madman
Super Moderator
For many years, virtually the only means of studying androgen metabolism in experimental animals or in human subjects was by examination of the urinary steroid metabolites which were presumed to originate from gonadal and adrenocortical androgens. Nevertheless, even though the techniques of urinary assays were refined extensively to enable the quantitative measurement of groups of steroids (such as 17-oxosteroids) or individual steroids, it was only rarely possible to deduce information of physiological or clinical importance from those measurements. It awaited the availability of a remarkably versatile biochemical tool before any real progress could be made, that tool being the radio-labelled chemical. In the steroid field, radio-isotopically labelled hormones have entirely revolutionised a situation that had progressed little over several years, and have made possible extensive and definitive studies of hormone metabolism. Although used initially as metabolic tracers, the use of high specific activity labelled steroids to develop extremely sensitive assay methods has at last made it possible to examine reasonably small quantities of blood quantitatively for the presence of various hormones. In the androgen field, this was particularly gratifying since these ar evirtually the only techniques which ar e capable of achieving the high sensitivity needed, particularly for studies of female plasma.
However, since this symposium is devoted to the endocrine function of the testes, it is appropriate for us to concentrate more upon the problems of androgen levels in male subjects, although even here , certain situations still require a degree of sensitivity in the assay which is normally appropriate to female plasma levels.
The techniques which have been successfully employed for the assay of testosterone have been the double isotope dilution derivative assay, gas liquid chromatography (usually with electron capture detectors), and saturation analysis (in which term I include competitive protein binding assays and radioimmunoassay). Unlike the situation with female plasma, the improved assay methods of the last few years have not greatly altered out concept of the normal range of plasma testosterone levels in normal male adult subjects. Thus some representative levels from different investigators are shown in Table 1.
We have not found any clear indication of a nyctohemeral rhythm in plasma testosterone levels, although in general, plasma levels tended in ail our subjects to be lower at night. In most of the studies reported in the literature, there has been only inconclusive evidence for a nyctohemeral rhythm, and since sampling has usually been at intervals of hours, rather than minutes, and bearing in mind the considerable fluctuations in plasma testosterone levels which can occur, it is not surprising that no clear conclusions have been drawn. Boon et al. (2), reviewing the existing literature, concluded that if any rhythm exists at all, it must be one which is easily ove-ridden by other stimuli.
The variations which occur in plasma testosterone levels pose a problem in defining what is a 'normal* plasma testosterone level, apart from the interesting but unsolved question of what mechanism is involved in causing these changes in steroid levels. Luteinising hormone (LH) can cause increased secretion of testosterone when administered to subjects with normal testicular function, but in human subjects, unlike other species, there is no direct relationship between plasma LH levels and plasma testosterone levels (3).
Testosterone is not the only steroid present in male plasma which is androgenic. It is now well recognised that a major metabolite, 5a-dihydrotestosterone, is of fundamental importance in the mode of action of testosterone at the cellular level, and it is likely that certain aspects of testosterone action are expressed through this metabolic product. Attempts have therefore been made to determine this steroid in human plasma, and it has been shown (4) that there is a progressive increase through puberty, reaching the adult male levels of approximately 50 ng/100 ml(5, 6). How far the assay of this steroid will produce useful clinical or physiological information remains to be seen.
More recently, methods have been developed for the determination of other C19 steroids, which may also have some physiological or clinical significance, since they have been demonstrated to have androgenic activity. These are androstanediol (5a-androstane-3ß, 17ß-diol) and androstenediol (androst5-ene-3ß, 170-diol). These compounds ar e present in relatively low concentrations in peripheral plasma and as yet, insufficient information is available to indicate what changes, if any, might occur in pathological conditions, or how far they might play any physiological role.
It is clear therefore, that testosterone is not the only potentially active C19 steroid which is present in human plasma in a free form, and all these compounds share with testosterone the property of binding to the testosterone-estradiol binding globulin (TEBG) which is present in plasma. This is a useful property since it provides a means of assaying these steroids by the competitive protein binding technique. This principle has been exploited by several workers who were seeking to assay testosterone (7,8) or androstenediol (9) and who described methods involving initial Chromatographie purification. Murphy, however, who has pioneered much of the work on competitive binding assay techniques, put forward a different philosophy in describing an assay (10) in which minimal purification was used, and the specificity of the binding protein for 17 ß-hydroxysteroids was invoked. On this basis, the assay would be regarded as a group analysis for steroids which would include testosterone and other 17ß-hydroxysteroids which might be potentially androgenic.Anderson, following the same reasoning, described in detail his method (11) for testosterone-like substances and demonstrated its considerable clinical value, and Horton, Kato and Sherins(12) have also described a simple assay of this type.
The virtue of these methods, is that they are relatively inexpensive in terms of reagents, and are less demanding technically and can be performed far more rapidly than methods which are more specific for testosterone. These qualities are valuable when there is a clinical demand for androgen assays, and as Anderson has shown, (11), used in conjunction with dynamic endocrine function tests, they offer clinical information which is probably as useful as the specific assays. The possibility that, as suggested by Murphy, more relevant information might be available from this group assay approach is also worth bearing in mind· We have therefore looked in more detail at this technique, in terms of its specificity and also in relation to changes in plasma testosterone levels. We wished also to improve the technique so as to make possible the use of relatively small volumes of plasma.
In our hands, ammonium sulphate precipitation has proved a more reliable method for the separation of bound and unbound steroid in saturation analyses than have adsorbents such as charcoal or florisil, and so our efforts were directed towards developing an assay along these lines. Initially though, crude plasma extracts caused considerable problems, since the ammonium sulphate invariably caused precipitation of lipid with subsequent poor separation. After various experimental approaches, it was found that hexane - ether produced a very satisfactory extract from plasma and the method ultimately adopted was as follows.
Plasma samples are stored deep frozen prior to analysis. After thawing overnight at 4 C the samples are mixed vigorously on a vortex mixer and 0.1 ml pipetted into glass test tube of approximate capacity 5 ml (Quickfit type MF 24/0/4) followed by the addition of 0.2 ml of 0. IN sodium hydroxide solution. The contents of the tubes are again mixed thoroughly on a vortex mixer and extracted twice with 2 mls of n-hexane :ether (8:2) by shaking horizontally for ten minutes. The extracts are transferred into a second set of Quickfit tubes, using a pasteur pipette, where the pooled extracts are washed twice with 1 ml of distilled water by shaking horizontally for five minutes. The first water wash is removed carefully with a pasteur pipette. After the second waterwash the tubes are centrifuged at 1, 000 g for ten minutes at 4 C. The extracts are then transferred, in two stages, to 10 x 50 mm glass specimen tubes, care being taken not to transfer any of the aqueous phase. The n-hexane :ether is evaporated to dryness in a vacuum oven at a temperature not greater than 40 C and evacuated by means of a filter pump.
Standards in duplicate of 0, 0.25, 0.5, 0.75, 1.0, 1.5 and 2.0 ng of testosterone are taken through the extraction procedure together with the plasma samples
A 2% solution of third-trimester pregnancy plasma is prepared by adding the appropriate volume of plasma to borate buffer pH 8.0 containing 2% methanol and approximately 50, 000 dpm/ml of (1,2- H) testosterone. 0.25 ml of this solution is added to each of the standard curve and test samples and also to two counting vials containing 0.05 ml of water. The specimen tubes are shaken gently but thoroughly on a vortex mixer and returned to their rack which is placed in a transparent plastic box with a tight fitting lid. The tubes are not stoppered individually.
After overnight incubation at 4 C the tubes are immersed in an ice bath and 0.25 ml of cold saturated ammonium sulphate added. The contents of each tube are mixed for approximately ten seconds on a vortex mixer immediately after the addition of the ammonium sulphate. When the addition of ammonium sulphate to all the tubes is complete each tube is again mixed for two more periods of ten seconds, returning them to the ice bath between each mix. The tubes are left for a further five minutes in the ice bath before centrifugation at 1,000 g for ten minutes at 4 C. The tubes are returned to the ice bath and 0.3 ml of the supernatant transferred to a counting vial. To all the counting vials 10 ml of scintillator is added. The scintillator is prepared by dissolving 6 gm of p-terphenyl and 0.08 gm of dimethyl P.O.P.O.P. in toluene, adding 40 ml of methanol, and making up to two litres with toluene. All the counting vials are capped securely and shaken mechanically for ten minutes before counting in a liquid scintillation spectrometer to an error of 2%. The activity of the standard curve and test samples is calculated as a percentage of the activity in 0.25 ml of the pregnancy plasma solution multiplied by 0.6. A curve is plotted of percentage free radioactivity against the mass of testosterone contained in the standard curve samples. From this curve the mass of Ί7ß-hydroxysteroids* in ng per 0.1 ml of plasma is read directly and the final results expressed as ng per 100 ml plasma.
The number of plasma samples analysed in one particular batch will of course depend upon the facilities available in each laboratory. Due to the fact that it is necessary to process the standard curve alongside the plasma samples fourteen places in each batch will be taken up with standards. The plasma extracts will however store, at least, overnight at 4 C with no apparent damage provided that they are not allowed to evaporate to dryness.This means that one can extract plasmas on two consecutive days and then incubate them all with one standard curve. Using this approach one technician can easily process 45 plasma samples plus the standard curve in two and a half days
The recovery of (1,2- 3H) testosterone from plasma when taken through the extraction and washing procedure was 102% with a range of 91 - 108, N = 9. The recovery of ( 3H)-androstanediol from three separate plasmas when taken through the extraction procedure was 89%, 91% and 89%. The recovery of 0.25 ng, 0.5 ng and 0. 75 ng of testosterone added to plasma when taken through the method was 104%, 105% and 106% respectively. Precision data were obtained from the assay of aliquots of a plasma pool in eleven separate batches. This gave a mean of 967 ng/100 ml with a coefficient of variation of + 6.7%. The analysis of within batch duplicates with a mean of 1075 ng/100 ml gave a coefficient of variation of + 5%.
Although it is possible to store plasma deep frozen without a great deal of difficulty, it is important to realise that the practise of centrifuging such plasma samples after they have been unfrozen (to remove precipitated protein) carries the risk of removing some part of the endogenous testosterone, which is presumably bound to the precipitate. This phenomenon is particularly marked if this happens repeatedly (e.g. with a frozen pool which is unfrozen on several occasions). Fig. 3 illustrates the progressive fall in steroid concentration (as shown by the radio-active tracer) when this is done, and shows that when the precipitate is resuspended, the plasma testosterone level returns towards the original value.
A group of 49 male control subjects was used to obtain normal data; their ages ranged from 15 to 59 years and the mean level of plasma 17-hydroxysteroids found was 614 ng/100 ml with a range of 270 - 1200. It is interesting to compare these data with those in Table 1. The values obtained are shown in Figure 2 in relation to age.
The question of specificity, or what is being measured in this assay is more difficult to answer. Figure 1 shows the relationship between the 17-hydroxysteroids measured by the method described above and testosterone as measured by radioimmunoassay. Although testosterone accounts for a major proportion of the 17-hydroxysteroids, this proportion is very variable, even in a series of plasma samples from one subject taken over a few hours. In some samples it accounts for almost all the total 17-hydroxysteroids; in others less than 50%. It is not likely that this is in any major part due to assay errors, since the precision is, as described above, quite good for both assays .This variability in the proportion of the 17-hydroxysteroids which is made up of testosterone is further illustrated by Figure 4, comparing testosterone measured by immunoassay and the 17-hydroxysteroid content of a number of plasma samples. The difference between the two assay results arises, at least in part, from the contribution of the 17-hydroxysteroids in plasma mentioned above, i.e . dihydrotestosterone, androstanediol and androstenediol. So far, we have no knowledge of how the plasma levels of these other steroids alters through the day, but it would not be surprising if they varied to some extent independently of testosterone, since some may originate from the adrenal cortex. Thus Wieland et al. (13) from their studies of adrenal venous blood, have found evidence for adrenocortical secretion of androstenediol in human subjects. Under basal conditions though, it is not easy to assess the significance of the small fluctuations in the levels of 17-hydroxysteroids although there is no obvious relationship to plasma cortisol levels and by implication, to corticotrophin secretion.
Under conditions in which changes of plasma testosterone are being studied, as in gonadal stimulation tests, the assay of 17-hydroxysteroids reveals virtually the same pattern as does the testosterone assay. To illustrate this point, figure 5 shows a comparison of the results of the two assays in two patients being treated by the administration of clomiphene.
The problem of deciding upon an appropriate methodology for plasma androgen assay must therefore depend very much upon the particular requirement to be met. Studies in which testosterone levels are primarily of interest still require fairly complicated assay methods to provide the necessary specificity. The double isotope and gas Chromato graphie methods, although excellent in highly skilled and experienced hands, are technically too demanding to attract unqualified recommendation, and most investigators would now prefer to employ radio-immunoassay after some initial purification stage. Separative techniques are also essential if dihydrotestosterone, androstenediol or androstanediol are to be measured. Inevitably, these methods prove relatively more difficult to employ in a busy diagnostic laboratory, and where the main need is for a method which rapidly reveals changes in testosterone levels, particularly as in the investigation of male gonadal dysfunction, then simpler assays, either as described by Anderson (11) or Horton (12) or the method used here , may prove entirely adequate.
However, since this symposium is devoted to the endocrine function of the testes, it is appropriate for us to concentrate more upon the problems of androgen levels in male subjects, although even here , certain situations still require a degree of sensitivity in the assay which is normally appropriate to female plasma levels.
The techniques which have been successfully employed for the assay of testosterone have been the double isotope dilution derivative assay, gas liquid chromatography (usually with electron capture detectors), and saturation analysis (in which term I include competitive protein binding assays and radioimmunoassay). Unlike the situation with female plasma, the improved assay methods of the last few years have not greatly altered out concept of the normal range of plasma testosterone levels in normal male adult subjects. Thus some representative levels from different investigators are shown in Table 1.
We have not found any clear indication of a nyctohemeral rhythm in plasma testosterone levels, although in general, plasma levels tended in ail our subjects to be lower at night. In most of the studies reported in the literature, there has been only inconclusive evidence for a nyctohemeral rhythm, and since sampling has usually been at intervals of hours, rather than minutes, and bearing in mind the considerable fluctuations in plasma testosterone levels which can occur, it is not surprising that no clear conclusions have been drawn. Boon et al. (2), reviewing the existing literature, concluded that if any rhythm exists at all, it must be one which is easily ove-ridden by other stimuli.
The variations which occur in plasma testosterone levels pose a problem in defining what is a 'normal* plasma testosterone level, apart from the interesting but unsolved question of what mechanism is involved in causing these changes in steroid levels. Luteinising hormone (LH) can cause increased secretion of testosterone when administered to subjects with normal testicular function, but in human subjects, unlike other species, there is no direct relationship between plasma LH levels and plasma testosterone levels (3).
Testosterone is not the only steroid present in male plasma which is androgenic. It is now well recognised that a major metabolite, 5a-dihydrotestosterone, is of fundamental importance in the mode of action of testosterone at the cellular level, and it is likely that certain aspects of testosterone action are expressed through this metabolic product. Attempts have therefore been made to determine this steroid in human plasma, and it has been shown (4) that there is a progressive increase through puberty, reaching the adult male levels of approximately 50 ng/100 ml(5, 6). How far the assay of this steroid will produce useful clinical or physiological information remains to be seen.
More recently, methods have been developed for the determination of other C19 steroids, which may also have some physiological or clinical significance, since they have been demonstrated to have androgenic activity. These are androstanediol (5a-androstane-3ß, 17ß-diol) and androstenediol (androst5-ene-3ß, 170-diol). These compounds ar e present in relatively low concentrations in peripheral plasma and as yet, insufficient information is available to indicate what changes, if any, might occur in pathological conditions, or how far they might play any physiological role.
It is clear therefore, that testosterone is not the only potentially active C19 steroid which is present in human plasma in a free form, and all these compounds share with testosterone the property of binding to the testosterone-estradiol binding globulin (TEBG) which is present in plasma. This is a useful property since it provides a means of assaying these steroids by the competitive protein binding technique. This principle has been exploited by several workers who were seeking to assay testosterone (7,8) or androstenediol (9) and who described methods involving initial Chromatographie purification. Murphy, however, who has pioneered much of the work on competitive binding assay techniques, put forward a different philosophy in describing an assay (10) in which minimal purification was used, and the specificity of the binding protein for 17 ß-hydroxysteroids was invoked. On this basis, the assay would be regarded as a group analysis for steroids which would include testosterone and other 17ß-hydroxysteroids which might be potentially androgenic.Anderson, following the same reasoning, described in detail his method (11) for testosterone-like substances and demonstrated its considerable clinical value, and Horton, Kato and Sherins(12) have also described a simple assay of this type.
The virtue of these methods, is that they are relatively inexpensive in terms of reagents, and are less demanding technically and can be performed far more rapidly than methods which are more specific for testosterone. These qualities are valuable when there is a clinical demand for androgen assays, and as Anderson has shown, (11), used in conjunction with dynamic endocrine function tests, they offer clinical information which is probably as useful as the specific assays. The possibility that, as suggested by Murphy, more relevant information might be available from this group assay approach is also worth bearing in mind· We have therefore looked in more detail at this technique, in terms of its specificity and also in relation to changes in plasma testosterone levels. We wished also to improve the technique so as to make possible the use of relatively small volumes of plasma.
In our hands, ammonium sulphate precipitation has proved a more reliable method for the separation of bound and unbound steroid in saturation analyses than have adsorbents such as charcoal or florisil, and so our efforts were directed towards developing an assay along these lines. Initially though, crude plasma extracts caused considerable problems, since the ammonium sulphate invariably caused precipitation of lipid with subsequent poor separation. After various experimental approaches, it was found that hexane - ether produced a very satisfactory extract from plasma and the method ultimately adopted was as follows.
Plasma samples are stored deep frozen prior to analysis. After thawing overnight at 4 C the samples are mixed vigorously on a vortex mixer and 0.1 ml pipetted into glass test tube of approximate capacity 5 ml (Quickfit type MF 24/0/4) followed by the addition of 0.2 ml of 0. IN sodium hydroxide solution. The contents of the tubes are again mixed thoroughly on a vortex mixer and extracted twice with 2 mls of n-hexane :ether (8:2) by shaking horizontally for ten minutes. The extracts are transferred into a second set of Quickfit tubes, using a pasteur pipette, where the pooled extracts are washed twice with 1 ml of distilled water by shaking horizontally for five minutes. The first water wash is removed carefully with a pasteur pipette. After the second waterwash the tubes are centrifuged at 1, 000 g for ten minutes at 4 C. The extracts are then transferred, in two stages, to 10 x 50 mm glass specimen tubes, care being taken not to transfer any of the aqueous phase. The n-hexane :ether is evaporated to dryness in a vacuum oven at a temperature not greater than 40 C and evacuated by means of a filter pump.
Standards in duplicate of 0, 0.25, 0.5, 0.75, 1.0, 1.5 and 2.0 ng of testosterone are taken through the extraction procedure together with the plasma samples
A 2% solution of third-trimester pregnancy plasma is prepared by adding the appropriate volume of plasma to borate buffer pH 8.0 containing 2% methanol and approximately 50, 000 dpm/ml of (1,2- H) testosterone. 0.25 ml of this solution is added to each of the standard curve and test samples and also to two counting vials containing 0.05 ml of water. The specimen tubes are shaken gently but thoroughly on a vortex mixer and returned to their rack which is placed in a transparent plastic box with a tight fitting lid. The tubes are not stoppered individually.
After overnight incubation at 4 C the tubes are immersed in an ice bath and 0.25 ml of cold saturated ammonium sulphate added. The contents of each tube are mixed for approximately ten seconds on a vortex mixer immediately after the addition of the ammonium sulphate. When the addition of ammonium sulphate to all the tubes is complete each tube is again mixed for two more periods of ten seconds, returning them to the ice bath between each mix. The tubes are left for a further five minutes in the ice bath before centrifugation at 1,000 g for ten minutes at 4 C. The tubes are returned to the ice bath and 0.3 ml of the supernatant transferred to a counting vial. To all the counting vials 10 ml of scintillator is added. The scintillator is prepared by dissolving 6 gm of p-terphenyl and 0.08 gm of dimethyl P.O.P.O.P. in toluene, adding 40 ml of methanol, and making up to two litres with toluene. All the counting vials are capped securely and shaken mechanically for ten minutes before counting in a liquid scintillation spectrometer to an error of 2%. The activity of the standard curve and test samples is calculated as a percentage of the activity in 0.25 ml of the pregnancy plasma solution multiplied by 0.6. A curve is plotted of percentage free radioactivity against the mass of testosterone contained in the standard curve samples. From this curve the mass of Ί7ß-hydroxysteroids* in ng per 0.1 ml of plasma is read directly and the final results expressed as ng per 100 ml plasma.
The number of plasma samples analysed in one particular batch will of course depend upon the facilities available in each laboratory. Due to the fact that it is necessary to process the standard curve alongside the plasma samples fourteen places in each batch will be taken up with standards. The plasma extracts will however store, at least, overnight at 4 C with no apparent damage provided that they are not allowed to evaporate to dryness.This means that one can extract plasmas on two consecutive days and then incubate them all with one standard curve. Using this approach one technician can easily process 45 plasma samples plus the standard curve in two and a half days
The recovery of (1,2- 3H) testosterone from plasma when taken through the extraction and washing procedure was 102% with a range of 91 - 108, N = 9. The recovery of ( 3H)-androstanediol from three separate plasmas when taken through the extraction procedure was 89%, 91% and 89%. The recovery of 0.25 ng, 0.5 ng and 0. 75 ng of testosterone added to plasma when taken through the method was 104%, 105% and 106% respectively. Precision data were obtained from the assay of aliquots of a plasma pool in eleven separate batches. This gave a mean of 967 ng/100 ml with a coefficient of variation of + 6.7%. The analysis of within batch duplicates with a mean of 1075 ng/100 ml gave a coefficient of variation of + 5%.
Although it is possible to store plasma deep frozen without a great deal of difficulty, it is important to realise that the practise of centrifuging such plasma samples after they have been unfrozen (to remove precipitated protein) carries the risk of removing some part of the endogenous testosterone, which is presumably bound to the precipitate. This phenomenon is particularly marked if this happens repeatedly (e.g. with a frozen pool which is unfrozen on several occasions). Fig. 3 illustrates the progressive fall in steroid concentration (as shown by the radio-active tracer) when this is done, and shows that when the precipitate is resuspended, the plasma testosterone level returns towards the original value.
A group of 49 male control subjects was used to obtain normal data; their ages ranged from 15 to 59 years and the mean level of plasma 17-hydroxysteroids found was 614 ng/100 ml with a range of 270 - 1200. It is interesting to compare these data with those in Table 1. The values obtained are shown in Figure 2 in relation to age.
The question of specificity, or what is being measured in this assay is more difficult to answer. Figure 1 shows the relationship between the 17-hydroxysteroids measured by the method described above and testosterone as measured by radioimmunoassay. Although testosterone accounts for a major proportion of the 17-hydroxysteroids, this proportion is very variable, even in a series of plasma samples from one subject taken over a few hours. In some samples it accounts for almost all the total 17-hydroxysteroids; in others less than 50%. It is not likely that this is in any major part due to assay errors, since the precision is, as described above, quite good for both assays .This variability in the proportion of the 17-hydroxysteroids which is made up of testosterone is further illustrated by Figure 4, comparing testosterone measured by immunoassay and the 17-hydroxysteroid content of a number of plasma samples. The difference between the two assay results arises, at least in part, from the contribution of the 17-hydroxysteroids in plasma mentioned above, i.e . dihydrotestosterone, androstanediol and androstenediol. So far, we have no knowledge of how the plasma levels of these other steroids alters through the day, but it would not be surprising if they varied to some extent independently of testosterone, since some may originate from the adrenal cortex. Thus Wieland et al. (13) from their studies of adrenal venous blood, have found evidence for adrenocortical secretion of androstenediol in human subjects. Under basal conditions though, it is not easy to assess the significance of the small fluctuations in the levels of 17-hydroxysteroids although there is no obvious relationship to plasma cortisol levels and by implication, to corticotrophin secretion.
Under conditions in which changes of plasma testosterone are being studied, as in gonadal stimulation tests, the assay of 17-hydroxysteroids reveals virtually the same pattern as does the testosterone assay. To illustrate this point, figure 5 shows a comparison of the results of the two assays in two patients being treated by the administration of clomiphene.
The problem of deciding upon an appropriate methodology for plasma androgen assay must therefore depend very much upon the particular requirement to be met. Studies in which testosterone levels are primarily of interest still require fairly complicated assay methods to provide the necessary specificity. The double isotope and gas Chromato graphie methods, although excellent in highly skilled and experienced hands, are technically too demanding to attract unqualified recommendation, and most investigators would now prefer to employ radio-immunoassay after some initial purification stage. Separative techniques are also essential if dihydrotestosterone, androstenediol or androstanediol are to be measured. Inevitably, these methods prove relatively more difficult to employ in a busy diagnostic laboratory, and where the main need is for a method which rapidly reveals changes in testosterone levels, particularly as in the investigation of male gonadal dysfunction, then simpler assays, either as described by Anderson (11) or Horton (12) or the method used here , may prove entirely adequate.
