Comparison of serum testosterone and estradiol measurements in 3174 European men using platform immunoassay and mass spectrometry; relevance for the diagnostics in aging men

in European Journal of Endocrinology
View More View Less
  • 1 Department of Surgery and Cancer, Arthritis Research UK Epidemiology Unit, Developmental and Regenerative Biomedicine Research Group, Department of Obstetrics, Department of Geriatric Medicine, Department of Medicine, CIBER de Fisiopatología Obesidad y Nutricion (CB06/03), Reproductive Medicine Centre, Department of Endocrinology, Department of Andrology and Reproductive Endocrinology, Laboratory of Molecular Endocrinology and Oncology, Department of Human Nutrition, School of Community Based Medicine, Andrology Unit, Arthritis Research UK, Department of Andrology and Endocrinology, Endocrinology Unit, Imperial College London, Hammersmith Campus, London, UK

(Correspondence should be addressed to I T Huhtaniemi; Email: ilpo.huhtaniemi@imperial.ac.uk; F C W Wu; Email: frederick.wu@manchester.ac.uk)

Background

The limitations of serum testosterone and estradiol (E2) measurements using non-extraction platform immunoassays (IAs) are widely recognized. Switching to more specific mass spectrometry (MS)-based methods has been advocated, but directly comparative data on the two methods are scarce.

Methods

We compared serum testosterone and E2 measurements in a large sample of middle-aged/elderly men using a common platform IA and a gas chromatography (GC)–MS method, in order to assess their limitations and advantages, and to diagnose male hypogonadism. Of subjects from the European Male Aging Study (n=3174; age 40–79 years), peripheral serum testosterone and E2 were analyzed using established commercial platform IAs (Roche Diagnostics E170) and in-house GC–MS methods.

Results

Over a broad concentration range, serum testosterone concentration measured by IA and MS showed high correlation (R=0.93, P<0.001), which was less robust in the hypogonadal range (<11 nmol/l; R=0.72, P<0.001). The IA/MS correlation was weaker in E2 measurements (R=0.32, P<0.001, at E2 <40.8 pmol/l, and R=0.74, P<0.001, at E2 >40.8 pmol/l). Using MS as the comparator method, IA ascertained low testosterone compatible with hypogonadism (<11 nmol/l), with 75% sensitivity and 96.3% specificity. The same parameters with IA for the detection of low E2 (<40.7 pmol/l) were 13.3 and 99.3%, and for high E2 (>120 pmol/l) 88.4 and 88.6%.

Conclusion

A validated platform IA is sufficient to detect subnormal testosterone concentrations in the diagnosis of male hypogonadism. The IA used for E2 measurements showed poor correlation with MS and may only be suitable for the detection of high E2 in men.

Abstract

Background

The limitations of serum testosterone and estradiol (E2) measurements using non-extraction platform immunoassays (IAs) are widely recognized. Switching to more specific mass spectrometry (MS)-based methods has been advocated, but directly comparative data on the two methods are scarce.

Methods

We compared serum testosterone and E2 measurements in a large sample of middle-aged/elderly men using a common platform IA and a gas chromatography (GC)–MS method, in order to assess their limitations and advantages, and to diagnose male hypogonadism. Of subjects from the European Male Aging Study (n=3174; age 40–79 years), peripheral serum testosterone and E2 were analyzed using established commercial platform IAs (Roche Diagnostics E170) and in-house GC–MS methods.

Results

Over a broad concentration range, serum testosterone concentration measured by IA and MS showed high correlation (R=0.93, P<0.001), which was less robust in the hypogonadal range (<11 nmol/l; R=0.72, P<0.001). The IA/MS correlation was weaker in E2 measurements (R=0.32, P<0.001, at E2 <40.8 pmol/l, and R=0.74, P<0.001, at E2 >40.8 pmol/l). Using MS as the comparator method, IA ascertained low testosterone compatible with hypogonadism (<11 nmol/l), with 75% sensitivity and 96.3% specificity. The same parameters with IA for the detection of low E2 (<40.7 pmol/l) were 13.3 and 99.3%, and for high E2 (>120 pmol/l) 88.4 and 88.6%.

Conclusion

A validated platform IA is sufficient to detect subnormal testosterone concentrations in the diagnosis of male hypogonadism. The IA used for E2 measurements showed poor correlation with MS and may only be suitable for the detection of high E2 in men.

Introduction

Testosterone and estradiol (E2) are the two most important sex steroids in men and women respectively, and their accurate determination in serum is of crucial importance in assessing gonadal function both in clinical management and research. Immunoassay (IA) methods have been the mainstay of sex steroid measurements since their advent in the late 1960s. Most of the time they provide rapid and economical information about circulating hormone concentrations. However, the accuracy and precision of testosterone IAs, especially at the low concentrations found in children, women, and hypogonadal men, remain a concern (1, 2, 3, 4). While the majority of IAs estimate high (adult male) concentrations sufficiently well, they usually overestimate low (female) concentrations (5), thus reducing the specificity and sensitivity of diagnosis of female hyperadrogenism and male hypogonadism. There is no consensus on whether IAs for testosterone are able to reliably discriminate between eugonadal and hypogonadal men (1, 5, 6). Professional societies and individual investigators have therefore emphasized the need for improved standardized methods, as well as traceability of the standards, to overcome these problems in sex steroid measurements (3, 4, 5, 6, 7, 8, 9).

Even during the years of IA dominance, mass spectrometry (MS) was regarded as the ‘gold standard’ of steroid analysis, but due to its technical complexity, cost, and suboptimal sensitivity, it has only recently reached the methodological refinement required for a routine clinical chemistry laboratory. The recent technical improvements in instrumentation and the wider availability due to falling costs of equipment have made MS a competitive method with IA, having reached sufficient sensitivity yet maintaining its superior specificity in steroid hormone measurements. Therefore, opinions are being expressed to promote MS as the standard method for steroid hormone measurements (6, 9, 10). However, MS remains more expensive and labour intensive (requiring solvent extraction), shows similar lack of between-laboratory standardisation to IA (11, 12), and is currently still not accessible to all practitioners. It is therefore important to investigate in which clinical situations MS is necessary for the measurement of testosterone and when it is still sufficient to rely on IAs. For instance, while it is clear that measurements of testosterone by IA in children and women are unreliable, it is uncertain whether MS or IA should be the method of choice for the quantitation of testosterone to diagnose adult male hypogonadism.

Measurement of E2 is more challenging than that of testosterone due to its much lower circulating concentration (50–100-fold less in men). Although E2 measurements are less often required for men, high concentrations are of diagnostic importance in gynecomastia and the rare cases of feminizing tumors and aromatase excess (13). Low E2 concentrations are important in the assessment of osteoporosis (14, 15) and cardiovascular diseases (16, 17), where the replacement of IAs with more specific and sensitive MS measurements is expected to be useful.

The European Male Aging Study research consortium (18) has measured testosterone and E2 concentrations in the serum samples of a large cohort (n=3174) of 40–79-year-old men using both an established IA and MS method. This provided a unique opportunity to compare the results obtained with these two methods and to assess the applicability of each technique for clinical diagnostics and research.

Subjects and methods

Subjects and study design

A total of 3369 community-dwelling men aged 40–79 (mean±s.d.: 60±11) years were recruited from population registers in eight European centers (Florence, Italy; Leuven, Belgium; Lodz, Poland; Malmö, Sweden; Manchester, UK; Santiago de Compostela, Spain; Szeged, Hungary; and Tartu, Estonia). Details of the research protocol have been published elsewhere (18). Ethical approval for the study was obtained in accordance with local institutional requirements in each center, and written informed consent was obtained from the study subjects.

Hormone measurements

A single fasting morning (before 1000 h) venous blood sample was obtained and separated serum was stored at −80 °C. Measurements of testosterone and E2 were carried out by the Modular E170 platform electrochemiluminescence IAs (Roche Diagnostics) and gas chromatography–MS (19, 20, 21, 22). Within- and between-assay coefficients of variation (CV) in IA measurements were 1.05 and 3.72% for testosterone (at 14.4 nmol/l human serum), and 5.2 and 9.1% for E2 (at 0.071 nmol/l human serum) respectively. The male reference range on IA for testosterone was 10.4–34.6 nmol/l and for E2 <200 pmol/l. In MS measurements, the intra- and interassay CV were 2.9 and 3.4% for testosterone (at 1.7 nmol/l human serum), and 3.5 and 3.7% for E2 (at 0.07 nmol/l human serum) respectively. The average recovery for steroids following extraction on MS was 102±3%, and the male reference ranges were 14.1–39.0 nmol/l for testosterone and 23–112 pmol/l for E2.

Sex hormone-binding globulin (SHBG) was measured by the Modular E170 platform electrochemiluminescence IAs (Roche Diagnostics). Free testosterone concentrations were derived from total testosterone, SHBG, and albumin concentrations (23).

Statistical analysis

From the total of 3369 participants, 150 were excluded because of prevalent pituitary or testicular diseases or current use of medications that could affect pituitary/testicular function (testosterone, dehydroepiandrosterone, antiandrogens, GnRH agonists, glucocorticoids, and psycholeptic agents) or interfere with sex steroid clearance or measurements (e.g. anticonvulsants). The reason for exclusion was their expected interference with the use of testosterone values in the diagnosis of late-onset hypogonadism (LOH; see Results). Of the remaining men, 3174 had complete data on testosterone and 3016 on E2 with both IA and MS and were included in this analysis.

The analysis consisted of descriptive statistics to assess subject characteristics, where the data were presented as mean and s.d. for continuous variables and count (percentage for discrete variables). Distributions of total testosterone and total E2 with both techniques were plotted via histogram. Spearman correlation measure was used to test the correlation for total testosterone and total E2 between the two assays and within each assay. Agreement between the two assays was explored using the Bland–Altman plot (24) for limits of agreement, and bias estimation was used; this plots the % difference between MS and IA (i.e. 100×(IA−MS)/MS) against the average of the two assays ((IA+MS)/2). Deviations from ±20% were used as the limits of bias.

Deming regression technique (25), which takes into account any measurement errors in the hormones, was used to additionally compare hormone concentrations between the two assays. Sensitivity and specificity of the IA measurement, using MS as the comparator method, were calculated to further explore the diagnostic accuracy of IA.

Results

Cohort characteristics

Characteristics of the analysis cohort of 3174 men are shown in Table 1. Mean (s.d.) age of the men was 59.7 (11.0) years. The recruitment was carried out from a random general population that was relatively healthy, as shown by a variety of characteristics. Of these, 21.4% were current smokers, and in 27.1% at least one co-morbid condition was reported, which included self-reported heart conditions, high blood pressure, bronchitis, asthma, peptic ulcer, epilepsy, diabetes, cancer, liver conditions, kidney conditions, prostate diseases, and thyroid disorders. The mean (s.d.) body mass index was 27.7 (4.1) kg/m2 and the mean (s.d.) waist circumference was 98.4 (11.1) cm.

Table 1

Cohort characteristics (n=3174).

Mean (s.d.)Count (%)95% CI(5–10–50–95–97.5)th Centiles
Age (years)59.67 (10.96)
BMI (kg/m2)27.66 (4.10)
Waist circumference (cm)98.42 (11.07)
Serum testosterone (nmol/l) MS16.58 (5.95)16.37, 16.78(8.3–9.7–15.9–27.8–30.6)
Serum testosterone (nmol/l) IA16.54 (5.80)16.34, 16.74(8.1–9.7–15.9–27.1–29.6)
Serum estradiol (pmol/l) MS74.13 (25.09)73.26, 75.01(40.8–46.7–70.3–119.8–133.1)
Serum estradiol (pmol/l) IA92.90 (28.73)91.90, 93.90(52.7–59.7–90.0–143.8–160.6)
Current smokers673 (21.40)
One morbidity845 (27.07)
Two or more morbidities711 (22.75)

MS, mass spectrometry; IA, immunoassay.

Mass spectrometry vs immunoassay

The mean (s.d.) testosterone concentrations were very similar between IA and MS: 16.5 (5.80) and 16.6 (5.95) nmol/l respectively. E2 concentrations were, on average, higher with IA than MS: 92.9 (28.7) and 74.1 (25.1) pmol/l respectively. There was a good agreement in the distribution of results between the two assays for testosterone (Fig. 1a). In contrast, with E2 (Fig. 1b), there were more samples with concentration below 70 pmol/l with MS than IA, and above this concentration there were more E2 samples by IA than by MS.

Figure 1
Figure 1

Distribution of testosterone (panel a), and E2 (panel b) concentrations, as measured with MS and IA.

Citation: European Journal of Endocrinology 166, 6; 10.1530/EJE-11-1051

Bland–Altman plot

For testosterone (Fig. 2a), there was little bias between the two methods at mean concentrations of the paired values (MS, IA) ranging from 0.175 to 46.21 nmol/l. The mean IA–MS difference (negative bias) was a low and nonsignificant −0.036 (95% confidence interval (95% CI), −0.113 to 0.040) nmol/l, with 95% limits of agreement of −4.36 to 4.29 nmol/l. Here, 9% of the testosterone concentrations by IA were more than 20% higher than those measured by MS, and 3% were over 20% lower. There was no significant trend in the relationship between the percentage bias and the average testosterone concentration of the two methods. Spearman correlation between the percentage bias and the average testosterone concentration of the two methods was 0.01 (P=0.508). This confirms that there was no concentration-dependent loss of agreement between the two methods of testosterone quantification.

Figure 2
Figure 2

Bland–Altman plots of testosterone (panel a) and E2 (panel b). Y-axis depicts the % difference between values of the two measurements (100×(IA−MS)/MS). The horizontal lines are 0 and −20 and +20%. The mean difference between the two assays in the testosterone values was 0.77% and of the E2 values, 30.1%.

Citation: European Journal of Endocrinology 166, 6; 10.1530/EJE-11-1051

For E2 (Fig. 2b), there was a significant mean percentage difference (positive bias for IA) of 18.77 (95% CI, 18.11 to 19.43) pmol/l between the IA and MS measurements, with 95% limits of agreement of −18.5 to 56.1 pmol/l. The range of mean concentrations of the paired values (MS, IA) was from 17.01 to 254.65 pmol/l. The average discrepancy between the concentrations of the two methods (bias) was high; 58% of the E2 concentrations by IA were over 20% higher than by MS, and 3% E2 concentrations by IA were more than 20% lower than by MS. Hence, IA grossly overestimated the E2 levels. Figure 2b shows also a trend of the relationship between the bias (percentage difference) and the average concentration of the MS and IA methods. The positive bias increased as the average of E2 decreased with a Spearman correlation of −0.07 (P<0.001), i.e. showing an inversely concentration-dependent positive bias for IA vs MS.

Deming regression

Figure 3 shows the scatter plots of testosterone (panel a) and E2 (panel b) for the two methods, as well as the results from the Deming regression. The agreement between testosterone concentrations measured by IA and MS was close to the line of best fit (y=x, i.e. the line of equality), 0.97 (95% CI, 0.96 to 0.99), and the intercept was 0.41 nmol/l (95% CI, 0.18 to 0.65). For E2, the agreement between the two techniques deviated considerably from the line of equality, with a slope of 1.19 (95% CI, 1.15 to 1.25) and an intercept of 4.28 pmol/l (95% CI, 0.75 to 7.81).

Figure 3
Figure 3

Scatter plots for testosterone (panel a) and E2 (panel b) with IA and MS, and the Deming regression results.

Citation: European Journal of Endocrinology 166, 6; 10.1530/EJE-11-1051

Table 2 presents the correlation coefficients between the MS and IA measurements of testosterone and E2 at different concentrations of the hormones. With testosterone, the correlation coefficient was 0.93 in the entire cohort, 0.92 with testosterone concentrations >8 nmol/l, and 0.69 at testosterone concentrations <8 nmol/l. Using Deming regression, the agreement between testosterone concentrations measured by IA and MS with testosterone levels >8 nmol/l was close to the line of best fit, with a slope of 0.97 (95% CI, 0.95 to 0.99), and an intercept of 0.42 nmol/l (95% CI, 0.15 to 0.70). The agreement between testosterone concentrations with testosterone levels <8 nmol/l deviated more from the line of best fit, with a slope of 1.71 (95% CI, 1.04 to 2.37), and an intercept of −4.27 nmol/l (95% CI, −8.52 to −0.02). With E2, the correlation coefficient between the MS and IA measurements was 0.76 in the entire cohort, and 0.74 at E2 levels above 40.8 pmol/l, but only 0.32 at concentrations below 40.8 pmol/l. Using Deming regression, the agreement between E2 concentrations measured by IA and MS with E2 concentrations either above 40.8 pmol/l or below 40.8 pmol/l was divergent from the line of best fit, i.e. with E2 >40.8 pmol/l, the slope was 1.24 (95% CI, 1.18 to 1.30) and the intercept was 0.04 pmol/l (95% CI, −4.13 to 4.21); with E2 ≤40.8 pmol/l, the slope was 7.89 (95% CI, 3.17 to 12.6) and the intercept was −214 pmol/l (95% CI, −382 to −46.5). Significant, though less robust, correlations were also found between the testosterone and E2 concentrations, which were weaker or nonsignificant in the IA/IA and IA/MS comparisons.

Table 2

Correlation coefficients between testosterone and E2 measurements by MS and IA.

Testosterone MSTestosterone IAE2 MSE2 IA
(A) Testosterone and E2; the entire analysis sample
 Testosterone MS10.93***0.49***0.32***
 Testosterone IA10.47***0.37***
 E2 MS10.76***
 E2 IA1
(B) Testosterone (<8 nmol/l)
 Testosterone MS10.69***0.39***0.19*
 Testosterone IA10.21*0.16
 E2 MS10.68***
 E2 IA1
(C) Testosterone (>8 nmol/l)
 Testosterone MS10.92***0.46***0.31***
 Testosterone IA10.44***0.37***
 E2 MS10.76***
 E2 IA1
(D) E2 (<40.8 pmol/l; lowest 5th centile)
 Testosterone MS10.91***0.41***0.01 (P=0.87)
 Testosterone IA10.30***0.08 (P=0.32)
 E2 MS10.32***
 E2 IA1
(E) E2 (>40.8 pmol/l)
 Testosterone MS10.93***0.45***0.28***
 Testosterone IA10.43***0.34***
 E2 MS10.74***
 E2 IA1

*P<0.05, ***P<0.001.

Using MS as the comparator method, we then assessed the sensitivity (% of true positives) and specificity (% of true negatives) of IA to detect low testosterone concentrations by MS at defined thresholds and to identify patients fulfilling the diagnostic criteria of LOH, i.e. low testosterone in combination with three sexual symptoms (reduced morning erections and sexual thoughts, and erectile dysfunction; Table 3) (26). The sensitivity and specificity of IA to detect total testosterone <11 nmol/l were 75.0 and 96.3% respectively. To detect total testosterone <8 nmol/l, the sensitivity and specificity of IA increased slightly to 76.9 and 98.3%. If the presence of symptoms of androgen deficiency (three sexual symptoms) in addition to a total testosterone <11 nmol/l and calculated free testosterone <220 pmol/l were used as the criteria, the sensitivity and specificity of identifying LOH with the IA testosterone measurements increased to 85.5 and 99.4%. If the threshold levels of testosterone were decreased to <8 nmol/l (together with the three sexual symptoms), the respective parameters increased even further to 92.3 and 99.8%.

Table 3

Sensitivity and specificity of IA in detection of low testosterone levels (panels A and B), diagnosing LOH (panels C and D), and detection of low and high E2 (panels E, F and G), using MS as the reference method.

MS
Low testosteroneOh by MSE2 (pmol/l)IA
YesNoTotalYesNoTotal>Sensitivity (%)Specificity (%)
(A) Detection of low testosteronea
 Low testosterone by IA75.096.3
 Yes40596501
 No13525382673
 Total54026343174
(B) Detection of low testosteroneb
 Low testosterone by IA75.498.5
 Yes9847145
 No3229973029
 Total13030443174
(C) Diagnosis of LOHc,h
 LOH by IA85.599.4
 Yes531770
 No928842893
 Total6229012963
(D) Diagnosis of LOHd,h
 LOH by IA88.599.8
 Yes23629
 No329312934
 Total2629372963
(E) Detetion of low E2 <40.8 pmol/le
 E2 by IA13.299.3
 ≤40.82120
 >40.81382995
 Total1593015
(F) Detection of low E2 <61.2 pmol/lf
 E2 by IA25.696.1
 ≤61.227182
 >61.27872034
 Total10582116
(G) Detection of high E2 >119.83 pmol/lg
 E2 by IA88.688.4
 <119.83266718
 >119.83349140
 Total3016158

Total testosterone <11 nmol/l.

Total testosterone <8 nmol/l.

Total testosterone <11 nmol+free testosterone <220 pmol/+3 sexual symptoms.

Total testosterone <8 nmol/+3 sexual symptoms.

The lowest 5th centile for E2 by MS.

The lowest tertile for E2 by MS.

The highest 5th centile for E2 by MS.

Total number of cases smaller because data on sexual symptoms were not available from 211 men.

In the E2 assay, the ability of IA to detect concentrations below 40.8 pmol/l (the lowest 5th centile for E2 by MS) had a sensitivity of only 13.3%, with a specificity of 99.3% (Table 3). The IA performance to detect a concentration below 61.2 pmol/l (the lowest tertile for E2 by MS) had a slightly better sensitivity of 25.6% and a specificity of 96.1%. In contrast, the sensitivity (88.6%) and specificity (88.4%) of E2 IA were clearly better to detect high E2 concentrations (>119.8 pmol/l; the highest 5th centile for E2 by MS). Hence, IA performed especially poorly at low E2 concentrations and grossly overestimated them, as also seen in Figs 2b and 3b.

Discussion

Our study provides thus far the largest comparative data on testosterone and E2 measurements by IA and MS in serum samples of over 3000 men. Using MS as the comparator method for testosterone and E2 measurements, we can conclude that testosterone measurements by IA offer good accuracy at all concentrations found in eugonadal as well as hypogonadal men. In contrast, IA provides acceptable estimates of E2 only at the higher concentrations detected. Importantly, our data do not confirm that our platform IA for testosterone lack sensitivity and specificity in the hypogonadal range. The correlation of testosterone values between IA and MS measurements was high in the entire assay cohort, 0.93, and when testosterone level by MS was >8 nmol/l, 0.92. However, when testosterone concentrations were <8 nmol/l, the correlation was clearly lower, 0.69, indicating poorer accuracy of one or both of the methods. As compared with MS, the sensitivity of testosterone IA to detect at low testosterone concentrations (either <11 or <8 nmol/l) was 75–77%, and the specificity to detect a normal testosterone concentration was 96–98%.

The above figures are probably underestimates because the comparator MS method is not free of variability either (11, 12). Hence, the performance of both assays may contribute to the degradation of correlation. IAs have the known problems with antibody specificity, matrix effects, and lack of linearity and functional sensitivity. The various MS methods are not identical, use diverse procedures, calibrations and technologies, and are not totally free of influence of interfering substances. Because we omitted 150 samples from men with pituitary or testicular diseases and their treatments from the analysis, our measurements do not take into account all potential interferences in the clinical samples.

Interestingly, in the context of clinical management, when the sexual symptoms were combined with low testosterone concentrations to diagnose symptomatic LOH, the sensitivity of detection by IA increased from 75 to 85.5% with testosterone <11 nmol/l (+free testosterone <220 pmol/l) and further to 92.3% with testosterone <8 nmol/l, probably by eliminating the impact of some functionally irrelevant borderline or erroneous testosterone concentrations (between 8 and 11 nmol/l). This emphasizes the importance of combining testosterone concentration and symptoms in the diagnosis of LOH. We can thus conclude that the IA used in our study is sufficiently sensitive and specific to discriminate between normal and low testosterone concentrations in men suspected to have LOH. However, it has to be emphasized that the testosterone IA we used was of good quality, having passed with acceptable accuracy a rigorous standardization procedure (http://www.cdc.gov/labstandards/hs.html). All IAs used in clinical testosterone measurements are unlikely to have the same high quality.

A similar assessment of the E2 measurements did not reveal as good correlations as with testosterone. In the entire cohort, the IA/MS correlation was 0.76, and it was 0.32 with E2 concentrations <40.8 pmol/l and 0.74 at E2 levels >40.8 pmol/l. It is expected that the assay performance for E2 is worse at molar levels that are, on average, 0.4% of those of testosterone. In particular, the sensitivity of IA to detect low E2 concentrations was poor, at 13–25%. Accordingly, IA grossly overestimated the low E2 values. This seriously hampers the usefulness of the IA data on E2 at low concentrations. However, the sensitivity and specificity of IA to detect E2 concentrations in the highest 5th centile (>120 pmol/l) were acceptable (88.6 and 88.4% respectively).

Testosterone is still considered the standard assessment tool in the diagnostic approach of men with low bone density. However, with serum E2 concentrations being more closely associated with BMD than those of testosterone (14, 15) in men, and with MS-based assays allowing more accurate and sensitive measurements at low concentrations of E2, their measurement is becoming increasingly useful. When comparing the clinical applicability of E2 data in studies of BMD, Khosla et al. (27) concluded that although the MS data provide more accurate measurements in men, the applicability of the E2 IA data for bone data is generally valid. Hence, the necessity of switching E2 measurements from IA to MS is somewhat relative, admitting that the latter technique yields more accurate, but not necessarily clinically more useful results. Our data, however, show that serum testosterone and E2 concentrations are not highly correlated. A case can be made to develop clinical algorithms incorporating accurate measurement of E2 as part of the evaluation of osteoporosis in men. Moreover, recent epidemiologic studies in men and women have demonstrated associations between low sex hormone concentrations (including E2) and the risk of cardiovascular disease in both sexes (16, 17, 28, 29), suggesting another indication where more reliable methods for E2 measurement should be used. It should be acknowledged that in everyday clinical practice, the higher concentrations of E2 in men that may occur in gynecomastia and the rare cases of feminising tumors and aromatase excess can be discriminated with sufficient accuracy by IA.

Although MS, in general, is more specific and has lower intra- and interassay variability than IA, it faces similar inter-laboratory variability issues as IA (11, 12). All MS methods are not equal; like IA, they represent a heterogeneous group of measurements with significant differences in performance. One study comparing several established MS methods for the determination of testosterone in serum found overall CV of up to 33% at low concentrations, up to 15% at >1.5 nmol/l, and 1.4–11.4% at concentrations >3.5 nmol/l (11). Nevertheless, the variability in testosterone results with MS methods in most comparisons is substantially smaller than those reported for platform IAs (1, 5, 30, 31). Whether this difference translates into improved clinical relevance requires additional data and experience. A very recent study comparing total testosterone assays in women concluded that the results obtained by IA and MS were comparable, and there is significant variability and poor precision also between various MS methods at low levels (12). Hence, switching from IA to MS is not a guaranteed solution to improve the quality of sex steroid measurements at low concentrations. Improvements in performance and standardization in platform IAs are feasible alternatives that are already being implemented by some manufacturers. It is a major investment to abandon IA technology in favor of MS, and the reasons for this must be tangible and supported by evidence rather than conjecture. Our results suggest that, at least in men, IA can be as good as MS in the clinically important discrimination between eugonadal and hypogonadal men, especially when combined with clinical signs of androgen deficiency. The variability and imprecision of E2 measurements by MS is smaller than by IA, and it is clear that MS is superior to IA in the measurement of this hormone, especially at low concentrations. It seems prudent to conclude that the selection of an assay should be driven by the measurement performance in light of the clinical need and not by assay technology.

In conclusion, the comparison of measurements of serum testosterone and E2 in the largest cohort so far of adult male samples indicates that clinically relevant results on serum testosterone for the diagnosis of hypogonadism can be obtained both with well-validated IA and MS assays. Our findings do not support a mandatory requirement, on either analytical or clinical grounds, to switch from good-quality IAs to MS in the measurements of testosterone in male subjects. In contrast, clinicians should be aware of the unreliability of apparently low E2 results in men obtained by IA. Finally, assay performance is more important than assay technology.

Declaration of interest

I T Huhtaniemi consulted for Ferring Pharmaceuticals, Denmark. F C W Wu consulted for Bayer-Schering Healthcare, Germany; Akzo-Nobel (Organon), The Netherlands; Ferring Pharmaceuticals, Denmark; Pierre-Fabre Medicaments, France; Ardana Biosciences, UK; Procter & Gamble, United States; and Lilly-ICOS, United States, and has also received research grant funding support from Bayer-Schering Healthcare, Germany; Bayer Schering; Lilly-ICOS; and other companies. All other authors have nothing to declare.

Funding

This work was supported by the Commission of the European Communities Fifth Framework Programme, ‘Quality of Life and Management of Living Resources’ grant QLK6-CT-2001-00258. Dr S Boonen is a senior clinical investigator of the Fund for Scientific Research (Flanders, Belgium; F.W.O. – Vlaanderen) and holder of the Novartis Leuven University Chair in Gerontology and Geriatrics. Dr Vanderschueren is a senior clinical investigator supported by the Clinical Research Fund of the University Hospitals Leuven, Belgium.

Acknowledgements

The authors wish to thank the men who participated in the eight countries, the research/nursing staff in the eight centers: C Pott (Manchester), E Wouters (Leuven), M Nilsson (Malmö), M del Mar Fernandez (Santiago de Compostela), M Jedrzejowska (Lodz), H-M Tabo (Tartu) and A Heredi (Szeged) for their meticulous data collection, and C Moseley (Manchester) for data entry and project co-ordination. Non-financial support from Arthritis Research UK and the NIHR Manchester Biomedical Research Centre is acknowledged. The EMAS Group: Florence (G Forti, Luisa Petrone, Giovanni Corona); Leuven (D Vanderschueren, S Boonen, Herman Borghs); Lodz (K Kula, Jolanta Slowikowska-Hilczer, Renata Walczak-Jedrzejowska); London (I T Huhtaniemi); Malmö (A Giwercman); Manchester (F C W Wu, A J Silman, T W O'Neill, J D Finn, Philip Steer, A Tajar, D M Lee, Stephen Pye); Santiago (F F Casanueva, Mary Lage, Ana I Castro); Szeged (G Bartfai, Imre Földesi, Imre Fejes); Tartu (M Punab, Paul Korrovitz); Turku (Min Jiang).

References

  • 1

    Wang C, Catlin DH, Demers LM, Starcevic B, Swerdloff RS. Measurement of total serum testosterone in adult men: comparison of current laboratory methods versus liquid chromatography–tandem mass spectrometry. Journal of Clinical Endocrinology and Metabolism 2004 89 534543. doi:10.1210/jc.2003-031287.

    • Search Google Scholar
    • Export Citation
  • 2

    Moal V, Mathieu E, Reynier P, Malthiery Y, Gallois Y. Low serum testosterone assayed by liquid chromatography–tandem mass spectrometry. Comparison with five immunoassay techniques. Clinica Chimica Acta 2007 386 1219. doi:10.1016/j.cca.2007.07.013.

    • Search Google Scholar
    • Export Citation
  • 3

    Rosner W, Vesper H. Toward excellence in testosterone testing: a consensus statement. Journal of Clinical Endocrinology and Metabolism 2010 95 45424548. doi:10.1210/jc.2010-1314.

    • Search Google Scholar
    • Export Citation
  • 4

    Vesper HW, Botelho JC. Standardization of testosterone measurements in humans. Journal of Steroid Biochemistry and Molecular Biology 2010 121 513519. doi:10.1016/j.jsbmb.2010.03.032.

    • Search Google Scholar
    • Export Citation
  • 5

    Taieb J, Mathian B, Millot F, Patricot MC, Mathieu E, Queyrel N, Lacroix I, Somma-Delpero C, Boudou P. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography–mass spectrometry in sera from 116 men, women, and children. Clinical Chemistry 2003 49 13811395. doi:10.1373/49.8.1381.

    • Search Google Scholar
    • Export Citation
  • 6

    Sikaris K, McLachlan RI, Kazlauskas R, de Kretser D, Holden CA, Handelsman DJ. Reproductive hormone reference intervals for healthy fertile young men: evaluation of automated platform assays. Journal of Clinical Endocrinology and Metabolism 2005 90 59285936. doi:10.1210/jc.2005-0962.

    • Search Google Scholar
    • Export Citation
  • 7

    Blair IA. Analysis of estrogens in serum and plasma from postmenopausal women: past present, and future. Steroids 2010 75 297306. doi:10.1016/j.steroids.2010.01.012.

    • Search Google Scholar
    • Export Citation
  • 8

    Bhasin S, Cunningham GR, Hayes FJ, Matsumoto AM, Snyder PJ, Swerdloff RS, Montori VM. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology and Metabolism 2010 95 25362559. doi:10.1210/jc.2009-2354.

    • Search Google Scholar
    • Export Citation
  • 9

    Wartofsky L, Handelsman DJ. Standardization of hormonal assays for the 21st century. Journal of Clinical Endocrinology and Metabolism 2010 95 51415143. doi:10.1210/jc.2010-2369.

    • Search Google Scholar
    • Export Citation
  • 10

    Bhasin S, Zhang A, Coviello A, Jasuja R, Ulloor J, Singh R, Vesper H, Vasan RS. The impact of assay quality and reference ranges on clinical decision making in the diagnosis of androgen disorders. Steroids 2008 73 13111317. doi:10.1016/j.steroids.2008.07.003.

    • Search Google Scholar
    • Export Citation
  • 11

    Vesper HW, Bhasin S, Wang C, Tai SS, Dodge LA, Singh RJ, Nelson J, Ohorodnik S, Clarke NJ, Salameh WA, Parker CR Jr, Razdan R, Monsell EA, Myers GL. Interlaboratory comparison study of serum total testosterone (corrected) measurements performed by mass spectrometry methods. Steroids 2009 74 498503. doi:10.1016/j.steroids.2009.01.004.

    • Search Google Scholar
    • Export Citation
  • 12

    Legro RS, Schlaff WD, Diamond MP, Coutifaris C, Casson PR, Brzyski RG, Christman GM, Trussell JC, Krawetz SA, Snyder PJ, Ohl D, Carson SA, Steinkampf MP, Carr BR, McGovern PG, Cataldo NA, Gosman GG, Nestler JE, Myers ER, Santoro N, Eisenberg E, Zhang M, Zhang H. Total testosterone assays in women with polycystic ovary syndrome: precision and correlation with hirsutism. Journal of Clinical Endocrinology and Metabolism 2010 95 53055313. doi:10.1210/jc.2010-1123.

    • Search Google Scholar
    • Export Citation
  • 13

    Braunstein GD. Aromatase and gynecomastia. Endocrine-Related Cancer 1999 6 315324. doi:10.1677/erc.0.0060315.

  • 14

    Mellstrom D, Vandenput L, Mallmin H, Holmberg AH, Lorentzon M, Oden A, Johansson H, Orwoll ES, Labrie F, Karlsson MK, Ljunggren O, Ohlsson C. Older men with low serum estradiol and high serum SHBG have an increased risk of fractures. Journal of Bone and Mineral Research 2008 23 15521560. doi:10.1359/jbmr.080518.

    • Search Google Scholar
    • Export Citation
  • 15

    Gennari L, Khosla S, Bilezikian JP. Estrogen and fracture risk in men. Journal of Bone and Mineral Research 2008 23 15481551. doi:10.1359/jbmr.0810c.

    • Search Google Scholar
    • Export Citation
  • 16

    Callou de Sa EQ, de Sa FC, Silva RD, de Oliveira KC, Guedes AD, Feres F, Verreschi IT. Endogenous oestradiol but not testosterone is related to coronary artery disease in men sex hormones and coronary artery disease. Clinical Endocrinology 2011 75 177183. doi:10.1111/j.1365-2265.2011.04017.x.

    • Search Google Scholar
    • Export Citation
  • 17

    Jeon GH, Kim SH, Yun SC, Chae HD, Kim CH, Kang BM. Association between serum estradiol level and coronary artery calcification in postmenopausal women. Menopause 2010 17 902907. doi:10.1097/gme.0b013e3181d76768.

    • Search Google Scholar
    • Export Citation
  • 18

    Lee DM, O'Neill TW, Pye SR, Silman AJ, Finn JD, Pendleton N, Tajar A, Bartfai G, Casanueva F, Forti G, Giwercman A, Huhtaniemi IT, Kula K, Punab M, Boonen S, Vanderschueren D, Wu FC. The European Male Ageing Study (EMAS): design, methods and recruitment. International Journal of Andrology 2009 32 1124. doi:10.1111/j.1365-2605.2008.00879.x.

    • Search Google Scholar
    • Export Citation
  • 19

    Labrie F, Bélanger A, Bélanger P, Bérubé R, Martel C, Cusan L, Gomez J, Candas B, Castiel I, Chaussade V, Deloche C, Leclaire J. Androgen glucuronides, instead of testosterone, as the new markers of androgenic activity in women. Journal of Steroid Biochemistry and Molecular Biology 2006 99 182188. doi:10.1016/j.jsbmb.2006.02.004.

    • Search Google Scholar
    • Export Citation
  • 20

    Labrie F, Bélanger A, Bélager P, Bérubé R, Martel C, Cusan L, Gomez J, Candas B, Chaussade V, Castiel I, Deloche C, Leclaire J. Metabolism of DHEA in postmenopausal women following percutaneous administration. Journal of Steroid Biochemistry and Molecular Biology 2007 103 178188. doi:10.1016/j.jsbmb.2006.09.034.

    • Search Google Scholar
    • Export Citation
  • 21

    Swanson C, Lorentzon M, Vanderput L, Labrie F, Rane A, Jakobsson J, Chouinard S, Bélanger A, Ohlsson C. Sex steroid levels and cortical bone size in young men are associated with uridine diphosphate glucuronyltransferase 2B7 polymorphism (H268Y). Journal of Clinical Endocrinology and Metabolism 2007 92 36973704. doi:10.1210/jc.2007-0359.

    • Search Google Scholar
    • Export Citation
  • 22

    Bhasin S, Pencina M, Jasuja GK, Travison TG, Coviello A, Orwoll E, Wang PY, Nielson C, Wu F, Tajar A, Labrie F, Vesper H, Zhang A, Ulloor J, Singh R, D'Agostino R, Vasan RS. Reference ranges for testosterone in men generated using liquid chromatography tandem mass spectrometry in a community-based sample of healthy nonobese young men in the Framingham Heart Study and applied to three geographically distinct cohorts. Journal of Clinical Endocrinology and Metabolism 2011 96 24302439. doi:10.1210/jc.2010-3012.

    • Search Google Scholar
    • Export Citation
  • 23

    Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. Journal of Clinical Endocrinology and Metabolism 1999 84 36663672. doi:10.1210/jc.84.10.3666.

    • Search Google Scholar
    • Export Citation
  • 24

    Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986 1 307310. doi:10.1016/S0140-6736(86)90837-8.

    • Search Google Scholar
    • Export Citation
  • 25

    Kelly G. The influence function in the error in variables problem. Annals of Statistics 1984 12 87100. doi:10.1214/aos/1176346394.

  • 26

    Wu FC, Tajar A, Beynon JM, Pye SR, Silman AJ, Finn JD, O'Neill TW, Bartfai G, Casanueva FF, Forti G, Giwercman A, Han TS, Kula K, Lean ME, Pendleton N, Punab M, Boonen S, Vanderschueren D, Labrie F, Huhtaniemi IT. Identification of late-onset hypogonadism in middle-aged and elderly men. New England Journal of Medicine 2010 363 123135. doi:10.1056/NEJMoa0911101.

    • Search Google Scholar
    • Export Citation
  • 27

    Khosla S, Amin S, Singh RJ, Atkinson EJ, Melton LJ III, Riggs BL. Comparison of sex steroid measurements in men by immunoassay versus mass spectroscopy and relationships with cortical and trabecular volumetric bone mineral density. Osteoporosis International 2008 19 14651471. doi:10.1007/s00198-008-0591-5.

    • Search Google Scholar
    • Export Citation
  • 28

    Chen Y, Zeleniuch-Jacquotte A, Arslan AA, Wojcik O, Toniolo P, Shore RE, Levitz M, Koenig KL. Endogenous hormones and coronary heart disease in postmenopausal women. Atherosclerosis 2011 216 414419. doi:10.1016/j.atherosclerosis.2011.01.053.

    • Search Google Scholar
    • Export Citation
  • 29

    Ahmed B, Bairey Merz CN, Johnson BD, Bittner V, Berga SL, Braunstein GD, Hodgson TK, Smith K, Shaw L, Kelsey SF, Sopko G. Diabetes mellitus, hypothalamic hypoestrogenemia, and coronary artery disease in premenopausal women (from the National Heart, Lung, and Blood Institute sponsored WISE study). American Journal of Cardiology 2008 102 150154. doi:10.1016/j.amjcard.2008.03.029.

    • Search Google Scholar
    • Export Citation
  • 30

    Steinberger E, Ayala C, Hsi B, Smith KD, Rodriguez-Rigau LJ, Weidman ER, Reimondo GG. Utilization of commercial laboratory results in management of hyperandrogenism in women. Endocrine Practice 1998 4 110.

    • Search Google Scholar
    • Export Citation
  • 31

    Hsing AW, Stanczyk FZ, Belanger A, Schroeder P, Chang L, Falk RT, Fears TR. Reproducibility of serum sex steroid assays in men by RIA and mass spectrometry. Cancer Epidemiology, Biomarkers & Prevention 2007 16 10041008. doi:10.1158/1055-9965.EPI-06-0792.

    • Search Google Scholar
    • Export Citation

(The details of the EMAS Group are presented in the acknowledgement section)

 

     European Society of Endocrinology

Sept 2018 onwards Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 2171 453 30
PDF Downloads 1075 236 28
  • View in gallery

    Distribution of testosterone (panel a), and E2 (panel b) concentrations, as measured with MS and IA.

  • View in gallery

    Bland–Altman plots of testosterone (panel a) and E2 (panel b). Y-axis depicts the % difference between values of the two measurements (100×(IA−MS)/MS). The horizontal lines are 0 and −20 and +20%. The mean difference between the two assays in the testosterone values was 0.77% and of the E2 values, 30.1%.

  • View in gallery

    Scatter plots for testosterone (panel a) and E2 (panel b) with IA and MS, and the Deming regression results.

  • 1

    Wang C, Catlin DH, Demers LM, Starcevic B, Swerdloff RS. Measurement of total serum testosterone in adult men: comparison of current laboratory methods versus liquid chromatography–tandem mass spectrometry. Journal of Clinical Endocrinology and Metabolism 2004 89 534543. doi:10.1210/jc.2003-031287.

    • Search Google Scholar
    • Export Citation
  • 2

    Moal V, Mathieu E, Reynier P, Malthiery Y, Gallois Y. Low serum testosterone assayed by liquid chromatography–tandem mass spectrometry. Comparison with five immunoassay techniques. Clinica Chimica Acta 2007 386 1219. doi:10.1016/j.cca.2007.07.013.

    • Search Google Scholar
    • Export Citation
  • 3

    Rosner W, Vesper H. Toward excellence in testosterone testing: a consensus statement. Journal of Clinical Endocrinology and Metabolism 2010 95 45424548. doi:10.1210/jc.2010-1314.

    • Search Google Scholar
    • Export Citation
  • 4

    Vesper HW, Botelho JC. Standardization of testosterone measurements in humans. Journal of Steroid Biochemistry and Molecular Biology 2010 121 513519. doi:10.1016/j.jsbmb.2010.03.032.

    • Search Google Scholar
    • Export Citation
  • 5

    Taieb J, Mathian B, Millot F, Patricot MC, Mathieu E, Queyrel N, Lacroix I, Somma-Delpero C, Boudou P. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography–mass spectrometry in sera from 116 men, women, and children. Clinical Chemistry 2003 49 13811395. doi:10.1373/49.8.1381.

    • Search Google Scholar
    • Export Citation
  • 6

    Sikaris K, McLachlan RI, Kazlauskas R, de Kretser D, Holden CA, Handelsman DJ. Reproductive hormone reference intervals for healthy fertile young men: evaluation of automated platform assays. Journal of Clinical Endocrinology and Metabolism 2005 90 59285936. doi:10.1210/jc.2005-0962.

    • Search Google Scholar
    • Export Citation
  • 7

    Blair IA. Analysis of estrogens in serum and plasma from postmenopausal women: past present, and future. Steroids 2010 75 297306. doi:10.1016/j.steroids.2010.01.012.

    • Search Google Scholar
    • Export Citation
  • 8

    Bhasin S, Cunningham GR, Hayes FJ, Matsumoto AM, Snyder PJ, Swerdloff RS, Montori VM. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology and Metabolism 2010 95 25362559. doi:10.1210/jc.2009-2354.

    • Search Google Scholar
    • Export Citation
  • 9

    Wartofsky L, Handelsman DJ. Standardization of hormonal assays for the 21st century. Journal of Clinical Endocrinology and Metabolism 2010 95 51415143. doi:10.1210/jc.2010-2369.

    • Search Google Scholar
    • Export Citation
  • 10

    Bhasin S, Zhang A, Coviello A, Jasuja R, Ulloor J, Singh R, Vesper H, Vasan RS. The impact of assay quality and reference ranges on clinical decision making in the diagnosis of androgen disorders. Steroids 2008 73 13111317. doi:10.1016/j.steroids.2008.07.003.

    • Search Google Scholar
    • Export Citation
  • 11

    Vesper HW, Bhasin S, Wang C, Tai SS, Dodge LA, Singh RJ, Nelson J, Ohorodnik S, Clarke NJ, Salameh WA, Parker CR Jr, Razdan R, Monsell EA, Myers GL. Interlaboratory comparison study of serum total testosterone (corrected) measurements performed by mass spectrometry methods. Steroids 2009 74 498503. doi:10.1016/j.steroids.2009.01.004.

    • Search Google Scholar
    • Export Citation
  • 12

    Legro RS, Schlaff WD, Diamond MP, Coutifaris C, Casson PR, Brzyski RG, Christman GM, Trussell JC, Krawetz SA, Snyder PJ, Ohl D, Carson SA, Steinkampf MP, Carr BR, McGovern PG, Cataldo NA, Gosman GG, Nestler JE, Myers ER, Santoro N, Eisenberg E, Zhang M, Zhang H. Total testosterone assays in women with polycystic ovary syndrome: precision and correlation with hirsutism. Journal of Clinical Endocrinology and Metabolism 2010 95 53055313. doi:10.1210/jc.2010-1123.

    • Search Google Scholar
    • Export Citation
  • 13

    Braunstein GD. Aromatase and gynecomastia. Endocrine-Related Cancer 1999 6 315324. doi:10.1677/erc.0.0060315.

  • 14

    Mellstrom D, Vandenput L, Mallmin H, Holmberg AH, Lorentzon M, Oden A, Johansson H, Orwoll ES, Labrie F, Karlsson MK, Ljunggren O, Ohlsson C. Older men with low serum estradiol and high serum SHBG have an increased risk of fractures. Journal of Bone and Mineral Research 2008 23 15521560. doi:10.1359/jbmr.080518.

    • Search Google Scholar
    • Export Citation
  • 15

    Gennari L, Khosla S, Bilezikian JP. Estrogen and fracture risk in men. Journal of Bone and Mineral Research 2008 23 15481551. doi:10.1359/jbmr.0810c.

    • Search Google Scholar
    • Export Citation
  • 16

    Callou de Sa EQ, de Sa FC, Silva RD, de Oliveira KC, Guedes AD, Feres F, Verreschi IT. Endogenous oestradiol but not testosterone is related to coronary artery disease in men sex hormones and coronary artery disease. Clinical Endocrinology 2011 75 177183. doi:10.1111/j.1365-2265.2011.04017.x.

    • Search Google Scholar
    • Export Citation
  • 17

    Jeon GH, Kim SH, Yun SC, Chae HD, Kim CH, Kang BM. Association between serum estradiol level and coronary artery calcification in postmenopausal women. Menopause 2010 17 902907. doi:10.1097/gme.0b013e3181d76768.

    • Search Google Scholar
    • Export Citation
  • 18

    Lee DM, O'Neill TW, Pye SR, Silman AJ, Finn JD, Pendleton N, Tajar A, Bartfai G, Casanueva F, Forti G, Giwercman A, Huhtaniemi IT, Kula K, Punab M, Boonen S, Vanderschueren D, Wu FC. The European Male Ageing Study (EMAS): design, methods and recruitment. International Journal of Andrology 2009 32 1124. doi:10.1111/j.1365-2605.2008.00879.x.

    • Search Google Scholar
    • Export Citation
  • 19

    Labrie F, Bélanger A, Bélanger P, Bérubé R, Martel C, Cusan L, Gomez J, Candas B, Castiel I, Chaussade V, Deloche C, Leclaire J. Androgen glucuronides, instead of testosterone, as the new markers of androgenic activity in women. Journal of Steroid Biochemistry and Molecular Biology 2006 99 182188. doi:10.1016/j.jsbmb.2006.02.004.

    • Search Google Scholar
    • Export Citation
  • 20

    Labrie F, Bélanger A, Bélager P, Bérubé R, Martel C, Cusan L, Gomez J, Candas B, Chaussade V, Castiel I, Deloche C, Leclaire J. Metabolism of DHEA in postmenopausal women following percutaneous administration. Journal of Steroid Biochemistry and Molecular Biology 2007 103 178188. doi:10.1016/j.jsbmb.2006.09.034.

    • Search Google Scholar
    • Export Citation
  • 21

    Swanson C, Lorentzon M, Vanderput L, Labrie F, Rane A, Jakobsson J, Chouinard S, Bélanger A, Ohlsson C. Sex steroid levels and cortical bone size in young men are associated with uridine diphosphate glucuronyltransferase 2B7 polymorphism (H268Y). Journal of Clinical Endocrinology and Metabolism 2007 92 36973704. doi:10.1210/jc.2007-0359.

    • Search Google Scholar
    • Export Citation
  • 22

    Bhasin S, Pencina M, Jasuja GK, Travison TG, Coviello A, Orwoll E, Wang PY, Nielson C, Wu F, Tajar A, Labrie F, Vesper H, Zhang A, Ulloor J, Singh R, D'Agostino R, Vasan RS. Reference ranges for testosterone in men generated using liquid chromatography tandem mass spectrometry in a community-based sample of healthy nonobese young men in the Framingham Heart Study and applied to three geographically distinct cohorts. Journal of Clinical Endocrinology and Metabolism 2011 96 24302439. doi:10.1210/jc.2010-3012.

    • Search Google Scholar
    • Export Citation
  • 23

    Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. Journal of Clinical Endocrinology and Metabolism 1999 84 36663672. doi:10.1210/jc.84.10.3666.

    • Search Google Scholar
    • Export Citation
  • 24

    Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986 1 307310. doi:10.1016/S0140-6736(86)90837-8.

    • Search Google Scholar
    • Export Citation
  • 25

    Kelly G. The influence function in the error in variables problem. Annals of Statistics 1984 12 87100. doi:10.1214/aos/1176346394.

  • 26

    Wu FC, Tajar A, Beynon JM, Pye SR, Silman AJ, Finn JD, O'Neill TW, Bartfai G, Casanueva FF, Forti G, Giwercman A, Han TS, Kula K, Lean ME, Pendleton N, Punab M, Boonen S, Vanderschueren D, Labrie F, Huhtaniemi IT. Identification of late-onset hypogonadism in middle-aged and elderly men. New England Journal of Medicine 2010 363 123135. doi:10.1056/NEJMoa0911101.

    • Search Google Scholar
    • Export Citation
  • 27

    Khosla S, Amin S, Singh RJ, Atkinson EJ, Melton LJ III, Riggs BL. Comparison of sex steroid measurements in men by immunoassay versus mass spectroscopy and relationships with cortical and trabecular volumetric bone mineral density. Osteoporosis International 2008 19 14651471. doi:10.1007/s00198-008-0591-5.

    • Search Google Scholar
    • Export Citation
  • 28

    Chen Y, Zeleniuch-Jacquotte A, Arslan AA, Wojcik O, Toniolo P, Shore RE, Levitz M, Koenig KL. Endogenous hormones and coronary heart disease in postmenopausal women. Atherosclerosis 2011 216 414419. doi:10.1016/j.atherosclerosis.2011.01.053.

    • Search Google Scholar
    • Export Citation
  • 29

    Ahmed B, Bairey Merz CN, Johnson BD, Bittner V, Berga SL, Braunstein GD, Hodgson TK, Smith K, Shaw L, Kelsey SF, Sopko G. Diabetes mellitus, hypothalamic hypoestrogenemia, and coronary artery disease in premenopausal women (from the National Heart, Lung, and Blood Institute sponsored WISE study). American Journal of Cardiology 2008 102 150154. doi:10.1016/j.amjcard.2008.03.029.

    • Search Google Scholar
    • Export Citation
  • 30

    Steinberger E, Ayala C, Hsi B, Smith KD, Rodriguez-Rigau LJ, Weidman ER, Reimondo GG. Utilization of commercial laboratory results in management of hyperandrogenism in women. Endocrine Practice 1998 4 110.

    • Search Google Scholar
    • Export Citation
  • 31

    Hsing AW, Stanczyk FZ, Belanger A, Schroeder P, Chang L, Falk RT, Fears TR. Reproducibility of serum sex steroid assays in men by RIA and mass spectrometry. Cancer Epidemiology, Biomarkers & Prevention 2007 16 10041008. doi:10.1158/1055-9965.EPI-06-0792.

    • Search Google Scholar
    • Export Citation