Women's Sexual Function and Dysfunction. Irwin Goldstein MD

Measurement of circulating levels of total and free testosterone

Frank Z Stanczyk

One of the key androgens in sexual dysfunction in women is testosterone (see Chapters 5.5, 6.1—6.3, and 13.1-13.3 of this book). In blood, testosterone is found predominantly in a protein-bound form. It is bound with high affinity but limited capacity to sex hormone-binding globulin, and with lower affinity but higher capacity to albumin; only about 1-2% is non-protein-bound (free).However, it is the free testosterone fraction that has physiologic importance, since it is available for androgen action or metabolism and subsequent clearance. Both total and free testosterone are measured in serum for diagnostic purposes. Total testosterone is measured routinely in diagnostic testing laboratories by immunoassay methods. Reliable measurement of free testosterone requires a more specialized assay that is carried out in relatively few laboratories.

The purpose of this chapter is to discuss the different assay methods that are used to quantify total and free testosterone levels. I will begin with a summary of the development and validation of the first steroid immunoassays, specifically, radioimmunoassays with preceding purification steps. This will be followed by a discussion of direct steroid immunoassays, which were developed to avoid the purification steps in the initial radioimmunoassay methods. After that, the current status of mass spectrometry assays to measure testosterone will be addressed. The final part of the chapter will deal with the different methods used to measure free testosterone. A particular emphasis in this chapter will be the advantages and limitations of the different assay methods used to quantify total and free testosterone (see Chapter 6.3).

Measurement of total testosterone

Historical aspects

In 1959, Yalow and Berson2 described an assay method that eventually revolutionized the entire field of endocrinology. This assay method arose from studies on the effects of 131I-labeled proteins in vivo, in which it was demonstrated that insulin- requiring diabetics usually have a circulating insulin-binding protein.3 The same study showed that 131I-insulin could be displaced from the insulin-binding protein by excess of unlabeled insulin. It also showed that the binding of 131I-insulin was quantitatively related to the total amount of insulin present. These observations formed the basis for the first radioimmunoassay. Subsequently, specific antisera to human insulin were prepared in guinea pigs and resulted in the first radioimmunoassay with sufficient sensitivity to detect endogenous insulin in human blood.4

In 1963, Greenwood et al.5 developed a method of coupling iodine to growth hormone that was far superior to any previous method. The coupling was accomplished by using a mild oxidizing agent, chloramine-T. This method made it possible to bind all proteins with radioactive iodine, with minimal damage to the protein. Along with the improvement in protein iodina- tion, antibody production methods were also improved, yielding good titers. Subsequently, a new era in reproductive endocrinology was launched in 1967 when Odell et al.6 developed the radioimmunoassay for luteinizing hormone and in 1969 Abraham7 did the same for estradiol. The success of the radioimmunoassay method can be attributed to the fact that it offers a general system for measurement of an immensely wide range of compounds of clinical and biologic importance.

First steroid radioimmunoassay

The estradiol assay method first reported by Abraham7,8 consisted of purification of estradiol in serum or plasma samples by organic solvent extraction and column chromatography prior to its quantification by radioimmunoassay. The rationale for the purification steps in the estradiol radioimmunoassay method was to separate estradiol from interfering metabolites. Table 10.7.1 shows numerous unconjugated metabolites of estradiol present in blood. Whenever a steroid molecule contains a hydroxyl group, it may be conjugated to form a sulfated or glucuronidated derivative. Thus, many of the metabolites shown in Table 10.7.1 are also found in a conjugated form in serum or plasma. The unconjugated and conjugated estradiol metabolites may interfere with the specificity of the estradiol radioimmunoassay if they are not separated from estradiol. In contrast to steroid hormones, protein hormones such as luteinizing hormone and follicle-stimulating hormone are transformed only to a relatively minor extent. Therefore, luteinizing hormone and follicle-stimulating hormone immunoassays do not require purification of the analytes.

In the estradiol radioimmunoassay method reported by Abraham,7,8 purification of estradiol in serum or plasma was carried out as follows. First, after taking an aliquot of each sample, a very small amount of tritiated estradiol, approximately 1000 disintegrations per minute (d.p.m.), with high specific activity to minimize the mass, was added to the aliquots. The tritiated estradiol served as the internal standard for monitoring losses during extraction procedures. Then, an organic solvent (diethyl ether) extraction was used to remove the conjugated steroids, which remain in the aqueous layer, and to dissociate the estradiol from sex hormone-binding globulin. This step was followed by LH-20 column chromatography to separate interfering unconjugated steroids from estradiol. After evaporating the organic solvent in which estradiol was eluted off the columns, the residues were reconstituted in assay buffer. Duplicate aliquots of each reconstituted residue were taken for radioimmunoassay, and a single aliquot was taken to determine procedural losses.

Radioimmunoassay of the purified estradiol in buffer was carried out, together with varying concentrations of authentic estradiol, also in assay buffer, and used to generate a standard curve. The procedure consisted of adding to each sample and standard a constant amount of tritiated estradiol (approximately 10,000 d.p.m.) and the appropriate dilution of antibody raised in rabbits against estradiol. After overnight incubation, antibody-bound estradiol was separated from unbound estradiol with dextran-coated charcoal. After centrifugation, an aliquot of each supernatant was taken for counting in a liquid scintillation counter. The aliquots taken previously for determining procedural losses were also counted. After the estradiol standard curve was plotted, the concentration of estradiol in each sample was determined by extrapolation from the standard curve, using a programmed calculator, and each value was subsequently corrected to account for procedural losses.

Table 10.7.1. Unconjugated metabolites of estradiol

Position on

Estrogen metabolite

Position on

Estrogen metabolite
















































15a-Hydroxyestriol (estetrol)












16-Ketoestrone 16a-Hydroxyestradiol (estriol)

























The extraction/chromatography radioimmunoassay method described for estradiol soon started to be used to measure other steroid hormones by a growing number of investigators in a variety of studies. In the early 1970s, this methodology was used at our institution to measure sex hormones, including testos- terone.9,10 Our assay methodology for quantifying steroid hormones has remained essentially the same since that time. Some changes made include replacement of radioligands with iodinated steroid derivatives, resulting in more sensitive radioimmunoassays, and use of ethylacetate:hexane instead of diethyl ether for extraction to obtain a purer extract. Moreover, although steroid hormones can be measured in either serum or plasma, serum is the preferred matrix, because sometimes clotting occurs in plasma after repeated freezing, thawing, and refreezing cycles.

Reliability of extraction/chromatography radioimmunoassays

Ever since the first steroid radioimmunoassay was developed, determination of assay sensitivity, accuracy, precision, and specificity has been deemed critical for validity and reliability of immunoassay methods.11 Each new immunoassay that is developed or an existing immunoassay that is being set up for the first time in a laboratory should be thoroughly validated before using it to measure an analyte in a sample.


Assay sensitivity can be defined as the minimal detection limit of an assay. From the practical standpoint, assay sensitivity is the lowest concentration of a steroid standard in a sample, which can be distinguished from a sample without a standard. The sensitivity can be calculated by first determining the mean±standard deviation of the antibody-bound counts of replicates of the zero standard, and subtracting two standard deviations from the mean counts.11 The difference in counts is then used to read the steroid concentration off the standard curve. The sensitivity of an immunoassay can be improved by making certain changes in the assay methodology, the most important of which are reducing the amount of antibody and increasing the time of incubation.


The precision of an assay refers to the variability in the concentration of a substance observed when multiple measurements are made on the same sample. In practice, the interassay precision and interassay precision are calculated as the per cent coefficient of variation of replicate measurements. Quality control samples containing low, medium, and high levels of the appropriate steroid hormone can be used in a single assay (intraassay) and in several assays to determine the coefficients of variation of the measurements. At least five replicates of each level of the steroid hormone in the quality control samples should be used for determination of assay precision.


Assay specificity refers to the degree of interference or crossreaction encountered from substances other than the one that is measured in the assay. Specificity can be divided into two types.11 One type refers to interference by identifiable compounds that are physicochemically similar to the analyte being measured and may therefore possibly react directly with the antibody in the radioimmunoassay. This type of specificity may be assessed by determining the percentage crossreaction from the amount of material required to produce 50% inhibition of the antibody-bound tracer, after plotting curves of different concentrations of the relevant standard and the potential interfering compounds, such as relevant metabolites.11 Crossreactivity of the antiserum in an immunoassay may lead to overestimation of an analyte being measured in a sample.

The other type of specificity-related interference in an assay involves factors other than those that can be clearly identified by their physicochemical similarity to the ligand. A commonly used synonym for this type of nonspecificity is “matrix effect”. It may lead to an underestimation or, more commonly, an overestimation of the amount of an analyte in a sample. A variety of materials can interfere with antibody-ligand reactions, and the effects can vary in different assay methods. High levels of serum lipids are particularly notorious for causing matrix effects. Other materials that may interfere include hemoglobin, heparin, salts, acids, alkalines, nucleic acids, and sodium azide.11


Assay accuracy can be defined as the extent to which a given measurement of a substance agrees with the actual value. Accuracy of an extraction/chromatographic radioimmunoassay method can be established by “spiking” known amounts of the appropriate steroid standard to serum and determining the amount of steroid recovered, after measuring the steroid by radioimmunoassay in the spiked and some unspiked samples. The expected recovery of the “spiked” steroid is generally ±10% for each of the amounts added. Moreover, regression analysis of the data should give a line that is parallel to that obtained with the standard curve.

Quality control of extraction/chromatographic steroid radioimmunoassays

Quality control parameters used to determine the reliability of each extraction/chromatographic radioimmunoassay include those related to the standard curve and the analyte being measured in samples. Parameters related to the standard curve include the following: total counts added to each radioimmunoassay tube, nonspecific binding, binding of the zero standard, slope, ^-intercept, and doses of the standard at which per cent bound/maximum bound (B/Bo) is equal to 20, 50, and 80. Parameters related to the analyte include the results of quantifying the analyte in quality-control samples and the performance characteristics of pipettes and instruments. However, the single most important quality-control parameter includes the result of repeated measurements of the low, medium, and high levels of the analyte in samples prepared to monitor the reliability of each assay. These quality-control samples are analyzed in the same manner as serum samples containing unknown concentrations of the analyte. Typically, one set of low, medium, and high quality-control samples and a buffer blank that is used to monitor potential assay contaminants are placed at the beginning and end of the samples with unknown concentrations of the analyte. Assays are accepted or rejected on the basis of the limits of rejection for different quality- control parameters, which are set by the individual laboratory.

Advantages and disadvantages of extraction/chromatographic steroid radioimmunoassays

Extraction/chromatographic radioimmunoassays have advantages and disadvantages. The advantages include the following: (a) steroid-binding proteins are denatured by the organic solvent in the extraction step, resulting in release of steroid hormones such as testosterone from sex hormone-binding globulin; (b) both conjugated and unconjugated metabolites are separated from the analyte prior to its quantification; (c) relatively large serum aliquots can be used, allowing the analyte in a sample to be measured on an accurate part of the standard curve. High assay reliability is achieved when the assay method is properly validated.

Disadvantages of the extraction/chromatographic radioimmunoassay include potential errors from counting very small amounts of tritiated labeled internal standard, such as tritiated testosterone in the testosterone radioimmunoassay, to follow procedural losses, and from use of factors to account for dilutions used in assay procedures. In addition, the assay method is somewhat cumbersome, takes about 2 days for analysis of around 40 samples, and is relatively costly.

Direct steroid immunoassays

Because extraction/chromatographic radioimmunoassays for quantifying steroids were so time-consuming and laborious, the need for rapid, direct immunoassay methods, i.e., without purification steps, became obvious. In the late 1970s, such assays started to become available commercially, initially using commercial kits that contained reagents to perform assays manually.

An iodinated steroid derivative, instead of a tritiated radioligand, was used in the assay methodology to reduce the time of counting the radioactivity. Subsequently, the kits were used in an instrument so that the assay could be performed automatically.

Automated immunoassay methods employ a chemiluminescent, fluorescent, or enzyme tag, which avoids use of radioactivity and prolonged counting times, as well as the high cost of radioactive material. The money saved from decreased labor costs and not having to pay for radioactivity disposal offsets the initial cost of purchasing an automated instrument. Thus, direct steroid immunoassays are simple, convenient, rapid, and relatively inexpensive.

Unfortunately, direct steroid immunoassays can have disadvantages. First, steroids that bind with high affinity to the plasma binding proteins, such as testosterone to sex hormonebinding globulin and cortisol to corticosteroid-binding globulin (CBG), sometimes may not be released efficiently from the protein. Manufacturers of steroid immunoassay kits often use a low pH in one of the reagents in the kit, such as the radioligand solution, to release the steroid from sex hormone-binding globulin or corticosteroid-binding globulin. If the pH is not correct, testosterone will not be dissociated from sex hormonebinding globulin, and a patient can be considered to be hypo- gonadal on the basis of a subnormal testosterone measurement.

A second deficiency of direct steroid immunoassay involves differences between the matrix containing the analyte being measured and the matrix containing the different concentrations of standard. The matrix often used by kit manufacturers for the different concentrations of steroid standard is defibrinated, delipidated, charcoal-stripped plasma, which is a clear, buffer-like solution. In contrast, serum samples often contain interfering materials that may cause a “matrix effect” and consequently an overestimation or underestimation of an analyte in a sample, as pointed out earlier. Serum samples that are hemolyzed should not be used in a direct assay.

A third deficiency of direct steroid immunoassays arises from the fact that there are several hundred different steroids, in an unconjugated or conjugated form, present in serum. Some of these steroids are closely related in chemical structure to the analyte being measured, and may therefore be recognized by the antiserum used in the assay. It is rare that an antiserum against a certain steroid is so highly specific that it measures solely the analyte that it is intended to measure. Due to the presence of crossreacting compounds, measures of analytes are often overestimated in direct immunoassays.

Evidence showing that direct steroid immunoassays carried out with reagents in commercial kits are not reliable can be found in several studies. In our study,12 we evaluated four different commercially available direct testosterone immunoassay kits and used our extraction/chromatographic testosterone radioimmunoassay as a standard for comparison. Our results show that the assays using the kits performed generally well for male serum samples, but gave poor intraclass correlations and/or failed validity for both premenopausal and postmenopausal serum samples. (The validity of an assay reflects its accuracy.) Our findings are consistent with those reported by Taieb et al.,13 who measured serum testosterone levels in women, men, and children by use of 10 different direct testosterone immunoassay kits and by isotope dilution gas chromatography-mass spectrometry. On the basis of their results, they concluded that the direct assays were acceptable for measuring testosterone in male samples, but not in samples from women or children. In an accompanying editorial14 on the article by Taieb et al.,13 the editors concluded that “guessing appears to be nearly as good as most commercially available immunoassays and clearly superior to some”.

Mass spectrometry assay methods

Gas chromatography-mass spectrometry and liquid chromatography-mass spectrometry are powerful analytic techniques that combine the resolving power of gas chromatography or liquid chromatography with the high sensitivity and specificity of the mass spectrometer. Separation of steroids by gas chromatography requires that they be first derivatized to increase their volatility, selectivity, and detectability. It is not necessary to derivatize steroids for separation by liquid chromatography, but derivatization is sometimes used to improve the ionization efficiency of steroids, a step which increases mass spectrometry assay sensitivity.

The mass spectrometer functions as a unique detector that provides structural information on individual solutes as they elute from the gas chromatography or liquid chromatography column. Addition of the appropriate internal standard to the sample enables accurate quantification in conjunction with a standard curve of known concentrations of the standard.

The mass spectrometry technique first involves ionization of the target compound at the ionization source. This is followed by separation and detection of the ions in the mass analyzer. A mass spectrum is produced, in which the relative abundance of a particular ion is plotted as a function of the mass-to-charge ratio.

Although a variety of gas chromatography-mass spectrometry and liquid chromatography-mass spectrometry assay methods exist, one of the most highly efficient of these methods is liquid chromatography combined with tandem mass spectrometry (liquid chromatography/tandem mass spectrometry). This technique involves use of a collision cell in which the ion of interest (precursor ion) undergoes collision-induced fragmentation into product ions. The mass of the product ion is then determined at the detector. Tandem mass spectrometry has the capability to achieve not only high sensitivity and specificity, but also high throughput. It is projected to become the reference standard for steroid hormone measurements. However, at present, measurement of testosterone by liquid chromatography/ tandem mass spectrometry for diagnostic testing is restricted to only a few laboratories. This is due to the high cost of the mass spectrometry instrument (nearly US$500,000), the requirement for a highly trained technician to operate the instrument, and the time required to develop the assay for quantifying a wide range of testosterone levels. Nevertheless, at least one laboratory (Quest Diagnostic Nichols Institute, San Juan Capistrano, CA, USA) is using liquid chromatography/tandem mass spectrometry to measure testosterone levels in women, children, and men for diagnostic testing. The initial high cost of the liquid chromatography/tandem mass spectrometry instrumentation and assay development is offset by the relatively rapid turnaround time of accurate testosterone results compared with the costly, cumbersome and time-consuming extraction/chromatography testosterone radioimmunoassay methodology.

Free and bioavailable testosterone

As mentioned earlier, testosterone is present in blood predominantly in a protein-bound form and only a very small portion is free. In premenopausal women, approximately 66% and 30% of total testosterone are bound to sex hormone-binding globulin and albumin, respectively, and the free fraction generally comprises under 2% of the total.1 Testosterone is bound with high affinity (Ka = 1.7 x 109 m-1) and low capacity to sex hormonebinding globulin, and it is bound with low affinity (Ka = 1 x 10m-1 to 1 x 105 m-1) but high capacity to albumin.1,15

For many years, it was accepted that only the free fraction of testosterone in the circulation can be taken up by tissues and that the protein-bound testosterone complex is inactive. However, some investigators observed that the fraction of testosterone bound to albumin dissociates rapidly and is taken up by tissues in a manner similar to that of the free steroid.16-18 Testosterone bound to the large pool of albumin, together with the small amount of the free steroid, probably forms the circulating pool of bioavailable (non-sex hormone-binding globulin- bound) testosterone. The bioavailable fraction of testosterone enters cells where it may undergo metabolism or binds to the androgen receptor and exerts biologic activity.

Commonly used methods for measuring free testosterone involve the addition of a small amount of 3H-testosterone to serum or plasma and, after a suitable incubation period, separation of the protein (sex hormone-binding globulin and albumin)-bound fractions from the free fraction of testosterone by means of a membrane (e.g., equilibrium dialysis) or filter (e.g., centrifugal ultrafiltration). These barriers retain the protein-bound fractions but allow free testosterone to pass through. The percentage of tritiated free testosterone is then calculated on the basis of the total tritiated testosterone added. Recovery of free components through a barrier is sometimes monitored with a small labeled molecule such as 14C-glucose.

Several technical limitations exist in the assays used to measure free testosterone. The equilibrium dialysis method is influenced by dilution of the serum sample. The centrifugal ultrafiltration method is complex and subject to adsorption of testosterone to the membrane. Both the dialysis and ultrafiltration methods can be affected by impurities of tritiated testosterone not bound by sex hormone-binding globulin or albumin; these impurities may increase the percentage of free testosterone. Moreover, the use of too large an amount of tritiated testosterone in the assays may increase the concentration of total testosterone and possibly disturb the equilibrium of endogenous testosterone.

Two methods used to determine the percentage of bioavailable testosterone in serum include centrifugal ultrafiltration with heat-treated serum and ammonium sulfate precipitation. In the centrifugal ultrafiltration method, the percentage of albumin-bound testosterone is determined after sex hormonebinding globulin is inactivated by heating the serum sample to 60°C for 1 h. After the temperature of the sample returns to 37°C, the testosterone dissociated from sex hormone-binding globulin is re-equilibrated in the serum, and the testosterone fraction bound to albumin can be determined by ultrafiltration. The fraction of testosterone bound to albumin along with the free testosterone fraction determined before heating the sample comprises the total bioavailable testosterone fraction. A much simpler method to determine bioavailable testosterone involves addition of a small amount of tritiated testosterone to serum and, after a suitable incubation period, precipitation of the globulins (including the sex hormone-binding globulin testosterone complex) with saturated ammonium sulfate, centrifugation, counting the tritium in the supernatant, and calculating the percentage of the total 3H-testosterone that is not sex hormonebinding globulin bound.

Certain technical difficulties are also encountered in the measurement of bioavailable testosterone. When this fraction is measured by the barrier method after inactivation of sex hormone-binding globulin, the same technical problems exist as described for the measurement of free testosterone. The most frequently encountered sources of error in the ammonium sulfate precipitation assay are use of impure tritiated testosterone, insufficient counting time of the small amount of radiolabeled testosterone, and incomplete precipitation of globulins. The deficiencies in both assays are often the cause of poor intraassay and interassay reproducibility.

By the methods described above, the concentration of free or bioavailable testosterone is usually calculated from the percentage of free or bioavailable testosterone multiplied by the total testosterone concentration, which is quantified separately by immunoassay. Free testosterone concentrations are sometimes measured directly in the dialysate after equilibrium dialysis. However, highly sensitive extraction/chromatography radioimmunoassays are required to measure the very low testosterone levels.

Because the assays described above for quantifying free or bioavailable testosterone are time-consuming and expensive, they are available in a limited number of reference laboratories. The most widely used assays for measurement of free testosterone in clinical laboratories are direct radioimmunoassays. In general, these assays use a 125I-labeled testosterone analog that has very low affinity for sex hormone-binding globulin and albumin, and competes with free testosterone for binding sites on an immobilized specific testosterone antibody. Although this approach provides a simple and rapid test for quantifying free testosterone, it has been pointed out that the assay method has several deficiencies: these include low antibody affinity, major biasing effects due to dilution of serum samples, significant binding of the analog to serum proteins, and lack of parallelism between measurements of serially diluted serum samples and free testosterone.19 For these reasons, the reliability of the analog-based free testosterone radioimmunoassay kit has been questioned.20,21 One study22 showed that plasma free testosterone levels in samples from normal women and patients with polycystic ovary syndrome were approximately three to four times higher when measured by a commercial analog-based radioimmunoassay kit than when measured by the equilibrium dialysis method. The results obtained with the latter method were comparable to published data. Nevertheless, good correlations were obtained between the results of the two methods. The investigators concluded that the free testosterone values measured by use of the kit had a mean bias of -76%, thereby making comparison with published data difficult. The higher levels of free testosterone measured by direct radioimmunoassay may result from the fact that the antibody in the radioimmunoassay system has a greater affinity for testosterone weakly bound to albumin than albumin does. This may allow the antibody to strip some of the testosterone that is bound to albumin. A subsequent study23 showed that the direct testosterone radioimmunoassay had unacceptably high systematic bias and random variability, and did not correlate well with equilibrium dialysis. A letter to the editor by Rosner24 about the direct free testosterone analog radioimmunoassay concluded, “the literature of science ought not to use a method so grossly inaccurate when better ones exist.” In addition, Rosner24 suggested, “The journal might choose to return manuscripts that use it without further evaluation to discourage its use.”

Some laboratories and investigators that have measured total testosterone and sex hormone-binding globulin have used the ratio of testosterone/sex hormone-binding globulin, referred to as the free androgen index, as an estimate of free testosterone. The validity of the free androgen index as an accurate reflection of free testosterone has been questioned. One small study20 showed the free androgen index to be unreliable, on the basis of its comparison with free testosterone quantified by equilibrium dialysis; the ratio of free androgen index to free testosterone determined by dialysis was 0.12-0.26. Another small study25 found a high correlation coefficient (0.858) between the free androgen index and free testosterone levels determined by centrifugal ultrafiltration in serum samples from women, whereas in male samples the correlation was only 0.435. A more recent study23 in women found a good correlation between free androgen index and equilibrium dialysis. However, the authors of that study pointed out that the free androgen index can be altered by changes in either testosterone or sex hormone-binding globulin, and that using this quotient alone can be misleading. Therefore, use of the free androgen index is limited.

Both free and bioavailable testosterone can also be calculated by an algorithm which requires the concentrations of total testosterone, sex hormone-binding globulin, and albumin, as well as the binding constants of testosterone to sex hormonebinding globulin and albumin obtained from published equations.20 Calculated free testosterone levels in men and women were found to be nearly identical to corresponding values measured by equilibrium dialysis.23,26,27

It is important to realize that when indirect methods, such as equilibrium dialysis or centrifugal ultrafiltration, are used to determine free testosterone concentrations, the accuracy of the total testosterone concentration is very important. This is because those methods determine the percentage of total testosterone that is free, and the percentage is multiplied by the total testosterone concentration to obtain the free testosterone concentration. Thus, direct immunoassay methods should not be used for quantifying total testosterone levels in female samples; radioimmunoassays with preceding organic solvent extraction and chromatography steps will provide reliable values. Similarly, if the free androgen index or algorithm is used to calculate free testosterone, the accuracy of both total testosterone and sex hormone-binding globulin values is essential. Although the concentration of albumin is also required in the algorithm method, an average normal albumin value can be used without any significant change in the calculated free testosterone concentration.

Differences in sex hormone-binding globulin concentrations obtained with different commercially available sex hormonebinding globulin kits have been reported.28 In one study,23 an approximately twofold greater absolute value was found with an immunoradiometric assay than a radioimmunoassay, and better accuracy was found with the former assay. The immunoradiometric method was calibrated against a dihydrotestosteronebinding capacity assay, which is considered to provide sex hormone-binding globulin values that reflect more physiologically relevant sex hormone-binding globulin concentrations in blood. Thus, it seems reasonable to use sex hormone-binding globulin assay methods that correlate well with assay methods based on testosterone- or dihydrotestosterone-binding capacity.

Normal ranges of total and free testosterone

Once an immunoassay for a hormone in serum is developed and validated in a laboratory, normal ranges for females and/or males should be established. Serum samples must be obtained from normal, healthy subjects who are not on any treatment that may affect the normal range. It is also important to establish normal ranges in a large number of individuals; this number should be over 100. In addition, blood samples for normal ranges should be obtained at the same time of day, such as 08:00-10:00. This is essential for those hormones that undergo diurnal variation.

An example of female and male normal ranges for total testosterone quantified by extraction/chromatography radioimmunoassay and free testosterone determined by equilibrium dialysis is shown in Table 10.7.2. The values were obtained from Quest Diagnostics Nichols Institute.

Table 10.7.2. Normal ranges of total and free testosterone in women and men


Total testosterone (ng/dl)

Free testosterone (pg/ml)













From the information presented in this chapter, the following conclusions can be made. (a) Extraction/chromatography radioimmunoassay methods for quantifying testosterone are highly reliable when thoroughly validated with respect to sensitivity, accuracy, precision, and specificity, and when the quality control of the assay is properly monitored. (b) It is essential that users of commercial direct immunoassay kits for measurement of any steroid hormone validate the assay methods thoroughly in their own laboratory; they should not merely accept the kit manufacturer’s validation. It is especially important that steroid hormone values measured in a direct assay be compared with corresponding values determined by a well-established assay, i.e., either extraction/chromatography radioimmunoassay, or a gas chromatography-mass spectrometry or liquid chromatography-mass spectrometry assay. (c) Tandem mass spectrometry is projected to become the reference standard for steroid hormone measurements. However, interlaboratory comparisons using tandem mass spectrometry to quantify steroid levels are essential before any measurement is accepted as the reference standard. (d) Measurement of free testosterone concentrations by the analog-based free testosterone radioimmunoassay is highly unreliable and should never be used. (e) When free testosterone concentrations are quantified experimentally, the first step is to determine the percentage of free testosterone by methods such as equilibrium dialysis or ultrafiltration. Because this percentage is then multiplied by the total testosterone concentration to obtain the free testosterone concentration, a highly reliable immunoassay, such as the extraction/chromatography radioimmunoassay, must be used to measure total testosterone. (f) Use of the calculated method to determine the free testosterone concentration gives values that are similar to the equilibrium dialysis method, which is considered the reference standard for determination of free testosterone concentrations. The calculated method requires accurate measurement of not only total testosterone but also sex hormone-binding globulin.

A recent editorial by Matsumoto and Bremner29 stated: “What is needed now is refocusing of attention to more rigorous validation and standardization of accuracy and normal ranges for these assays to alleviate the confusion that has arisen in the clinical and research community as a result of the variability and discrepancies in testosterone assays. We hope that assay vendors, endocrinologists, clinical chemists, and regulatory agencies can act together to achieve better standardization of hormone measurements, including testosterone assays.”


1 Westphal U. Steroid-Protein Interactions. Berlin: Springer-Verlag, 1986.

2. Yalow RS, Berson SA. Assay of plasma insulin in human subjects by immunological methods. Nature 1959; 184: 1648-9.

3. Berson SA, Yalow R, Bauman Aetal. Insulin-I-131 metabolism in human subjects: demonstration of insulin binding globulin in the circulation of insulin treated subjects. J Clin Invest 1956; 35: 170-90.

4. Yalow RS, Berson SA. Immunoassay of endogenous plasma insulin in man. J Clin Invest 1960; 38: 1157-75.

5. Greenwood FC, Hunter WM, Grove JS. The preparation of 125I-labelled human growth hormone high specific radioactivity. Biochem J 1963; 89: 114.

6. Odell WD, Ross GT, Rayford RL. RIA of luteinizing hormone in human plasma or serum: physiological studies. J Clin Invest 1967; 46: 248.

7. Abraham GE. Solid-phase radioimmunoassay of estradiol-17 p. J Clin Endocrinol Metab 1969; 29: 866-70.

8. Abraham GE, Odell WD. Solid-phase radioimmunoassay of serum estradiol-17 P: a semi-automated approach. In FG Peron, BV Caldwell, eds. Immunologic Methods in Steroid Determination. New York: Appleton-Century-Crofts, 1970: 87-112.

9. Thorneycroft IH, Sribyatta B, Tom WK et al. Measurement of serum LH, FSH, progesterone, 17-hydroxyprogesterone, and estradiol- 17 P levels at 4-hour intervals during the periovulatory phase of the menstrual cycle. J Clin Endocrinol Metab 1974; 39: 754-8.

10. Goebelsmann U, Arce JJ, Thorneycroft IH et al. Serum testosterone concentrations in women throughout the menstrual cycle and following hCG administration. Am J Obstet Gynecol 1974; 119: 445-52.

11. Chard T. An introduction to radioimmunoassay and related technique. In PC Van der Vliet, ed. Laboratory Techniques in Biochemistry and Molecular Biology. Amsterdam: Elsevier, 1995.

12. Stanczyk FZ, Cho MM, Endres DB et al. Limitations of direct estradiol and testosterone immunoassay kits. Steroids 2003; 68: 1173-8.

13. Taieb J, Mathian B, Millot F et al. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography - mass spectrometry in sera from 116 men, women, and children. Clin Chem 2003; 49: 1381-95.

14. Fitzgerald RL, Herold DA. Ciba Corning ACS: 180 direct total testosterone assay can be used on female sera. Clin Chem 1997; 43: 1466-7.

15. Westphal U. Steroid—Protevnlntemçtions. Berlin: Springer-Verlag, 1971.

16. Manni A, Pardrige WM, Cefalus W et al. Bioavailability of albumin-bound testosterone. J Clin Endocrinol Metab 1985; 61: 705-10.

17. Pardridge W, Landaw EM. Tracer kinetics model of blood-brain barrier transport of plasma protein-bound ligands: empiric testing of free hormone hypothesis. JdÿJnv£2± 1984; 74: 745-52.

18. Pardridge WM. Transport of protein bound hormones into tissues in vivo. EüdocïRv 1981; 2: 103-23.

19. Ekins R. Hirsutism: free and bound testosterone. Ann Clin Biochem 1990; 27: 91-4.

20. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 1999; 84: 3666-72.

21. Winters SJ, Kelley DE, Goodpaster B. The analog free testosterone assay: are the results in man clinically useful? Clin Chem 1998; 44: 2178-82 [Erratum 1999; 45: 444].

22. Cheng RN, Reed MJ, James VHT. Plasma free testosterone: equilibrium dialysis vs direct radioimmunoassay. Clin Chem 1986; 32: 1411.

23. Miller KK, Rosner W, Lee H et al. Measurement of free testosterone in normal women and women with androgen deficiency: comparison of methods. JdÿEsdüüdÿdMeüL 2004; 89: 525-33.

24. Rosner W. An extraordinary inaccurate assay for free testosterone is still with us [Letter]. J Clin Endocrinol Metab 2001; 86: 2903.

25. Kapoor P, Luttrell B, Williams D. The free androgen index is not valid for adult males. JSeoidBschemMûLBid 1993; 45: 325-6.

26. Morley JE, Patrick P, Perry HM III. Evaluation of assays available to measure free testosterone. Metabolism 2002; 51: 554-9.

27. Emadi-Konjin P, Bain J, Bromberg IL. Evaluation of an algorithm for calculation of serum “bioavailable” testosterone (BAT). Clin Biochem 2003; 36: 591-6.

28. Bukowski C, Grigg MA, Longcope C. Sex-hormone-binding globulin concentrations: differences among commercially available methods. Clin Chem 2000; 46: 1415-16.

29. Matsumoto AM, Bremner WJ. Editorial: serum testosterone assays - accuracy matters. JdnEndocrnriMeat 2004; 89: 520-4.