Abstract
Objective
Male aging is characterized by a decline in testosterone (TS) levels with a substantial variability between subjects. However, it is unclear whether differences in age-related changes in TS are associated with general health. We investigated associations between mortality and intra-individual changes in serum levels of total TS, SHBG, free TS and LH during a ten-year period with up to 18 years of registry follow-up.
Design
1167 men aged 30–60 years participating in the Danish Monitoring Trends and Determinants of Cardiovascular Disease (MONICA1) study and who had a follow-up examination ten years later (MONICA10) were included. From MONICA10, the men were followed up to 18 years (mean: 15.2 years) based on the information from national mortality registries via their unique personal ID numbers.
Methods
Cox proportional hazard models were used to investigate the association between intra-individual hormone changes and all-cause, CVD and cancer mortalities.
Results
A total of 421 men (36.1%) died during the follow-up period. Men with most pronounced decline in total TS (<10th percentile) had a higher all-cause mortality risk compared to men within the 10th to 90th percentile (hazard ratio (HR): 1.60; 95% confidence interval (CI): 1.08–2.36). No consistent associations were seen in cause-specific mortality analyses.
Conclusion
Our study showed that higher mortality rates were seen among the men who had the most pronounced age-related decline in TS, independent of their baseline TS levels.
Introduction
Testosterone (TS) is essential for male reproductive function. Besides stimulating Sertoli cell function and spermatogenesis, TS exerts effects in nonreproductive organs and stimulates bone mineralization, muscle growth, erythropoiesis and cognitive function (1, 2, 3). In men, aging is characterized by a physiological decline in the TS level with a substantial variability between subjects (4). The decline is paralleled by an age-related increase in sex hormone-binding globulin (SHBG), the primary binding protein for TS; thus, a more pronounced decline in the free fraction of TS occurs with increasing age (5, 6).
Recently, a number of large prospective studies have shown that lifestyle and health changes are associated with changes in TS levels (4, 7, 8). Weight gain is closely associated with an accelerated decline in total TS, free TS and SHBG, and deterioration in health also contributes to this decline (4). Interestingly, smoking men have higher TS levels compared to nonsmoking men, and smoking cessation is associated with a decrease in the TS level (4, 7). Thus, age-related changes in TS seem to be modified by changes in health and lifestyle factors indicating that the physiological decline in TS with aging is to some degree preventable.
Previously, two studies in elderly men have investigated whether changes in TS levels and related hormone levels were associated with all-cause mortality, however with conflicting results (9, 10). In Australian men, a decrease in TS levels was associated with all-cause mortality (9), whereas no consistent association was seen in American men (10).
We hypothesized that the degree of age-related changes in testosterone may be of clinical relevance and a possible risk marker of health also in younger men. In the present study, we therefore examined whether intra-individual changes in levels of total and free TS, SHBG and LH over a ten-year period were associated with all-cause mortality in 1167 men aged 30–60 years at baseline and on whom we had follow-up information from national mortality registries up to 18 years later. This is, to our knowledge, the first population-based study to relate intra-individual changes in reproductive hormone levels to mortality in men includingthose as young as 30 years of age.
Subjects and methods
Study population
The study included 1930 men who participated in the Danish Monitoring Trends and Determinants of Cardiovascular Disease 1 (MONICA1) study conducted between November 1982 and February 1984 and for whom serum samples were available for hormone analyses. The MONICA1 study was part of an international project initiated by the World Health Organization with the aim of monitoring trends and determinants of cardiovascular disease (11, 12). Men born in 1922, 1932, 1942 or 1952 were randomly selected from 11 municipalities in the southwestern part of the Copenhagen area and the participation rate was 81%. The men were invited to participate in a follow-up study between 1993 and 1994 (MONICA10). Follow-up serum samples were available for hormone analyses from 1167 men agreeing to participate (Fig. 1). The study protocol for the health examination was the same at baseline and follow-up and included a physical examination, and a self-administered questionnaire with detailed information on lifestyle, sociodemographic characteristics and health status. Furthermore, a blood sample was drawn in the morning after an overnight fast. Serum samples were stored in aliquots at −20°C until the hormone analysis was performed. Informed consent was obtained from all participants.
Lexis diagram reflecting the study design.
Citation: European Journal of Endocrinology 178, 1; 10.1530/EJE-17-0280
Lexis diagram reflecting the study design.
Citation: European Journal of Endocrinology 178, 1; 10.1530/EJE-17-0280
Lexis diagram reflecting the study design.
Citation: European Journal of Endocrinology 178, 1; 10.1530/EJE-17-0280
Hormone measurements
To be able to include as many reproductive hormones as possible, all samples were analyzed using the same assay platform due to a limited amount of serum available; TS was measured by time-resolved fluorimmunoassay (DELFIA; Wallac Oy, Turku, Finland) and SHBG and LH were measured by time-resolved immunofluorometric assay (DELFIA; Wallac Oy). Limits of detection were 0.23 nmol/L for TS and SHBG and 0.05 IU/L for LH. Intra- and interassay coefficients of variation were less than 12% for TS, less than 8% for SHBG and less than 6% for LH. All samples were analyzed for the three hormones in the same laboratory, within the same period in 2004. Thus, samples from the different surveys had been stored for a varying period (10–22 years) before the hormone analysis. Furthermore, the samples from the different surveys have, to a varying degree, been thawed and analyzed for other purposes prior to the hormone analyses, whereas others have been kept in the freezer exclusively. Therefore, the hormone levels were initially corrected for evaporation according to a general correction factor based on median serum Na+ levels measured in a subset of samples from each of the two surveys as previously extensively validated (13). Free TS was calculated based on the TS and SHBG concentrations according to the method suggested by Vermeulen et al. with the assumption of an average albumin concentration of 43.0 g/L (14).
Follow-up and outcome definition
In Denmark, every citizen has a unique personal identification number (CPR number) through which information from national registries can be obtained. Based on the CPR numbers of the participants, we obtained information on outcomes from The Central Office of Civil Registration for information on vital status with complete follow-up. Follow-up time was calculated as the time from the follow-up examination MONICA10 until the time of death, censoring or end of follow-up (December 17, 2012), whichever occurred first (Fig. 1). Participants who emigrated during the follow-up period (n = 8) contributed with person time at risk until date of emigration, after which they were censored. Information about specific causes of death was obtained from The Danish Register of Causes of Death. A minor proportion of deaths (n = 56) was not registered with a specific cause of death and was therefore censored in the analyses of CVD and cancer mortalities. Finally, information on hospital admissions related to CVD prior to baseline was obtained from the National Patient Register in which all admissions to Danish hospitals since 1977 are registered. Thus, men who had been diagnosed with either ischemic heart disease, stroke or atherosclerosis (ICD-8: 410–414, 430–438, 440; ICD-10: G45, I20–I25, I60–I70) prior to the baseline examination were excluded in the analysis of CVD mortality (n = 66).
Statistical analysis
Initially, the changes (delta) per year in total and free TS, SHBG and LH were calculated as the difference between the hormone levels at the follow-up examination and the baseline examination divided by the time span between the two examinations (e.g. TSMONICA10 – TSMONICA1)/time). Thus, negative delta (Δ) values reflect a decline, whereas positive numbers reflect an increase over time. The marginal association between the hormone changes per year and the baseline hormone levels was investigated in plots and the Pearson correlation coefficient was reported. Furthermore, mean differences in baseline and delta Δ hormone values per ten years stratified by lifestyle characteristics were investigated using general linear regression. The baseline level of LH was log-transformed to obtain normality and geometric means were reported.
To investigate the association of Δ hormone levels in relation to all-cause, CVD and cancer mortalities, we used Cox proportional hazard models with age as the underlying time scale. To accommodate potential nonlinear associations, Δ hormone levels were analyzed in tertiles and in percentiles (<10th, 10th–90th, >90th). Potential confounders were identified based on prior literature and confirmed in our data. All final models were adjusted for the baseline hormone level, baseline age (30, 40, 50 or 60 years), smoking habits in MONICA1 to MONICA10 (nonsmoker–nonsmoker, nonsmoker–smoker, smoker–smoker or smoker–nonsmoker), baseline BMI (<20.0, 20.0–24.9, 25.0–29.9 or >30), changes in weight from MONICA1 to MONICA10 (>−5 kg, +/−5 kg or >+5 kg), exercise level in MONICA1 to MONICA10 (sedentary–sedentary, sedentary–active, active–active, active–sedentary), baseline alcohol intake per week (0, 1–14 or +14) and changes in the number of alcohol units per week between MONICA1 and MONICA10. In the fully adjusted models, 29 (2.5%) individuals were excluded because of one or more missing values and the tertiles and percentiles were recalculated. P-values <0.05 were considered statistically significant. All models were tested for the assumption of proportional hazards. Furthermore, we tested for interaction between the Δ hormone values and each of the covariates including the baseline hormone level. In sensitivity analyses, we restricted the population to include only eugonadal men at baseline defined as men with TS level >10.5 nmol/L and LH <9.4 IU/L (n = 1089) as suggested in the European Male Aging Study (EMAS) (15) and repeated the Cox proportional hazard models and compared the results with the models without restrictions. The statistical analyses were performed using R (version 3.2.3).
Results
A detailed description of baseline hormone levels and changes in hormone levels in relation to lifestyle characteristics is shown in Table 1. Mean time span between the two examinations (MONICA1 and MONICA10) was 10.9 years (min–max: 9.7–12.0). The estimated mean (SEM) intra-individual percentage changes in hormone levels per year were −1.5% (0.16) for total TS, 0.9% (0.18) for SHBG, −1.9% (0.39) for free TS and 1.0% (0.27) for LH. In comparison, when estimated cross-sectionally, based on baseline hormone levels, mean percentage changes in hormone levels per year were −0.4% (0.06) for total TS, 1.2% (0.08) for SHBG, −1.1% for free TS (0.05) and 1.1% (0.12) for LH.
Mean baseline levels and absolute changes per 10 years (Δ) in hormones stratified according to lifestyle characteristics (n = 1167).
| Total TS (nmol/L) | SHBG (nmol/L) | Free TS (pmol/L) | LH (IU/L) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Baseline | Δb per 10 year | Baseline | Δb per 10 year | Baseline | Δb per 10 year | Baselinea | Δb per 10 year | |
| Age (years) | |||||||||
| 30c | 305 | 22.7 | −5.1 | 28.3 | 0.6 | 545.9 | −146.3 | 3.0 | 0.0 |
| 40 | 329 | 20.4d | −4.1d | 29.7 | 1.2 | 464.7d | −116.7d | 3.0 | 0.1 |
| 50 | 304 | 19.3d | −2.9d | 34.0d | 3.3d | 401.1d | −89.1d | 3.4d | 0.2 |
| 60 | 229 | 20.1d | −3.1d | 40.1d | 3.3d | 383.3d | −81.6d | 3.9d | 0.5d |
| Baseline BMI (kg/m2) | |||||||||
| <20 | 42 | 25.7d | −4.1 | 40.9d | 1.0 | 520.5 | −109.8 | 3.3 | 0.7 |
| 20–25c | 581 | 22.4 | −4.5 | 34.7 | 1.1 | 484.2 | −121.6 | 3.2 | 0.1 |
| 25–30 | 441 | 18.9d | −3.3d | 30.2d | 2.8d | 424.2d | −100.5d | 3.4 | 0.2 |
| 30+ | 103 | 16.1d | −2.6d | 26.4d | 3.7d | 377.6d | −89.1d | 3.1 | 0.0 |
| Absolute changes in weight (kg) from baseline to follow-up examination | |||||||||
| >−5 | 82 | 18.5d | −1.8d | 31.3 | 9.5d | 405.4d | −97.2 | 3.6 | 0.7d |
| +/−5c | 663 | 20.6 | −3.2 | 33.4 | 3.3 | 444.9 | −99.7 | 3.3 | 0.1 |
| >+5 | 420 | 21.2 | −5.4d | 31.3d | −1.6d | 476.4d | −129.8d | 3.1 | 0.1 |
| Changes in smoking habits from baseline to follow-up examination | |||||||||
| ns-nsc | 497 | 19.4 | −3.4 | 30.8 | 1.9 | 431.7 | −98.0 | 3.2 | 0.2 |
| ns-s | 34 | 18.4 | −1.5d | 27.6 | 4.9d | 427.9 | −74.8 | 3.2 | 0.5 |
| s-s | 483 | 22.2d | −4.0d | 34.6d | 2.9 | 477.0d | −117.6d | 3.4d | 0.1 |
| s-ns | 153 | 20.6 | −5.5d | 32.6 | −1.1d | 455.3 | −135.6d | 3.1 | 0.2 |
| Alcohol intake (units/week) | |||||||||
| 0 | 59 | 18.9d | −2.5d | 34.5 | 2.4 | 394.8d | −70.5d | 3.1 | 0.7 |
| 1–14c | 749 | 21.0 | −4.0 | 33.2 | 1.5 | 455.8 | −108.0 | 3.3 | 0.2 |
| <15 | 359 | 20.3 | −3.9 | 30.8d | 2.9d | 458.0 | −121.7 | 3.3 | 0.0 |
| Leisure time physical activity | |||||||||
| Sedentary | 220 | 20.3 | −3.6 | 31.1 | 3.5 | 450.4 | −114.0 | 3.4 | 0.2 |
| Activec | 947 | 20.8 | −3.9 | 32.8 | 1.6d | 454.1 | −109.5 | 3.2 | 0.1 |
| Known CVD at baseline | |||||||||
| Noc | 1101 | 20.7 | −3.9 | 32.5 | 1.8 | 455.9 | −110.7 | 3.7 | 0.1 |
| Yes | 66 | 19.6 | −3.5 | 33.2 | 4.7d | 412.4d | −104.1 | 3.9 | 0.2 |
Δ hormone levels in relation to baseline hormone levels
Higher baseline TS level was associated with a more pronounced absolute decline in TS per year (r = −0.52, P < 0.01, Supplementary Fig. 1, see section on supplementary data given at the end of this article) which seemed to be explained by age as the youngest men were characterized by higher baseline level and experienced the largest decline (Table 1). Similarly, higher level of free TS was strongly associated with a more pronounced decline in free TS per year (r = −0.69, P < 0.01). A weak correlation was seen for SHBG (r = −0.12, P < 0.01) and LH (r = −0.20, P < 0.01).
Baseline and Δ hormone levels in relation to lifestyle factors
At baseline, an inverse association between BMI and levels of total and free TS and SHBG was found. Men with higher BMI had a less pronounced decline in total and free TS. Men who gained more than 5 kg from baseline to follow-up examination had the most pronounced decline in total and free TS compared to men who had a stable weight (+/−5 kg) and men who lost more than 5 kg. Furthermore, men who gained more than 5 kg in weight showed a decrease in SHBG in contrast to the other two groups who were characterized by an increase in SHBG with aging.
Men who were smokers at both time points were characterized by significantly higher baseline hormone levels compared to the other groups. Smoking cessation was associated with the most pronounced decline in total and free TS and was associated with a minor decrease in SHBG (whereas mean SHBG increased from baseline to follow-up among the other groups). The higher the number of alcohol units per week, the lower the baseline SHBG level, and the more pronounced decline in free TS was seen. No significant differences in baseline hormone levels were seen according to baseline leisure time physical activity. Finally, the baseline level of free TS was lower among men with a cardiovascular disease at baseline and they experienced a larger increase in SHBG within the time period.
Mortality
Mean follow-up from MONICA10 until the end of follow-up was 15.2 years and up to 18.6 years. A total of 421 men (36.1%) died during the follow-up period of which 106 cases were cancer-related and 119 were CVD-related.
Figure 2 shows the Cox proportional hazard models for intra-individual changes in hormone levels in relation to all-cause mortality, where Δ hormone levels were categorized into either tertiles or percentiles (<10th, 10th–90th, >90th). In both models, the middle category served as the reference category. A significant increased risk in all-cause mortality was observed for the tertile of men with the most pronounced decline in total TS, corresponding to an annual decline of at least −0.6 nmol/L (HR, 1.34, 95% CI, 1.02–1.78; for specific model estimates, see Supplementary Table 1). A more pronounced effect was seen for men in the lowest 10th percentile (HR, 1.60, 95% CI, 1.08–2.38) compared to the reference category. No significant difference in all-cause mortality was seen for tertiles of Δ SHBG levels. However, men with Δ SHBG levels below the 10th percentile (<−0.7 nmol/L/year) or above the 90th percentile (>1.1 nmol/L/year) had an increased mortality compared to the middle group (HR, 1.51, 95% CI, 1.05–2.17; HR, 1.42, 95% CI 1.08–1.88, respectively). In accordance with the findings for total TS, men with the most pronounced decline in free TS had an increased risk of all-cause mortality, however only significant in the tertile model (HR, 1.45, 95% CI, 1.09–1.92). No association was seen for differences in Δ LH levels and Δ TS/LH levels in relation to all-cause mortality.
Hazard ratios and 95% confidence intervals for tertiles and percentiles (<10th, 10th–90th, >90th) of delta hormone levels in relation to all-cause mortality. All models are adjusted for baseline hormone level, baseline age, changes in smoking status, baseline BMI, weight changes, exercise level and baseline and changes in alcohol consumption with age as the underlying time scale.
Citation: European Journal of Endocrinology 178, 1; 10.1530/EJE-17-0280
Hazard ratios and 95% confidence intervals for tertiles and percentiles (<10th, 10th–90th, >90th) of delta hormone levels in relation to all-cause mortality. All models are adjusted for baseline hormone level, baseline age, changes in smoking status, baseline BMI, weight changes, exercise level and baseline and changes in alcohol consumption with age as the underlying time scale.
Citation: European Journal of Endocrinology 178, 1; 10.1530/EJE-17-0280
Hazard ratios and 95% confidence intervals for tertiles and percentiles (<10th, 10th–90th, >90th) of delta hormone levels in relation to all-cause mortality. All models are adjusted for baseline hormone level, baseline age, changes in smoking status, baseline BMI, weight changes, exercise level and baseline and changes in alcohol consumption with age as the underlying time scale.
Citation: European Journal of Endocrinology 178, 1; 10.1530/EJE-17-0280
In analyses with CVD mortality as the outcome, no significant associations were seen (Supplementary Table 2). Similarly, in analyses with cancer as the outcome, no significant associations were seen (Supplementary Table 3). Finally, similar results were seen when we restricted the population to men who at baseline were eugonadal (data not shown).
Discussion
In this large population-based follow-up study of men aged 30 or older, we observed that higher mortality rates were seen among those who had the most pronounced decline in TS levels. This association was independent of age, baseline hormone levels and lifestyle factors.
Few other studies have looked at intra-individual declines in TS levels in aging men in relation to mortality. In line with our findings, a study of older Australian men observed that progressive declines in testosterone were associated with increased all-cause mortality (9), whereas a study of elderly men participating in the Framingham Heart Study did not find any consistent association between changes in total TS and all-cause mortality (10). However, the findings of the latter study may be explained by the older age range in this population and a reduced statistical power caused by a limited study size (n = 254) and number of cases (n = 104). Furthermore, no adjustments for weight changes and changes in smoking habits were performed, yet both these lifestyle factors are known to have effects on circulating levels of TS.
The association between serum TS, measured at a single time point, and all-cause mortality has been investigated in a number of population-based studies, where some (15, 16, 17, 18, 19, 20, 21, 22, 23) but not all (24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35) concluded that low TS is associated with an increased all-cause mortality risk. Based on 11 community-based studies, a meta-analysis from 2011 found a significant association between a single point measurement of low TS and all-cause mortality but also identified significant heterogeneity between studies (36). In a population of 5350 men from the age of 30 years with hormone measurements at a single time point, we did not see an association between low TS and all-cause mortality but showed that compensated Leydig cell function, characterized by elevated LH levels and LH/TS levels and normal TS levels, was associated with increased all-cause mortality (30).
It has been suggested that men with a compensated Leydig cell function may be at a greater risk of developing an overt primary hypogonadism (characterized by low TS levels and elevated LH levels) with age (15). A link between impaired testicular function and increased morbidity and mortality was recently also indicated in three studies showing associations between lower semen quality and increased morbidity and mortality risk (37, 38, 39). However, while we, in the present study, observed an association between the decline in the age-related TS and free TS and mortality, we did not observe any association between age-related changes in LH levels and all-cause mortality as would be expected if the association for TS was linked to a primary Leydig cell insufficiency.
In accordance with previous longitudinal studies, we observed a close link between weight changes and changes in levels of total TS and free TS where men experiencing weight loss were characterized by the least pronounced decline in TS and weight-gaining men were characterized by an accelerated decline in TS compared to age-matched men with stable weight. Interestingly, weight-gaining men were characterized by a decline in SHBG compared to stable weight men and men who lost weight for whom an age-related increase in SHBG was observed.
In our analyses of all-cause mortality, a U-shaped association was seen when categorizing Δ SHBG levels into percentiles. However, no association was seen when categorizing Δ SHBG levels into tertiles in line with the findings by Hsu et al., where no association was seen after adjustment for confounders (9). It is well known that SHBG levels increase with age but also smokers have higher levels, whereas lower levels are seen in the case of obesity. However, in fully adjusted models, the inclusion of age, BMI, smoking and other lifestyle factors as confounders did not change the estimates significantly. We cannot rule out that men with the most extreme SHBG trajectories, either below the 10th percentile or above the 90th percentile, have a higher mortality. Alternatively, the lower power when categorizing according to percentiles could have resulted in chance findings. Thus, more studies are needed to confirm these findings.
Our study has a number of strengths including the comprehensive study population with repeated measurements on 1167 men. In addition, besides being part of a longitudinal study with a ten-year interval, the men were further followed via registries for up to 18 years. Finally, the same hormone assays were used to measure the hormone levels from both the baseline and the follow-up examination reducing potential interassay variation. One of the major methodological limitations inherent in longitudinal studies is subject attrition which can lead to selection bias and reduced generalizability. In our study, 1167 men attended the follow-up examination out of the 1930 men (60.5%) participating in the baseline examination. It is possible that men who participated in the follow-up examination differ in lifestyle and health characteristics compared to men who did not participate in the follow-up examination. As expected, in subanalyses (Supplementary Table 4), we found that men who did not survive until the follow-up examination (n = 208) were older compared to the men who participated in the follow-up examination and also more likely to be smokers. Therefore, the men who participated in our longitudinal study population to some degree represent a slightly younger and possibly healthier part of the original study population which could have biased our findings toward less pronounced changes in TS levels compared to the original baseline population. TS levels were analyzed using fluorimmunoassays (DELFIA). For sex steroids, tandem mass spectrometry has been recommended because of higher specificity than immunoassays (40). We have subsequently compared the testosterone DELFIA assay to a turboflow-liquid chromatography tandem mass spectrometry method (established in our laboratory in 2012) and a strong correlation between the two methods was found in the relevant ranges (data not published). Thus, methods of measurements are highly unlikely to have any influence of the obtained results.
In conclusion, our findings suggest that a more pronounced age-related decline in TS is associated with mortality in men independent of baseline hormone level. A possible causal link between an increased tempo in age-related TS decline and subsequent health is unknown and remains to be investigated.
Supplementary data
This is linked to the online version of the paper at https://doi.org/10.1530/EJE-17-0280.
Declaration of interest
The authors declare that there is no conflicts of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
This study was financially supported by the Kirsten and Freddy Johansen’s Foundation and financial support for part of the study was obtained from the following foundations: ‘Snedkermester Sophus Jacobsen og hustru Astrid Jacobsens Fond’, ‘Arvid Nilssons Fond’ and ‘Grosserer Valdemar Foersom og hustru Thyra Foersoms Fond’.
References
- 1↑
Weinbauer GF, Luetjens CM, Simoni M & Nieschlag E. Physiology of testicular function. In Andrology, Male Reproductive Health and Dysfunction, edn 3rd, ch 2, pp 11–60. Eds Nieschlag E, Behre HM, Nieschlag S. Berlin, Heidelberg: Springer, 2009. (https://doi.org/10.1007/978-3-540-78355-8)
- 2↑
Finkelstein JS, Lee H, Burnett-Bowie SAM, Pallais JC, Yu EW, Borges LF, Jones BF, Barry CV, Wulczyn KE & Thomas BJ et al. Gonadal steroids and body composition, strength, and sexual function in men. New England Journal of Medicine 2013 369 1011–1022. (https://doi.org/10.1056/NEJMoa1206168)
- 3↑
Finkelstein JS, Lee H, Leder BZ, Burnett-Bowie SAM, Goldstein DW, Hahn CW, Hirsch SC, Linker A, Perros N & Servais AB et al. Gonadal steroid-dependent effects on bone turnover and bone mineral density in men. Journal of Clinical Investigation 2016 126 1114–1125. (https://doi.org/10.1172/JCI84137)
- 4↑
Travison TG, Araujo AB, Kupelian V, O’Donnell AB & McKinlay JB. The relative contributions of aging, health, and lifestyle factors to serum testosterone decline in men. Journal of Clinical Endocrinology and Metabolism 2007 92 549–555. (https://doi.org/10.1210/jc.2006-1859)
- 5↑
Feldman HA, Longcope C, Derby CA, Johannes CB, Araujo AB, Coviello AD, Bremner WJ & McKinlay JB. Age trends in the level of serum testosterone and other hormones in middle-aged men: longitudinal results from the Massachusetts Male Aging Study. Journal of Clinical Endocrinology and Metabolism 2002 87 589–598. (https://doi.org/10.1210/jcem.87.2.8201)
- 6↑
Travison TG, Araujo AB, O’Donnell AB, Kupelian V & McKinlay JB. A population-level decline in serum testosterone levels in American men. Journal of Clinical Endocrinology and Metabolism 2007 92 196–202. (https://doi.org/10.1210/jc.2006-1375)
- 7↑
Camacho EM, Huhtaniemi IT, O’Neill TW, Finn JD, Pye SR, Lee DM, Tajar A, Bartfai G, Boonen S & Casanueva FF et al. Age-associated changes in hypothalamic-pituitary-testicular function in middle-aged and older men are modified by weight change and lifestyle factors: longitudinal results from the European Male Ageing Study. European Journal of Endocrinology 2013 168 445–455. (https://doi.org/10.1530/EJE-12-0890)
- 8↑
Holmboe SA, Priskorn L, Jørgensen N, Skakkebaek NE, Linneberg A, Juul A & Andersson AM. Influence of marital status on testosterone levels-A ten year follow-up of 1113 men. Psychoneuroendocrinology 2017 80 155–161. (https://doi.org/10.1016/j.psyneuen.2017.03.010)
- 9↑
Hsu B, Cumming RG, Naganathan V, Blyth FM, Le Couteur DG, Hirani V, Waite LM, Seibel MJ & Handelsman DJ. Temporal changes in androgens and estrogens are associated with all-cause and cause-specific mortality in older men. Journal of Clinical Endocrinology and Metabolism 2016 101 2201–2210. (https://doi.org/10.1210/jc.2016-1025)
- 10↑
Haring R, Teng Z, Xanthakis V, Coviello A, Sullivan L, Bhasin S, Murabito JM, Wallaschofski H & Vasan RS. Association of sex steroids, gonadotrophins, and their trajectories with clinical cardiovascular disease and all-cause mortality in elderly men from the Framingham Heart Study. Clinical Endocrinology 2013 78 629–634. (https://doi.org/10.1111/cen.12013)
- 11↑
Gerdes LU, Brønnum-Hansen H, Madsen M, Borch-Johnsen K, Jørgensen T, Sjøl A & Schroll M. Trends in selected biological risk factors of cardiovascular diseases in the Danish MONICA population, 1982–1992. Journal of Clinical Epidemiology 2000 53 427–434. (https://doi.org/10.1016/S0895-4356(99)00193-6)
- 12↑
Osler M, Linneberg A, Glumer C & Jorgensen T. The cohorts at the Research Centre for Prevention and Health, formerly ‘The Glostrup Population Studies’. International Journal of Epidemiology 2011 40 602–610. (https://doi.org/10.1093/ije/dyq041)
- 13↑
Andersson AM, Jensen TK, Juul A, Petersen JH, Jørgensen T & Skakkebaek NE. Secular decline in male testosterone and sex hormone binding globulin serum levels in Danish population surveys. Journal of Clinical Endocrinology and Metabolism 2007 92 4696–4705. (https://doi.org/10.1210/jc.2006-2633)
- 14↑
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 3666–3672.
- 15↑
Tajar A, Forti G, O’Neill TW, Lee DM, Silman AJ, Finn JD, Bartfai G, Boonen S, Casanueva FF & Giwercman A et al. Characteristics of secondary, primary, and compensated hypogonadism in aging men: evidence from the European Male Ageing Study. Journal of Clinical Endocrinology and Metabolism 2010 95 1810–1818. (https://doi.org/10.1210/jc.2009-1796)
- 16↑
Khaw KT, Dowsett M, Folkerd E, Bingham S, Wareham N, Luben R, Welch A & Day N. Endogenous testosterone and mortality due to all causes, cardiovascular disease, and cancer in men: European prospective investigation into cancer in Norfolk (EPIC-Norfolk) prospective population study. Circulation 2007 116 2694–2701. (https://doi.org/10.1161/CIRCULATIONAHA.107.719005)
- 17↑
Haring R, John U, Völzke H, Nauck M, Dörr M, Felix SB & Wallaschofski H. Low testosterone concentrations in men contribute to the gender gap in cardiovascular morbidity and mortality. Gender Medicine 2012 9 557–568. (https://doi.org/10.1016/j.genm.2012.10.007)
- 18↑
Shores MM, Biggs ML, Arnold AM, Smith NL, Longstreth WT, Kizer JR, Hirsch CH, Cappola AR & Matsumoto AM. Testosterone, dihydrotestosterone and incident cardiovascular disease and mortality in the cardiovascular health study. Journal of Clinical Endocrinology and Metabolism 2014 99 jc20133576.
- 19↑
Pye SR, Huhtaniemi IT, Finn JD, Lee DM, O’Neill TW, Tajar A, Bartfai G, Boonen S, Casanueva FF & Forti G et al. Late-onset hypogonadism and mortality in aging men. Journal of Clinical Endocrinology and Metabolism 2014 99 1357–1366. (https://doi.org/10.1210/jc.2013-2052)
- 20↑
Laughlin GA, Barrett-Connor E & Bergstrom J. Low serum testosterone and mortality in older men. Journal of Clinical Endocrinology and Metabolism 2008 93 68–75. (https://doi.org/10.1210/jc.2007-1792)
- 21↑
Lehtonen A, Huupponen R, Tuomilehto J, Lavonius S, Arve S, Isoaho H, Huhtaniemi I & Tilvis R. Serum testosterone but not leptin predicts mortality in elderly men. Age and Ageing 2008 37 461–464. (https://doi.org/10.1093/ageing/afn048)
- 22↑
Tivesten A, Vandenput L, Labrie F, Karlsson MK, Ljunggren O, Mellström D & Ohlsson C. Low serum testosterone and estradiol predict mortality in elderly men. Journal of Clinical Endocrinology and Metabolism 2009 94 2482–2488. (https://doi.org/10.1210/jc.2008-2650)
- 23↑
Shores MM, Matsumoto AM, Sloan KL & Kivlahan DR. Low serum testosterone and mortality in male veterans. Archives of Internal Medicine 2006 166 1660. (https://doi.org/10.1001/archinte.166.15.1660)
- 24↑
Yeap BB, Alfonso H, Chubb SAP, Handelsman DJ, Hankey GJ, Almeida OP, Golledge J, Norman PE & Flicker L. In older men an optimal plasma testosterone is associated with reduced all-cause mortality and higher dihydrotestosterone with reduced ischemic heart disease mortality, while estradiol levels do not predict mortality. Journal of Clinical Endocrinology and Metabolism 2014 99 E9–E18. (https://doi.org/10.1210/jc.2013-3272)
- 25↑
Araujo AB, Kupelian V, Page ST, Handelsman DJ, Bremner WJ & McKinlay JB. Sex steroids and all-cause and cause-specific mortality in men. Archives of Internal Medicine 2007 167 1252–1260. (https://doi.org/10.1001/archinte.167.12.1252)
- 26↑
Barrett-Connor E & Khaw KT. Endogenous sex hormones and cardiovascular disease in men. A prospective population-based study. Circulation 1988 78 539–545. (https://doi.org/10.1161/01.CIR.78.3.539)
- 27↑
Hyde Z, Norman PE, Flicker L, Hankey GJ, Almeida OP, McCaul KA, Chubb SAP & Yeap BB. Low free testosterone predicts mortality from cardiovascular disease but not other causes: the health in men study. Journal of Clinical Endocrinology and Metabolism 2012 97 179–189. (https://doi.org/10.1210/jc.2011-1617)
- 28↑
Phillips AC, Gale CR & Batty GD. Sex hormones and cause-specific mortality in the male veterans: the vietnam experience study. QJM 2012 105 241–246. (https://doi.org/10.1093/qjmed/hcr204)
- 29↑
Vikan T, Schirmer H, Njølstad I & Svartberg J. Endogenous sex hormones and the prospective association with cardiovascular disease and mortality in men: the Tromsø Study. European Journal of Endocrinology 2009 161 435–442. (https://doi.org/10.1530/EJE-09-0284)
- 30↑
Holmboe SA, Vradi E, Jensen TK, Linneberg A, Husemoen LLN, Scheike T, Skakkebæk NE, Juul A & Andersson AM. The association of reproductive hormone levels and all-cause, cancer, and cardiovascular disease mortality in men. Journal of Clinical Endocrinology and Metabolism 2015 100 4472–4480. (https://doi.org/10.1210/jc.2015-2460)
- 31↑
Menke A, Guallar E, Rohrmann S, Nelson WG, Rifai N, Kanarek N, Feinleib M, Michos ED, Dobs A & Platz EA. Sex steroid hormone concentrations and risk of death in US men. American Journal of Epidemiology 2010 171 583–592. (https://doi.org/10.1093/aje/kwp415)
- 32↑
Shores MM, Arnold AM, Biggs ML, Longstreth WT, Smith NL, Kizer JR, Cappola AR, Hirsch CH, Marck BT & Matsumoto AM. Testosterone and dihydrotestosterone and incident ischaemic stroke in men in the Cardiovascular Health Study. Clinical Endocrinology 2014 81 746–753. (https://doi.org/10.1111/cen.12452)
- 33↑
Smith GD, Ben-Shlomo Y, Beswick A, Yarnell J, Lightman S & Elwood P. Cortisol ,testosterone, and coronary heart disease: prospective evidence from the Caerphilly study. Circulation 2005 112 332–340. (https://doi.org/10.1161/CIRCULATIONAHA.104.489088)
- 34↑
Szulc P, Claustrat B & Delmas PD. Serum concentrations of 17beta-E2 and 25-hydroxycholecalciferol (25OHD) in relation to all-cause mortality in older men – the MINOS study. Clinical Endocrinology 2009 71 594–602. (https://doi.org/10.1111/j.1365-2265.2009.03530.x)
- 35↑
Chan YX, Knuiman MW, Hung J, Divitini ML, Beilby JP, Handelsman DJ, Beilin J, McQuillan B & Yeap BB. Neutral associations of testosterone, dihydrotestosterone and estradiol with fatal and non-fatal cardiovascular events, and mortality in men aged 17–97 years. Clinical Endocrinology 2016 85 575–582. (https://doi.org/10.1111/cen.13089)
- 36↑
Araujo AB, Dixon JM, Suarez EA, Murad MH, Guey LT & Wittert GA. Endogenous testosterone and mortality in men: a systematic review and meta-analysis. Journal of Clinical Endocrinology and Metabolism 2011 96 3007–3019. (https://doi.org/10.1210/jc.2011-1137)
- 37↑
Jensen TK, Jacobsen R, Christensen K, Nielsen NC & Bostofte E. Good semen quality and life expectancy: a cohort study of 43,277 men. American Journal of Epidemiology 2009 170 559–565. (https://doi.org/10.1093/aje/kwp168)
- 38↑
Eisenberg ML, Li S, Behr B, Cullen MR, Galusha D, Lamb DJ & Lipshultz LI. Semen quality, infertility and mortality in the USA. Human Reproduction 2014 29 1567–1574. (https://doi.org/10.1093/humrep/deu106)
- 39↑
Latif T, Jensen TK, Mehlsen J, Holmboe SA, Brinth L, Pors K, Skouby SO, Jørgensen N & Lindahl-jacobsen R. Original contribution semen quality as a predictor of subsequent morbidity : a Danish cohort study of 4, 712 men with long-term follow-up. American Journal of Epidemiology 2017 186 910–917. (https://doi.org/10.1093/aje/kwx067)
- 40↑
Rosner W, Auchus RJ, Azziz R, Sluss PM & Raff H. Position statement: utility, limitations, and pitfalls in measuring testosterone: an endocrine society position statement. Journal of Clinical Endocrinology and Metabolism 2007 92 405–413. (https://doi.org/10.1210/jc.2006-1864)
