Abstract
Context
There is a lack of understanding of what is normal in terms of sex steroid levels in older women.
Objective
To determine whether sex steroid levels vary with age in and establish reference ranges for women >70 years of age.
Design and Setting
Cross-sectional, community-based study.
Participants
Included 6392 women ≥70 years of age.
Main Outcome Measures
Sex steroids measured by liquid chromatography–tandem mass spectrometry. A reference group, to establish sex steroid age-specific reference ranges, excluded women using systemic or topical sex steroid, antiandrogen or glucocorticoid therapy, or an antiglycemic agent.
Results
The reference group of 5326 women had a mean age of 75.1 (±4.2) years, range of 70 to 94.7 years. Median values (range) were 181.2 pmol/L (3.7 to 5768.9) for estrone (E1), 0.38 nmol/L (0.035 to 8.56) for testosterone (T), 2.60 nmol/L (0.07 to 46.85) for dehydroepiandrosterone (DHEA), and 41.6 nmol/L (2.4 to 176.6) for SHBG. Estradiol and DHT were below method sensitivity in 66.1% and 72.7% of the samples, respectively. Compared with women aged 70 to 74 years, women aged ≥85 years had higher median levels of E1 (11.7%, P = 0.01), T (11.3%, P = 0.02), and SHBG (22.7%, P < 0.001) and lower DHEA (30% less, P < 0.001). Women with overweight and obesity had higher E1 (P < 0.001) and T (P < 0.03) and lower SHBG (P < 0.001) than did women with normal body mass index. Smokers had 17.2% higher median T levels (P = 0.005).
Conclusion
From the age of 70 years, T and E1 increase with age, despite a steady decline in DHEA. Whether E1 and T are biomarkers for longevity or contribute to healthy aging merits investigation.
This study shows that estrone and testosterone increase in women after 70 years of age and that testosterone levels in women aged 70 years or more are similar to levels measured in premenopausal women.
With the steady increase in the life expectancy of older women during the past four decades (1), understanding the factors that keep women healthy as they age is imperative to reduce the number of years lived with disability. Sex hormones are implicated as having a critical role in the development and evolution of age-associated disease, including cardiovascular disease and cancer, with the focus mainly on estrogens in women. Counterintuitively, testosterone (T), which circulates in higher concentrations than estrogens in women of all ages, may be as, or even more, important than estrogens in determining disease risk in elderly women (2).
The first step in advancing the understanding of androgens in older women’s health is establishing normal levels in a community-based sample of women. Until now, the available data have been limited by small sample sizes and/or the use of direct immunoassays that lack sensitivity and specificity for the measurement of T at the concentrations occurring in women, compared with the higher levels seen in men (3–6).
The Aspirin in Reducing Events in the Elderly (ASPREE) study was a placebo-controlled randomized clinical trial of daily low-dose aspirin vs placebo in older people, free of cardiovascular disease (CVD) events, with unimpaired cognition at recruitment. The large sample of women in this cohort has enabled us to determine whether sex steroids vary beyond the age of 70 years and, to our knowledge for the first time, establish age-specific reference ranges for each of the main sex steroids for community-dwelling women aged ≥70 years, using liquid chromatography–tandem mass spectrometry (LC-MS/MS) (7).
Methods
Study participants
ASPREE was a randomized clinical trial of aspirin (100 mg enteric coated daily) vs placebo in healthy older people. Details of the trial design have been published elsewhere (7, 8). In brief, Australian recruitment to ASPREE was achieved through partnerships with >2500 general practitioners across the southern states of Victoria, South Australia, and Tasmania. Australian participants were at least 70 years of age. Exclusion criteria included any chronic illnesses likely to limit survival to <5 years, documented CVD or cerebrovascular disease, dementia or a score of <78 on the Modified Mini-Mental State Examination (9), disability [severe difficulty or inability to perform any of the six Katz activities of daily living (10)], any condition associated with a high current or recurrent risk of bleeding, anemia, or uncontrolled high blood pressure (systolic ≥180 mm Hg and/or diastolic ≥105 mm Hg).
Clinical parameters
Date of birth, all concomitant medication use, and smoking and alcohol consumption were documented at randomization. Clinical measurements included weight, height, waist circumference, and systolic and diastolic blood pressure.
Sex steroid measurement
Blood samples were drawn at recruitment (or within 12 months) and plasma was stored under nitrogen vapor. Sex steroids and SHBG were measured in a single sample of plasma by LC-MS/MS at the ANZAC Research Institute, University of Sydney. T, DHT, dehydroepiandrosterone (DHEA), estradiol (E2), and estrone (E1) were quantified within a single run without derivatization as previously described (11) and with modifications (12).
Briefly, plasma (200 µL) spiked with 50 µL of internal standard (d3-T, d4-E2, d3-DHT, d2-DHEA) was extracted with 1 mL of methyl tert-butyl ether, separated by freezing to allow removal of the organic layer. After evaporation, the extract was reconstituted in 75 µL of 20% methanol so that 50 µL was injected into the LC-MS/MS system. Extracts of samples, standards, and quality controls were injected into a Shimadzu Nexera ultra-HPLC system comprising a Phenomenex Kinetex 1.7-µm XB C18 100 Å (50 × 2.1 mm) column with a Phenomenex C18 guard cartridge. The elution solvents were water (A) and methanol (B). A gradient elution was performed at a flow rate of 0.5 mL/min with 25% B (0 to 0.10 minutes), 52% to 62% B (0.11 to 4.30 minutes), 100% B (4.31 to 5.45 minutes), 25% B (5.46 to 7.00 minutes). The column temperature and autosampler were set at 40°C and 4°C, respectively. An AB Sciex API 5000 triple quad mass spectrometer was used with PhotoSpray (atmospheric pressure photoionization) ion source in positive (androgens) and negative (estrogens) polarity. Toluene was used as dopant, delivered at 0.05 mL/min. The ion source, curtain, and collision gas was nitrogen. Multiple reaction monitoring was used with both quadrupoles at unit resolution. Two mass transitions were monitored for each analyte (13). Certified reference materials were used for assay standards for T, DHT, and DHEA (National Measurement Institute, Sydney) and E2 and E1 (Cerillant). Internal standards used were stable isotopes: d3-T, d3-DHT, and d2-DHEA (National Measurement Institute) and d4-E2 (Cambridge Isotopes). For E1, the internal standard d4-E2 was used for quantitation purpose.
The assay limits of detection, limits of quantification, and within-run and between-run coefficients of variation are T (35 pmol/L, 0.09 nmol/L, 2.0%, 3.9% to 6.5%), DHT (0.17 nmol/L, 0.34 nmol/L, 8.1%, 6.7% to 13.4%), E2 (11 pmol/L, 18 pmol/L, 6.6%, 4.8% to 8.6%), E1 (3.7 pmol/L, 11 pmol/L, 4.7%, 4.6% to 7.5%), and DHEA (0.07 nmol/L, 0.17 nmol/L, <10%, <10%) (14). SHBG was measured in batches by automated immunoassay (Roche Diagnostics Australia) with a coefficient of variation of 1.0% to 2.0%. The median (range) for T and DHEA for menstruating, premenopausal women, aged 18 to 39 years (n = 602) (15) performed using the same assay are 0.34 nmol/L (0.04 to 1.0) and DHEA 4.91 nmol/L (0.08 to 23.51), respectively.
This study was approved by the Monash Human Research Ethics Committee (CF16/10—2016000001) and the Alfred Hospital Human Research Ethics Committee (616/15). All participants provided written informed consent to contribute biospecimens to the ASPREE Healthy Ageing Biobank.
Sample size and statistical analysis
We defined a reference group of women from within the study population to establish normative sex steroid values by age. ASPREE participants were excluded from the reference group when they were using any of the following at recruitment: any form of systemic or topical sex steroid therapy (estrogen, progestogen, tibolone, DHEA, or T), tamoxifen, or other selective estrogen receptor modulator, aromatase inhibitors, antiandrogen therapy (spironolactone or cyproterone acetate), glucocorticoid therapy, or any antiglycemic agent.
It was anticipated that the distribution of women with recruitment sex steroid values measured in the age groups 70 to 74 years, 75 to 79 years, 80 to 84 years, and ≥85 years and older would be approximately n = 3787, 1840, 824, and 268, respectively. The 80 to 84 year strata would have adequate precision to estimate sex steroid ranges with the mean estimated with a 95% CI width of +/−0.07 SD and the 2.5th and 97.5th percentiles would be estimated with approximate 95% CI widths of +/−0.12 SD. Means, SD, medians, and 10th to 90th percentile ranges are reported for each sex steroid in each of the age strata 70 to 74, 75 to 79, 80 to 84, and ≥85 years.
The Kruskal–Wallis test was used for comparison of the distribution of continuous variables across groups, and the χ2 test was used for categorical variables. Linear regression for each of E1, T, DHEA, and SHBG was done for the ln(analyte) with age, body mass index (BMI), and smoking as factors in the model and coefficients were exponentiated to have interpretation as ratios of adjusted geometric means (with a geometric mean being similar in value to a median).
Results
Recruitment to ASPREE commenced in 2010 and was completed by December 2014. Of the 16,703 Australian participants, 9180 (55%) were women. Of the women, 6392 had biobank samples available for measurement of sex steroids and SHBG (Fig. 1). The characteristics of these women and the subset of 5326 women in the reference group are shown in Table 1. The included and excluded participants were similar in age, ethnicity, smoking status, and blood pressure. The included women were aged 70 to 94.7 years and 98.9% were of European ancestry. They had a lower mean BMI (28.0 ± 5.0 vs 28.6 ± 5.0 kg/m2, P = 0.001) and smaller mean waist circumference than did the excluded women (92.6 ± 12.5 vs 94.4 ± 13.2 cm, P = 0.001).
Figure 1.
Flowchart of study participants. *Not including women who reported estrogen use.
Table 1.
Descriptive Statistics of Women Included in, and Excluded From, the Reference Group
| Included in Analysis (n = 5326, 83.3%) | Excluded From Analysis (n = 1066, 16.7%) | P Value | |
|---|---|---|---|
| Age, y | |||
| Mean (SD), range | 75.1 (4.2), 70–94.7 | 75.0 (4.2), 70.2–91.7 | |
| 70–74, n (%) | 3122 (58.6) | 654 (61.45) | 0.25 |
| 75–79, n (%) | 1442 (27.1) | 257 (24.1) | |
| 80-84, n (%) | 592 (11.1) | 121 (11.4) | |
| ≥85, n (%) | 170 (3.2) | 34 (3.2) | |
| Ethnicity, n (%) | 0.61 | ||
| European | 5267 (98.9) | 1053 (98.8) | |
| Asian | 30 (0.6) | 8 (0.7) | |
| Aboriginal/Torres Strait Islander | 5 (0.1) | 2 (0.2) | |
| Other | 24 (0.4) | 3 (0.3) | |
| Smoking status, n (%) | 0.37 | ||
| Current | 156 (2.93) | 23 (2.16) | |
| Former | 1683 (31.60) | 344 (32.27) | |
| Never | 3487 (65.47) | 699 (65.57) | |
| Weight, kg, mean (SD) | 71.0 (13.3) | 72.6 (13.9) | 0.0005 |
| Height, cm, mean (SD) | 1.59 (0.061) | 1.59 (0.058) | 0.77 |
| BMI, kg/m2, mean (SD) | 28.0 (5.0) | 28.6 (5.3) | 0.001a |
| <18.5, n (%) | 46 (0.87) | 7 (0.66) | 0.005b |
| 18.5 to <25, n (%) | 1533 (28.91) | 277 (26.01) | |
| 25 to <30, n (%) | 2128 (40.13) | 402 (37.75) | |
| ≥30, n (%) | 1596 (30.10) | 379 (35.59) | |
| Waist circumference, cm, mean (SD) | 92.6 (12.5) | 94.4 (13.2) | <0.001 |
| Systolic blood pressure, mm Hg mean (SD) | 141 (18) | 141 (18) | 0.71 |
| Diastolic blood pressure, mm Hg, mean (SD) | 78 (11) | 77 (11) | 0.18 |
Kruskal–Wallis test.
χ 2 test.
For the 5326 women included in the reference group, E2 and DHT were below the sensitivity of the assay method in 3522 (66.1%) and 3873 (72.7%) participants, respectively (Table 2). The proportion of women with serum E2 below the measurement limit increased with age from 64.8% in the 70 to 74 years group to 72.3% in women ≥85 years of age (P < 0.001). The proportion of women with serum DHT below the assay sensitivity declined with age from 73.8% in 70- to 74-year-old women to 58.2% in the ≥85 years of age group (P < 0.001). For the 1083 women with measurable E2, the median value was 22.03 pmol/L (range, 11.0 to 1373.0). Because of the relatively small numbers of women with values above the assay sensitivity for E2 and DHT, these steroid levels were not analyzed further.
Table 2.
Proportion of Women With Measurable Sex Steroid Levels by Age
| 70–74 y of Age | 75–79 y of Age | 80–85 y of Age | ≥85 y of Age | Total | |
|---|---|---|---|---|---|
| E1 | |||||
| Result available | 3093 (99.1%) | 1425 (98.8%) | 589 (99.5%) | 170 (100.0%) | 5277 (99.1%) |
| Below LOD | 29 (0.9%) | 17 (1.2%) | 3 (0.5%) | 0 (0.0%) | 49 (0.9%) |
| Estradiola | |||||
| Result available | 1097 (35.2%) | 480 (33.3%) | 179 (30.2%) | 47 (27.7%) | 1803 (33.9%) |
| Below LOD | 2024 (64.8%) | 962 (66.7%) | 413 (69.8%) | 123 (72.3%) | 3522 (66.1%) |
| Total T | |||||
| Result available | 3073 (98.4%) | 1417 (98.3%) | 588 (99.3%) | 170 (100.0%) | 5248 (98.5%) |
| Below LOD | 49 (1.6%) | 25 (1.7%) | 4 (0.7%) | 0 (0.0%) | 78 (1.5%) |
| DHT | |||||
| Result available | 817 (26.2%) | 390 (27.1%) | 175 (29.6%) | 71 (41.8%) | 1453 (27.3%) |
| Below LOD | 2305 (73.8%) | 1052 (72.9%) | 417 (70.4%) | 99 (58.2%) | 3873 (72.7%) |
| DHEA | |||||
| Result available | 3107 (99.5%) | 1424 (98.8%) | 590 (99.7%) | 170 (100.0%) | 5291 (99.3%) |
| Not detected | 15 (0.5%) | 18 (1.2%) | 2 (0.3%) | 0 (0.0%) | 35 (0.7%) |
| SHBGb | |||||
| Result available | 3113 (99.8%) | 1441 (99.9%) | 592 (100.0%) | 169 (100.0%) | 5315 (99.9%) |
| Below LOD | 5 (0.2%) | 1 (0.1%) | 0 (0.0%) | 0 (0.0%) | 6 (0.1%) |
Abbreviation: LOD, limit of detection.
One sample missing.
Five samples missing.
Details for E1, total T, DHEA, and SHBG by age groups are provided in Table 3 and shown graphically in Fig. 2. Median steroid values (range) for the total reference group were 181.2 pmol/L (3.7 to 5768.9) for E1, 0.38 nmol/L (0.035 to 8.56) for total T, 2.60 nmol/L (0.069 to 46.85) for DHEA, and 41.6 nmol/L (2.4 to 176.6) for SHBG. As seen in Fig. 2, there were a number of outliers for each steroid in each age group. We examined whether any of the measured variables (BMI, weight, waist circumference, smoking) and any reported medication use predicted extreme outliers for any of the steroids, but we could not identify a common explanation for the outliers.
Table 3.
Sex Steroids by Age Group in the Reference Group of Women
| 70–74 y of Age | 75–79 y of Age | 80–85 y of Age | ≥85 y of Age | Total | |
|---|---|---|---|---|---|
| E1, n (pmol/L)a | 3093 | 1425 | 589 | 170 | 5277 |
| Mean (SD) | 210.8 (214.2) | 213.7 (194.3) | 220.1 (114.9) | 223.4 (121.5) | 213.0 (197.6) |
| Median | 181.2 | 181.2 | 196.0 | 205.2 | 181.2 |
| Minimum | 3.7 | 3.7 | 3.7 | 14.8 | 3.7 |
| 10th percentile | 88.7 | 85.0 | 99.8 | 83.2 | 88.7 |
| 90th percentile | 340.2 | 351.3 | 362.4 | 369.8 | 347.6 |
| Maximum | 5768.9 | 4733.4 | 894.9 | 717.4 | 5768.9 |
| T, n (nmol/L)b | 3073 | 1417 | 588 | 170 | 5248 |
| Mean (SD) | 0.45 (0.45) | 0.46 (0.37) | 0.49 (0.44) | 0.52 (0.50) | 0.47 (0.43) |
| Median | 0.35 | 0.35 | 0.38 | 0.38 | 0.38 |
| Minimum | 0.03 | 0.03 | 0.03 | 0.03 | 0.03 |
| 10th percentile | 0.17 | 0.17 | 0.17 | 0.17 | 0.17 |
| 90th percentile | 0.76 | 0.87 | 0.90 | 1.11 | 0.83 |
| Maximum | 8.56 | 4.75 | 5.37 | 4.16 | 8.56 |
| DHEA, n (nmol/L)b | 3107 | 1424 | 590 | 170 | 5291 |
| Mean (SD) | 3.25 (2.19) | 3.02 (2.20) | 2.87 (1.99) | 2.35 (1.61) | 3.11 (2.16) |
| Median | 2.74 | 2.50 | 2.32 | 2.05 | 2.60 |
| Minimum | 0.07 | 0.07 | 0.07 | 0.07 | 0.07 |
| 10th percentile | 1.11 | 0.94 | 0.97 | 0.82 | 1.04 |
| 90th percentile | 6.11 | 5.73 | 6.07 | 4.15 | 6.00 |
| Maximum | 46.85 | 27.38 | 10.03 | 9.82 | 46.85 |
| SHBG, n (nmol/L) | 3113 | 1441 | 592 | 169 | 5315 |
| Mean (SD) | 42.9 (18.2) | 45.5 (19.3) | 49.3 (19.5) | 54.6 (21.9) | 44.7 (19.0) |
| Median | 39.8 | 42.8 | 46.3 | 51.5 | 41.6 |
| Minimum | 2.8 | 2.4 | 13.4 | 16.6 | 2.4 |
| 10th percentile | 23.5 | 24.3 | 27.2 | 29.4 | 24.2 |
| 90th percentile | 66.2 | 70.0 | 75.2 | 83.2 | 68.9 |
| Maximum | 167.6 | 151.2 | 155.3 | 124.2 | 167.6 |
To convert pmol/L to pg/ml, divide by 3.699.
To convert nmol/L to ng/dL, divide by 0.0347.
Figure 2.
Relationship between age and individual sex steroids for the reference group. In the box-and-whisker plots (outliers not included), the box represents the interquartile range (IQR) and the line in the box is the median. The whiskers extend to the upper and lower adjacent values. The upper adjacent value is defined as the largest data point less than or equal to the 75th percentile + 1.5 × IQR. The lower adjacent value is defined as the smallest data point greater than or equal to the 25th percentile − 1.5 × IQR. Outliers are any values beyond the whiskers. The raw data are represented as scatter graphs with fitted locally weighted scatterplot smoothing curves on a log (base 10) scale. DHEAS, DHEA sulfate.
The following multivariable analysis results are derived from the ratios of adjusted geometric means, as described in Methods. Compared with women aged 70 to 74 years, women aged 80 to 84 and ≥85 years had higher E1 levels [respectively higher by 9.2% (P = 0.001) and 11.7% (P = 0.01)]. Women with overweight and obesity had higher E1 levels than did normal weight women (respectively higher by 14.6% and 34.1%, P < 0.001). T levels were significantly higher in women 80 to 84 years and ≥85 years of age than for women 70 to 74 years of age [higher, respectively, by 9.3% (P = 0.004) and 11.3% (P = 0.02)] (Table 4). Women with overweight and obesity also had higher T levels [5.2% (P = 0.03) and 5.5% (P = 0.03), respectively]. Current smokers had 17.2% higher T levels than did nonsmokers (P = 0.005). SHBG increased with age from 70 to 74 years, being 5.6% higher in women aged 75 to 79 years, 13.6% higher in women 80 to 84 years of age, and 22.7% higher in women ≥85 years of age (P < 0.001 for all). Being overweight was associated with a 15% lower SHBG and being obese a 27% lower SHBG level (P < 0.001). Current smokers had SHBG levels 7.6% higher than those of nonsmokers (P = 0.023). In multivariable linear regression, E1 was not independently associated with SHBG when BMI and age were taken into account. T was independently, positively associated with SHBG (P < 0.001) when BMI, age, and smoking were included in the model (data not shown).
Table 4.
Factors Contributing to Sex Steroid Levels
| Determinants Evaluated | E1 Univariable Analysis | E1 Multivariable Analysis | T Univariable Analysis | T Multivariable Analysis | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Effect of Determ. | 95% CI | P Value | Effect of Determ. | 95% CI | P Value | Effect of Determ. | 95% CI | P Value | Effect of Determ. | 95% CI | P Value | |
| Age, y | ||||||||||||
| 70–74 | 1.00 | 1.00 | 1.00 | |||||||||
| 75–79 | 1.0008 | 0.965 to 1.038 | 0.96 | 0.999 | 0.964 to 1.036 | 0.99 | 1.016 | 0.973 to 1.061 | 0.48 | 1.014 | 0.971 to 1.059 | 0.527 |
| 80–84 | 1.070 | 1.017 to 1.127 | 0.010 | 1.092 | 1.038 to 1.148 | 0.001 | 1.085 | 1.021 to 1.153 | 0.008 | 1.093 | 1.029 to 1.162 | 0.004 |
| ≥85 | 1.072 | 0.980 to 1.174 | 0.13 | 1.117 | 1.022 to 1.220 | 0.014 | 1.119 | 1.007 to 1.244 | 0.037 | 1.133 | 1.019 to 1.260 | 0.021 |
| BMI, kg/m2 | ||||||||||||
| <18.5 | 1.015 | 0.856 to 1.202 | 0.87 | 1.002 | 0.846 to 1.188 | 0.98 | 0.941 | 0.769 to 1.150 | 0.553 | 0.925 | 0.757 to 1.131 | 0.446 |
| 18.5 to <25 | 1.00 | 1.00 | 1.00 | 1.00 | ||||||||
| 25 to <30 | 1.142 | 1.100 to 1.186 | <0.001 | 1.146 | 1.103 to 1.190 | <0.001 | 1.048 | 1.001 to 1.096 | 0.045 | 1.052 | 1.006 to 1.101 | 0.028 |
| ≥30 | 1.331 | 1.279 to 1.386 | <0.001 | 1.341 | 1.288 to 1.396 | <0.001 | 1.045 | 0.996 to 1.097 | 0.072 | 1.055 | 1.005 to 1.108 | 0.029 |
| Current smoker | ||||||||||||
| Yes | 1.017 | 0.926 to 1.118 | 0.722 | 1.058 | 0.964 to 1.160 | 0.233 | 1.153 | 1.032 to 1.287 | 0.012 | 1.172 | 1.049 to 1.309 | 0.005 |
| No | 1.00 | 1.00 | 1.00 | |||||||||
| DHEA Univariable Analysis | DHEA Multivariable Analysis | SHBG Univariable Analysis | SHBG Multivariable Analysis | |||||||||
| Effect of Determ. | 95% CI | P Value | Effect of Determ. | 95% CI | P Value | Effect of Determ. | 95% CI | P Value | Effect of Determ. | 95% CI | P Value | |
| Age, y | ||||||||||||
| 70–74 | 1.00 | 1.00 | 1.00 | 1.00 | ||||||||
| 75–79 | 0.897 | 0.858 to 0.937 | <0.001 | 0.894 | 0.856 to 0.935 | <0.001 | 1.056 | 1.029 to 1.084 | <0.001 | 1.056 | 1.030 to 1.082 | <0.001 |
| 80–84 | 0.858 | 0.807 to 0.912 | <0.001 | 0.859 | 0.808 to 0.914 | <0.001 | 1.161 | 1.119 to 1.204 | <0.001 | 1.136 | 1.097 to 1.177 | <0.001 |
| ≥85 | 0.706 | 0.633 to 0.786 | <0.001 | 0.705 | 0.633 to 0.786 | <0.001 | 1.277 | 1.198 to 1.362 | <0.001 | 1.227 | 1.154 to 1.305 | <0.001 |
| BMI, kg/m2 | ||||||||||||
| <18.5 | 1.221 | 0.990 to 1.504 | 0.061 | 1.249 | 1.015 to 1.538 | 0.036 | 1.094 | 0.973 to 1.231 | 0.133 | 1.073 | 0.955 to 1.206 | 0.238 |
| 18.5 to <25 | 1.00 | 1.00 | 1.00 | 1.00 | ||||||||
| 25 to <30 | 1.055 | 1.007 to 1.105 | 0.024 | 1.052 | 1.004 to 1.101 | 0.032 | 0.844 | 0.823 to 0.867 | <0.001 | 0.848 | 0.826 to 0.870 | <0.001 |
| ≥30 | 1.039 | 0.989 to 1.092 | 0.127 | 1.027 | 0.978 to 1.079 | 0.289 | 0.722 | 0.702 to 0.743 | <0.001 | 0.730 | 0.710 to 0.751 | <0.001 |
| Current smoker | ||||||||||||
| Yes | 1.048 | 0.936 to 1.174 | 0.418 | 1.028 | 0.918 to 1.151 | 0.630 | 1.103 | 1.032 to 1.179 | 0.004 | 1.076 | 1.010 to 1.147 | 0.023 |
| No | 1.00 | 1.00 | ||||||||||
Abbreviation: Determ., determinant.
DHEA levels showed a steady decline with age such that women aged ≥85 years had 30% less DHEA than did women aged 70 to 74 years (P < 0.001). DHEA levels were significantly higher in women who were underweight (25% higher, P = 0.036) and overweight (5.2% higher, P = 0.032).
When compared with the reference group, current glucocorticoid users (n = 188) had significantly lower levels of DHEA (median, 1.04 nmol/L; range, 0.07 to 8.61; P = 0.0001), T (median, 0.024 nmol/L; range, 0.03 to 1.84; P = 0.0001) and E1 (median 107.1; range, 3.7 to 6212.6; P = 0.0001). Glucocorticoid users had a lower median SHBG of borderline statistical significance (median, 39.2 nmol/L; range, 1.4 to 237.6; P = 0.055).
Discussion
This large, cross-sectional study of circulating sex steroid levels in older women free of severe illness demonstrates small, but steady, increases in circulating T and E1 in women from the age of 70 years, despite a steady decline in DHEA, and that circulating T levels in women aged ≥70+ years are similar to levels in healthy premenopausal women measured by the same assay. This study also highlights the importance of E1 as the major circulating estrogen in older, postmenopausal women and possibly a detrimental impact of glucocorticoid use.
The T levels in the women in this study did not differ from levels seen in premenopausal women, measured by the same LC-MS/MS assay, and exhibited a small, but statistically significant, difference across age groups. We previously observed, in a study of women aged 18 to 75 years, a nadir for total T, measured by sensitive immunoassay, for women between 62 and 63 years of age, followed by a small increase beyond that in women up to age 75 years (3). Laughlin et al. (16) observed a similar, statistically significant, positive association between T and age in their study of women with a mean age of 73.8 years. In contrast, an apparent decline in total T, measured by immunoassay, in women between 65 and 80 years was reported in a sample of 347 women with a mean age of 74 years (6). Having not included women <70 years of age, the current study neither supports or refutes our prior observation of a nadir in total T in women in their early 60s. It does, however, establish that T levels in older women are maintained in the setting of a 30% decline in the primary precursor of T, DHEA. As in younger women, circulating levels of DHEA are several fold greater than that of T (4), indicating that biosynthesis of T is dependent on enzyme activity, not precursor availability. The determinants of extragonadal T biosynthesis by conversion of circulating precursors of adrenal origin in women are not known. However, as in our previous study (3), and as has been reported for men (17), tobacco smokers had significantly higher T levels than did nonsmokers. Whether this is due to an effect of smoking on T biosynthesis, secretion, or metabolism is not known.
In postmenopausal women, DHEA and its derivative androstenedione are the major source of circulating E1 through aromatization of androstenedione in extragonadal tissues, primarily adipose (18). Adipose aromatase gene expression increases with age in women (19, 20). Therefore, the capacity for adipose tissue to biosynthesize E1 increases with age. The steady increase in E1 levels with increasingly older groups of women in the current study, independent of BMI and SHBG, is, to our knowledge, a new finding that indicates aromatase activity continues to increase with age even in elderly women. With the loss of ovarian E2 production, in postmenopausal women E1 is an important precursor for peripheral E2 biosynthesis, and T is further converted to DHT, with both E2 and DHT being further metabolized intracellularly (18, 21). Hence, in most women concentrations of E2 and DHT were below the limit of detection, consistent with the intracrine production and metabolism of both hormones occurring within the tissues in which they act, with serum levels arising from unregulated spillover from tissues into the circulation (18).
The increasing proportion of women with unmeasurable E2 in the older groups most likely reflects different effects of age on the enzymatic pathways essential for the biosynthesis of these hormones. Regardless, a key message from the findings is that studies investigating the association between estrogens and diseases of aging in postmenopausal women must measure E1 to provide meaningful findings.
The positive association between age and SHBG seen in the current study has previously been reported in a study of SHBG across the lifespan (22). As in the current study, we have previously reported a positive, independent association between SHBG and T in postmenopausal women (23). This most likely reflects the high proportion of circulating T that is SHBG bound. SHBG appears to be metabolically important, with low SHBG identified as an independent marker of insulin resistance and type 2 diabetes risk (24). SHBG has also been implicated in the pathogenesis of type 2 diabetes and CVD (24, 25). We have previously demonstrated strong, independent, and highly statistically significant inverse associations between both insulin resistance, estimated by the homeostatic model of insulin resistance, and SHBG, and between BMI and SHBG (23). These associations were independent of sex steroids. The greater SHBG levels in older women may reflect a metabolic survivorship advantage of women who have higher SHBG.
Agreeing with previous reports, we found that glucocorticoid users had significantly lower levels of T (6), as well as E1 and DHEA, than did nonusers. The low level of DHEA in glucocorticoid users reflects adrenal suppression, but it might also be a consequence of their underlying disease. This finding does, however, support the importance of DHEA for E1 and T biosynthesis.
Strengths of this study include the large community-based study sample providing high statistical precision for estimating sex steroid ranges. This cohort is likely to be as representative as any other healthy volunteer-based study. Measurement of sex steroids by gold standard LC-MS/MS is an important study strength. Sensitivity for E2 was a limitation. Although a more sensitive method to measure E2 was available, given the time and labor demands of the more sensitive technique, it was not feasible for a study of this magnitude. The LC-MS/MS method used for this study does not allow for androstenedione (analyzed in positive polarity) and E2 (analyzed in negative polarity) to be measured in the same run, as the two steroids elute at virtually the same retention time. Our method includes fast polarity switching but cannot cope with coeluting compounds with opposite polarity. However, in postmenopausal women virtually all androstenedione is of adrenal origin. Therefore, its parent steroid DHEA provides a strong overall index of adrenal androgen production.
Our earlier study (3) and that of Cappola (6) reported much higher T levels in older women than those measured with LC-MS/MS in the current study, consistent with known issues of cross-reactivity and less specificity of the older RIAs used previously (26, 27). To explore the physiology of aging, we excluded women who reported use of medications that would influence sex steroid measurement or metabolism. We excluded women taking antiglycemic medications to rule out women with type 2 diabetes requiring treatment, to minimize the effects of insulin resistance on sex steroid levels through effects on SHBG. A number of women in the reference group had sex steroid levels suggestive of exogenous hormone use. We examined all medication use in the outliers for each steroid to exclude any unexpected cross-reactivity in assays and found none. Although the higher sex steroid levels found in a few women did not fit our expectations, without evidence of exogenous steroid use, we could not justify their exclusion from the analysis. This decision is supported by prior smaller studies that have reported unexpectedly high E1 and androgen levels in otherwise well postmenopausal women (3, 6, 28–30). The high levels observed in some women in our study reaffirm the wide range of “normality” within a community-based population, and they support the representativeness of our study sample. A limitation of our study is that we are inferring the association between hormone levels and age from cross-sectional data. Whereas the optimal study design would be longitudinal, it would be prohibitively difficult and vulnerable to attrition bias. Our sample, being mostly of European ancestry, matches that of the Australian population of this age (31). However, our findings cannot be extrapolated to women of non-European ancestries. Information about past hysterectomy or oophorectomy was not collected, such that the impact of oophorectomy on hormone levels in older women could not be examined.
In summary, we have successfully established normal ranges for sex steroids in women aged ≥70 years of age. Important observations are that in women aged ≥70 years circulating T levels do not differ meaningfully from those of premenopausal women, and together with E1, appear to increase from the age of 70 years. Concurrently, DHEA levels exhibit a progressive decline from the age of 70 years. E1, T, and SHBG may be biomarkers for longevity or contribute to healthy aging.
Acknowledgments
Bayer AG provided aspirin and matching placebo and had no other role in the trial. The authors acknowledge the dedicated and skilled staff in Australia and the United States for the conduct of the trial and the ASPREE Investigator Group listed at www.aspree.org. The authors are most grateful to the ASPREE participants, who so willingly volunteered for this study, and the general practitioners and medical clinics who supported the participants in the ASPREE study.
Financial Support: The ASPREE trial was supported by the National Institute on Aging and the National Cancer Institute at the National Institutes of Health (Grant U01 AG029824); the National Health and Medical Research Council (NHMRC) of Australia (Grants 34047 and 1127060); Monash University (Australia); and the Victorian Cancer Agency (Australia). The ASPREE Healthy Ageing Biobank was funded by the Commonwealth Scientific and Industrial Research Organisation (Flagship Grant), the National Cancer Institute (Grant U01 AG029824), and Monash University. This analysis of sex hormones was funded by NHMRC of Australia Project Grant 1105305. S.R.D. is an Australian NHMRC Senior Principal Research Fellow (Grant 1135843).
Clinical Trial Information: International Standard Randomized Controlled Trial Number Register ISRCTN83772183 (registered 14 July 2005) and ClinicalTrials.gov no. NCT01038583 (registered 24 December 2009).
Author Contributions: Literature search: S.R.D. Figures: S.R.D. and P.J.R. Study design: S.R.D., R.J.B., C.M.R., M.R.N., J.J.M., R.L.W., A.M.M., R.S.W., and J.E.L. Data collection and biobank management: T.G., J.P., R.L.W., and J.E.L. Biochemical analysis: D.J.H. and R.D. Data analysis: P.J.R., R.J.B., and R.S.W. Data interpretation: S.R.D., R.J.B., D.J.H., and R.L.W. Writing: S.R.D., P.J.R., and R.J.B. Manuscript review: all co-authors.
Additional Information
Disclosure Summary: S.R.D. reports having received honoraria from Besins Healthcare and Pfizer Australia and has been a consultant to Mayne Pharmaceuticals, Lawley Pharmaceuticals, and Que Oncology. D.J.H. has received institutional grant funding (but no personal income) for investigator-initiated clinical testosterone pharmacology studies (Lawley, Besins Healthcare) and has provided expert testimony to anti-doping and professional standards tribunals and testosterone litigation. M.R.N. reports receiving travel support from Bayer and fees for serving on an advisory board from Sanofi. A.M.M. has received travel funds from Bayer Pharmaceuticals. The remaining authors have nothing to disclose.
Data Availability: Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.
Glossary
Abbreviations:
- ASPREE
Aspirin in Reducing Events in the Elderly
- BMI
body mass index
- CVD
cardiovascular disease
- DHEA
dehydroepiandrosterone
- E1
estrone
- E2
estradiol
- LC-MS/MS
liquid chromatography–tandem mass spectrometry
- T
testosterone
References and Notes
- 1. Australian Institute of Health and Welfare. Life expectancy and death in Autstraila. Available at: www.aihw.gov.au/reports/life-expectancy-death/deaths-in-australia/contents/life-expectancy. Accessed 27 September 2019.
- 2. Wahlin-Jacobsen S, Pedersen AT, Kristensen E, Laessoe NC, Lundqvist M, Cohen AS, Hougaard DM, Giraldi A. Is there a correlation between androgens and sexual desire in women? J Sex Med. J Sex Med. 2015;12(2):358–373. [DOI] [PubMed] [Google Scholar]
- 3. Davison SL, Bell R, Donath S, Montalto JG, Davis SR. Androgen levels in adult females: changes with age, menopause, and oophorectomy. J Clin Endocrinol Metab. 2005;90(7):3847–3853. [DOI] [PubMed] [Google Scholar]
- 4. Labrie F, Bélanger A, Cusan L, Gomez J-L, Candas B. Marked decline in serum concentrations of adrenal C19 sex steroid precursors and conjugated androgen metabolites during aging. J Clin Endocrinol Metab. 1997;82(8):2396–2402. [DOI] [PubMed] [Google Scholar]
- 5. Haring R, Hannemann A, John U, Radke D, Nauck M, Wallaschofski H, Owen L, Adaway J, Keevil BG, Brabant G. Age-specific reference ranges for serum testosterone and androstenedione concentrations in women measured by liquid chromatography-tandem mass spectrometry. J Clin Endocrinol Metab. 2012;97(2):408–415. [DOI] [PubMed] [Google Scholar]
- 6. Cappola AR, Ratcliffe SJ, Bhasin S, Blackman MR, Cauley J, Robbins J, Zmuda JM, Harris T, Fried LP. Determinants of serum total and free testosterone levels in women over the age of 65 years. J Clin Endocrinol Metab. 2007;92(2):509–516. [DOI] [PubMed] [Google Scholar]
- 7. McNeil JJ, Woods RL, Nelson MR, Murray AM, Reid CM, Kirpach B, Storey E, Shah RC, Wolfe RS, Tonkin AM, Newman AB, Williamson JD, Lockery JE, Margolis KL, Ernst ME, Abhayaratna WP, Stocks N, Fitzgerald SM, Trevaks RE, Orchard SG, Beilin LJ, Donnan GA, Gibbs P, Johnston CI, Grimm RH; ASPREE Investigator Group. Baseline characteristics of participants in the ASPREE (ASPirin in Reducing Events in the Elderly) study. J Gerontol A Biol Sci Med Sci. 2017;72(11):1586–1593. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. ASPREE Investigator Group. Study design of ASPirin in Reducing Events in the Elderly (ASPREE): a randomized, controlled trial. Contemp Clin Trials. 2013;36(2):555–564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Teng EL, Chui HC. The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry. 1987;48(8):314–318. [PubMed] [Google Scholar]
- 10. Katz S, Akpom CA. A measure of primary sociobiological functions. Int J Health Serv. 1976;6(3):493–508. [DOI] [PubMed] [Google Scholar]
- 11. Harwood DT, Handelsman DJ. Development and validation of a sensitive liquid chromatography–tandem mass spectrometry assay to simultaneously measure androgens and estrogens in serum without derivatization. Clin Chim Acta. 2009;409(1–2):78–84. [DOI] [PubMed] [Google Scholar]
- 12. Keski-Rahkonen P, Desai R, Jimenez M, Harwood DT, Handelsman DJ. Measurement of estradiol in human serum by LC-MS/MS using a novel estrogen-specific derivatization reagent. Anal Chem. 2015;87(14):7180–7186. [DOI] [PubMed] [Google Scholar]
- 13. Desai R, Harwood DT, Handelsman DJ. Simultaneous measurement of 18 steroids in human or mouse serum by liquid chromatography–mass spectrometry without derivatization to profile the classical and alternate pathways of androgen synthesis and metabolism. Clin Mass Spectrom. 2019;11:42–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hsu B, Cumming RG, Hirani V, Blyth FM, Naganathan V, Le Couteur DG, Seibel MJ, Waite LM, Handelsman DJ. Temporal trend in androgen status and androgen-sensitive outcomes in older men. J Clin Endocrinol Metab. 2016;101(4):1836–1846. [DOI] [PubMed] [Google Scholar]
- 15. Skiba MA, Bell RJ, Islam RM, Davis SR. Challenges in recruitment to an epidemiological study of young Australian women: the Grollo-Ruzzene Foundation Young Women’s Health Study. Aust N Z J Public Health. 2019;43(2):131–136. [DOI] [PubMed] [Google Scholar]
- 16. Laughlin GA, Goodell V, Barrett-Connor E. Extremes of endogenous testosterone are associated with increased risk of incident coronary events in older women. J Clin Endocrinol Metab. 2010;95(2):740–747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Shiota M, Kashiwagi E, Murakami T, Takeuchi A, Imada K, Inokuchi J, Tatsugami K, Eto M. Serum testosterone before and during androgen-deprivation therapy, and prognosis between cigarette smokers and nonsmokers with metastatic prostate cancer. Andrologia. 2018;50(10):e13119. [DOI] [PubMed] [Google Scholar]
- 18. Simpson ER, Davis SR. Minireview: aromatase and the regulation of estrogen biosynthesis—some new perspectives. Endocrinology. 2001;142(11):4589–4594. [DOI] [PubMed] [Google Scholar]
- 19. Bulun SE, Simpson E. Competitive reverse transcription-polymerase chain reaction analysis indicates that levels of aromatase cytochrome P450 transcripts in adipose tissue of buttocks, thighs, and abdomen of women increase with advancing age. J Clin Endocrinol Metab. 1994;78:428–432. [DOI] [PubMed] [Google Scholar]
- 20. Misso ML, Jang C, Adams J, Tran J, Murata Y, Bell R, Boon WC, Simpson ER, Davis SR. Adipose aromatase gene expression is greater in older women and is unaffected by postmenopausal estrogen therapy. Menopause. 2005;12(2):210–215. [DOI] [PubMed] [Google Scholar]
- 21. Labrie F, Luu-The V, Labrie C, Bélanger A, Simard J, Lin SX, Pelletier G. Endocrine and intracrine sources of androgens in women: inhibition of breast cancer and other roles of androgens and their precursor dehydroepiandrosterone. Endocr Rev. 2003;24(2):152–182. [DOI] [PubMed] [Google Scholar]
- 22. Handelsman DJ, Sikaris K, Ly LP. Estimating age-specific trends in circulating testosterone and sex hormone-binding globulin in males and females across the lifespan. Ann Clin Biochem. 2016;53(Pt 3):377–384. [DOI] [PubMed] [Google Scholar]
- 23. Davis SR, Robinson PJ, Moufarege A, Bell RJ. The contribution of SHBG to the variation in HOMA-IR is not dependent on endogenous oestrogen or androgen levels in postmenopausal women. Clin Endocrinol (Oxf). 2012;77(4):541–547. [DOI] [PubMed] [Google Scholar]
- 24. Ding EL, Song Y, Manson JE, Hunter DJ, Lee CC, Rifai N, Buring JE, Gaziano JM, Liu S. Sex hormone-binding globulin and risk of type 2 diabetes in women and men. N Engl J Med. 2009;361(12):1152–1163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Peter A, Kantartzis K, Machann J, Schick F, Staiger H, Machicao F, Schleicher E, Fritsche A, Häring HU, Stefan N. Relationships of circulating sex hormone-binding globulin with metabolic traits in humans. Diabetes. 2010;59(12):3167–3173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Krasowski MD, Drees D, Morris CS, Maakestad J, Blau JL, Ekins S. Cross-reactivity of steroid hormone immunoassays: clinical significance and two-dimensional molecular similarity prediction. BMC Clin Pathol. 2014;14(1):33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Groenestege WM, Bui HN, ten Kate J, Menheere PP, Oosterhuis WP, Vader HL, Heijboer AC, Janssen MJ. Accuracy of first and second generation testosterone assays and improvement through sample extraction. Clin Chem. 2012;58(7):1154–1156. [DOI] [PubMed] [Google Scholar]
- 28. Fuhrman BJ, Xu X, Falk RT, Dallal CM, Veenstra TD, Keefer LK, Graubard BI, Brinton LA, Ziegler RG, Gierach GL. Assay reproducibility and interindividual variation for 15 serum estrogens and estrogen metabolites measured by liquid chromatography–tandem mass spectrometry. Cancer Epidemiol Biomarkers Prev. 2014;23(12):2649–2657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Falk RT, Brinton LA, Dorgan JF, Fuhrman BJ, Veenstra TD, Xu X, Gierach GL. Relationship of serum estrogens and estrogen metabolites to postmenopausal breast cancer risk: a nested case-control study. Breast Cancer Res. 2013;15(2):R34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Trabert B, Brinton LA, Anderson GL, Pfeiffer RM, Falk RT, Strickler HD, Sliesoraitis S, Kuller LH, Gass ML, Fuhrman BJ, Xu X, Wentzensen N. Circulating estrogens and postmenopausal ovarian cancer risk in the women’s health initiative observational study. Cancer Epidemiol Biomarkers Prev. 2016;25(4):648–656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Australian Institute of Health and Welfare. Older Australia at a glance. Available at: www.aihw.gov.au/reports/older-people/older-australia-at-a-glance/contents/diverse-groups-of-older-australians/culturally-and-linguistically-diverse-people. Accessed 27 September 2019.