madman
Super Moderator
Dr. Handlesman and Dr. Anawalt two highlly respected and well known endos in the field!
Abstract:
In this debate, two clinicians discuss the merits and demerits of the free testosterone hypothesis. Although most clinical guidelines recommend the measurement of free testosterone in the evaluation for male hypogonadism (at least in men with suspected low serum sex hormone binding globulin), there remains controversy about underlying hypothesis that serum free testosterone is a significant contributor to androgen effect. In this debate, Dr. Anawalt presents evidence to support the free testosterone hypothesis and advocates for the usefulness of accurate assessment of free testosterone in the evaluation of male hypogonadism. Dr. Handelsman argues that the evidence is unconvincing, the hypothesis is flawed and biologically invalid, and free testosterone cannot be accurately assessed in clinical practice.
In this debate, Dr. Anawalt advocates for the scientific validity and clinical value of the free testosterone hypothesis, but Dr. Handelsman argues that the hypothesis is unfounded, nonsensical, and unnecessary for clinical care.
Free Testosterone Hypothesis: True (Dr. Anawalt)
The FT hypothesis posits that the biological activity of testosterone depends principally on the concentration of the unbound (“free”) hormone and not its bound form.1 The hypothesis does not stipulate that serum FT is the sole determinant of its biological activity.1 Thus, the FT hypothesis must be distinguished from the FT transport hypothesis that only the free hormone is transported into a target tissue cell. For the FT hypothesis to be true, healthy individuals with low serum concentrations of testosterone binding protein (and total testosterone)—but normal FT—will have no or minimal phenotype of testosterone deficiency.
Mathematical basis of the FT hypothesis
Mathematical modeling of the FT hypothesis indicates that it is valid for many or most tissues in the body, but not for tissues with high rates of intracellular metabolic elimination of testosterone.1 Thus, the FT hypothesis is likely not true for the liver because 40-50% of the net testosterone delivered to the liver is eliminated by hepatocytes.1,3 9 The FT hypothesis is plausible (and likely) in most tissues (that have much lower intracellular testosterone metabolism) of women and men.1,3
About 40-45% of circulating testosterone is tightly bound by SHBG; about 2% is unbound.1,4 Most of the remainder of circulating testosterone is bound to albumin and other proteins.1,4 The amount of FT delivered to tissues is defined by blood flow, the rates of disassociation from binding proteins, tissue influx , and intracellular metabolism.1 Tissues that are very vascular and have relatively slow blood flow (e.g. hepatic sinusoids) are exposed to greater amounts of FT. SHBG-bound testosterone is “protected” from intrahepatocyte metabolic clearance of testosterone because only ~25% of testosterone that is tightly bound to SHBG disassociates (and enters hepatocytes) during passage through hepatic sinusoids.1 19 Compared to SHBG-bound testosterone, a much greater percentage of testosterone that is weakly bound to albumin disassociates in the hepatic sinusoids and is metabolized.1,2 Because testosterone bound to albumin disassociates more quickly, it has been hypothesized that this “weakly bound” testosterone also contributes to intracellular testosterone concentration and androgenic bioactivity.4,5 23 The amount of testosterone bound to serum albumin does not vary greatly in patients with serum albumin concentrations within the normal range, and typically there is a strong correlation between the serum concentrations of albumin-bound testosterone and FT.6
Laboratory-based studies that support the FT hypothesis
Experimental manipulation of serum testosterone-binding proteins concentrations causes reciprocal changes of FT and androgenic effect.7-12 Normal rats and mice do not produce SHBG, and albumin is the principal testosterone-binding protein. In a study of castrated rats treated with testosterone implants, repeated infusions of SHBG modestly decreased serum FT, but more than doubled total testosterone concentrations compared to animals not infused with SHBG.7 Serum luteinizing hormone concentrations were not suppressed in the SHBG + testosterone-treated rats despite a serum total testosterone greater three-fold higher than the typical threshold for suppression—suggesting that FT and not total testosterone exerts negative feedback on the pituitary gonadotropes. Similarly, addition of human serum with a high SHBG concentration blocks cellular uptake of dihydrotestosterone and its action in a human prostatic cell line.8 Male mice that are transfected with overexpressing human SHBG have markedly elevated serum total testosterone, but they have slightly decreased serum FT, elevated LH (indicated that feedback is mediated by FT—not total testosterone), and a normal male phenotype except decreased weight of the most androgen-sensitive organs (seminal vesicles and levator ani/bulbocavernosal muscles).9
On the other hand, a mutant rat strain that has very low serum albumin concentrations (and no SHBG) has decreased albumin-bound testosterone, smaller testes, but normal free T, and otherwise normal male phenotype including fertility compared to healthy normal rats.10-12 A study demonstrated that in vitro androgen effect is inversely related to the SHBG concentrations in sera from hypogonadal men being treated with testosterone, and androgen bioactivity correlated more strongly with FT than total testosterone.13 Another study demonstrated that novel compounds that competitively inhibit the binding of testosterone to SHBG cause a dose-dependent increase o fandrogen activity (by increasing FT).14
Human studies supporting the FT hypothesis
Epidemiological studies support the FT hypothesis. There are many SHBG polymorphisms in normal, healthy men.15 Although these polymorphisms are associated with significant differences of serum total testosterone, serum FT does not differ significantly (except for an uncommon genotype—rs6259 homozygosity that is associated with higher FT).15 The only reported man with an SHBG missense mutation and no detectable circulating SHBG had a normal physical examination (including normal genitalia and normal-sized testes), very low total testosterone, low normal FT, normal LH, and normal seminal fluid analysis. He had some possible symptoms of hypogonadism including low libido and decreased early morning erections, but his overall eugonadal phenotype corroborated in the absence of SHBG-bound testosterone.16
Two epidemiological studies have reported cross-sectional and longitudinal data for the association between symptoms and outcomes of hypogonadism and serum total testosterone (measured by mass spectrometry), calculated FT (using a validated formula), and SHBG. In the European Male Aging Study (EMAS) of over 3300 community-dwelling men (40-79 years; mean BMI 27-30), men with baseline low FT were significantly more likely to report low libido and decreased erectile function than men with low or normal total testosterone but normal FT; men with low totaltestosterone and normal FT did not differ from those with normal total and FT.17 In the four-year longitudinal follow-up of EMAS, men with low total and FT were more likely to report decreased libido [OR = 2.67 (1.27-5.60)], erectile dysfunction [OR = 4.53 (2.05-10.01)] and/or infrequent morning erections [OR = 3.40 (1.48-7.84)] than men with normal total and FT.18 Men with low total testosterone and normal FT did not differ in reported sexual symptoms from men with normal total and FT. For the Concord Health and Aging in Men Project of ~1700 community-dwelling men (> 704 years old; mean BMI ~ 27-28), the authors have reported that FT offered little additional independent clinical value over the measurement of total testosterone.19 However, similar to EMAS, the men who had concordant low total and FT concentrations had high odds ratios for low libido in the cross-sectional and longitudinal studies: cross-sectional unadjusted: 2.20; CI 1.52-3.28 and after multivariable adjustment 2.11; CI 1.42-3.13; and longitudinal unadjusted 1.41; CI 1.11-9 1.80 and after multivariable adjustment 1.36; CI 1.06-1.74. However, the men with low serum total testosterone and normal FT had non-significant odds ratios for low libido that approximated 1.011 compared to men with normal total and FT. There are several other outcomes (that are less specific to hypogonadism) that have been independently associated with low serum FT; osteoporosis (or low bone density), frailty, and major cardiovascular events.20-25
Overall, the evidence indicates that the FT hypothesis is true. Concordant low total testosterone and FT concentrations buttress the diagnosis of hypogonadism. Importantly, concordance of low total and FT concentrations is most strongly associated with low libido and sexual dysfunction—the clinical outcome that is most specific for hypogonadism and most consistently responsive to testosterone therapy for hypogonadal men.26-28 19 The most common diagnostic dilemma that clinicians face in the biochemical assessment of hypogonadism is the man with a low serum total testosterone due to obesity-mediated suppression of SHBG. In these men, the finding of a normal FT is useful to avoid misclassification (e.g., over-diagnosis in obese men) as hypogonadal and unnecessary lifelong testosterone therapy.28
The Free Testosterone Hypothesis: False (Dr Handelsman)
(The Misguidance of a Mirage: The Many Fictions of Free Testosterone)
Generations of Endocrinology trainees have had the FT hypothesis embedded into clinical endocrinology thinking as intuitively obvious, needing no proof. FT dogma developed a life of its own, as a sophisticated sounding concept regularly invoked to evaluate androgen effects when measured testosterone yielded no meaningful finding. So, this debate is directed towards the proverbial Endocrinologist from Mars whose open mind is focused on evidence, untainted by received wisdom or the straight jacket of dogma. The distinctive instinctive appeal of the FT hypothesis is classically reminiscent of Kahneman’s fast thinking – intuitive, emotional, irrational -compared with slow thinking, which is logical and scientific.29
Flawed concept of the FT hypothesis
The FT hypothesis asserts that the small fraction of non-protein bound testosterone has greater biological activity because it is more accessible to target tissues.4 An immediate problem is that unbound testosterone is also more accessible to sites of its degradation. 16 30 Hence, unbound testosterone may be more, or less, biologically active depending on the balance between its tissuetargets; regardless, no unidirectional interpretation of bioactivity can be deduced from such a shaky implausible hypothesis. Similar reservations apply to son-of-free FT, “bioavailable” testosterone, the sum of unbound plus loosely bound (mainly to albumin) testosterone.
Flawed methodology of FT measurement
FT is inherently a laboratory measurement using dialysis to separate bound from free testosterone.However, dialysis methods are non-robust, laborious and largely unavailable. Instead, formulae estimate lab-measured FT as calculated FT (cFT). However, cFT is not a valid analytical variable as it fails the fundamental principle of analytical chemistry, to compare like with like. The absence (indeed impossibility) of a certified reference standard for FT means there is no sound quality control nor, in turn, any valid pooling or comparing cFT results between different laboratories/centres, making constructing a validated reference range unattainable.
These cFT formulae have built-in flaws. Most (Sodergard, Vermeulen) are based on equilibrium8binding equation theory* although capillary transit time is too short to reach equilibrium. Furthermore, binding theory involves SHBG’s testosterone binding capacity (not its mass*Sometimes amusingly referred to as by the “Law of Mass Action”, a grandiloquence comparable to measuring weight by the Law of Gravity. measured by one of four non-harmonized immunoassays) and there are two binding sites with different testosterone-SHBG affinities.31,32 Nevertheless, binding theory formulae assume a unitary binding affinity with a single “affinity” spanning a five-fold range so formulae employing a unitary, fixed, plug-in binding affinity create systematic formulaic bias diverging from lab-measured FT.33,34 Genetic polymorphisms influence both circulating SHBG concentration, its testosterone binding affinity, or both; these variables are not accounted for in cFT formulae. 35 Ultimately, cFT is not a valid analytical variable, is calculated by erroneous formulae that cannot provide meaningful guidance on androgen status.
Lack of biological validity of the FT hypothesis
A high road to evaluate a postulated pathophysiological mechanism is a genetic mouse knock-out model, which form high-fidelity models for human reproductive pathophysiology.36 However, mature rodents represent a natural SHBG knockout as they have no circulating SHBG, indicating the dispensability of SHBG for reproductive function (including defining FT).
Conversely, a transgenic mouse model over-expressing human SHBG at physiological levels might be an interesting experiment; however, one model expressing SHBG at 200 times the human blood SHBG concentration defies physiological interpretation.9 In the closest analogy, genetic suppression of circulating corticosteroid binding globulin is detrimental to survival from septic shock despite increased free glucocorticoid fraction.37. Direct clinical testing of the FT hypothesis is remarkably scarce for such a ubiquitous go-to concept. One 5-year prospective study in older men using cFT, calculated by either binding theory or assumption-free empirical method, showed that cFT did not enhance prediction of mortality, morbidity or validated clinical outcomes overaccurately measured testosterone by LCMS.19
FT hypothesis: Superfluous for clinical care, fosters testosterone misuse and androgen dependence
Decades of experience in clinical management of testosterone deficiency verifies that appeal to cFT is unnecessary and there is no evidence excluding it from consideration compromises anyone’s health. Indeed, the quasi-epidemic of excessive testosterone prescribing over three recent decades, comprising a 100-fold increase without any new approved indication, indicates a pressingneed to resolve its driving forces, including guidelines that allow cFT as a valid analytical variable.38 These issues converge on a strong need for a return to the classical definition of “hypogonadism” as a pathological disorder of the hypothalamus-pituitary-testicular (HPT) axis. Over recent decades, the term “hypogonadism”, has been hijacked by re-definition to any lowering of testosterone below an arbitrary, inflated “low” threshold, thereby vindicating testosterone prescription for “hypogonadism” without any authentic organic basis. That new definition creates an invented disorder, known by a cavalcade of neologisms seeking medical gravitas (“Andropause”, “late-onset or functional hypogonadism” etc) but better described as non-gonadal illness or sick eugonadal syndrome.28,39 In that, cFT has played a major supporting role by further defining “low” testosterone4 when blood testosterone was still not “low” enough even for the confected, but inflated, threshold. Rather than trying to squeeze more meaning from inaccurate testosterone immunoassays, a better physiological appeal is to using the endogenous androgen biosensors we already have, the cheaply available, serum LH and FSH used in an analogous fashion to TSH in thyroidal disorders.
Lack of explanatory power of the FT hypothesis
Many trained into FT dogma as an Endocrinology Commandment, consider FT a convenient explanation for patients, usually to deter with a scientific flourish, medically unjustified demands for testosterone prescription in a form of noble cause corruption. But when we come to medical science, evidence is required before convenient beliefs gain credence; ultimately, health literate patients will see through the façade of flimsy evidence. Just as there is no free lunch, confected evidence-free explanations do not hold water or nourish us.
What shall we do?
Diagnosis of hypogonadism needs straightening out. With an identifiable pathological cause (structural or genetic) then testosterone replacement therapy may have a sound basis. But otherwise, in the absence of a pathological cause, a low serum testosterone alone requires a proper diagnosis but not usually testosterone treatment. When such usually futile treatment inevitably ceases, it can create androgen dependence even without there ever having been genuine androgen deficiency in the first place. In this situation, iatrogenic androgen deficiency withdrawal symptoms foster a return to testosterone treatment to alleviate the symptoms creating a cycle of androgen dependence.40 Well known in androgen abusers, this is also evident in men without pathological hypogonadism receiving standard testosterone doses for prolonged periods.41
Practical steps are (1) reject the FT hypothesis as unsound conceptually and empirically, (2) greater reliance on clinical evaluation for recognized HPT axis pathology (structural or genetic) with hormonal testing to confirm, not make, the clinical diagnosis, (3) take multiple serum LH and FSH (androgen sensors) and SHBG (testosterone carrier protein) measurements to interpret serum testosterone, and (4) be vigilant about androgen dependence, thinking of androgens like glucocorticoids - powerful hormones but dangerous to (mis)over use without sound cause.
Dr. Anawalt’s Rebuttal:
My original argument shows that the FT hypothesis is a sound concept that has biological validity. Although I agree that the methods of measurement of free testosterone need improvement and standardization, I do not agree that all available methods are invalid. More importantly, this debate is not about the accuracy of free testosterone measurement; it is about the validity of the FT hypothesis. Dr. Handelsman advances three additional arguments that I rebut:
1. “Flawed methodology [FT measurement]”: Genetic polymorphisms do not generally affect the cFT.15 19 The most commonly used cFT formula (Vermeulen) yields results that are reasonably concordant with equilibrium dialysis with tandem mass spectrometry (the gold standard) across a wide range of serum total testosterone concentrations except at very low levels (where there is little debate about the assessment).4,34,42 When using a total testosterone assay that has been harmonized and certified by an external quality controlinstitution, a validated, useful FT reference range is possible.28,42,43
2. “FT… fosters testosterone misuse”: FT is useful to reduce and prevent testosterone misuse and androgen dependence. The worldwide escalation of testosterone prescriptions paralleled a marked global increase of prevalent obesity (resulting in large number of eugonadal men with low serum total testosterone and normal serum FT). Decision-making based on accurate assessment of FT would have mitigated this quasi-epidemic).28
3. “FT [does not have] explanatory power”: Many eugonadal men will want an explanation for a low serum total testosterone. Analogous to finding a normal free thyroxine in a patient with low total thyroxine, telling them that their free (“active”) testosterone is useful. For most of these men, their free testosterone is normal, they are eugonadal, and their total testosterone is low due to obesity-induced suppression of serum SHBG. That explanation motivates some men to make healthy lifestyle changes to lose weight.
Dr. Handelman’s Rebuttal
In rebutting five arguments, Dr Anawalt tackles four, largely ignoring the assertion that the concept lacks analytical measurement validity. He cites multiple studies using cFT to defend its utility without apparent concern about the baseless nature of these arbitrary figures, nor for what to interpret from such conjured numbers. Under the World Antidoping Code, elite athletes seeking testosterone treatment for hypogonadism require a Therapeutic Use Exemption (TUE) to permit usage of a banned substance to continue working in their profession; however, FT results are specifically excluded from consideration because their analytical invalidity make them unlikely to survive crucial scrutiny by the Court of Arbitration for Sport (CAS), the sports Supreme Court. cFT is not an analytical variable at all, not just lacking any certified standard or quality control but constitutively incapable of ever having one, which disallows pooling or comparison between labs/centers or creating reference ranges to evaluate its utility. That makes belief in its unprovable utility a shining exemplar of Kahneman’s fast thinking – intuitive, emotional and irrational – the contrary of slow thinking scientific logic. It is also congruent with an excellent update review of HOW to “measure” FT, without any cognizance of WHY to do it and what it might mean, an issue taken up in a detailed critique.4,30 To paraphrase JM Keynes, practical evidence-wielding4endocrinologists who believe themselves to be exempt from academic or scholarly analysis of the variables they employ are usually unwitting slaves of some defunct theory, distilling their thoughts from obsolete discussions of decades ago.
Abstract:
In this debate, two clinicians discuss the merits and demerits of the free testosterone hypothesis. Although most clinical guidelines recommend the measurement of free testosterone in the evaluation for male hypogonadism (at least in men with suspected low serum sex hormone binding globulin), there remains controversy about underlying hypothesis that serum free testosterone is a significant contributor to androgen effect. In this debate, Dr. Anawalt presents evidence to support the free testosterone hypothesis and advocates for the usefulness of accurate assessment of free testosterone in the evaluation of male hypogonadism. Dr. Handelsman argues that the evidence is unconvincing, the hypothesis is flawed and biologically invalid, and free testosterone cannot be accurately assessed in clinical practice.
In this debate, Dr. Anawalt advocates for the scientific validity and clinical value of the free testosterone hypothesis, but Dr. Handelsman argues that the hypothesis is unfounded, nonsensical, and unnecessary for clinical care.
Free Testosterone Hypothesis: True (Dr. Anawalt)
The FT hypothesis posits that the biological activity of testosterone depends principally on the concentration of the unbound (“free”) hormone and not its bound form.1 The hypothesis does not stipulate that serum FT is the sole determinant of its biological activity.1 Thus, the FT hypothesis must be distinguished from the FT transport hypothesis that only the free hormone is transported into a target tissue cell. For the FT hypothesis to be true, healthy individuals with low serum concentrations of testosterone binding protein (and total testosterone)—but normal FT—will have no or minimal phenotype of testosterone deficiency.
Mathematical basis of the FT hypothesis
Mathematical modeling of the FT hypothesis indicates that it is valid for many or most tissues in the body, but not for tissues with high rates of intracellular metabolic elimination of testosterone.1 Thus, the FT hypothesis is likely not true for the liver because 40-50% of the net testosterone delivered to the liver is eliminated by hepatocytes.1,3 9 The FT hypothesis is plausible (and likely) in most tissues (that have much lower intracellular testosterone metabolism) of women and men.1,3
About 40-45% of circulating testosterone is tightly bound by SHBG; about 2% is unbound.1,4 Most of the remainder of circulating testosterone is bound to albumin and other proteins.1,4 The amount of FT delivered to tissues is defined by blood flow, the rates of disassociation from binding proteins, tissue influx , and intracellular metabolism.1 Tissues that are very vascular and have relatively slow blood flow (e.g. hepatic sinusoids) are exposed to greater amounts of FT. SHBG-bound testosterone is “protected” from intrahepatocyte metabolic clearance of testosterone because only ~25% of testosterone that is tightly bound to SHBG disassociates (and enters hepatocytes) during passage through hepatic sinusoids.1 19 Compared to SHBG-bound testosterone, a much greater percentage of testosterone that is weakly bound to albumin disassociates in the hepatic sinusoids and is metabolized.1,2 Because testosterone bound to albumin disassociates more quickly, it has been hypothesized that this “weakly bound” testosterone also contributes to intracellular testosterone concentration and androgenic bioactivity.4,5 23 The amount of testosterone bound to serum albumin does not vary greatly in patients with serum albumin concentrations within the normal range, and typically there is a strong correlation between the serum concentrations of albumin-bound testosterone and FT.6
Laboratory-based studies that support the FT hypothesis
Experimental manipulation of serum testosterone-binding proteins concentrations causes reciprocal changes of FT and androgenic effect.7-12 Normal rats and mice do not produce SHBG, and albumin is the principal testosterone-binding protein. In a study of castrated rats treated with testosterone implants, repeated infusions of SHBG modestly decreased serum FT, but more than doubled total testosterone concentrations compared to animals not infused with SHBG.7 Serum luteinizing hormone concentrations were not suppressed in the SHBG + testosterone-treated rats despite a serum total testosterone greater three-fold higher than the typical threshold for suppression—suggesting that FT and not total testosterone exerts negative feedback on the pituitary gonadotropes. Similarly, addition of human serum with a high SHBG concentration blocks cellular uptake of dihydrotestosterone and its action in a human prostatic cell line.8 Male mice that are transfected with overexpressing human SHBG have markedly elevated serum total testosterone, but they have slightly decreased serum FT, elevated LH (indicated that feedback is mediated by FT—not total testosterone), and a normal male phenotype except decreased weight of the most androgen-sensitive organs (seminal vesicles and levator ani/bulbocavernosal muscles).9
On the other hand, a mutant rat strain that has very low serum albumin concentrations (and no SHBG) has decreased albumin-bound testosterone, smaller testes, but normal free T, and otherwise normal male phenotype including fertility compared to healthy normal rats.10-12 A study demonstrated that in vitro androgen effect is inversely related to the SHBG concentrations in sera from hypogonadal men being treated with testosterone, and androgen bioactivity correlated more strongly with FT than total testosterone.13 Another study demonstrated that novel compounds that competitively inhibit the binding of testosterone to SHBG cause a dose-dependent increase o fandrogen activity (by increasing FT).14
Human studies supporting the FT hypothesis
Epidemiological studies support the FT hypothesis. There are many SHBG polymorphisms in normal, healthy men.15 Although these polymorphisms are associated with significant differences of serum total testosterone, serum FT does not differ significantly (except for an uncommon genotype—rs6259 homozygosity that is associated with higher FT).15 The only reported man with an SHBG missense mutation and no detectable circulating SHBG had a normal physical examination (including normal genitalia and normal-sized testes), very low total testosterone, low normal FT, normal LH, and normal seminal fluid analysis. He had some possible symptoms of hypogonadism including low libido and decreased early morning erections, but his overall eugonadal phenotype corroborated in the absence of SHBG-bound testosterone.16
Two epidemiological studies have reported cross-sectional and longitudinal data for the association between symptoms and outcomes of hypogonadism and serum total testosterone (measured by mass spectrometry), calculated FT (using a validated formula), and SHBG. In the European Male Aging Study (EMAS) of over 3300 community-dwelling men (40-79 years; mean BMI 27-30), men with baseline low FT were significantly more likely to report low libido and decreased erectile function than men with low or normal total testosterone but normal FT; men with low totaltestosterone and normal FT did not differ from those with normal total and FT.17 In the four-year longitudinal follow-up of EMAS, men with low total and FT were more likely to report decreased libido [OR = 2.67 (1.27-5.60)], erectile dysfunction [OR = 4.53 (2.05-10.01)] and/or infrequent morning erections [OR = 3.40 (1.48-7.84)] than men with normal total and FT.18 Men with low total testosterone and normal FT did not differ in reported sexual symptoms from men with normal total and FT. For the Concord Health and Aging in Men Project of ~1700 community-dwelling men (> 704 years old; mean BMI ~ 27-28), the authors have reported that FT offered little additional independent clinical value over the measurement of total testosterone.19 However, similar to EMAS, the men who had concordant low total and FT concentrations had high odds ratios for low libido in the cross-sectional and longitudinal studies: cross-sectional unadjusted: 2.20; CI 1.52-3.28 and after multivariable adjustment 2.11; CI 1.42-3.13; and longitudinal unadjusted 1.41; CI 1.11-9 1.80 and after multivariable adjustment 1.36; CI 1.06-1.74. However, the men with low serum total testosterone and normal FT had non-significant odds ratios for low libido that approximated 1.011 compared to men with normal total and FT. There are several other outcomes (that are less specific to hypogonadism) that have been independently associated with low serum FT; osteoporosis (or low bone density), frailty, and major cardiovascular events.20-25
Overall, the evidence indicates that the FT hypothesis is true. Concordant low total testosterone and FT concentrations buttress the diagnosis of hypogonadism. Importantly, concordance of low total and FT concentrations is most strongly associated with low libido and sexual dysfunction—the clinical outcome that is most specific for hypogonadism and most consistently responsive to testosterone therapy for hypogonadal men.26-28 19 The most common diagnostic dilemma that clinicians face in the biochemical assessment of hypogonadism is the man with a low serum total testosterone due to obesity-mediated suppression of SHBG. In these men, the finding of a normal FT is useful to avoid misclassification (e.g., over-diagnosis in obese men) as hypogonadal and unnecessary lifelong testosterone therapy.28
The Free Testosterone Hypothesis: False (Dr Handelsman)
(The Misguidance of a Mirage: The Many Fictions of Free Testosterone)
Generations of Endocrinology trainees have had the FT hypothesis embedded into clinical endocrinology thinking as intuitively obvious, needing no proof. FT dogma developed a life of its own, as a sophisticated sounding concept regularly invoked to evaluate androgen effects when measured testosterone yielded no meaningful finding. So, this debate is directed towards the proverbial Endocrinologist from Mars whose open mind is focused on evidence, untainted by received wisdom or the straight jacket of dogma. The distinctive instinctive appeal of the FT hypothesis is classically reminiscent of Kahneman’s fast thinking – intuitive, emotional, irrational -compared with slow thinking, which is logical and scientific.29
Flawed concept of the FT hypothesis
The FT hypothesis asserts that the small fraction of non-protein bound testosterone has greater biological activity because it is more accessible to target tissues.4 An immediate problem is that unbound testosterone is also more accessible to sites of its degradation. 16 30 Hence, unbound testosterone may be more, or less, biologically active depending on the balance between its tissuetargets; regardless, no unidirectional interpretation of bioactivity can be deduced from such a shaky implausible hypothesis. Similar reservations apply to son-of-free FT, “bioavailable” testosterone, the sum of unbound plus loosely bound (mainly to albumin) testosterone.
Flawed methodology of FT measurement
FT is inherently a laboratory measurement using dialysis to separate bound from free testosterone.However, dialysis methods are non-robust, laborious and largely unavailable. Instead, formulae estimate lab-measured FT as calculated FT (cFT). However, cFT is not a valid analytical variable as it fails the fundamental principle of analytical chemistry, to compare like with like. The absence (indeed impossibility) of a certified reference standard for FT means there is no sound quality control nor, in turn, any valid pooling or comparing cFT results between different laboratories/centres, making constructing a validated reference range unattainable.
These cFT formulae have built-in flaws. Most (Sodergard, Vermeulen) are based on equilibrium8binding equation theory* although capillary transit time is too short to reach equilibrium. Furthermore, binding theory involves SHBG’s testosterone binding capacity (not its mass*Sometimes amusingly referred to as by the “Law of Mass Action”, a grandiloquence comparable to measuring weight by the Law of Gravity. measured by one of four non-harmonized immunoassays) and there are two binding sites with different testosterone-SHBG affinities.31,32 Nevertheless, binding theory formulae assume a unitary binding affinity with a single “affinity” spanning a five-fold range so formulae employing a unitary, fixed, plug-in binding affinity create systematic formulaic bias diverging from lab-measured FT.33,34 Genetic polymorphisms influence both circulating SHBG concentration, its testosterone binding affinity, or both; these variables are not accounted for in cFT formulae. 35 Ultimately, cFT is not a valid analytical variable, is calculated by erroneous formulae that cannot provide meaningful guidance on androgen status.
Lack of biological validity of the FT hypothesis
A high road to evaluate a postulated pathophysiological mechanism is a genetic mouse knock-out model, which form high-fidelity models for human reproductive pathophysiology.36 However, mature rodents represent a natural SHBG knockout as they have no circulating SHBG, indicating the dispensability of SHBG for reproductive function (including defining FT).
Conversely, a transgenic mouse model over-expressing human SHBG at physiological levels might be an interesting experiment; however, one model expressing SHBG at 200 times the human blood SHBG concentration defies physiological interpretation.9 In the closest analogy, genetic suppression of circulating corticosteroid binding globulin is detrimental to survival from septic shock despite increased free glucocorticoid fraction.37. Direct clinical testing of the FT hypothesis is remarkably scarce for such a ubiquitous go-to concept. One 5-year prospective study in older men using cFT, calculated by either binding theory or assumption-free empirical method, showed that cFT did not enhance prediction of mortality, morbidity or validated clinical outcomes overaccurately measured testosterone by LCMS.19
FT hypothesis: Superfluous for clinical care, fosters testosterone misuse and androgen dependence
Decades of experience in clinical management of testosterone deficiency verifies that appeal to cFT is unnecessary and there is no evidence excluding it from consideration compromises anyone’s health. Indeed, the quasi-epidemic of excessive testosterone prescribing over three recent decades, comprising a 100-fold increase without any new approved indication, indicates a pressingneed to resolve its driving forces, including guidelines that allow cFT as a valid analytical variable.38 These issues converge on a strong need for a return to the classical definition of “hypogonadism” as a pathological disorder of the hypothalamus-pituitary-testicular (HPT) axis. Over recent decades, the term “hypogonadism”, has been hijacked by re-definition to any lowering of testosterone below an arbitrary, inflated “low” threshold, thereby vindicating testosterone prescription for “hypogonadism” without any authentic organic basis. That new definition creates an invented disorder, known by a cavalcade of neologisms seeking medical gravitas (“Andropause”, “late-onset or functional hypogonadism” etc) but better described as non-gonadal illness or sick eugonadal syndrome.28,39 In that, cFT has played a major supporting role by further defining “low” testosterone4 when blood testosterone was still not “low” enough even for the confected, but inflated, threshold. Rather than trying to squeeze more meaning from inaccurate testosterone immunoassays, a better physiological appeal is to using the endogenous androgen biosensors we already have, the cheaply available, serum LH and FSH used in an analogous fashion to TSH in thyroidal disorders.
Lack of explanatory power of the FT hypothesis
Many trained into FT dogma as an Endocrinology Commandment, consider FT a convenient explanation for patients, usually to deter with a scientific flourish, medically unjustified demands for testosterone prescription in a form of noble cause corruption. But when we come to medical science, evidence is required before convenient beliefs gain credence; ultimately, health literate patients will see through the façade of flimsy evidence. Just as there is no free lunch, confected evidence-free explanations do not hold water or nourish us.
What shall we do?
Diagnosis of hypogonadism needs straightening out. With an identifiable pathological cause (structural or genetic) then testosterone replacement therapy may have a sound basis. But otherwise, in the absence of a pathological cause, a low serum testosterone alone requires a proper diagnosis but not usually testosterone treatment. When such usually futile treatment inevitably ceases, it can create androgen dependence even without there ever having been genuine androgen deficiency in the first place. In this situation, iatrogenic androgen deficiency withdrawal symptoms foster a return to testosterone treatment to alleviate the symptoms creating a cycle of androgen dependence.40 Well known in androgen abusers, this is also evident in men without pathological hypogonadism receiving standard testosterone doses for prolonged periods.41
Practical steps are (1) reject the FT hypothesis as unsound conceptually and empirically, (2) greater reliance on clinical evaluation for recognized HPT axis pathology (structural or genetic) with hormonal testing to confirm, not make, the clinical diagnosis, (3) take multiple serum LH and FSH (androgen sensors) and SHBG (testosterone carrier protein) measurements to interpret serum testosterone, and (4) be vigilant about androgen dependence, thinking of androgens like glucocorticoids - powerful hormones but dangerous to (mis)over use without sound cause.
Dr. Anawalt’s Rebuttal:
My original argument shows that the FT hypothesis is a sound concept that has biological validity. Although I agree that the methods of measurement of free testosterone need improvement and standardization, I do not agree that all available methods are invalid. More importantly, this debate is not about the accuracy of free testosterone measurement; it is about the validity of the FT hypothesis. Dr. Handelsman advances three additional arguments that I rebut:
1. “Flawed methodology [FT measurement]”: Genetic polymorphisms do not generally affect the cFT.15 19 The most commonly used cFT formula (Vermeulen) yields results that are reasonably concordant with equilibrium dialysis with tandem mass spectrometry (the gold standard) across a wide range of serum total testosterone concentrations except at very low levels (where there is little debate about the assessment).4,34,42 When using a total testosterone assay that has been harmonized and certified by an external quality controlinstitution, a validated, useful FT reference range is possible.28,42,43
2. “FT… fosters testosterone misuse”: FT is useful to reduce and prevent testosterone misuse and androgen dependence. The worldwide escalation of testosterone prescriptions paralleled a marked global increase of prevalent obesity (resulting in large number of eugonadal men with low serum total testosterone and normal serum FT). Decision-making based on accurate assessment of FT would have mitigated this quasi-epidemic).28
3. “FT [does not have] explanatory power”: Many eugonadal men will want an explanation for a low serum total testosterone. Analogous to finding a normal free thyroxine in a patient with low total thyroxine, telling them that their free (“active”) testosterone is useful. For most of these men, their free testosterone is normal, they are eugonadal, and their total testosterone is low due to obesity-induced suppression of serum SHBG. That explanation motivates some men to make healthy lifestyle changes to lose weight.
Dr. Handelman’s Rebuttal
In rebutting five arguments, Dr Anawalt tackles four, largely ignoring the assertion that the concept lacks analytical measurement validity. He cites multiple studies using cFT to defend its utility without apparent concern about the baseless nature of these arbitrary figures, nor for what to interpret from such conjured numbers. Under the World Antidoping Code, elite athletes seeking testosterone treatment for hypogonadism require a Therapeutic Use Exemption (TUE) to permit usage of a banned substance to continue working in their profession; however, FT results are specifically excluded from consideration because their analytical invalidity make them unlikely to survive crucial scrutiny by the Court of Arbitration for Sport (CAS), the sports Supreme Court. cFT is not an analytical variable at all, not just lacking any certified standard or quality control but constitutively incapable of ever having one, which disallows pooling or comparison between labs/centers or creating reference ranges to evaluate its utility. That makes belief in its unprovable utility a shining exemplar of Kahneman’s fast thinking – intuitive, emotional and irrational – the contrary of slow thinking scientific logic. It is also congruent with an excellent update review of HOW to “measure” FT, without any cognizance of WHY to do it and what it might mean, an issue taken up in a detailed critique.4,30 To paraphrase JM Keynes, practical evidence-wielding4endocrinologists who believe themselves to be exempt from academic or scholarly analysis of the variables they employ are usually unwitting slaves of some defunct theory, distilling their thoughts from obsolete discussions of decades ago.
