madman
Super Moderator
Sex Steroids and Androgen Biomarkers in the Healthy Man Study: Within-Person Variability and Impact of Fasting
AbstractObjective. Serum testosterone measurements in clinical practice mostly utilize “direct” (non-extraction) immunoassays which have method-specific bias du
academic.oup.com
Abstract
Objective
Serum testosterone measurements in clinical practice mostly utilize “direct” (non-extraction) immunoassays which have a method-specific bias due to steroid cross-reactivity and non-specific matrix artefacts. Although more accurate, sensitive, and specific liquid chromatography-mass spectrometry (LCMS) dominates in clinical research, the within-person variability of serum testosterone in healthy men using LCMS measurement is not reported. Design: Longitudinal multi-sampling observational study of men in excellent health over three months.
Methods
Elite healthy men (n=325) over 40 years of age in excellent, asymptomatic health provided nine blood samples over 3 months with serum testosterone, dihydrotestosterone (DHT), estradiol (E2), and estrone (E1) measured by validated LCMS with conventional biochemical and anthropometric variables.
Results
Quantitative estimates of within-person variability within day and between day, week,14 months, and quarter were stable other than an increase due to fasting. The androgen biomarkers most sensitive to age and testosterone among widely used biochemical and anthropometric variables in middle-aged and older men were identified.
Conclusions
This study provides estimates of variability in serum testosterone and the best androgen biomarkers that may prove useful for future studies of androgen action in male aging.
Introduction
The precision and variability of testosterone measurement are important in the evaluation of circulating testosterone clinically as well as for power and sample size estimation for the design of prospective clinical trials. Clinically, testosterone measurements are used neonatally to evaluate ambiguous genitalia, during puberty to manage precocious or delayed male puberty, and among adults to define pathological androgen deficiency in men or androgen excess in women (1). In clinical practice, most testosterone measurements are based on single blood samples and, although multiple measurements are recommended by most clinical guidelines, they are less practical and not always implemented.
Accurate, sensitive, and specific mass spectrometry-based testosterone measurements increasingly dominate high-quality clinical research (2, 3) and are likely to become more widely available in clinical practice. Yet systematic studies of the variability of serum testosterone measurements by the reference methodology of liquid chromatography-mass spectrometry (LCMS) and its impact on androgen-sensitive biomarkers of male aging have been little studied. Hence this study analyzed data from an investigation of men over 40 years of age in excellent asymptomatic health who provided nine blood samples over 3 months (4) aiming to partition the within-subject variability of serum testosterone and related sex steroids into within-day as well as between day, week, month and quarterly periods. These estimates will be valuable for power and sample size estimates for clinical trials employing steroid LCMS measurements. In addition, we investigated the impact of accurate serum testosterone measurement on a range of potential biomarkers of androgen action in male aging.
Analyte measurement
Serum testosterone, dihydrotestosterone (DHT), estradiol, and estrone were measured by liquid chromatography-mass spectrometry (6, 7) in a single batch at the end of the study as described previously (4). Serum LH, FSH, and SHBG were measured by immunoassays, full blood counts, and biochemical variables (lipids, prostate-specific antigen (PSA), renal and liver function tests, iron status) by routine hospital-based autoanalyzer methods subject to external and internal quality control.
Discussion
The present study provides the first estimates to our knowledge of between and within-person variability of the three bioactive sex steroids, testosterone, DHT, and estradiol, measured by LCMS in a reasonably large sample of healthy asymptomatic men undergoing repeated sampling over three months. These provide potentially useful estimates for interpreting the measurement of individual sex steroids, most often testosterone, in men based on the variability of any single estimate. Additionally, they provide useful estimates for power and sample size calculation for planning clinical trials that have a measurement of testosterone or other bioactive sex steroids in men as significant endpoints. A potential limitation of the present data is its restriction to very healthy asymptomatic men means that the extrapolation to less healthy or unselected men who may be subject to disorders with non-specific impact on male reproductive functions is untested as to whether similar variability occurs. We previously reported the effects of fasting to increase serum testosterone measured by LCMS in the original report of this study (4). Other investigators have subsequently focused on various nutrient components of a meal as reducing blood testosterone, measured by less accurate or specific “direct” (non-extraction) testosterone immunoassay in small studies (11-17 16) whereas our more detailed findings suggest it is fasting per se that increases serum testosterone and estradiol without necessary regard to nutrient composition. The lowest variability of within-person serum testosterone measurement was in the non-fasting state and at a week interval suggesting that if multiple blood samples are required, spacing them at weekly intervals would provide the most consistent estimates.
Although, as reported previously, this study showed no significant change in circulating testosterone with male aging among men in excellent asymptomatic health, nevertheless the men in this cohort did exhibit changes in some variables associated with aging. This indicates that even older men with excellent asymptomatic health do exhibit some progressive changes with aging, mostly over the age of 70 years and mostly unrelated to testosterone exposure.
We conclude that the present estimates, based on LCMS measurements of serum testosterone, of within-person variability indicate that, aside from the fasting state increasing serum testosterone, repeated sampling in the non-fasting state at a week interval was optimal to minimize sampling error. Furthermore, we derive a set of variables that best reflect androgen action in male aging that may be relied upon for further analytical studies evaluating the impact of declining androgen exposure due to aging co-morbidities among older men.
