The Model T

Tin Lizzy was the affectionate moniker for Henry Ford's Model T automobile. As the first automobile made using assembly line methodology, the Model T was the earliest car to be affordable to the working class. It was prized for its low cost, ease of maintenance, and durability. It was the most popular automobile for nearly 20 years. Similarly, for the endocrinologist, the total T assay has been a low-cost, easy to use, and reliable test for the assessment of male gonadal function for decades. The 2010 Endocrine Society guideline for the assessment of male hypogonadism recommended measurement of total T as the best initial diagnostic test for male hypogonadism (1). The guideline also recommended the measurement of free or bioavailable T in some men whose serum total T concentrations are near the lower limit of the normal range and who have conditions associated with alterations in SHBG.

In humans, most T circulates in three forms: avidly bound to SHBG, weakly bound to albumin, and unbound (“free”). According to the free hormone hypothesis first posited by Robbins and Rall, only unbound T is “active” and able to bind to the androgen receptor of target tissues of the body (2). The free hormone hypothesis has remained the dominant dogma for circulating androgen effects for decades, but it has been questioned episodically (2–6). Because T is weakly bound to albumin, an alternative hypothesis is that free and weakly bound T both contribute to androgen effects. The sum of the free T and free weakly bound T is often referred to as “bioavailable T” or “non-SHBG bound T.” Some experts have posited that free T is available to all target tissues, whereas weakly bound T availability might depend on transit time through the target tissue; longer transit time would allow for more time for disassociation of T from albumin (3). A third hypothesis stipulates that T bound to SHBG is available to some tissues. This hypothesis is based largely on the observation that male megalin knockout mice are born with unilateral cryptorchidism, a sign of partial androgen deficiency or resistance (7). In mice, the megalin receptor appears to mediate endocytosis of T bound to SHBG.

The free hormone hypothesis is clinically important because a number of medications, diseases, and conditions affect serum binding proteins and may lead to discordance between total hormone and free hormone concentrations. For example, conditions that result in higher than normal hepatic estrogen concentrations (eg, pregnancy, thyrotoxicosis, and oral estrogenic drugs) increase binding proteins for T, cortisol, and thyroid hormone. In men with normal endocrine function, high hepatic estrogen exposure results in higher serum total T, cortisol, and thyroid hormone concentrations but normal serum free hormone concentrations. On the other hand, in patients with endocrinopathies such as primary hypothyroidism, high hepatic estrogen exposure results in no change in total hormone concentration (because the patient cannot make more hormone) but lower free hormone concentration.

For thyroid hormone, the free hormone hypothesis appears to be correct (3, 8). Patients with congenital abnormalities in thyroid binding globulin concentrations have normal phenotypes (8). Furthermore, when women with primary hypothyroidism become pregnant or use oral estrogens, they often require significantly higher dosages of levothyroxine to maintain clinical and biochemical euthyroidism (with euthyrotropinemia) (9).

For T, there are scant human data to determine whether the free hormone is correct. Although there are many clinically important causes of perturbations in SHBG (eg, diabetes mellitus, aging, hepatopathy, and medications such as anti-epileptics), there is little evidence to determine the validity of the free T hypothesis. In the single published case report of a man with congenital SHBG deficiency, the man had a normal phenotype except for subtle manifestations of hypogonadism (including low libido, decreased morning erections, and weakness) (10).

The serum gonadotropin response to perturbations in serum SHBG concentrations is not useful in the treatment of men with primary or secondary hypogonadism. Unlike primary hypothyroidism, where serum thyrotropin is a reliable indicator of the correct thyroid hormone replacement dosage, serum gonadotropins are not a reliable measure of gonadal status in men with hypogonadism who are on T therapy. Thus, it is not possible to use serum gonadotropins to determine whether perturbations in SHBG concentrations alter the necessary T replacement dosage to achieve eugonadism.

One epidemiological study has suggested that serum free T or bioavailable T might correlate better with classic signs of androgen effect than serum total T concentrations in older men. In a cross-sectional study of 403 men over age 70 living independently, calculated free T and bioavailable T concentrations both correlated more strongly than total T concentrations with increases in leg strength and bone density and favorable changes in lean and fat body mass (11). However, a preliminary analysis of a recent clinical trial of the dose-ranging effects of T on bone mineral density, strength, body composition, and sexual function showed no differences in the observed outcomes when the analyses were performed using serum total T vs free T concentrations (J. Finkelstein, personal communication) (12, 13).

With this background of uncertainty about the free T hypothesis, the latest data from the European Male Aging Study (EMAS) are clinically important (14, 15). Over 3000 European men were recruited to participate in this multicenter trial. Men with classical hypogonadism due to a known cause of hypothalamo-pituitary or testicular dysfunction were excluded. After a median follow-up of 4.3 years, 2736 of the original cohort participated in phase 2. At both phases, all subjects completed sexual, physical, and psychological questionnaires and underwent single early morning blood phlebotomy, anthropometry, and a quantitative heel calcaneal ultrasonography for bone densitometry. The men were divided into four groups based on serum total T (measured by tandem mass spectrometry) and calculated free T concentrations: 1) a referent group with normal total T (TT) and normal calculated free T (cFT); 2) normal TT and low cFT; 3) low TT and normal cFT; and 4) low TT and low cFT. The threshold for a low cFT was 220 pmol/L (63.5 pg/mL), based on the previous EMAS findings demonstrating that a triad of symptoms of sexual dysfunction (decreased frequency of sexual thoughts, morning erections, and erectile function) occurred at higher frequency below this threshold (14). The lower limit of TT was 10.5 nmol/L (303 ng/dL) and was based on a series of receiver operator curves to identify the TT threshold that was predictive of a low cFT. This defined lower limit of TT is also the threshold when the triad of sexual symptoms increases, and it conforms to most experts' definition of the lower limit of normal.

The principal clinical findings from this latest study from the EMAS group include the following: 1) men with low TT and low cFT have the severest sexual and physical symptoms and signs of hypogonadism; 2) men with normal TT but low cFT often have multiple symptoms and signs of hypogonadism; and 3) whereas men with low TT but normal cFT generally lack specific clinical characteristics of hypogonadism.

The authors acknowledge that their study is not designed to assess the benefits and risks of T therapy in men with low TT and low cFT or normal TT and low cFT, and it is not directly applicable to clinical practice. Specifically, the design of the study does not address the question of whether low serum TT or cFT is causally related to symptoms in men without classical hypogonadism. It remains likely that low serum TT or cFT is a marker, not a cause, of poor health in many men without classical hypogonadism. However, this study supports the Endocrine Society guidelines to measure serum TT in men who present with symptoms and signs of hypogonadism and to assess serum free T concentrations in men who have conditions that are associated with abnormal SHBG concentrations. Based on this new EMAS analysis, measurement of TT is still useful because men with concordance of low TT and low free T are most likely to be hypogonadal and are most likely to benefit from T therapy.

This study does not prove that the free T hypothesis is correct. However, based on the aggregate (albeit limited) knowledge that we have from animal and human data, it is likely that the free T accounts for much of the physiological effects of circulating T (4). This conclusion is consistent with the relatively normal phenotype of the man with congenital absence of SHBG and with the recent EMAS epidemiological findings. It is also plausible that T bound to SHBG may exert endocrine effects in specific tissues, including the developing male gonad in utero, but these effects appear to be modest. Much work must be done to further elucidate the physiological role of T bound SHBG or albumin. In the meantime, it seems likely that the free T hypothesis will account for most, but not all, of the effects of circulating T.

What tests should endocrinologists order in the assessment of T deficiency? Total T measurement remains useful because accurate assays for total T are widely available. Widespread use of tandem mass spectrometry would further increase the accuracy of total T measurement. In most men with classical hypogonadism due to a known cause of hypothalamo-pituitary or testicular dysfunction (eg, a pituitary tumor or Klinefelter syndrome), total T concentrations are unequivocally low, and assessment of serum free T is not necessary.

Symptoms of male hypogonadism are nonspecific and might be due to depression, system illness, poor physical conditioning, or drugs. Conditions that decrease serum SHBG concentrations are common. Thus, in men with low serum total T concentrations without a clearly identifiable cause of hypogonadism or in men with conditions known to affect serum SHBG concentrations, assessment of serum free T concentration is often useful. In an endocrinologist's practice, obesity, diabetes mellitus, and other conditions associated with abnormal SHBG concentrations are so common that assessment of serum free T might be warranted in a majority of patients.

Measurement of free T by equilibrium dialysis remains the “gold standard,” but it is an expensive and laborious assay performed by only a few commercial laboratories. Calculation of free T with the measurement of T, SHBG, and albumin is inexpensive and readily available. Although there is controversy about the most accurate formula to use, some formulae for calculated free T concentrations have excellent concordance with values obtained by equilibrium dialysis (16–18). Measurement of free testosterone by direct measurement by an analog method is inaccurate and should not be used (1). Establishment of a normal range is a major issue with any method for total and free (either direct or calculated) T measurements. Regrettably, there is no uniformly accepted normal range, and local laboratories often use published normal ranges or create normal ranges with a convenience sample of the population (eg, blood donors) that do not represent healthy young men with a normal gonadal axis and reproductive function. Endocrinologists must work with their local laboratory to establish a reliable normal range or to use one of the commercial laboratories that follow standard principles for normal ranges.

The ideal or “model” T assay does not exist yet, but our understanding of the transportation of T and its physiological effects in the human body inches forward.

Acknowledgments

Disclosure Summary: The author has nothing to disclose.

Abbreviations

 
• cFT

  calculated free T

 
• TT

  total T.

References