madman
Super Moderator
Neson's Excel u heard!
* Approximately 98% of circulating T is protein-bound, predominantly to SHBG and albumin, leaving only 2% to circulate in an unbound form. The free hormone hypothesis asserts that the latter nonprotein-bound T, or free T, rather than total T, represents the fraction most readily accessible to target tissues and therefore more accurately reflects the biological activity of T [1,2]. Robust experimental evidence exists to support the physiological role of free T [2–4], reinforced by human experiments showing the impact of SHBG gene variants and even complete SHBG deficiency (Figure 1) [5,6]. Additionally, the added diagnostic value of cFT, alongside total T, has been demonstrated to provide a more accurate assessment of androgen status in various clinical settings [7–10].
* Incorporating free T assessment into clinical practice is essential to improve diagnostic accuracy, prevent misclassification, and avoid inappropriate treatment in men with suspected hypogonadism.
^ As outlined above, analog-type immunoassays have so far not performed adequately for free T quantification, so any future direct free T method will need to rely on alternative technologies and undergo rigorous validation.
* Surprisingly, the use of cFT has continued to be questioned, with some even advocating its abandonment [23,29]. This summary dismissal of cFT, which some have labeled‘scientifically unsound’ [29], is, in our view, not supported by the available evidence and risks reversing progress toward more physiologically based and comprehensive (free) T assessment. Given that methodological advances now permit more robust estimation, or even direct measurement, of free T, it is no longer acceptable to dismiss a physiologically based metric, such as cFT, as a mere artifact.
* While total T remains the default biochemical entry-level test to evaluate gonadal status in men, international societies and expert guidelines increasingly recognize the importance of assessing free T [28,30–33]. In conditions that affect SHBG levels, or whenever total T concentrations are near the lower limit of the normal range, an assessment of free T, either directly measured after ED or estimated by calculation, is warranted [28].
* Direct comparison of cFT derived from these different formulas against free T measured by ED with LC-MS/MS showed that Vermeulen-derived cFT had a very strong and linear correlation with measured free T (mFT) across a wide range of T, SHBG, and albumin concentrations, albeit with a systematic overestimation of approximately 20–30% [39]. The Södergård equation, an analogous mass-action model, yields cFT values that are highly concordant with Vermeulen-derived cFT values, with only modest differences attributable to different parameter choices [40,41]. In contrast, the Ly equation yielded median cFT values close to mFT, but its agreement was highly dependent on SHBG and T concentrations, leading to increasing underestimation at low SHBG or low T levels, precisely the context of potential male hypogonadism in which accurate free T assessment is most needed [39]. The allosteric Zakharov model, on the other hand, produced substantially higher cFT values than both Vermeulen and Ly formulas and showed pronounced dependence on SHBG [39]. Taken together, these data support the use of Vermeulen-derived cFT as the most pragmatic and physiologically grounded approach in clinical practice, and its widespread adoption in routine clinical laboratories/practice across European centers further underlines this position [42].
* Discrepancies between cFT and mFT are primarily technical rather than conceptual and are reminiscent of the refinement process that improved free T4 assays in the past. Critics, however, have argued that measuring or calculating free T is not a valid analytical variable due to the current lack of certified reference materials, standardized quality control, and reference intervals [29]. Importantly, methodological work is underway to address several of these highlighted gaps. ED combined with LC-MS/MS is currently considered the gold standard for accurate measurement of free T, although its technical complexity and resource demands render it generally inaccessible for everyday clinical practice [37]. Despite these practical limitations, ED–LC-MS/MS-derived age-stratified reference ranges for mFT were recently developed using a large cohort of healthy, and community-dwelling men [43]. Furthermore, access to mFT may be expected to broaden in the future, given the rapid expansion of LC-MS/MS applications (e.g., steroid hormones and drug levels/detection) in clinical chemistry laboratories, the introduction of LC-MS/MS automation [44], and technical advances that simplify ED procedures, such as reductions in dialysis time. Until such advances become widely available, carefully derived cFT remains the most practical way to reliably evaluate the clinically relevant free T fraction in routine endocrine diagnostics.
* However, several methodological criticisms of cFT deserve explicit consideration. Because cFT is mathematically derived from total T and SHBG, it is inherently subject to mathematical coupling and propagates any analytical error in its input measurements. Additionally, residual interlaboratory variation in total T and SHBG, together with the use of different formulas, can still lead to between-center differences in cFT values. These challenges do not invalidate the concept of free T but instead highlight that the path forward should involve further harmonizing different SHBG assays, transparent reporting of the cFT calculation method applied, refining specific cFT algorithms (e.g., Vermeulen), and establishing reference intervals traceable to ED-based mFT [42,43,45].
* For total T values in the intermediate or ‘gray’ zone (e.g., between 8 and 12 nmol/L, as adopted by several guidelines [28,30]), or whenever SHBG levels are likely altered (e.g., in cases of obesity, diabetes, aging, HIV, thyrotoxicosis, estrogen or glucocorticoid therapy, liver or renal disease), a measurement of SHBG and calculation of cFT should be strongly considered. Notably, except when total T levels are below 5.2 nmol/L, total T measurement has low specificity for the biochemical diagnosis of hypogonadism , implying that additional information from SHBG and/or cFT becomes increasingly important above this threshold [24]. Importantly, cFT results should be interpreted against method- and/or laboratory-specific reference ranges to account for local assay variability [42]. If there is discordance between clinical features and laboratory results, testing should initially be repeated, and the use of ED-based mFT considered. From the laboratory perspective, cFT should systematically be reported whenever total T and SHBG are requested, particularly when total T concentrations are borderline (e.g., in the gray zone) or when SHBG is clearly abnormal. Laboratories must explicitly identify the formula used for cFT calculation (e.g., the most commonly used being the Vermeulen method [42]). Reference intervals should ideally be established from validated or locally representative population data, with a traceable link to ED data.
* Future research in the field of T assessment should prioritize multicenter validation of cFT against ED measurements of free T in diverse populations to confirm cFT as a robust and clinically accessible tool (see Outstanding questions).
* Additionally, the growing availability of mFT data after ED offers an opportunity to revise currently available cFT algorithms (such as the Vermeulen equation) (Box 1). By recalibrating these formulas against direct free T measurements, derived free T values will become more accurate, bringing calculated estimates of free T closer to the measured concentrations. Furthermore, expanding access to mass spectrometry- based ED in routine clinical laboratories, by simplifying dialysis procedures and increasing throughput, enhances the capacity for accurate direct free T measurement in clinical diagnostics.
* In parallel, analogous to the historical transition from total to free T4, improved access to reliable mFT—either through broader implementation of simplified ED–LC-MS/MS or through genuinely robust and properly validated automated free T assays standardized against ED—would further enhance clinical applicability and diagnostic precision.
Calculated free testosterone: rightful heir to the free thyroxine index?
Nick Narinx, Leen Antonio, Bruno Lapauw, Frederick C.W. Wu and Dirk Vanderschueren
Revisiting the case for free testosterone
The diagnosis of male hypogonadism still hinges largely on serum total t estosterone (T) (see Glossary), despite growing recognition that a rather small fraction, free T, better reflects biological activity. Accumulating physiological, genetic, and clinical evidence supports the concept of ‘free T’ and shows that calculated free T (cFT) is more closely related to androgen-sensitive outcomes than total T alone, particularly when sex hormone-binding globulin (SHBG) levels are altered. Yet, in contrast to thyroid practice, the routine use of cFT remains contentious, with some guidelines offering only vague advice to‘interpret total T in relation to SHBG’ or questioning the role of cFT altogether.
In this opinion article, we argue that such skepticism is not evidence-based and that cFT is the logical successor to total T, just as the free thyroxine index (FTI) was to total thyroxine (T4), and not a mere ‘artifact’ to be dismissed. We first highlight (pre-) clinical parallels between the T-SHBG and T4-binding protein systems, then discuss current strengths and limitations of free T methodology, and finally propose a pragmatic framework that integrates total T, SHBG, and free T in routine diagnostics. In parallel, we outline a research agenda to further validate cFT, aiming to consolidate it as a robust, clinically accessible tool. Not incorporating free T systematically risks repeating diagnostic errors long resolved in thyroid disease and may contribute to both over- and under-diagnosis of hypogonadism in clinical practice.
Clinical and preclinical parallels between T and T4
Biochemical confirmation of the diagnosis of male hypogonadism remains anchored in the measurement of serum total T, resembling the outdated approach for thyroid disease based on total T4 decades ago. Both circulating total hormone concentrations vary with their respective carrier binding proteins, SHBG and thyroxine-binding globulin (TBG). Thus, as is well accepted in the case of T4, reliance on total hormone concentration alone risks diagnostic misclassification when binding protein levels fluctuate. A parallel exists between the SHBG–T and TBG–T4 paradigms, supported by clinical data, human experiments, and preclinical animal models, as illustrated in Figure 1. By juxtaposing these two axes, Figure 1 depicts the striking similarity between T and T4 transport and hormone action. Approximately 98% of circulating T is protein-bound, predominantly to SHBG and albumin, leaving only 2% to circulate in an unbound form. The free hormone hypothesis asserts that the latter nonprotein-bound T, or free T, rather than total T, represents the fraction most readily accessible to target tissues and therefore more accurately reflects the biological activity of T [1,2]. Robust experimental evidence exists to support the physiological role of free T [2–4], reinforced by human experiments showing the impact of SHBG gene variants and even complete SHBG deficiency (Figure 1) [5,6]. Additionally, the added diagnostic value of cFT, alongside total T, has been demonstrated to provide a more accurate assessment of androgen status in various clinical settings [7–10]. Further support for the free hormone hypothesis comes from other endocrine axes, such as cortisol, where changes in binding proteins modify total but not free hormone [11] and from pharmacological data showing that unbound drug concentrations are the principal drivers of tissue exposure and effect. This is closely analogous to the way dissolved (unbound) oxygen, rather than hemoglobin-bound oxygen, is available for tissue oxygenation [12,13]. Notwithstanding, greater bioavailability may be accompanied by increased metabolic clearance; this effect is counterbalanced by rapid equilibrium with the much larger bound T pool. Thus, circulating free T concentrations are directly representative of tissue androgen exposure independent of SHBG [1,2]. In contrast, variations in SHBG levels (common in obesity, aging, liver disease, thyrotoxicosis, HIV, or drug-induced conditions) can substantially alter total T values without affecting free or tissue T levels [2], a scenario that mirrors the infuence of TBG on total T4.
Historically, the FTI, a calculated metric like cFT, was gainfully used to correct total T4 for TBG fluctuations for many years until reliable free T4 assays became widely available in routine clinical laboratories. FTI correlated well with free T4 and improved diagnostic accuracy over total T4 alone. Indeed, in certain high-TBG conditions, such as pregnancy, the American Thyroid Association guidelines caution against uncritical reliance on direct free T4 assays, noting that FTI better preserves the expected inverse relationship with thyroid-stimulating hormone (TSH) [14,15]. Over time, equilibrium dialysis (ED) methods and analog-based free T4 immunoassays were developed and implemented on automated platforms. Although method-dependent and requiring specific reference intervals, they provided a clinically reliable estimation of the free T4 fraction that correlated well with TSH-based assessment of thyroid status [16,17]. In parallel, liquid chromatography-tandem mass spectrometry (LC-MS/MS) free T4 measurement methods following ED or ultrafiltration were developed as candidate reference procedures to define traceable reference intervals and to benchmark routine assays, thereby consolidating free T4 in routine thyroid diagnostics [18,19]. In contrast, despite attempts to procure a similar analog-based direct immunoassay for free T, they have thus far performed poorly [20,21]. Consequently, advances in T diagnostics have lagged behind those achieved in thyroid hormone assessment [20,22].
The parallel between SHBG/T and TBG/T4 is compelling. No endocrinologist would diagnose hyperthyroidism in pregnancy based solely on elevated total T4, nor attribute low total T4 under androgen therapy to thyroid failure. Similarly, diagnosing hypogonadism based only on low total T in men with low SHBG (common in obesity, type 2 diabetes, or glucocorticoid excess) risks erroneous (over)diagnosis and unnecessary treatment. This is a strong argument in favor of free T assessment, which is overlooked and wrongly criticized by the opponents of cFT [23–25]. Conversely, elevated SHBG levels (encountered in liver disease, hyperthyroidism, estrogen therapy, HIV, or aging) may potentially mask underlying androgen deficiency by inflating (falsely normal) total T despite low free T concentrations [26]. An important difference, however, is that most ‘functional’ thyroid disorders are primary, so that serum levels of TSH provide a highly sensitive upstream marker of thyroid dysfunction [27]. In contrast, male hypogonadism encountered in clinical practice is most often secondary or functional, and, additionally, luteinizing hormone does not act as a simple ‘TSH-like’ integrative marker of androgen status in men treated with T [28]. Consequently, the gonadal axis lacks a robust upstream biomarker comparable to TSH, which increases reliance on accurate measurements of circulating T and strengthens the case for assessing the biologically active T fr action, free T, especially when SHBG concentration is altered.
Surprisingly, the use of cFT has continued to be questioned, with some even advocating its abandonment [23,29]. This summary dismissal of cFT, which some have labeled‘scientifically unsound’ [29], is, in our view, not supported by the available evidence and risks reversing progress toward more physiologically based and comprehensive (free) T assessment. Given that methodological advances now permit more robust estimation, or even direct measurement, of free T, it is no longer acceptable to dismiss a physiologically based metric, such as cFT, as a mere artifact. Therefore, the striking inconsistency between thyroid practice and androgen testing needs to be addressed, and we advocate a more constructive, rational, evidence-based approach to improve the in-tandem use of total and free T, measured and/or calculated, in the diagnosis and management of male hypogonadism
Status of cFT
While total T remains the default biochemical entry-level test to evaluate gonadal status in men, international societies and expert guidelines increasingly recognize the importance of assessing free T [28,30–33]. In conditions that affect SHBG levels, or whenever total T concentrations are near the lower limit of the normal range, an assessment of free T, either directly measured after ED or estimated by calculation, is warranted [28]. However, whereas some guidelines explicitly endorse free T [28,30], others advise only that ‘T levels must be interpreted in relation to SHBG levels’, without specifying practical scales or metrics to do so [34]. By dismissing cFT as‘unscientific’, the latter’s vague guidance leaves clinicians without a clear modus operandi, thereby inviting subjective interpretation rather than evidence-based decision-making. Furthermore, the ongoing criticism confuses technical limitations with conceptual invalidity and continues to recycle the same unsubstantiated arguments to undermine the physiological basis of the free hormone hypothesis [23]. However, methodological imperfections of free T assessment must indeed be acknowledged [35]. Direct measurement of free T by LC-MS/MS after ED currently represents the gold standard method, but it is technically demanding, costly, and requires specialized expertise. Consequently, reliance on cFT has become the pragmatic clinical standard, as it is accessible in virtually any setting with access to immunoassay testing. Several mass-action-based and more complex allosteric or empirical algorithms have been proposed to calculate cFT, including, among others, the Vermeulen, Södergård, Ly, and Zakharov equations [2,36–38]. Direct comparison of cFT derived from these different formulas against free T measured by ED with LC-MS/MS showed that Vermeulen-derived cFT had a very strong and linear correlation with measured free T (mFT) across a wide range of T, SHBG, and albumin concentrations, albeit with a systematic overestimation of approximately 20–30% [39]. The Södergård equation, an analogous mass-action model, yields cFT values that are highly concordant with Vermeulen-derived cFT values, with only modest differences attributable to different parameter choices [40,41]. In contrast, the Ly equation yielded median cFT values close to mFT, but its agreement was highly dependent on SHBG and T concentrations, leading to increasing underestimation at low SHBG or low T levels, precisely the context of potential male hypogonadism in which accurate free T assessment is most needed [39]. The allosteric Zakharov model, on the other hand, produced substantially higher cFT values than both Vermeulen and Ly formulas and showed pronounced dependence on SHBG [39]. Taken together, these data support the use of Vermeulen-derived cFT as the most pragmatic and physiologically grounded approach in clinical practice, and its widespread adoption in routine clinical laboratories/practice across European centers further underlines this position [42].
Discrepancies between cFT and mFT are primarily technical rather than conceptual and are reminiscent of the refinement process that improved free T4 assays in the past. Critics, however, have argued that measuring or calculating free T is not a valid analytical variable due to the current lack of certified reference materials, standardized quality control, and reference intervals [29]. Importantly, methodological work is underway to address several of these highlighted gaps. ED combined with LC-MS/MS is currently considered the gold standard for accurate measurement of free T, although its technical complexity and resource demands render it generally inaccessible for everyday clinical practice [37]. Despite these practical limitations, ED–LC-MS/MS-derived age-stratified reference ranges for mFT were recently developed using a large cohort of healthy, and community-dwelling men [43]. Furthermore, access to mFT may be expected to broaden in the future, given the rapid expansion of LC-MS/MS applications (e.g., steroid hormones and drug levels/detection) in clinical chemistry laboratories, the introduction of LC-MS/MS automation [44], and technical advances that simplify ED procedures, such as reductions in dialysis time. Until such advances become widely available, carefully derived cFT remains the most practical way to reliably evaluate the clinically relevant free T fraction in routine endocrine diagnostics. Concerns raised about interassay variability in SHBG measurement and its impact on cFT, as well as heterogeneity between formulas, are valid but can nevertheless be addressed. Recent work comparing SHBG assays shows that, on average, different SHBG methods produce broadly similar SHBG concentrations and only minor differences in cFT estimates, so assay brand alone is unlikely to drive large shifts in clinical decisions [45]. However, inter-laboratory variation in SHBG measurement and the choice of cFT formula can still substantially influence cFT values and their interpretation, underscoring the need for harmonization and transparent reporting of the equations used [45]. Collectively, these developments underscore that, although substantial progress has been made in free T assessment and further refinement is still required, the essential methodological elements provide a rational foundation for the pragmatic use of cFT in clinical practice.
Toward pragmatic standardization
Labeling cFT as unscientific dismisses decades of biochemical physiology and substantial clinical evidence. Moreover, with the increasing prevalence of obesity (high body mass index is associated with lower SHBG) and the aging population (age-related increase i n SHBG), cFT has become critically important to adjust for alterations in SHBG in the diagnostic workup of possible hypogonadism [28,30–33]. This is particularly relevant in the diagnostic pitfall of pseudo-hypogonadism in obesity, in which proportionate reductions in total T and SHBG (e.g., low total T due to low SHBG) reflect a eugonadal state rather than true hypogonadism, and where incorrectly diagnosing hypogonadism on total T alone may drive inappropriate T prescriptions [46], fueling the epidemic misuse of T [23,29,47,48].
In contrast to these critics’ vague and noncommittal advice [34], a clear, standardized approach based on cFT offers a transparent and systematic way to incorporate SHBG into the diagnostic algorithm and helps prevent misclassifying men with obesity-related low total T but preserved free T as hypogonadal, as well as men with other disease-related increases in SHBG (e.g., HIV or hyperthyroidism). Outright rejection of cFT would reintroduce systematic diagnostic errors reminiscent of the thyroid era before the FTI and free T4 assays became standard measurements. However, several methodological criticisms of cFT deserve explicit consideration. Because cFT is mathematically derived from total T and SHBG, it is inherently subject to mathematical coupling and propagates any analytical error in its input measurements. Additionally, residual interlaboratory variation in total T and SHBG, together with the use of different formulas, can still lead to between-center differences in cFT values. These challenges do not invalidate the concept of free T but instead highlight that the path forward should involve further harmonizing different SHBG assays, transparent reporting of the cFT calculation method applied, refining specific cFT algorithms (e.g., Vermeulen), and establishing reference intervals traceable to ED-based mFT [42,43,45].
Diagnostic workflow combining total and free T
We propose a rational and pragmatic approach that integrates clinical practice with a focused research agenda. In accordance with current guidelines, serum total T remains the first-line biochemical test to evaluate suspected hypogonadism. Clinicians should procure two independent morning samples in a fasted state for total T measurement [28]. Total T levels above 12 nmol/L can be considered normal, whereas levels below 8 nmol/L warrant further investig ation to exclude organic hypogonadism [28]. For total T values in the intermediate or ‘gray’ zone (e.g., between 8 and 12 nmol/L, as adopted by several guidelines [28,30]), or whenever SHBG levels are likely altered (e.g., in cases of obesity, diabetes, aging, HIV, thyrotoxicosis, estrogen or glucocorticoid therapy, liver or renal disease), a measurement of SHBG and calculation of cFT should be strongly considered. Notably, except when total T levels are below 5.2 nmol/L, total T measurement has low specificity for the biochemical diagnosis of hypogonadism , implying that additional information from SHBG and/or cFT becomes increasingly important above this threshold [24]. Importantly, cFT results should be interpreted against method- and/or laboratory-specific reference ranges to account for local assay variability [42]. If there is discordance between clinical features and laboratory results, testing should initially be repeated, and the use of ED-based mFT considered. From the laboratory perspective, cFT should systematically be reported whenever total T and SHBG are requested, particularly when total T concentrations are borderline (e.g., in the gray zone) or when SHBG is clearly abnormal. Laboratories must explicitly identify the formula used for cFT calculation (e.g., the most commonly used being the Vermeulen method [42]). Reference intervals should ideally be established from validated or locally representative population data, with a traceable link to ED data. In parallel, research efforts should identify specific populations and clinical contexts in which cFT performs suboptimally, to delineate where direct free T measurement or alternative approaches are required. To ensure the reliability of results, laboratories should be encouraged to participate in external quality assessment programs and initiatives aimed at the further harmonization of SHBG assays. Importantly, while the process described above aims for reliable and precise evaluation of biochemical T status, it is crucial to remember that the diagnosis of hypogonadism should only be established in the presence of suggestive clinical features, and not based on T measurements alone. Thus, in elderly men who may have nonspecific symptoms suggestive of androgen deficiency and a high prevalence of low T compared to younger men [10], the putative diagnosis of hypogonadism does not, by definition, equate to an indication for T therapy. Given that low free T becomes increasingly common with advancing age, age-specific reference ranges may ultimately improve the classiification of androgen status; however, such approaches remain investigational at present [43].
Concluding remarks and future perspectives
Future research in the field of T assessment should prioritize multicenter validation of cFT against ED measurements of free T in diverse populations to confirm cFT as a robust and clinically accessible tool (see Outstanding questions). Importantly, the clinical utility of free T assessment is not limited to men; in women, free T provides clear added value for the diagnosis of hyperandrogenism and polyendocrine metabolic ovarian syndrome, where cFT offers practical advantages given the analytical challenges of measuring very low free T concentrations [49,50]. Additionally, the growing availability of mFT data after ED offers an opportunity to revise currently available cFT algorithms (such as the Vermeulen equation) (Box 1). By recalibrating these formulas against direct free T measurements, derived free T values will become more accurate, bringing calculated estimates of free T closer to the measured concentrations. Furthermore, expanding access to mass spectrometry- based ED in routine clinical laboratories, by simplifying dialysis procedures and increasing throughput, enhances the capacity for accurate direct free T measurement in clinical diagnostics. In parallel, analogous to the historical transition from total to free T4, improved access to reliable mFT—either through broader implementation of simplified ED–LC-MS/MS or through genuinely robust and properly validated automated free T assays standardized against ED—would further enhance clinical applicability and diagnostic precision. As outlined above, analog-type immunoassays have so far not performed adequately for free T quantification, so any future direct free T method will need to rely on alternative technologies and undergo rigorous validation. This constitutes a clear call to action for diagnostic manufacturers and represents a key research goal as well as a clinical priority to improve the diagnostic accuracy of androgen status assessments in men.
In conclusion, T assessment today stands at a crossroads, strikingly similar to that of T4 measurement decades ago. Just as reliance on total T4 alone proved inaccurate without accounting for binding protein variation—leading to the adoption of the FTI and subsequently free T4 assays as the diagnostic standard, so too must T evaluation now embrace free T assessment. cFT, grounded in clear biochemical principles and validated against ED, offers a nuanced, clinically relevant, and readily accessible metric of androgen status that total T alone cannot provide. Dismissing free T disregards this compelling parallel and risks repeating the same diagnostic errors that were overcome in thyroid practice before the adoption of free T4 assays. Incorporating free T assessment into clinical practice is essential to improve diagnostic accuracy, prevent misclassification, and avoid inappropriate treatment in men with suspected hypogonadism. Ultimately, the imperative to enhance patient care should outweigh any tenuous theoretical objections.
* Approximately 98% of circulating T is protein-bound, predominantly to SHBG and albumin, leaving only 2% to circulate in an unbound form. The free hormone hypothesis asserts that the latter nonprotein-bound T, or free T, rather than total T, represents the fraction most readily accessible to target tissues and therefore more accurately reflects the biological activity of T [1,2]. Robust experimental evidence exists to support the physiological role of free T [2–4], reinforced by human experiments showing the impact of SHBG gene variants and even complete SHBG deficiency (Figure 1) [5,6]. Additionally, the added diagnostic value of cFT, alongside total T, has been demonstrated to provide a more accurate assessment of androgen status in various clinical settings [7–10].
* Incorporating free T assessment into clinical practice is essential to improve diagnostic accuracy, prevent misclassification, and avoid inappropriate treatment in men with suspected hypogonadism.
^ As outlined above, analog-type immunoassays have so far not performed adequately for free T quantification, so any future direct free T method will need to rely on alternative technologies and undergo rigorous validation.
* Surprisingly, the use of cFT has continued to be questioned, with some even advocating its abandonment [23,29]. This summary dismissal of cFT, which some have labeled‘scientifically unsound’ [29], is, in our view, not supported by the available evidence and risks reversing progress toward more physiologically based and comprehensive (free) T assessment. Given that methodological advances now permit more robust estimation, or even direct measurement, of free T, it is no longer acceptable to dismiss a physiologically based metric, such as cFT, as a mere artifact.
* While total T remains the default biochemical entry-level test to evaluate gonadal status in men, international societies and expert guidelines increasingly recognize the importance of assessing free T [28,30–33]. In conditions that affect SHBG levels, or whenever total T concentrations are near the lower limit of the normal range, an assessment of free T, either directly measured after ED or estimated by calculation, is warranted [28].
* Direct comparison of cFT derived from these different formulas against free T measured by ED with LC-MS/MS showed that Vermeulen-derived cFT had a very strong and linear correlation with measured free T (mFT) across a wide range of T, SHBG, and albumin concentrations, albeit with a systematic overestimation of approximately 20–30% [39]. The Södergård equation, an analogous mass-action model, yields cFT values that are highly concordant with Vermeulen-derived cFT values, with only modest differences attributable to different parameter choices [40,41]. In contrast, the Ly equation yielded median cFT values close to mFT, but its agreement was highly dependent on SHBG and T concentrations, leading to increasing underestimation at low SHBG or low T levels, precisely the context of potential male hypogonadism in which accurate free T assessment is most needed [39]. The allosteric Zakharov model, on the other hand, produced substantially higher cFT values than both Vermeulen and Ly formulas and showed pronounced dependence on SHBG [39]. Taken together, these data support the use of Vermeulen-derived cFT as the most pragmatic and physiologically grounded approach in clinical practice, and its widespread adoption in routine clinical laboratories/practice across European centers further underlines this position [42].
* Discrepancies between cFT and mFT are primarily technical rather than conceptual and are reminiscent of the refinement process that improved free T4 assays in the past. Critics, however, have argued that measuring or calculating free T is not a valid analytical variable due to the current lack of certified reference materials, standardized quality control, and reference intervals [29]. Importantly, methodological work is underway to address several of these highlighted gaps. ED combined with LC-MS/MS is currently considered the gold standard for accurate measurement of free T, although its technical complexity and resource demands render it generally inaccessible for everyday clinical practice [37]. Despite these practical limitations, ED–LC-MS/MS-derived age-stratified reference ranges for mFT were recently developed using a large cohort of healthy, and community-dwelling men [43]. Furthermore, access to mFT may be expected to broaden in the future, given the rapid expansion of LC-MS/MS applications (e.g., steroid hormones and drug levels/detection) in clinical chemistry laboratories, the introduction of LC-MS/MS automation [44], and technical advances that simplify ED procedures, such as reductions in dialysis time. Until such advances become widely available, carefully derived cFT remains the most practical way to reliably evaluate the clinically relevant free T fraction in routine endocrine diagnostics.
* However, several methodological criticisms of cFT deserve explicit consideration. Because cFT is mathematically derived from total T and SHBG, it is inherently subject to mathematical coupling and propagates any analytical error in its input measurements. Additionally, residual interlaboratory variation in total T and SHBG, together with the use of different formulas, can still lead to between-center differences in cFT values. These challenges do not invalidate the concept of free T but instead highlight that the path forward should involve further harmonizing different SHBG assays, transparent reporting of the cFT calculation method applied, refining specific cFT algorithms (e.g., Vermeulen), and establishing reference intervals traceable to ED-based mFT [42,43,45].
* For total T values in the intermediate or ‘gray’ zone (e.g., between 8 and 12 nmol/L, as adopted by several guidelines [28,30]), or whenever SHBG levels are likely altered (e.g., in cases of obesity, diabetes, aging, HIV, thyrotoxicosis, estrogen or glucocorticoid therapy, liver or renal disease), a measurement of SHBG and calculation of cFT should be strongly considered. Notably, except when total T levels are below 5.2 nmol/L, total T measurement has low specificity for the biochemical diagnosis of hypogonadism , implying that additional information from SHBG and/or cFT becomes increasingly important above this threshold [24]. Importantly, cFT results should be interpreted against method- and/or laboratory-specific reference ranges to account for local assay variability [42]. If there is discordance between clinical features and laboratory results, testing should initially be repeated, and the use of ED-based mFT considered. From the laboratory perspective, cFT should systematically be reported whenever total T and SHBG are requested, particularly when total T concentrations are borderline (e.g., in the gray zone) or when SHBG is clearly abnormal. Laboratories must explicitly identify the formula used for cFT calculation (e.g., the most commonly used being the Vermeulen method [42]). Reference intervals should ideally be established from validated or locally representative population data, with a traceable link to ED data.
* Future research in the field of T assessment should prioritize multicenter validation of cFT against ED measurements of free T in diverse populations to confirm cFT as a robust and clinically accessible tool (see Outstanding questions).
* Additionally, the growing availability of mFT data after ED offers an opportunity to revise currently available cFT algorithms (such as the Vermeulen equation) (Box 1). By recalibrating these formulas against direct free T measurements, derived free T values will become more accurate, bringing calculated estimates of free T closer to the measured concentrations. Furthermore, expanding access to mass spectrometry- based ED in routine clinical laboratories, by simplifying dialysis procedures and increasing throughput, enhances the capacity for accurate direct free T measurement in clinical diagnostics.
* In parallel, analogous to the historical transition from total to free T4, improved access to reliable mFT—either through broader implementation of simplified ED–LC-MS/MS or through genuinely robust and properly validated automated free T assays standardized against ED—would further enhance clinical applicability and diagnostic precision.
Calculated free testosterone: rightful heir to the free thyroxine index?
Nick Narinx, Leen Antonio, Bruno Lapauw, Frederick C.W. Wu and Dirk Vanderschueren
Revisiting the case for free testosterone
The diagnosis of male hypogonadism still hinges largely on serum total t estosterone (T) (see Glossary), despite growing recognition that a rather small fraction, free T, better reflects biological activity. Accumulating physiological, genetic, and clinical evidence supports the concept of ‘free T’ and shows that calculated free T (cFT) is more closely related to androgen-sensitive outcomes than total T alone, particularly when sex hormone-binding globulin (SHBG) levels are altered. Yet, in contrast to thyroid practice, the routine use of cFT remains contentious, with some guidelines offering only vague advice to‘interpret total T in relation to SHBG’ or questioning the role of cFT altogether.
In this opinion article, we argue that such skepticism is not evidence-based and that cFT is the logical successor to total T, just as the free thyroxine index (FTI) was to total thyroxine (T4), and not a mere ‘artifact’ to be dismissed. We first highlight (pre-) clinical parallels between the T-SHBG and T4-binding protein systems, then discuss current strengths and limitations of free T methodology, and finally propose a pragmatic framework that integrates total T, SHBG, and free T in routine diagnostics. In parallel, we outline a research agenda to further validate cFT, aiming to consolidate it as a robust, clinically accessible tool. Not incorporating free T systematically risks repeating diagnostic errors long resolved in thyroid disease and may contribute to both over- and under-diagnosis of hypogonadism in clinical practice.
Clinical and preclinical parallels between T and T4
Biochemical confirmation of the diagnosis of male hypogonadism remains anchored in the measurement of serum total T, resembling the outdated approach for thyroid disease based on total T4 decades ago. Both circulating total hormone concentrations vary with their respective carrier binding proteins, SHBG and thyroxine-binding globulin (TBG). Thus, as is well accepted in the case of T4, reliance on total hormone concentration alone risks diagnostic misclassification when binding protein levels fluctuate. A parallel exists between the SHBG–T and TBG–T4 paradigms, supported by clinical data, human experiments, and preclinical animal models, as illustrated in Figure 1. By juxtaposing these two axes, Figure 1 depicts the striking similarity between T and T4 transport and hormone action. Approximately 98% of circulating T is protein-bound, predominantly to SHBG and albumin, leaving only 2% to circulate in an unbound form. The free hormone hypothesis asserts that the latter nonprotein-bound T, or free T, rather than total T, represents the fraction most readily accessible to target tissues and therefore more accurately reflects the biological activity of T [1,2]. Robust experimental evidence exists to support the physiological role of free T [2–4], reinforced by human experiments showing the impact of SHBG gene variants and even complete SHBG deficiency (Figure 1) [5,6]. Additionally, the added diagnostic value of cFT, alongside total T, has been demonstrated to provide a more accurate assessment of androgen status in various clinical settings [7–10]. Further support for the free hormone hypothesis comes from other endocrine axes, such as cortisol, where changes in binding proteins modify total but not free hormone [11] and from pharmacological data showing that unbound drug concentrations are the principal drivers of tissue exposure and effect. This is closely analogous to the way dissolved (unbound) oxygen, rather than hemoglobin-bound oxygen, is available for tissue oxygenation [12,13]. Notwithstanding, greater bioavailability may be accompanied by increased metabolic clearance; this effect is counterbalanced by rapid equilibrium with the much larger bound T pool. Thus, circulating free T concentrations are directly representative of tissue androgen exposure independent of SHBG [1,2]. In contrast, variations in SHBG levels (common in obesity, aging, liver disease, thyrotoxicosis, HIV, or drug-induced conditions) can substantially alter total T values without affecting free or tissue T levels [2], a scenario that mirrors the infuence of TBG on total T4.
Historically, the FTI, a calculated metric like cFT, was gainfully used to correct total T4 for TBG fluctuations for many years until reliable free T4 assays became widely available in routine clinical laboratories. FTI correlated well with free T4 and improved diagnostic accuracy over total T4 alone. Indeed, in certain high-TBG conditions, such as pregnancy, the American Thyroid Association guidelines caution against uncritical reliance on direct free T4 assays, noting that FTI better preserves the expected inverse relationship with thyroid-stimulating hormone (TSH) [14,15]. Over time, equilibrium dialysis (ED) methods and analog-based free T4 immunoassays were developed and implemented on automated platforms. Although method-dependent and requiring specific reference intervals, they provided a clinically reliable estimation of the free T4 fraction that correlated well with TSH-based assessment of thyroid status [16,17]. In parallel, liquid chromatography-tandem mass spectrometry (LC-MS/MS) free T4 measurement methods following ED or ultrafiltration were developed as candidate reference procedures to define traceable reference intervals and to benchmark routine assays, thereby consolidating free T4 in routine thyroid diagnostics [18,19]. In contrast, despite attempts to procure a similar analog-based direct immunoassay for free T, they have thus far performed poorly [20,21]. Consequently, advances in T diagnostics have lagged behind those achieved in thyroid hormone assessment [20,22].
The parallel between SHBG/T and TBG/T4 is compelling. No endocrinologist would diagnose hyperthyroidism in pregnancy based solely on elevated total T4, nor attribute low total T4 under androgen therapy to thyroid failure. Similarly, diagnosing hypogonadism based only on low total T in men with low SHBG (common in obesity, type 2 diabetes, or glucocorticoid excess) risks erroneous (over)diagnosis and unnecessary treatment. This is a strong argument in favor of free T assessment, which is overlooked and wrongly criticized by the opponents of cFT [23–25]. Conversely, elevated SHBG levels (encountered in liver disease, hyperthyroidism, estrogen therapy, HIV, or aging) may potentially mask underlying androgen deficiency by inflating (falsely normal) total T despite low free T concentrations [26]. An important difference, however, is that most ‘functional’ thyroid disorders are primary, so that serum levels of TSH provide a highly sensitive upstream marker of thyroid dysfunction [27]. In contrast, male hypogonadism encountered in clinical practice is most often secondary or functional, and, additionally, luteinizing hormone does not act as a simple ‘TSH-like’ integrative marker of androgen status in men treated with T [28]. Consequently, the gonadal axis lacks a robust upstream biomarker comparable to TSH, which increases reliance on accurate measurements of circulating T and strengthens the case for assessing the biologically active T fr action, free T, especially when SHBG concentration is altered.
Surprisingly, the use of cFT has continued to be questioned, with some even advocating its abandonment [23,29]. This summary dismissal of cFT, which some have labeled‘scientifically unsound’ [29], is, in our view, not supported by the available evidence and risks reversing progress toward more physiologically based and comprehensive (free) T assessment. Given that methodological advances now permit more robust estimation, or even direct measurement, of free T, it is no longer acceptable to dismiss a physiologically based metric, such as cFT, as a mere artifact. Therefore, the striking inconsistency between thyroid practice and androgen testing needs to be addressed, and we advocate a more constructive, rational, evidence-based approach to improve the in-tandem use of total and free T, measured and/or calculated, in the diagnosis and management of male hypogonadism
Status of cFT
While total T remains the default biochemical entry-level test to evaluate gonadal status in men, international societies and expert guidelines increasingly recognize the importance of assessing free T [28,30–33]. In conditions that affect SHBG levels, or whenever total T concentrations are near the lower limit of the normal range, an assessment of free T, either directly measured after ED or estimated by calculation, is warranted [28]. However, whereas some guidelines explicitly endorse free T [28,30], others advise only that ‘T levels must be interpreted in relation to SHBG levels’, without specifying practical scales or metrics to do so [34]. By dismissing cFT as‘unscientific’, the latter’s vague guidance leaves clinicians without a clear modus operandi, thereby inviting subjective interpretation rather than evidence-based decision-making. Furthermore, the ongoing criticism confuses technical limitations with conceptual invalidity and continues to recycle the same unsubstantiated arguments to undermine the physiological basis of the free hormone hypothesis [23]. However, methodological imperfections of free T assessment must indeed be acknowledged [35]. Direct measurement of free T by LC-MS/MS after ED currently represents the gold standard method, but it is technically demanding, costly, and requires specialized expertise. Consequently, reliance on cFT has become the pragmatic clinical standard, as it is accessible in virtually any setting with access to immunoassay testing. Several mass-action-based and more complex allosteric or empirical algorithms have been proposed to calculate cFT, including, among others, the Vermeulen, Södergård, Ly, and Zakharov equations [2,36–38]. Direct comparison of cFT derived from these different formulas against free T measured by ED with LC-MS/MS showed that Vermeulen-derived cFT had a very strong and linear correlation with measured free T (mFT) across a wide range of T, SHBG, and albumin concentrations, albeit with a systematic overestimation of approximately 20–30% [39]. The Södergård equation, an analogous mass-action model, yields cFT values that are highly concordant with Vermeulen-derived cFT values, with only modest differences attributable to different parameter choices [40,41]. In contrast, the Ly equation yielded median cFT values close to mFT, but its agreement was highly dependent on SHBG and T concentrations, leading to increasing underestimation at low SHBG or low T levels, precisely the context of potential male hypogonadism in which accurate free T assessment is most needed [39]. The allosteric Zakharov model, on the other hand, produced substantially higher cFT values than both Vermeulen and Ly formulas and showed pronounced dependence on SHBG [39]. Taken together, these data support the use of Vermeulen-derived cFT as the most pragmatic and physiologically grounded approach in clinical practice, and its widespread adoption in routine clinical laboratories/practice across European centers further underlines this position [42].
Discrepancies between cFT and mFT are primarily technical rather than conceptual and are reminiscent of the refinement process that improved free T4 assays in the past. Critics, however, have argued that measuring or calculating free T is not a valid analytical variable due to the current lack of certified reference materials, standardized quality control, and reference intervals [29]. Importantly, methodological work is underway to address several of these highlighted gaps. ED combined with LC-MS/MS is currently considered the gold standard for accurate measurement of free T, although its technical complexity and resource demands render it generally inaccessible for everyday clinical practice [37]. Despite these practical limitations, ED–LC-MS/MS-derived age-stratified reference ranges for mFT were recently developed using a large cohort of healthy, and community-dwelling men [43]. Furthermore, access to mFT may be expected to broaden in the future, given the rapid expansion of LC-MS/MS applications (e.g., steroid hormones and drug levels/detection) in clinical chemistry laboratories, the introduction of LC-MS/MS automation [44], and technical advances that simplify ED procedures, such as reductions in dialysis time. Until such advances become widely available, carefully derived cFT remains the most practical way to reliably evaluate the clinically relevant free T fraction in routine endocrine diagnostics. Concerns raised about interassay variability in SHBG measurement and its impact on cFT, as well as heterogeneity between formulas, are valid but can nevertheless be addressed. Recent work comparing SHBG assays shows that, on average, different SHBG methods produce broadly similar SHBG concentrations and only minor differences in cFT estimates, so assay brand alone is unlikely to drive large shifts in clinical decisions [45]. However, inter-laboratory variation in SHBG measurement and the choice of cFT formula can still substantially influence cFT values and their interpretation, underscoring the need for harmonization and transparent reporting of the equations used [45]. Collectively, these developments underscore that, although substantial progress has been made in free T assessment and further refinement is still required, the essential methodological elements provide a rational foundation for the pragmatic use of cFT in clinical practice.
Toward pragmatic standardization
Labeling cFT as unscientific dismisses decades of biochemical physiology and substantial clinical evidence. Moreover, with the increasing prevalence of obesity (high body mass index is associated with lower SHBG) and the aging population (age-related increase i n SHBG), cFT has become critically important to adjust for alterations in SHBG in the diagnostic workup of possible hypogonadism [28,30–33]. This is particularly relevant in the diagnostic pitfall of pseudo-hypogonadism in obesity, in which proportionate reductions in total T and SHBG (e.g., low total T due to low SHBG) reflect a eugonadal state rather than true hypogonadism, and where incorrectly diagnosing hypogonadism on total T alone may drive inappropriate T prescriptions [46], fueling the epidemic misuse of T [23,29,47,48].
In contrast to these critics’ vague and noncommittal advice [34], a clear, standardized approach based on cFT offers a transparent and systematic way to incorporate SHBG into the diagnostic algorithm and helps prevent misclassifying men with obesity-related low total T but preserved free T as hypogonadal, as well as men with other disease-related increases in SHBG (e.g., HIV or hyperthyroidism). Outright rejection of cFT would reintroduce systematic diagnostic errors reminiscent of the thyroid era before the FTI and free T4 assays became standard measurements. However, several methodological criticisms of cFT deserve explicit consideration. Because cFT is mathematically derived from total T and SHBG, it is inherently subject to mathematical coupling and propagates any analytical error in its input measurements. Additionally, residual interlaboratory variation in total T and SHBG, together with the use of different formulas, can still lead to between-center differences in cFT values. These challenges do not invalidate the concept of free T but instead highlight that the path forward should involve further harmonizing different SHBG assays, transparent reporting of the cFT calculation method applied, refining specific cFT algorithms (e.g., Vermeulen), and establishing reference intervals traceable to ED-based mFT [42,43,45].
Diagnostic workflow combining total and free T
We propose a rational and pragmatic approach that integrates clinical practice with a focused research agenda. In accordance with current guidelines, serum total T remains the first-line biochemical test to evaluate suspected hypogonadism. Clinicians should procure two independent morning samples in a fasted state for total T measurement [28]. Total T levels above 12 nmol/L can be considered normal, whereas levels below 8 nmol/L warrant further investig ation to exclude organic hypogonadism [28]. For total T values in the intermediate or ‘gray’ zone (e.g., between 8 and 12 nmol/L, as adopted by several guidelines [28,30]), or whenever SHBG levels are likely altered (e.g., in cases of obesity, diabetes, aging, HIV, thyrotoxicosis, estrogen or glucocorticoid therapy, liver or renal disease), a measurement of SHBG and calculation of cFT should be strongly considered. Notably, except when total T levels are below 5.2 nmol/L, total T measurement has low specificity for the biochemical diagnosis of hypogonadism , implying that additional information from SHBG and/or cFT becomes increasingly important above this threshold [24]. Importantly, cFT results should be interpreted against method- and/or laboratory-specific reference ranges to account for local assay variability [42]. If there is discordance between clinical features and laboratory results, testing should initially be repeated, and the use of ED-based mFT considered. From the laboratory perspective, cFT should systematically be reported whenever total T and SHBG are requested, particularly when total T concentrations are borderline (e.g., in the gray zone) or when SHBG is clearly abnormal. Laboratories must explicitly identify the formula used for cFT calculation (e.g., the most commonly used being the Vermeulen method [42]). Reference intervals should ideally be established from validated or locally representative population data, with a traceable link to ED data. In parallel, research efforts should identify specific populations and clinical contexts in which cFT performs suboptimally, to delineate where direct free T measurement or alternative approaches are required. To ensure the reliability of results, laboratories should be encouraged to participate in external quality assessment programs and initiatives aimed at the further harmonization of SHBG assays. Importantly, while the process described above aims for reliable and precise evaluation of biochemical T status, it is crucial to remember that the diagnosis of hypogonadism should only be established in the presence of suggestive clinical features, and not based on T measurements alone. Thus, in elderly men who may have nonspecific symptoms suggestive of androgen deficiency and a high prevalence of low T compared to younger men [10], the putative diagnosis of hypogonadism does not, by definition, equate to an indication for T therapy. Given that low free T becomes increasingly common with advancing age, age-specific reference ranges may ultimately improve the classiification of androgen status; however, such approaches remain investigational at present [43].
Concluding remarks and future perspectives
Future research in the field of T assessment should prioritize multicenter validation of cFT against ED measurements of free T in diverse populations to confirm cFT as a robust and clinically accessible tool (see Outstanding questions). Importantly, the clinical utility of free T assessment is not limited to men; in women, free T provides clear added value for the diagnosis of hyperandrogenism and polyendocrine metabolic ovarian syndrome, where cFT offers practical advantages given the analytical challenges of measuring very low free T concentrations [49,50]. Additionally, the growing availability of mFT data after ED offers an opportunity to revise currently available cFT algorithms (such as the Vermeulen equation) (Box 1). By recalibrating these formulas against direct free T measurements, derived free T values will become more accurate, bringing calculated estimates of free T closer to the measured concentrations. Furthermore, expanding access to mass spectrometry- based ED in routine clinical laboratories, by simplifying dialysis procedures and increasing throughput, enhances the capacity for accurate direct free T measurement in clinical diagnostics. In parallel, analogous to the historical transition from total to free T4, improved access to reliable mFT—either through broader implementation of simplified ED–LC-MS/MS or through genuinely robust and properly validated automated free T assays standardized against ED—would further enhance clinical applicability and diagnostic precision. As outlined above, analog-type immunoassays have so far not performed adequately for free T quantification, so any future direct free T method will need to rely on alternative technologies and undergo rigorous validation. This constitutes a clear call to action for diagnostic manufacturers and represents a key research goal as well as a clinical priority to improve the diagnostic accuracy of androgen status assessments in men.
In conclusion, T assessment today stands at a crossroads, strikingly similar to that of T4 measurement decades ago. Just as reliance on total T4 alone proved inaccurate without accounting for binding protein variation—leading to the adoption of the FTI and subsequently free T4 assays as the diagnostic standard, so too must T evaluation now embrace free T assessment. cFT, grounded in clear biochemical principles and validated against ED, offers a nuanced, clinically relevant, and readily accessible metric of androgen status that total T alone cannot provide. Dismissing free T disregards this compelling parallel and risks repeating the same diagnostic errors that were overcome in thyroid practice before the adoption of free T4 assays. Incorporating free T assessment into clinical practice is essential to improve diagnostic accuracy, prevent misclassification, and avoid inappropriate treatment in men with suspected hypogonadism. Ultimately, the imperative to enhance patient care should outweigh any tenuous theoretical objections.
