madman
Super Moderator
Again where is that natty male FT 30-60 ng/dL and we are talking a daily short-lived peak here LMFAO!
Healthy men 18 to 29 years old hitting a high peak FT 25.3 ng/dL are those outliers that fall in the 95th percentile!
Better yet look at where the median FT 12 ng/dL sits LOL!
* Serum from 20 healthy male blood donors (non-fasting) was analyzed to investigate free hormone concentrations and %FTe in healthy males (Table 3). FT3 ranged from 4.5 to 6.7 pM (median 5.5 pM; Fig 5A) and FT4 from 13.8 to 24.1 pM (median 18.6 pM; Fig 5B). FTe ranged from 155 to 368 pM (median 251 pM; Fig 5C) and the calculated %FTe ranged from 1.1% to 2.4%, with a median of 1.6 % (Fig 5D).
* Mass spectrometry (MS)-based methods following equilibrium dialysis (ED) are considered “gold standard” for free hormone quantification.
* Interpretation of studies is further complicated by methodological limitations, since free testosterone (FTe) - the biologically active fraction - is difficult to measure reliably, and reliance on total testosterone (TTe) or calculated values for free testosterone (cFTe) may misclassify androgen status.
* Mass spectrometry-based methods that directly quantitate the free concentration of hormones are widely regarded as the most accurate approach for measuring free hormones and least affected by interferences from patient-derived substances (e.g., medications, supplements, or endogenous antibodies) or binding protein variants. Physical separation of free from bound hormones in serum using equilibrium dialysis (ED; or less ideally ultrafiltration) performed at 37 °C followed by isotope-dilution liquid chromatography tandem mass spectrometry (LC-MS/MS) or gas chromatography (GC)-MS/MS quantitation is generally considered “gold standard” and is used for reference methods for FT3, FT4 and FTe.(13), (14) However, these methods are technically more complex to perform and costly compared to immunoassays or calculations and are today largely restricted to highly specialized laboratories.
* Measured FTe correlated well with cFTe(V) (r = 0.93) and PaBa-regression indicated a linear relationship (Fig 4B, see S3A for residuals). However, cFTe(V) overestimated FTe compared to measured FTe (Figs 4B and S3B), as also reported by others.(36), (37)
Abstract
Background
Testosterone (Te) and the thyroid hormones (THs) thyroxine (T4) and triiodothyronine (T3) are key endocrine hormones regulating development, metabolism and reproductive function. The free (not protein-bound) hormone concentrations more accurately reflect biological activity than total levels. Free hormone measurements are therefore used for clinical decision making in many contexts. Mass spectrometry (MS)-based methods following equilibrium dialysis (ED) are considered “gold standard” for free hormone quantification. Unfortunately, few clinical laboratories offer such measurements today.
Results
A 96-well format ED-LC-MS/MS method incorporating isotope-dilution for simultaneous quantitation of free Te, free T4 and free T3 in clinical samples that used offline (parallel) sample-preparation was successfully developed and validated. The method covered clinically relevant ranges, and clinical samples with a wide concentration range were analyzed to perform a comparison to existing (immunoassay or calculation) methods. Free hormone concentration in 20 healthy blood donors is reported.
Significance
This is the first method to directly quantitate both free testosterone and free thyroid hormones in clinical samples. It can be implemented alongside existing LC-MS/MS methods in specialized clinical laboratories to support improved endocrine assessment and clinical decision making.
Introduction
Here, we report an ED-LC-MS/MS method incorporating isotope-dilution for the concurrent quantitation of FTe, FT4 and FT3 in human samples, developed to be suitable for specialized clinical laboratories: Serum (or plasma) is dialyzed at 37 °C, the free hormones extracted by solid phase extraction (SPE) and then quantitated by analytical flow LC-MS/MS. The method uses 96-well format, avoids derivatization and employs electrospray ionization (ESI+). The method was extensively validated including comparison to the clinical analyzers Abbott Alinity (serum; FT4 and FT3) and Roche Cobas (lithium heparin plasma; FT4 and FT3) or to calculated FTe across a wide concentration range. The method covered clinically relevant concentration ranges for both males and females.
Routine clinical analyses (FT4, FT3, SHBG and TTe)
The immunological FT4 and FT3 assays were performed on an Alinity i instrument (Abbott Diagnostics; reagent kits: Free T4 and Free T3) at the Hormone Laboratory (Oslo University Hospital, Oslo, Norway) or on a Cobas 8000 instrument (Roche Diagnostics; kits Elecsys FT4 III and Elecsys FT3 III) at Vejle Hospital. SHBG concentration was measured on an Immulite 2000xpi system (Siemens Healthineers, Kit IMMULITE 2000 SHBG) at the Hormone Laboratory.
TTe was measured by LC-MS/MS at the Hormone Laboratory using a previously reported method in clinical use:30 In brief, isotopically labelled ISTDs were added to 250 μL serum, analytes extracted using supported liquid extraction, separated using reversed-phase LC and detected using MS/MS.
The clinical assays at the Hormone Laboratory were all accredited methods (NS-EN ISO/IEC 17025) in routine clinical use and were verified to perform satisfactorily by evaluation of precision, carry-over, and participation in an external quality assessment (EQA; Labquality or UK NEQAS). The Roche Cobas methods (Vejle Hospital) were accredited according to ISO 15189 and performed satisfactory in an EQA (Labquality or Reference Institute for Bioanalytics).
Calculation free testosterone
Calculation (or estimation) of free testosterone was performed using the Vermeulen-algorithm (cFTe(V)) solved for FTe as previously reported:31
where TTe is measured concentration of total testosterone (in nM), SHBG is measured concentration of SHBG (in nM) and N is (0.5217 * Ca + 1). Ca is the concentration of albumin (in g/L) and was assumed to be normal and set to 43 g/L for all samples. The equation output was converted from nM to pM by multiplying with 1000.
For method comparison, ED-LC-MS/MS results were compared to routine immunoassay platforms results or calculated values using regression and by constructing Bland-Altman plots in R: Passing-Bablok regression (PaBa) was performed using mcr::mcreg with setting method.reg = “PaBa”33. Second-order polynomial (quadratic) regression was performed lm with 1/x weighting, and fit and confidence intervals predicted by predict.lm with settings interval = "confidence" and level = 0.95.
Comparison of ED-LCMS measurements to clinical or calculated values
Free T3 and Free T4
FT4 and FT3 concentrations measured by ED-LCMS are known to be higher than results from most automated clinical immunoassays.(22), (34), (35)Serum samples covering TH deficiency (biochemical hypothyroidism) to TH excess (hyperthyroidism) were used to investigate how the ED-LCMS method compared to measurements from routine clinical analyzers (immunoassays).
ED-LCMS FT3 (Fig 3A) and FT4 (Fig 3B) correlated well with Abbott Alinity (Pearson’s r > 0.9). For FT3 the two methods showed a linear or close-to-linear relationship (Fig 3A; see Fig. S2A for regression residuals). In contrast, for FT4 the methods appeared non-linearly related (Figs 3C and S2A). As expected, TH concentrations (especially FT4) were markedly higher using ED-LCMS compared to the immunoassay methods (Figs 3D, 3E, S2B), consistent with previous reports.(22), (34)
We also investigated Li-Hep plasma samples, another common clinical sample matrix. ED-LCMS was compared to results measured on the Roche Cobas clinical analyzer: Good correlations (r > 0.97) and linear or close-to-linear relationship were observed for both FT3 and FT4 (Fig. S2C), indicating that Li-Hep plasma is also a possible matrix for the ED-LCMS method.
Free Testosterone
FTe is known to depend on TTe and binding proteins, especially SHBG. Samples from 79 adult males aged 18 years or older were selected from our routine analysis to cover a wide range of TTe and SHBG concentrations, and TTe/SHBG ratios. Measured FTe in two samples were outside the quantitative range (extrapolated to 9.6 and 2008 pM) and were excluded from subsequent analysis. Measured FTe in this diverse group ranged from 62.0 to 932 pM (Table S7).
Measured FTe exhibited trends expected based on known testosterone and SHBG-dynamics,8 e.g. high FTe for males with high TTe and low SHBG (Fig 4A).
Measured FTe correlated well with cFTe(V) (r = 0.93) and PaBa-regression indicated a linear relationship (Fig 4B, see S3A for residuals). However, cFTe(V)overestimated FTe compared to measured FTe (Figs 4B and S3B), as also reported by others.(36), (37)
Free hormones in healthy male blood donors
Serum from 20 healthy male blood donors (non-fasting) was analyzed to investigate free hormone concentrations and %FTe in healthy males (Table 3). FT3 ranged from 4.5 to 6.7 pM (median 5.5 pM; Fig 5A) and FT4 from 13.8 to 24.1 pM (median 18.6 pM; Fig 5B). FTe ranged from 155 to 368 pM (median 251 pM; Fig 5C) and the calculated %FTe ranged from 1.1% to 2.4%, with a median of 1.6 % (Fig 5D).
Advantages of ED-LCMS compared to immunoassays
MS-based methods for free hormone measurements are particularly well suited as complementary methods to automated immunoassay and direct FT4 measurement using ED is recommended when interference is suspected in clinical samples.(15), (16) ED-LCMS overcomes most (but not all) sources of interference that can occur in clinical routine measurements, particularly for FT4: A variety of endogenous antibodies (e.g. human anti-mouse-, anti-thyroid hormone-, anti-reagent- and other heterophile antibodies) or supplement intake (especially biotin) are well-known sources of interference causing cross-reactions with components in the immunoassay resulting in erroneous results.(12), (46) Furthermore, many immunoassays systematically produce falsely elevated FT4 concentration for individuals with familial dysalbuminemic hyperthyroxinemia (FDH).47 In contrast, MS-based methods measure normal concentrations,(48), (49) although exceptions have been reported.50 ED-LCMS is therefore a useful complementary clinical method, despite being more expensive and cumbersome to perform than automated immunoassays.
For testosterone, ED- or ultrafiltration-based MS methods have been considered the only reliable and accurate measurement method to determine free serum concentrations. Aavailable immunoassays have been discouraged8, however in Jan 2025 Revity announced U.S. Food and Drug Administration (FDA)-clearance of their automated antibody-based serum FTe test, possibly.51 Testosterone is present in saliva, and this has been used as a proxy for serum FTe. For males, salivary testosterone appears to compare well to ED-LCMS when collected properly. However, female salivary testosterone was reported to not be directly comparable to measured serum FTe.52
Measured free thyroid hormone concentrations agree with reported RI
Determining the true accuracy of a method to measure free TH or FTe is difficult, as there exists no certified reference material. However, a “spike and recovery” experiment indicated good accuracy (Table 1). Furthermore, our blood donor FT4 concentrations (Table 3, Fig 5B; range 13.8 to 24.1 pM) were comparable to published reference intervals (RIs; “normal ranges”) (16.5 – 28.6 pM, measured by ED-LCMS)23, adult RIs from the clinical laboratories LabCorp (10.3 – 21.9 pM, ED-LCMS)53, ARUP Laboratories (14.2 – 30.9 pM, ED-LCMS)54 and Mayo Clinic Laboratories (11.6 – 28.3 pM, ED-unreported chromatography-MS)55 and the upper (but not lower) limit of a published RI (6.2 – 24.8 pM)21.
FT3 concentrations for the donors ranged from 4.5 to 6.7 (Table 3, Fig 5A). This was generally consistent with RIs provided by Labcorp (2.8 – 6.2 pM, ED-LCMS)53 and one ultrafiltration-LCMS (1.1 – 6.4 pM)21, but slightly lower than the RI (5.6-10.4 pM) reported for another ED-LCMS method23. We do not know the reason for this discrepancy between the RIs from different sources, but this highlights the importance of determining or at least verifying RIs for clinical methods.
While FT4 is arguably more clinically informative than FT3 in most cases, we note that patients taking T3-medication (Liothyronine) can have low or normal FT4 but high FT3, making FT3 measurements highly informative. As an anecdotal example, for one patient with a FT4 concentration of 5.4 pM (by ED-LCMS), the measured FT3 was 16.2 pM (the pattern was confirmed using Alinity). Our method comparison between ED-LCMS and Alinity indicated a non-linearity between the two methods (Fig 3C). This is consistent with our previous report that Alinity FT4 was non-linearly related to the two clinical analyzers Roche Cobas and Siemens Centaur and to an ED-LCMS method.22
Measured free testosterone and comparison to expected values
The male blood donor FTe and %FTe (Table 3, Figs 5C and 5D) were in general agreement with expected concentrations: Walravens et al. recently reported age-stratified 95% RI for men (showing the expected decline with age) with RI for 18-29 years olds from 239 to 877 pM (median 406 pM; %FTe 2.3%), 40 to 49 years 147 – 499 pM (281 pM; 1.6%) and 70-79 years 76 – 309 pM (187 pM; 1.1%).56 Several other studies or clinical labs have also reported RI or ranges in line with our results.(25), (27), (52), (57) In contrast, Jasuja et al. reported RI for males that spanned much higher concentrations.26
Our method was not sensitive enough to quantitate FTe in all female samples investigated but could quantitate 38.5% of our presumed healthy females (Table 4), indicating that the quantitative range most likely covers the upper limit of a 95% females RI. The method should therefore be suitable for use in a clinical test of biochemical hyperandrogenism. Furthermore, our measured female concentrations appears to align well with recently reported healthy female FTe concentrations: Fiers et al. reported a RI from 2.01 to 16.8 pM (95%; median 8.3 pM),36 Rhea et al. reported a RI from 8.7 to 22.2 pM,27 and Rao et al. reported a range of approx. 4 to 20 pM (median 9.4 pM, n=100).58 In contrast, Huang et al. reported a much higher female median concentration of approx. 87 pM (max. approx. 173 pM; n=19).25
The Vermeulen-algorithm is commonly used to estimate FTe mathematically. It depends on SHBG and albumin, but the latter is often assumed to be normal and not measured (set constant to 43 g/L). Using this approach, the cFTe(V) for two of our samples deviated markedly from measured concentrations (Fig 4C, arrows), indicating the potential issues of relying on simplified calculations (at least by assuming normal albumin). Users should also be aware that cFTe(V) also overestimate FTe, as we (Fig 4C) and others have shown.(36), (58) The reason for this may be because the Vermeulen-algorithm was initially validated based on TTe measured by a (older) radioimmunoassay and FTe was measured using an indirect ED-scintillation (3H-testosterone) method using more diluted plasma samples (1:5 plasma:saline).59 In contrast, most specialized endocrine laboratories today measure testosterone (total or free) using LC-MS/MS and avoid ED dilutions above 1:1.
In summary, we have reported a novel 96-well format ED-LCMS method that allows for simultaneous quantitation of FT3, FT4 and FTe in human serum at a medium throughput suitable for specialized clinical laboratories. The method was validated and covered clinically relevant concentrations in men and women.
Healthy men 18 to 29 years old hitting a high peak FT 25.3 ng/dL are those outliers that fall in the 95th percentile!
Better yet look at where the median FT 12 ng/dL sits LOL!
* Serum from 20 healthy male blood donors (non-fasting) was analyzed to investigate free hormone concentrations and %FTe in healthy males (Table 3). FT3 ranged from 4.5 to 6.7 pM (median 5.5 pM; Fig 5A) and FT4 from 13.8 to 24.1 pM (median 18.6 pM; Fig 5B). FTe ranged from 155 to 368 pM (median 251 pM; Fig 5C) and the calculated %FTe ranged from 1.1% to 2.4%, with a median of 1.6 % (Fig 5D).
* Mass spectrometry (MS)-based methods following equilibrium dialysis (ED) are considered “gold standard” for free hormone quantification.
* Interpretation of studies is further complicated by methodological limitations, since free testosterone (FTe) - the biologically active fraction - is difficult to measure reliably, and reliance on total testosterone (TTe) or calculated values for free testosterone (cFTe) may misclassify androgen status.
* Mass spectrometry-based methods that directly quantitate the free concentration of hormones are widely regarded as the most accurate approach for measuring free hormones and least affected by interferences from patient-derived substances (e.g., medications, supplements, or endogenous antibodies) or binding protein variants. Physical separation of free from bound hormones in serum using equilibrium dialysis (ED; or less ideally ultrafiltration) performed at 37 °C followed by isotope-dilution liquid chromatography tandem mass spectrometry (LC-MS/MS) or gas chromatography (GC)-MS/MS quantitation is generally considered “gold standard” and is used for reference methods for FT3, FT4 and FTe.(13), (14) However, these methods are technically more complex to perform and costly compared to immunoassays or calculations and are today largely restricted to highly specialized laboratories.
* Measured FTe correlated well with cFTe(V) (r = 0.93) and PaBa-regression indicated a linear relationship (Fig 4B, see S3A for residuals). However, cFTe(V) overestimated FTe compared to measured FTe (Figs 4B and S3B), as also reported by others.(36), (37)
Abstract
Background
Testosterone (Te) and the thyroid hormones (THs) thyroxine (T4) and triiodothyronine (T3) are key endocrine hormones regulating development, metabolism and reproductive function. The free (not protein-bound) hormone concentrations more accurately reflect biological activity than total levels. Free hormone measurements are therefore used for clinical decision making in many contexts. Mass spectrometry (MS)-based methods following equilibrium dialysis (ED) are considered “gold standard” for free hormone quantification. Unfortunately, few clinical laboratories offer such measurements today.
Results
A 96-well format ED-LC-MS/MS method incorporating isotope-dilution for simultaneous quantitation of free Te, free T4 and free T3 in clinical samples that used offline (parallel) sample-preparation was successfully developed and validated. The method covered clinically relevant ranges, and clinical samples with a wide concentration range were analyzed to perform a comparison to existing (immunoassay or calculation) methods. Free hormone concentration in 20 healthy blood donors is reported.
Significance
This is the first method to directly quantitate both free testosterone and free thyroid hormones in clinical samples. It can be implemented alongside existing LC-MS/MS methods in specialized clinical laboratories to support improved endocrine assessment and clinical decision making.
Introduction
Here, we report an ED-LC-MS/MS method incorporating isotope-dilution for the concurrent quantitation of FTe, FT4 and FT3 in human samples, developed to be suitable for specialized clinical laboratories: Serum (or plasma) is dialyzed at 37 °C, the free hormones extracted by solid phase extraction (SPE) and then quantitated by analytical flow LC-MS/MS. The method uses 96-well format, avoids derivatization and employs electrospray ionization (ESI+). The method was extensively validated including comparison to the clinical analyzers Abbott Alinity (serum; FT4 and FT3) and Roche Cobas (lithium heparin plasma; FT4 and FT3) or to calculated FTe across a wide concentration range. The method covered clinically relevant concentration ranges for both males and females.
Routine clinical analyses (FT4, FT3, SHBG and TTe)
The immunological FT4 and FT3 assays were performed on an Alinity i instrument (Abbott Diagnostics; reagent kits: Free T4 and Free T3) at the Hormone Laboratory (Oslo University Hospital, Oslo, Norway) or on a Cobas 8000 instrument (Roche Diagnostics; kits Elecsys FT4 III and Elecsys FT3 III) at Vejle Hospital. SHBG concentration was measured on an Immulite 2000xpi system (Siemens Healthineers, Kit IMMULITE 2000 SHBG) at the Hormone Laboratory.
TTe was measured by LC-MS/MS at the Hormone Laboratory using a previously reported method in clinical use:30 In brief, isotopically labelled ISTDs were added to 250 μL serum, analytes extracted using supported liquid extraction, separated using reversed-phase LC and detected using MS/MS.
The clinical assays at the Hormone Laboratory were all accredited methods (NS-EN ISO/IEC 17025) in routine clinical use and were verified to perform satisfactorily by evaluation of precision, carry-over, and participation in an external quality assessment (EQA; Labquality or UK NEQAS). The Roche Cobas methods (Vejle Hospital) were accredited according to ISO 15189 and performed satisfactory in an EQA (Labquality or Reference Institute for Bioanalytics).
Calculation free testosterone
Calculation (or estimation) of free testosterone was performed using the Vermeulen-algorithm (cFTe(V)) solved for FTe as previously reported:31
where TTe is measured concentration of total testosterone (in nM), SHBG is measured concentration of SHBG (in nM) and N is (0.5217 * Ca + 1). Ca is the concentration of albumin (in g/L) and was assumed to be normal and set to 43 g/L for all samples. The equation output was converted from nM to pM by multiplying with 1000.
For method comparison, ED-LC-MS/MS results were compared to routine immunoassay platforms results or calculated values using regression and by constructing Bland-Altman plots in R: Passing-Bablok regression (PaBa) was performed using mcr::mcreg with setting method.reg = “PaBa”33. Second-order polynomial (quadratic) regression was performed lm with 1/x weighting, and fit and confidence intervals predicted by predict.lm with settings interval = "confidence" and level = 0.95.
Comparison of ED-LCMS measurements to clinical or calculated values
Free T3 and Free T4
FT4 and FT3 concentrations measured by ED-LCMS are known to be higher than results from most automated clinical immunoassays.(22), (34), (35)Serum samples covering TH deficiency (biochemical hypothyroidism) to TH excess (hyperthyroidism) were used to investigate how the ED-LCMS method compared to measurements from routine clinical analyzers (immunoassays).
ED-LCMS FT3 (Fig 3A) and FT4 (Fig 3B) correlated well with Abbott Alinity (Pearson’s r > 0.9). For FT3 the two methods showed a linear or close-to-linear relationship (Fig 3A; see Fig. S2A for regression residuals). In contrast, for FT4 the methods appeared non-linearly related (Figs 3C and S2A). As expected, TH concentrations (especially FT4) were markedly higher using ED-LCMS compared to the immunoassay methods (Figs 3D, 3E, S2B), consistent with previous reports.(22), (34)
We also investigated Li-Hep plasma samples, another common clinical sample matrix. ED-LCMS was compared to results measured on the Roche Cobas clinical analyzer: Good correlations (r > 0.97) and linear or close-to-linear relationship were observed for both FT3 and FT4 (Fig. S2C), indicating that Li-Hep plasma is also a possible matrix for the ED-LCMS method.
Free Testosterone
FTe is known to depend on TTe and binding proteins, especially SHBG. Samples from 79 adult males aged 18 years or older were selected from our routine analysis to cover a wide range of TTe and SHBG concentrations, and TTe/SHBG ratios. Measured FTe in two samples were outside the quantitative range (extrapolated to 9.6 and 2008 pM) and were excluded from subsequent analysis. Measured FTe in this diverse group ranged from 62.0 to 932 pM (Table S7).
Measured FTe exhibited trends expected based on known testosterone and SHBG-dynamics,8 e.g. high FTe for males with high TTe and low SHBG (Fig 4A).
Measured FTe correlated well with cFTe(V) (r = 0.93) and PaBa-regression indicated a linear relationship (Fig 4B, see S3A for residuals). However, cFTe(V)overestimated FTe compared to measured FTe (Figs 4B and S3B), as also reported by others.(36), (37)
Free hormones in healthy male blood donors
Serum from 20 healthy male blood donors (non-fasting) was analyzed to investigate free hormone concentrations and %FTe in healthy males (Table 3). FT3 ranged from 4.5 to 6.7 pM (median 5.5 pM; Fig 5A) and FT4 from 13.8 to 24.1 pM (median 18.6 pM; Fig 5B). FTe ranged from 155 to 368 pM (median 251 pM; Fig 5C) and the calculated %FTe ranged from 1.1% to 2.4%, with a median of 1.6 % (Fig 5D).
Advantages of ED-LCMS compared to immunoassays
MS-based methods for free hormone measurements are particularly well suited as complementary methods to automated immunoassay and direct FT4 measurement using ED is recommended when interference is suspected in clinical samples.(15), (16) ED-LCMS overcomes most (but not all) sources of interference that can occur in clinical routine measurements, particularly for FT4: A variety of endogenous antibodies (e.g. human anti-mouse-, anti-thyroid hormone-, anti-reagent- and other heterophile antibodies) or supplement intake (especially biotin) are well-known sources of interference causing cross-reactions with components in the immunoassay resulting in erroneous results.(12), (46) Furthermore, many immunoassays systematically produce falsely elevated FT4 concentration for individuals with familial dysalbuminemic hyperthyroxinemia (FDH).47 In contrast, MS-based methods measure normal concentrations,(48), (49) although exceptions have been reported.50 ED-LCMS is therefore a useful complementary clinical method, despite being more expensive and cumbersome to perform than automated immunoassays.
For testosterone, ED- or ultrafiltration-based MS methods have been considered the only reliable and accurate measurement method to determine free serum concentrations. Aavailable immunoassays have been discouraged8, however in Jan 2025 Revity announced U.S. Food and Drug Administration (FDA)-clearance of their automated antibody-based serum FTe test, possibly.51 Testosterone is present in saliva, and this has been used as a proxy for serum FTe. For males, salivary testosterone appears to compare well to ED-LCMS when collected properly. However, female salivary testosterone was reported to not be directly comparable to measured serum FTe.52
Measured free thyroid hormone concentrations agree with reported RI
Determining the true accuracy of a method to measure free TH or FTe is difficult, as there exists no certified reference material. However, a “spike and recovery” experiment indicated good accuracy (Table 1). Furthermore, our blood donor FT4 concentrations (Table 3, Fig 5B; range 13.8 to 24.1 pM) were comparable to published reference intervals (RIs; “normal ranges”) (16.5 – 28.6 pM, measured by ED-LCMS)23, adult RIs from the clinical laboratories LabCorp (10.3 – 21.9 pM, ED-LCMS)53, ARUP Laboratories (14.2 – 30.9 pM, ED-LCMS)54 and Mayo Clinic Laboratories (11.6 – 28.3 pM, ED-unreported chromatography-MS)55 and the upper (but not lower) limit of a published RI (6.2 – 24.8 pM)21.
FT3 concentrations for the donors ranged from 4.5 to 6.7 (Table 3, Fig 5A). This was generally consistent with RIs provided by Labcorp (2.8 – 6.2 pM, ED-LCMS)53 and one ultrafiltration-LCMS (1.1 – 6.4 pM)21, but slightly lower than the RI (5.6-10.4 pM) reported for another ED-LCMS method23. We do not know the reason for this discrepancy between the RIs from different sources, but this highlights the importance of determining or at least verifying RIs for clinical methods.
While FT4 is arguably more clinically informative than FT3 in most cases, we note that patients taking T3-medication (Liothyronine) can have low or normal FT4 but high FT3, making FT3 measurements highly informative. As an anecdotal example, for one patient with a FT4 concentration of 5.4 pM (by ED-LCMS), the measured FT3 was 16.2 pM (the pattern was confirmed using Alinity). Our method comparison between ED-LCMS and Alinity indicated a non-linearity between the two methods (Fig 3C). This is consistent with our previous report that Alinity FT4 was non-linearly related to the two clinical analyzers Roche Cobas and Siemens Centaur and to an ED-LCMS method.22
Measured free testosterone and comparison to expected values
The male blood donor FTe and %FTe (Table 3, Figs 5C and 5D) were in general agreement with expected concentrations: Walravens et al. recently reported age-stratified 95% RI for men (showing the expected decline with age) with RI for 18-29 years olds from 239 to 877 pM (median 406 pM; %FTe 2.3%), 40 to 49 years 147 – 499 pM (281 pM; 1.6%) and 70-79 years 76 – 309 pM (187 pM; 1.1%).56 Several other studies or clinical labs have also reported RI or ranges in line with our results.(25), (27), (52), (57) In contrast, Jasuja et al. reported RI for males that spanned much higher concentrations.26
Our method was not sensitive enough to quantitate FTe in all female samples investigated but could quantitate 38.5% of our presumed healthy females (Table 4), indicating that the quantitative range most likely covers the upper limit of a 95% females RI. The method should therefore be suitable for use in a clinical test of biochemical hyperandrogenism. Furthermore, our measured female concentrations appears to align well with recently reported healthy female FTe concentrations: Fiers et al. reported a RI from 2.01 to 16.8 pM (95%; median 8.3 pM),36 Rhea et al. reported a RI from 8.7 to 22.2 pM,27 and Rao et al. reported a range of approx. 4 to 20 pM (median 9.4 pM, n=100).58 In contrast, Huang et al. reported a much higher female median concentration of approx. 87 pM (max. approx. 173 pM; n=19).25
The Vermeulen-algorithm is commonly used to estimate FTe mathematically. It depends on SHBG and albumin, but the latter is often assumed to be normal and not measured (set constant to 43 g/L). Using this approach, the cFTe(V) for two of our samples deviated markedly from measured concentrations (Fig 4C, arrows), indicating the potential issues of relying on simplified calculations (at least by assuming normal albumin). Users should also be aware that cFTe(V) also overestimate FTe, as we (Fig 4C) and others have shown.(36), (58) The reason for this may be because the Vermeulen-algorithm was initially validated based on TTe measured by a (older) radioimmunoassay and FTe was measured using an indirect ED-scintillation (3H-testosterone) method using more diluted plasma samples (1:5 plasma:saline).59 In contrast, most specialized endocrine laboratories today measure testosterone (total or free) using LC-MS/MS and avoid ED dilutions above 1:1.
In summary, we have reported a novel 96-well format ED-LCMS method that allows for simultaneous quantitation of FT3, FT4 and FTe in human serum at a medium throughput suitable for specialized clinical laboratories. The method was validated and covered clinically relevant concentrations in men and women.
