, who has had her own difficulties with this method. "As anybody knows who's tried to develop one, steroid assays for mass spectrometry are very difficult. I spent two years trying to do it here and we failed," she said. "We have to be a little careful about the new kid on the block, which is expensive and labor intensive. We have to prove that it's better before we jump on it." Wierman is a professor of medicine and neuroscience at the University of Colorado School of Medicine and chief of endocrinology at the Denver VA Medical Center.

Fine Variances Make a Big Difference

Wierman went on to explain that regardless of method, a host of analytical issues—intra- and inter-assay variability, inadequate sensitivity and specificity, and interferences—combine to yield inaccurate and inconsistent estradiol results, complicating the diagnostic process. "The current platform assays can't distinguish between 10 and 60 pg/mL. That can lead you in two directions. Very low numbers cause you to think something's wrong so you do an extra set of tests to evaluate for what you think might be wrong. On the other hand, you don't realize something is wrong because the assay you're using doesn't have sensitivity at the low levels or sensitivity or specificity at the high levels," she said. "I just can't tell you how many wasted evaluations I've done over the last four-to-five years because of available assays. In endocrinology practice, we rely on hormone tests, and if we can't depend on the results to help diagnose normal and pathologic states, we're sort of stuck."

She elaborated that for many endocrine disorders, fine differences in reported estradiol concentrations substantially influence clinicians' thoughts about differential diagnoses. For instance, a woman menstruating irregularly could be under stress or experiencing early menopause. In a workup for this patient, Wierman would evaluate both follicle stimulating hormone (FSH) and estradiol results. Small concentration disparities in either hormone could lead her to draw different conclusions, with profound consequences for the patient. "If the FSH is high, but the estradiol is low, that suggests early menopause. But if this is a young woman, this is heart-breaking news, so I wouldn't want to tell her that unless I was certain. I have to decide: do I tell her, do I not, and how much more do I evaluate before I know for sure?" she said.

Minor estradiol testing inaccuracies can spell trouble for managing diseases, too. Wierman cited as an example managing disorders of puberty, where the transition to sexual maturity starts too early or too late. "You need to have the right numbers to target, not too low, not too high," she said.

Lack of analytical accuracy not only impacts current clinical care but also is holding back a better understanding of emerging science around estradiol's role beyond female reproduction and sexual maturity. "One of the conundrums we see—and this is not 100 percent correct but is true in a general sense—is that women don't have heart attacks until they're menopausal. The question is, is that because their estradiol disappears?" said Rosner. "On the other hand, there are some papers to indicate that men with higher estradiol are at greater risk for heart attacks. That literature is quite confusing, but it's pretty clear that estradiol is involved with the heart and arterial health. There's a lot to be learned."

The panel cited a number of areas where there is incomplete or conflicting data, including estradiol's role in men not only in normal physiology and pathophysiology, but also in bone health, muscle strength, hypogonadism, and hormone-dependent cancers.

Given all the challenges with measuring estradiol, the panel called first and foremost for a universally recognized estradiol standard to which all other measurements can be traced. In addition, the group recommended development of age- sex- and biologically-specific reference ranges. The committee emphasized the need for wider recognition that estradiol measurements using whole serum without prior extraction are unlikely to be reliable, especially in populations known to have low values. Finally, the authors called for new methods to measure estradiol in the 0.2–2 pg/mL range in routine clinical specimens.

The PATH Forward

If the panel's analysis of the current state of estradiol testing seemed downbeat, their forecast for its future was equally optimistic. This in no small part is due to strides made recently in standardizing testosterone measurements. Much of the heavy lifting for this has been accomplished by CDC's hormone standardization project (HoST), which seeks to ensure accurate and comparable results across assay methods, labs, and time. In the case of testosterone, HoST within just a few years has developed a reference method for HPLC-MS/MS calibrated to a reference material, NMI-M914, with another reference material, SRM 971, used as a trueness control. This process assured metrological traceability. The agency also is working with both manufacturers and labs to calibrate assays, and it collaborates with proficiency testing programs, including the College of American Pathologists and New York State.

Beyond CDC's efforts is the Partnership for the Accurate Testing of Hormones (PATH), a consortium of at least 14 organizations, including AACC, with a goal of advancing standardization and harmonization of hormone assays. PATH has taken several crucial steps toward achieving testosterone standardization, including supporting a male reference interval project and planned female and pediatric projects, undertaking several educational initiatives, and urging medical journal editors to encourage and eventually require the use of accuracy-based testosterone assays in any studies they publish.

As a long-standing, tireless advocate of improved hormone testing, this has been music to Rosner's ears. "I can't tell you how pleased I am at how well HoST and PATH efforts have been accepted and how they have changed the whole landscape of testosterone testing," he said. He and other members of the panel now hope to pass these same milestones on the road to standardize estradiol testing.

AACC also is banking on PATH's success with testosterone measurements to guide the way for estradiol, according to Gary Myers, PhD, vice president of programs and policy. "The success of PATH again emphasizes the need for standardization of clinical laboratory results to improve patient care. This is really important because it will take a combined partnership effort between the laboratory community, the diagnostic manufacturers, and clinical groups to implement standardization improvement in clinical laboratory results. With the experience gained in standardizing testosterone measurement, the effort to improve estradiol measurement is a natural next step." Myers added that PATH's success is important to AACC because it directly supports the Association's strategic initiative around harmonization, the International Consortium for Harmonization of Clinical Laboratory Results.

Already, CDC has taken several important steps, according to Vesper. The agency has developed a LC-MS/MS reference method for estradiol, which it is using to assign target values to single donor serum materials that labs can use to assess the calibration and accuracy of their assays. CDC also has other efforts underway.

The Here and Now for Labs

Before these initiatives take root, laboratorians can do much to improve the use and understanding of estradiol measurements. Wierman and Sluss urged labs as a starter to communicate clearly to physicians when they change their assay methods. "Our methods are driven by lots of things, including economics. With multi-analyte platforms we very often are making method decisions that move from one vendor to another, and with something like estradiol where it's very method-dependent it creates a lot of use problems," explained Sluss. "The most important thing for laboratorians is to develop good communications with endocrinologists and to work together on how those tests are being ordered, used, and interpreted."

Wierman also implored laboratorians to listen carefully to physicians when they have questions about test results. From her own experience, labs are variable in their interest in physician concerns.

Focusing on communication, Sluss takes a time-intensive, multilevel approach to ensure clinicians make the most of estradiol testing. He meets with incoming medical house staff and discusses in detail the biochemistry of steroid hormones, reviews the different assay methods, their pros and cons, and presents case studies in which there are disparities between test results and clinical presentations. He also makes himself available to consult with clinicians as needed and spends considerable time reviewing test utilization. "I try to understand ordering patterns and reach out to the people who have a high test volume to see if it's meeting their clinical needs," he explained.

Rosner extoled the power of these collective efforts. "If we insist as a community on accuracy-based standards, it can be done. It doesn't happen overnight and it's labor- and money-intensive but it's what we should be doing."