Abstract
Context:
Previous studies have demonstrated lower testosterone concentrations in men with type 2 diabetes mellitus. Data in men with type 1 diabetes mellitus (T1DM) are limited.
Objective:
Our objective was to determine the prevalence of low testosterone in men with T1DM and identify predisposing factors.
Design, Setting, and Participants:
This was a cross-sectional study of men with T1DM participating in UroEDIC (n = 641), an ancillary study of urologic complications in the Epidemiology of Diabetes Interventions and Complications (EDIC).
Main Outcome Measures:
Total serum testosterone levels were measured using mass spectrometry, and SHBG levels were measured using sandwich immunoassay on samples from EDIC year 17/18. Calculated free testosterone was determined using an algorithm incorporating binding constants for albumin and SHBG. Low testosterone was defined as total testosterone <300 mg/dL. Multivariate regression models were used to compare age, body mass index, factors related to diabetes treatment and control, and diabetic complications with testosterone levels.
Results:
Mean age was 51 years. Sixty-one men (9.5%) had testosterone <300 mg/dL. Decreased testosterone was significantly associated with obesity (P < .01), older age (P < .01) and decreased SHBG (P < .001). Insulin dose was inversely associated with calculated free testosterone (P = .02). Hypertension retained a significant adjusted association with lower testosterone (P = .05). There was no observed significant relationship between lower testosterone and nephropathy, peripheral neuropathy, and autonomic neuropathy measures.
Conclusion:
The men with T1DM in the EDIC cohort do not appear to have a high prevalence of androgen deficiency. Risk factors associated with low testosterone levels in this population are similar to the general population.
An association between diabetes mellitus and decreased testosterone concentrations was first suggested in the 1970s (1). Although it is now widely accepted that men with type 2 diabetes mellitus (T2DM) tend to have lower testosterone levels compared with the general population (2), reports describing testosterone levels in men with T1DM are limited to small studies (3–7). How insulin resistance and body fat composition, which are linked to low testosterone, are related to the hormone levels within the T1DM population has not been systematically studied.
The goal of the current study was to describe testosterone levels in a large cohort of men with T1DM, to identify predisposing factors associated with low testosterone levels, and to determine whether complications linked to T1DM were associated with testosterone levels. We analyzed hormone data from men who participated in UroEDIC, an ancillary study of urological complications of T1DM in the Diabetes Control and Complications Trial (DCCT) and its observational follow up, the Epidemiology of Diabetes Interventions and Complications (EDIC).
Subjects and Methods
Detailed descriptions of DCCT/EDIC procedures and baseline characteristics have been published (8). Briefly, DCCT had 2 treatment arms, intensive insulin therapy vs conventional therapy. Intensive therapy was aimed at achieving normal hemoglobin A1c (HbA1c) levels using 3 or more daily insulin injections or continuous sc insulin infusion with dose adjustments, whereas conventional therapy was aimed at maintaining clinical well-being with 1 to 2 daily insulin injections and no specified glucose targets. At study conclusion, intensive therapy was recommended for the entire cohort and subjects were enrolled into the EDIC follow-up study. By EDIC year 10, 97% of intensive and 94% of conventional treatment arm subjects were implementing intensive therapy and similar levels of HbA1c (∼8%) had been attained (9). At EDIC year 17/18, in 2010/2011, 641 men (96% of the active EDIC male participants) provided serum for testosterone assays as part of an ancillary UroEDIC protocol. The institutional review board of each participating center approved the study, participants provided informed consent, and the U.S. Federal Government issued a Certificate of Confidentiality for UroEDIC.
Diabetes and clinical measurements
HbA1c was measured at baseline, quarterly during DCCT, and annually in EDIC. The weighted-mean HbA1c over the entire study period weighted each DCCT value by 0.25 and each EDIC value by 1 (9). Peripheral neuropathy, autonomic neuropathy, retinopathy, and nephropathy definitions have been published (9). Briefly, peripheral neuropathy was determined by physical examination and the Michigan Neuropathy Screening Instrument (10). Autonomic neuropathy was determined using R-R variation, Valsalva ratio, and diastolic blood pressure (BP). Retinopathy was assessed using the Early Treatment Diabetic Retinopathy Study scale (11). Nephropathy was defined using albumin excretion rate (milligrams per 24 hours) at EDIC year 15/16. Hypertension was defined as sitting systolic BP ≥140 mm Hg and/or diastolic BP ≥90 mm Hg or use of antihypertensive medications. Erectile dysfunction (ED) was defined as a yes response to a single yes/no question querying presence of impotence.
Hormone measurements
Total testosterone (TT) was measured at the University of Minnesota using a rapid liquid chromatography-tandem mass spectrophotometry platform. Lower and upper limits of detection for the TT assay are <0.25 and 50 nmol/L. The assay was certified by the Centers for Disease Control (CDC) Hormone Standardization Program (HoST, http://www.cdc.gov/labstandards/hs.html) and passed the required intra-assay performance criterion of 6.4%. Specifically, intra-assay mean bias was 4.6% at 6.9 mL and 1.8% at 34.9 mL. Low testosterone was defined as serum TT levels below 10.4 nmol/L or 300 ng/dL, the lower limit of the normal range for healthy young men as defined by the Endocrine Society Guidelines (12). SHBG was measured on a Roche Elecsys 2010 analyzer (Roche Diagnostics Corporation) using a sandwich immunoassay method. This method has been standardized against the first International Standard for SHBG from the National Institute for Biological Standards and Control code 95/560. The laboratory CV is 3.0% at a concentration of 25.38 nmol/L. Calculated free testosterone (cFT) was determined using the Vermeulen formula with an association constant of testosterone binding to albumin of 3.6 × 104 L/mol and testosterone binding to SHBG of 109 L/mol (13).
Statistical analysis
Mean hormone levels by sociodemographic, clinical, and diabetes-related characteristics were assessed at EDIC year 17/18. Univariate association with testosterone at the continuous level was evaluated using unadjusted linear regression. For both body mass index (BMI) and HbA1c, time-weighted mean values were used, representing the running mean in each study visit in the DCCT and EDIC up to year 17/18. Use of nonweighted BMI and HbA1c variables did not appreciably alter regression coefficients (data not shown). Multivariate linear regression was used to estimate association of testosterone levels with the following a priori defined variables: age (continuous), DCCT/EDIC time-weighted BMI (continuous), treatment arm (conventional vs intensive therapy), DCCT/EDIC time-weighted HbA1c (continuous), insulin dose in units per kilogram per day (continuous), and SHBG (continuous). Each diabetic complication (ie, retinopathy, nephropathy, peripheral neuropathy, autonomic neuropathy, and hypertension) was evaluated in separate regression models using the same variable set. Testosterone levels dichotomized into clinically low (TT <300 ng/dL) vs normal was examined in logistic regression models using the same variable sets. All analyses were performed using SAS version 9.2 statistical analysis software.
Results
Table 1 shows hormone levels stratified by demographic and clinical characteristics of the 641 participants. Clinically low testosterone was classified in 61 men (9.5%). After multivariate adjustment, older age, obesity, and lower SHBG retained a significant association with decreased testosterone levels (Table 2 and Supplemental Figure 1). These same relationships were observed in association with clinically classified low testosterone (data not shown). An inverse relationship with testosterone levels and insulin dose was evident but did not retain significance in the multivariate model (P = .08). When cFT was examined in an adjusted model with the same covariates, this inverse association with insulin dose remained significant (P = .02). Hypertension retained a significant adjusted association with lower testosterone (P = .05). In those reporting ED, 14.4% were categorized as low testosterone as compared with 7.1% with no ED; the adjusted odds of having clinically low testosterone for those with ED was 2.0-fold greater (95% confidence interval 1.1–3.7). Although testosterone levels were lower in patients exhibiting nephropathy, peripheral neuropathy, and autonomic neuropathy, this finding was not significant in the adjusted models (Tables 1 and 2).
Table 1.
Hormone Distributions by Sociodemographic/Clinical and Diabetes Characteristics at EDIC Year 17/18
| Characteristic | n (%)a | Mean (SD) |
||
|---|---|---|---|---|
| TT, ng/dL | cFT, ng/dL | SHBG, nmol/L | ||
| Total | 641 | |||
| Sociodemographic/clinical | ||||
| Age, y | ||||
| 35–44 | 107 (16.7) | 514.7 (204.5) | 8.7 (2.3) | 47.5 (22.0) |
| 45–54 | 315 (49.1) | 561.2 (208.7) | 8.8 (2.5) | 54.2 (23.8) |
| 55–67 | 219 (34.2) | 545.9 (220.5) | 8.0 (2.4) | 59.0 (25.0) |
| BMI category, kg/m2 | ||||
| Normal, BMI <25 | 136 (21.2) | 651.1 (218.0) | 8.6 (2.5) | 71.0 (28.1) |
| Overweight, BMI 25–30 | 272 (42.4) | 577.5 (206.3) | 8.8 (2.4) | 56.3 (21.5) |
| Obese, BMI ≥30 | 230 (35.9) | 451.6 (174.2) | 8.0 (2.3) | 43.5 (18.4) |
| Diabetes treatment and control | ||||
| Treatment arm | ||||
| Conventional | 321 (50.1) | 557.6 (204.7) | 8.6 (2.4) | 56.0 (24.7) |
| Intensive | 320 (49.9) | 540.9 (220.2) | 8.4 (2.5) | 53.7 (23.8) |
| Insulin dose, U/kg/d | ||||
| Tertile 1 (0.22–0.59) | 210 (32.8) | 610.6 (201.7) | 8.7 (2.3) | 62.7 (24.4) |
| Tertile 2 (0.60–0.82) | 210 (32.8) | 560.0 (209.0) | 8.6 (2.5) | 55.4 (22.7) |
| Tertile 3 (0.83–8.28) | 210 (32.8) | 471.6 (203.0) | 8.1 (2.5) | 45.9 (22.6) |
| Time-weighted DCCT/EDIC HbA1c, % | ||||
| Tertile (5.55–7.47) | 210 (32.8) | 572.7 (210.0) | 8.7 (2.6) | 55.8 (21.9) |
| Tertile (7.48–8.30) | 211 (32.8) | 537.2 (207.7) | 8.2 (2.3) | 56.0 (25.5) |
| Tertile (8.31–10.98) | 211 (32.8) | 532.9 (215.3) | 8.6 (2.4) | 52.1 (24.3) |
| Erectile dysfunction | ||||
| No | 410 (64.0) | 556.3 (200.0) | 8.6 (2.3) | 55.3 (23.4) |
| Yes | 222 (34.6) | 531.8 (234.0) | 8.3 (2.6) | 53.7 (25.9) |
| Diabetes complications | ||||
| Hypertensionb | ||||
| No | 190 (30.0) | 573.5 (213.1) | 8.5 (2.3) | 58.5 (26.9) |
| Yes | 444 (70.0) | 537.2 (211.5) | 8.5 (2.5) | 53.2 (22.9) |
| Retinopathyc | ||||
| Nonproliferative or none | 510 (79.6) | 548.8 (212) | 8.5 (2.4) | 55 (24.4) |
| Proliferative | 131 (20.4) | 551 (215.5) | 8.6 (2.6) | 54.4 (23.9) |
| Nephropathy measuresd | ||||
| None (AER <30) | 473 (74.5) | 563.1 (213.9) | 8.6 (2.5) | 56.2 (23.7) |
| Microalbuminuria (30 ≤ AER < 300) | 122 (19.2) | 511.4 (209.8) | 8.3 (2.4) | 51.1 (26.2) |
| Albuminuria (AER ≥300) | 40 (6.3) | 519.1 (187.8) | 8.3 (1.9) | 53.1 (23.3) |
| Sustained AER <30 | 517 (80.7) | 558.7 (214.3) | 8.5 (2.5) | 55.8 (24.3) |
| Sustained AER ≥30 | 124 (19.3) | 509.9 (201.4) | 8.3 (2.1) | 50.9 (23.9) |
| eGFR ≥60 | 613 (95.6) | 553.6 (211.4) | 8.5 (2.4) | 55.3 (24.3) |
| eGFR <60 | 28 (4.4) | 454.4 (220.9) | 7.8 (2.3) | 44.6 (21.2) |
| Peripheral neuropathye | ||||
| No | 357 (56.8) | 551.7 (202.1) | 8.7 (2.4) | 53.3 (21.6) |
| Yes | 272 (43.2) | 544.9 (227.3) | 8.2 (2.5) | 56.7 (27.4) |
| Autonomic neuropathyf | ||||
| No | 390 (62.3) | 557.9 (212.9) | 8.6 (2.5) | 54.8 (24.3) |
| Yes | 236 (37.7) | 535.3 (211.7) | 8.2 (2.4) | 55.1 (24.1) |
Abbreviations: AER, albumin excretion rate; eGFR, estimated glomerular filtration rate.
Variable categories that do not add up to n = 641 have missing values not shown.
Hypertension defined as sitting systolic BP ≥140 mm Hg and/or diastolic BP ≥90 mm Hg or use of antihypertensive medication.
Defined through EDIC year 14 using the Early Treatment Diabetic Retinopathy Study on a scale of 0 to 23 (<12, nonproliferative or none; ≥12, proliferative).
AER (mg/24 h) at EDIC years 15/16.
Defined at EDIC year 17/18 by Michigan Neuropathy Screening Instrument >6 responses or a score of >2 on the exam.
Autonomic testing completed in EDIC year 16/17 and abnormal finding defined as R-R variation <15 or R-R variation of15 to 20 in combination with Valsalva ratio ≤1.5 or a decrease of >10 mm Hg in diastolic BP.
Table 2.
Unadjusted and Adjusted Correlates of Sociodemographic/Clinical and Diabetes Characteristics with Testosterone Levels at EDIC Year 17/18
| Characteristic | Unadjusted Coefficient (95% CI) | P | Adjusted Coefficient (95% CI) | Pa |
|---|---|---|---|---|
| Age | 1.41 (−1.10 to 3.92) | .27 | −2.74 (−4.40 to −1.08) | <.01 |
| Time-weighted BMI | −21.81 (−26.26 to −17.37) | <.001 | −5.01 (−8.32 to −1.70) | <.01 |
| Treatment arm: intensive | −16.65 (−49.63 to 16.32) | .32 | −5.53 (−27.53 to 16.47) | .62 |
| Insulin dose, U/kg/d | −122.39 (−159.46 to −85.31) | <.001 | −22.97 (−48.45 to 2.51) | .08 |
| Time-weighted DCCT/EDIC HbA1c, % | −14.48 (−31.56 to 2.60) | .1 | −2.63 (−14.07 to 8.81) | .65 |
| SHBG, nmol/L | 6.61 (6.17–7.05) | <.001 | 6.40 (5.91–6.89) | <.001 |
| Diabetes complications | ||||
| Hypertensionb | ||||
| No | Reference | Reference | ||
| Yes | −36.24 (−72.33 to −0.15) | .05 | 24.10 (−0.15 to 48.34) | .05 |
| Retinopathyc | ||||
| Nonproliferative or none | Reference | Reference | ||
| Proliferative | 2.20 (−38.72 to 43.12) | .92 | 12.77 (−15.77 to 41.32) | .38 |
| Nephropathy measuresd | ||||
| None (AER <30) | Reference | Reference | ||
| Microalbuminuria(30 ≤ AER < 300) | −51.70 (−93.90 to −9.50) | .02 | −10.46 (−38.97 to 18.06) | .47 |
| Albuminuria (AER ≥300) | −44.05 (−112.48 to 24.38) | .21 | −27.09 (−74.04 to 19.87) | .26 |
| Sustained AER <30 | Reference | Reference | ||
| Sustained AER ≥30 | −48.77 (−90.37 to −7.17) | .02 | −12.24 (−41.67 to 17.18) | .41 |
| eGFR ≥60 | Reference | Reference | ||
| eGFR <60 | −99.23 (−179.59 to −18.86) | .02 | −17.38 (−71.74 to 36.98) | |
| Peripheral neuropathye | ||||
| No | Reference | Reference | ||
| Yes | −6.73 (−40.45 to 26.99) | .70 | −12.46 (−35.61 to 10.69) | .29 |
| Autonomic neuropathyf | ||||
| No | Reference | Reference | ||
| Yes | −22.66 (−57.07 to 11.75) | .20 | −5.69 (−29.52 to 18.15) | .64 |
Abbreviations: AER, albumin excretion rate; CI, confidence interval; eGFR, estimated glomerular filtration rate.
Sociodemographic and diabetic treatment/control variables were adjusted for age (continuous), time-weighted BMI (continuous), treatment arm, time-weighted DCCT/EDIC HbA1c (continuous), and insulin dose at EDIC year 17/18 in units per kilogram per day (continuous). Each complication (ie, retinopathy, nephropathy, peripheral neuropathy, autonomic neuropathy, and hypertension) was used in a separate model controlling for age, BMI, treatment arm, weighted HbA1c, and insulin dose.
Hypertension is defined as sitting systolic BP ≥140 mm Hg and/or diastolic BP ≥90 mm Hg or the use of antihypertensive medication.
Defined through EDIC year 14 using the Early Treatment Diabetic Retinopathy Study on a scale of 0 to 23 (<12, nonproliferative or none; ≥12, proliferative).
AER (mg/24 h) at EDIC years 15/16.
Defined at EDIC year 17/18 by the Michigan Neuropathy Screening Instrument >6 responses on the questionnaire or a score of >2 on the exam.
Autonomic testing completed in EDIC year 16/17 and abnormal finding defined as R-R variation <15 or R-R variation of 15 of 20 in combination with a Valsalva ratio ≤1.5 or a decrease of >10 mm Hg in diastolic BP.
Discussion
We found that in a cohort of middle-aged men (mean age of 51 years) with longstanding T1DM (mean duration of 29.3 years), low testosterone was relatively uncommon (<10%) and TT levels were inversely associated with age, BMI, and hypertension. We also observed a positive association with cFT and insulin dose. Whereas the association of low testosterone with T2DM is well documented, reports regarding androgen levels in men with T1DM have generally been in small cohorts (3–7). Our results suggest that testosterone levels in men with T1DM are more comparable to the general population than to men with T2DM (14, 15) (Supplemental Table 2).
Factors that have been consistently shown to be associated with low testosterone in T2DM are insulin resistance and obesity (16). This was true in our study, as men with decreased testosterone levels were on average heavier and used higher insulin doses, an indicator of insulin resistance. Previous reports support a bidirectional relationship between BMI and testosterone, with lower testosterone levels preceding development of obesity in men and high BMI and central adiposity contributing to low testosterone levels (17). This phenomenon is probably a consequence of the higher rate of aromatase conversion of testosterone to estrogen by the adipose tissue, triggering inhibition of the hypothalamic-pituitary-gonadal axis and subsequent hypogonadism (4). In addition, because obesity leads to an inflammatory state, subsequent increased systemic levels of free fatty acids, cytokines, and adipokines from the visceral adipose tissue interact with low levels of testosterone to contribute to insulin resistance and vascular damage (16). There is evidence that decreased testosterone levels precede and predict insulin resistance and development of T2DM; however, this has not been established for T1DM (18). Cross-sectional measures preclude assertions regarding temporality of the association between low testosterone with higher BMI and insulin dose in this study.
Both large study size and detailed collection of outcomes allowed for analysis of correlations with specific diabetic complications. The relationship between hypertension and lower testosterone observed in this study has been shown in both normal and diabetic populations (16). Once again, the cross-sectional design makes it difficult to assess the causal relationship between hypertension and testosterone. Both a moderate decrease in renal function and end-stage renal disease have been associated with low testosterone levels (19). Although the mechanism of this association remains unknown, we did observe lower testosterone with measures of nephropathy, albeit not in the adjusted models.
This is the largest study to date to investigate the association between testosterone levels in T1DM and both predisposing factors and diabetic complications. The long duration of follow-up allowed for calculation of weighted measures of glycemic control and BMI. In addition, use of a CDC-certified liquid chromatography-tandem mass spectrophotometry platform leant validity to assay results. To capture biologically active testosterone, we used calculated rather than direct estimation of serum FT, because these tests were not feasible. It should be noted that FT might not necessarily reflect exposure of cells to androgen because there are endocytic receptors that actively uptake the entire steroid hormone complex (20). In addition, although the Vermeulen formula is based on an equilibrium-binding theory accounting for albumin, cFT has been shown to overestimate FT and have reduced predictive accuracy in some settings (21). Thus, the focus of the cFT results is not to make clinical recommendations, rather to note how a TT measurement that accounts for SHBG could be associated with factors related to diabetes.
There are a number of limitations of the study. Testosterone sampling conditions were not entirely uniform; 92% were drawn in the morning, but only 55% were fasting. These factors, however, were not found to be associated with testosterone levels or to alter adjusted models. The DCCT/EDIC cohort may not be generalizable to the general population of middle-aged men with T1DM because DCCT/EDIC participants are generally a highly motivated group with good glycemic control. There was no control group of men without diabetes, so analyses do not allow a direct comparison with nondiabetic men. We did review previous large cohort studies examining androgen levels in the general population and found similar reported mean testosterone levels and percentages of androgen-deficient men (14, 15) (Supplemental Table 2). We did not have data on androgen replacement therapy, although self-reported medication lists did not show use of testosterone replacement therapy. We were not able to perform a comprehensive evaluation of other symptoms of hypogonadism outside of diagnosis of ED by questionnaire. In an attempt to account for testicular and pituitary disease, we measured LH and FSH in the 61 men with classified low testosterone. A sensitivity analysis excluding men with extremely high LH/FSH levels (n = 4) and extremely low LH/FSH levels (n = 2) did not appreciably alter the findings (Supplemental Table 1).
In summary, results of this study do not support a high prevalence of low testosterone in men with T1DM. Those subjects with low levels of testosterone were more likely to have comorbidities such as obesity and hypertension. T1DM appears not to affect androgen levels adversely in middle-aged men.
Acknowledgments
A complete list of participants in the DCCT/EDIC research group can be found in Ref. 22.
Support for this DCCT/EDIC collaborative study (uroEDIC) was provided by Grant R01 DK083927. The DCCT/EDIC has been supported by U01 Cooperative Agreement grants (1982–1993 and 2011–2016) and contracts (1982–2011) with the Division of Diabetes Endocrinology and Metabolic Diseases of the National Institute of Diabetes and Digestive and Kidney Disease (current grant numbers U01 DK094176 and U01 DK094157) and through support by the National Eye Institute, the National Institute of Neurologic Disorders and Stroke, the General Clinical Research Centers Program (1993–2007), and the Clinical Translational Science Center Program (2006 to present). Additional support for this DCCT/EDIC collaborative study (uroEDIC) was provided by Grant R01 DK083927.
The following industry contributors had no role in the DCCT/EDIC study but have provided free or discounted supplies or equipment to support participants' adherence to the study: Abbott Diabetes Care (Alameda, CA), Animas (Westchester, PA), Bayer Diabetes Care (North America Headquarters, Tarrytown, NY), Becton Dickinson (Franklin Lakes, NJ), Eli Lilly (Indianapolis, IN), Lifescan (Milpitas, CA), Medtronic Diabetes (Minneapolis, MI), Nova Diabetes Care (Billerica, MA), Omron (Shelton, CT), OmniPod Insulin Management System (Bedford, MA), Roche Diabetes Care (Indianapolis, IN), Extend Nutrition (St. Louis, MO), Nipro Home Diagnostics (Ft Lauderdale, FL), Perrigo Diabetes Care (Allegan, MI), and Sanofi-Aventis (Bridgewater NJ).
Disclosure Summary: The authors have nothing to disclose. There are no conflicts of interest.
Footnotes
- BMI
- body mass index
- BP
- blood pressure
- cFT
- calculated free testosterone
- DCCT
- Diabetes Control and Complications Trial
- ED
- erectile dysfunction
- EDIC
- Epidemiology of Diabetes Interventions and Complications
- HbA1c
- hemoglobin A1c
- T2DM
- type 2 diabetes mellitus
- TT
- total testosterone.
References
- 1. Cooper AJ. Diagnosis and management of “endocrine impotence.” Br Med J. 1972;2:34–36 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Corona G, Monami M, Rastrelli G, et al. Type 2 diabetes mellitus and testosterone: a meta-analysis study. Int J Androl. 2011;34:528–540 [DOI] [PubMed] [Google Scholar]
- 3. van Dam EW, Dekker JM, Lentjes EG, et al. Steroids in adult men with type 1 diabetes: a tendency to hypogonadism. Diabetes Care. 2003;26:1812–1188 [DOI] [PubMed] [Google Scholar]
- 4. Grossmann M, Thomas MC, Panagiotopoulos S, et al. Low testosterone levels are common and associated with insulin resistance in men with diabetes. J Clin Endocrinol Metab. 2008;93:1834–1840 [DOI] [PubMed] [Google Scholar]
- 5. Tomar R, Dhindsa S, Chaudhuri A, Mohanty P, Garg R, Dandona P. Contrasting testosterone concentrations in type 1 and type 2 diabetes. Diabetes Care. 2006;29:1120–1122 [DOI] [PubMed] [Google Scholar]
- 6. Biswas M, Hampton D, Newcombe RG, Rees DA. Total and free testosterone concentrations are strongly influenced by age and central obesity in men with type 1 and type 2 diabetes but correlate weakly with symptoms of androgen deficiency and diabetes-related quality of life. Clin Endocrinol (Oxf). 2012;76:665–673 [DOI] [PubMed] [Google Scholar]
- 7. Ng Tang Fui M, Hoermann R, Cheung AS, Gianatti EJ, Zajac JD, Grossmann M. Obesity and age as dominant correlates of low testosterone in men irrespective of diabetes status. Andrology. 2013;1:906–912 [DOI] [PubMed] [Google Scholar]
- 8. Epidemiology of Diabetes Interventions and Complications (EDIC). Design, implementation, and preliminary results of a long-term follow-up of the Diabetes Control and Complications Trial cohort. Diabetes Care 1999;22:99–111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. de Boer IH, Rue TC, Cleary PA, et al. Long-term renal outcomes of patients with type 1 diabetes mellitus and microalbuminuria: an analysis of the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications cohort. Arch Intern Med. 2011;171:412–420 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Feldman EL, Stevens MJ, Thomas PK, Brown MB, Canal N, Greene DA. A practical two-step quantitative clinical and electrophysiological assessment for the diagnosis and staging of diabetic neuropathy. Diabetes Care. 1994;17:1281–1289 [DOI] [PubMed] [Google Scholar]
- 11. Early Treatment Diabetic Retinopathy Study design and baseline patient characteristics. ETDRS report number 7. Ophthalmology 1991;98:741–756 [DOI] [PubMed] [Google Scholar]
- 12. Bhasin S, Cunningham GR, Hayes FJ, et al. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2010;95:2536–2559 [DOI] [PubMed] [Google Scholar]
- 13. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab. 1999;84:3666–3672 [DOI] [PubMed] [Google Scholar]
- 14. Bhasin S, Pencina M, Jasuja GK, et al. 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. J Clin Endocrinol Metab. 2011;96:2430–2439 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Lee DM, Pye SR, Tajar A, et al. Cohort profile: the European Male Ageing Study. Int J Epidemiol. 2013;42:391–401 [DOI] [PubMed] [Google Scholar]
- 16. Wang C, Jackson G, Jones TH, et al. Low testosterone associated with obesity and the metabolic syndrome contributes to sexual dysfunction and cardiovascular disease risk in men with type 2 diabetes. Diabetes Care. 2011;34:1669–1675 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. MacDonald AA, Herbison GP, Showell M, Farquhar CM. The impact of body mass index on semen parameters and reproductive hormones in human males: a systematic review with meta-analysis. Hum Reprod Update. 2010;16:293–311 [DOI] [PubMed] [Google Scholar]
- 18. Oh JY, Barrett-Connor E, Wedick NM, Wingard DL. Endogenous sex hormones and the development of type 2 diabetes in older men and women: the Rancho Bernardo study. Diabetes Care. 2002;25:55–60 [DOI] [PubMed] [Google Scholar]
- 19. Haring R, Nauck M, Völzke H, et al. Low serum testosterone is associated with increased mortality in men with stage 3 or greater nephropathy. Am J Nephrol. 2011;33:209–217 [DOI] [PubMed] [Google Scholar]
- 20. Hammes A, Andreassen TK, Spoelgen R, et al. Role of endocytosis in cellular uptake of sex steroids. Cell. 2005;122:751–762 [DOI] [PubMed] [Google Scholar]
- 21. Ly LP, Sartorius G, Hull L, et al. Accuracy of calculated free testosterone formulae in men. Clin Endocrinol (Oxf). 2010;73:382–388 [DOI] [PubMed] [Google Scholar]
- 22. DCCT/EDIC Research Group, de Boer IH, Sun W, et al. Intensive diabetes therapy and glomerular filtration rate in type 1 diabetes. N Engl J Med 2011;365:2366–2376 [DOI] [PMC free article] [PubMed] [Google Scholar]