Determining what constitutes a low level of testosterone is a major point of confusion in the field of testosterone deficiency — also known as hypogonadism. A recent article adds to the controversy by proposing a new, lower threshold for normal testosterone concentrations on the basis of harmonized reference ranges.
Refers to Travison, T. G. et al. Harmonized reference ranges for circulating testosterone levels in men of four cohort studies in the USA and Europe. J. Clin. Endocrinol. Metab. http://dx.doi.org/10.1210/jc.2016-2935 (2017)
The study by Travison et al.1 is the result of a collaboration between the Endocrine Society and the Centers for Disease Control and Prevention (CDC) with a stated objective being “...to generate consensus reference ranges for total testosterone levels in men.” Around 100 frozen blood samples from randomly selected men who participated in one of four epidemiological studies (Framingham Heart Study (FHS), European Male Ageing Study (EMAS), Osteoporotic Fractures in Men Study (MrOS), and Belgian Siblings Study of Osteoporosis (SIBLOS)) underwent testing at the CDC using liquid chromatography with tandem mass spectrometry (LC-MS/MS). Previous testosterone level measurements in each of the four study groups had not been made centrally, so the researchers developed transformational equations to create harmonized results for all cohorts. A reference range was then generated based on healthy, nonobese men aged 19–39 years. Age-specific ranges were also reported.
The primary result was that statistical modelling successfully reduced the variability in results between studies. The harmonized 2.5th, 50th, and 97.5th percentile values for a reference population of nonobese men aged 19–39 years were 264 ng/dl, 531 ng/dl, and 916 ng/dl, representing the lower limit of normal, the median, and the upper limit of normal, respectively. Harmonized age-specific 2.5th percentile values for men aged 40–49, 50–59, 60–69, 70–79, and 80–89 years were 208 ng/dl, 192 ng/dl, 190 ng/dl, 190 ng/dl, and 119 ng/dl, respectively1.
At first glance, this study seems to recommend an unusually low threshold for low testosterone values of 264 ng/dl. In comparison, 300 ng/dl and ∼350 ng/dl (12 nmol/l) are frequently cited as the lower limit of normal in the USA and in Europe, respectively2,3. Far fewer men would be diagnosed with testosterone deficiency if this new lower limit was to be widely adopted. However, these harmonized values only have meaning within the specific experimental conditions of this study. As the accuracy and reliability of these results are unknown, it would be inappropriate to apply them clinically.
“Clinical decision making requires reliable information, which unfortunately cannot be provided by experimental models”
Supporting this view is the fact that a similar study by many of the same authors was published in 2011 (Ref. 4), involving three of the four datasets included in the new study1. The 2011 study4 reported the 2.5th percentile as 348 ng/dl, a value 32% higher than the 264 ng/dl reported in the 2017 paper1. Testosterone measurements were performed via LC-MS/MS in both studies. This large difference seems to result from the fact that the 348 ng/dl value was derived from actual results from a reference population, whereas the new harmonized results are based on statistical modelling. Clinical decision making requires reliable information, which unfortunately cannot be provided by experimental models. The remainder of the data is notable for demonstrating a negligible decline in median total testosterone concentrations with increasing age in men ≥40 years of age.
Travison and colleagues1 assert, “Rigorously derived reference ranges are essential for distinguishing “healthy” from “diseased” individuals and constitute the foundation of our contemporary approach to making the diagnosis of clinical disorders.” This notion might be true in an ideal world, but clinical medicine is not nearly so tidy. The problem becomes obvious when considering that, by convention, the central 95% of results from a reference population are categorized as normal, whereas the lowest and highest 2.5% of values are categorized as abnormal. The authors' faith in reference ranges would be justified if, by remarkable coincidence, the prevalence of testosterone deficiency (or any other condition) was exactly 2.5%. However, in the example of a medical condition with 5% prevalence, fully half of affected individuals would be falsely categorized as healthy.
The 2010 Endocrine Society guidelines recommend clinicians use their own laboratory's reference range to determine whether a patient's testosterone value is low5. This recommendation is problematic owing to the enormous variability in ranges between laboratories. Lazarou et al.6 reported 17 different testosterone reference ranges among 25 laboratories, with the lower limit of normal ranging from 130 ng/dl to 450 ng/dl, a 350% difference. For example, a man with a testosterone concentration of 280 ng/dl would be categorized as having testosterone deficiency by one laboratory or to be within the normal range by another.
Men with characteristic symptoms and documented low testosterone levels are candidates for testosterone therapy. The challenge is determining what constitutes a low testosterone level. Reference values have been used extensively, yet, experts agree that no testosterone threshold exists that reliably separates men with testosterone deficiency from men without the condition5. Clinical presentation simply varies too much between individuals, and interpretation of total testosterone concentrations is affected by too many confounders. Much of this disparity is caused by substantial variability in individual sex-hormone-binding globulin (SHBG) levels. High SHBG concentrations can raise total testosterone levels into the normal range despite low free testosterone concentrations; however, free testosterone concentration corresponds more closely to clinical presentation than total testosterone levels7. In my practice, I rely on free testosterone more than total testosterone values, and consider calculated free testosterone values <100 pg/ml as low8. Another confounding factor is the number of CAG repeats in the androgen receptor gene; increased numbers of CAG repeats are associated with reduced androgen sensitivity9. For these reasons, the rigid application of total testosterone thresholds to diagnose testosterone deficiency is illogical and contrary to good medical practice.
Travison and colleagues1 have successfully performed a difficult statistical challenge by harmonizing testosterone results across four diverse datasets. Unfortunately, although their results might be internally consistent, they seem to be inaccurate in comparison with actual, measured results previously published by the same group of authors4. Thus, these harmonized reference values should not be applied clinically.
Reference ranges can aid clinicians in the interpretation of testosterone measurements, but they have important limitations and are not designed to provide a definitive clinical diagnosis. Awareness is growing that free testosterone might be a more reliable indicator of androgen status than total testosterone, and that symptoms should be the primary consideration in the diagnosis of testosterone deficiency10. For those who still choose to consider the 2.5th percentile of a young healthy reference population as the lower limit of normal, the most reliable value to date is 348 ng/dl (Ref. 4).
Travison, T. G. et al. Harmonized reference ranges for circulating testosterone levels in men of four cohort studies in the USA and Europe. J. Clin. Endocrinol. Metab. http://dx.doi.org/10.1210/jc.2016-2935 (2017).
Bhasin, S. et al. Testosterone therapy in adult men with androgen deficiency syndromes: an endocrine society clinical practice guideline. J. Clin. Endocrinol. Metab. 91, 1995–2010 (2006).
Wang, C. et al. Investigation, treatment, and monitoring of late-onset hypogonadism in males: ISA, ISSAM, EAU, EAA, and ASA recommendations. Eur. Urol. 55, 121–130 (2009).
Bhasin, S. et al. Reference ranges for testosterone in men generated using liquid chromatography tandem mass spectrometry in a community-based sample of healthy nonobese young men in the Framingham Heart Study and applied to three geographically distinct cohorts. J. Clin. Endocrinol. Metab. 96, 2430–2439 (2011).
Bhasin, S. et al. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J. Clin. Endocrinol. Metab. 95, 2536–2559 (2010).
Lazarou, S., Reyes-Vallejo, L. & Morgentaler, A. Wide variability in laboratory reference values for serum testosterone. J. Sex. Med. 3, 1085–1089 (2006).
Antonio, L. et al. Low free testosterone is associated with hypogonadal signs and symptoms in men with normal total testosterone. J. Clin. Endocrinol. Metab. 101, 2647–2657 (2016).
Aversa, A. & Morgentaler, A. The practical management of testosterone deficiency in men. Nat. Rev. Urol. 12, 641–650 (2015).
Morgentaler, A. et al. Fundamental concepts regarding testosterone deficiency and treatment: international expert consensus resolutions. Mayo Clin. Proc. 91, 881–896 (2016).
Morgentaler, A., Khera, M., Maggi, M. & Zitzmann, M. Commentary: who is a candidate for testosterone therapy? A synthesis of international expert opinions. J. Sex. Med. 11, 1636–1645 (2014).
A.M. has received payments for consulting or lecture honoraria from AbbVie, Aytu, Bayer and BioTE.
About this article
Cite this article
Morgentaler, A. Testosterone reference ranges and diagnosis of testosterone deficiency. Nat Rev Urol 14, 263–264 (2017). https://doi.org/10.1038/nrurol.2017.35
International Journal of Impotence Research (2019)