Accepted Preprint first posted on 22 June 2015 as Manuscript EJE PREVALENCE OF SUBNORMAL TESTOSTERONE CONCENTRATIONS IN MEN

Transcription

1 Page 1 of 23 Accepted Preprint first posted on 22 June 2015 as Manuscript EJE PREVALENCE OF SUBNORMAL TESTOSTERONE CONCENTRATIONS IN MEN WITH TYPE 2 DIABETES AND CHRONIC KIDNEY DISEASE Sandeep Dhindsa, MD 1 Anand Reddy, MD 1, 2 Jyotheen Sukhmoy Karam, MD 3 Sayeeda Bilkis, MD 1 Archana Chaurasia, MBBS 2 Aditya Mehta, MBBS 4 Keerthi P Raja 2 Manav Batra, MD 4 Paresh Dandona, MD, PhD Texas Tech University Health Sciences Center, Department of Medicine, Division of Endocrinology and Metabolism, Odessa, TX 2 Permian Basin Kidney Center, Odessa, TX 16 3 Nephrology Associates, Murfreesboro, TN State University of New York at Buffalo, Department of Medicine, Division of Endocrinology and Metabolism AND the Diabetes and Endocrinology Center of Western New York, Buffalo, NY Abbreviated title: Testosterone in Chronic Kidney Disease Word count: Copyright 2015 European Society of Endocrinology.

2 Page 2 of Tables: 2 Figures: 0 Keywords: Testosterone, renal failure, chronic kidney disease, diabetes, hypogonadism Disclosure Summary: SD and PD have received lecture fees from Abbvie. AR, JK, SB, AC, AM, KR and MB have nothing to disclose Correspondence To: SANDEEP DHINDSA, M.D. Associate Professor of Medicine Chief, Division of Endocrinology and Metabolism Texas Tech University Health Sciences Center- Permian Basin Campus 701 W 5 th Street, Odessa, TX sandeep.dhindsa@ttuhsc.edu 2

3 Page 3 of Abstract Background: One third of men with type 2 diabetes have subnormal testosterone concentrations along with inappropriately normal LH and FSH concentrations. It is not known if the presence of renal insufficiency affects free testosterone concentrations in men with type 2 diabetes Hypothesis: We hypothesized that type 2 diabetic men with chronic renal disease (CKD; estimated glomerular filtration rate (egfr)<60 ml/min/1.73m 2 ) have lower free testosterone concentrations than men with normal renal function (egfr 60 ml/min/1.73m 2 ) Study Design and Setting: This is a retrospective chart review of patients attending diabetes and nephrology clinics. Men with type 2 diabetes who had the following information available were included in the study: testosterone (total and free) done by LC/MS-MS followed by equilibrium dialysis, sex hormone binding globulin, LH, FSH and prolactin concentrations Participants: We present data on T and gonadotropin concentrations in 111 men with type 2 diabetes and CKD (stages 3-5) and 182 type 2 diabetic men without CKD Results: The prevalence of subnormal free testosterone concentrations was higher in men with type 2 diabetes and CKD as compared to those without CKD (66% vs 37%, p<0.001). Men with CKD had a higher prevalence of hypergonadotropic hypogonadism (26% vs. 5%, p<0.001) but not of hypogonadotropic hypogonadism (40% vs. 32%, p=0.22). There was an increase in the prevalence of hypergonadotropic hypogonadism with worsening CKD. 52% of men with renal failure (CKD stage 5) had hypergonadotropic hypogonadism and 25% had hypogonadotropic hypogonadism. In men with CKD, the hemoglobin concentrations were lower in those with 3

4 Page 4 of subnormal free T concentrations as compared to men with normal free T concentrations (119 ±19 vs. 128±19 g/l, P=0.04) Conclusions: Two-thirds of men with type 2 diabetes and CKD have subnormal free T concentrations. The hypogonadism associated with CKD is predominantly hypergonadotropic. 61 4

5 Page 5 of INTRODUCTION We have demonstrated, in a sequence of publications over the past decade, that one third of patients with type 2 diabetes have subnormal plasma concentrations of free testosterone (T)(1-4). These studies also demonstrated that T concentrations are inversely related to body mass index (BMI). These studies were based on accurate measurements of free T. This is important because half of the total T in serum is bound to sex hormone binding globulin (SHBG). SHBG concentrations are lower in obesity and type 2 diabetes which would lower total T concentrations(3). Thus measurement of free T concentrations is necessary to accurately assess gonadal status in these patients(5). The subnormal free T concentrations in type 2 diabetes are associated with low normal or normal LH and FSH concentrations and a high prevalence of symptoms suggestive of hypogonadism(2, 6). Hence this common syndrome can be termed hypogonadotropic hypogonadism (HH). 10.6% of patients with type 2 diabetes have chronic kidney disease (estimated glomerular filtrate rate (egfr) <60 ml/min/1.73m 2 ) (7). Total T concentrations are decreased in men with chronic kidney disease (CKD) (8-13). Endocrine society considers CKD and type 2 diabetes as conditions associated with a high prevalence of low T concentrations(5). However, no study has documented the prevalence of subnormal free T concentrations in type 2 diabetic men with CKD. Neither is it well understood whether the subnormal T concentrations in CKD are associated with elevated, normal or subnormal gonadotropin concentrations. It is our clinical practice to measure T concentrations on all men with type 2 diabetes. This is consistent with The Endocrine Society guidelines(5). To perform a comprehensive assessment, we measure T (total and free), SHBG, LH, FSH and prolactin concentrations. We hypothesized 5

6 Page 6 of that type 2 diabetic men with CKD will have a higher prevalence of subnormal free T concentrations as compared to men without CKD. We present data on T and gonadotropin concentrations in 111 men with type 2 diabetes and CKD (stages 3-5) and 182 type 2 diabetic men without CKD MATERIALS AND METHODS This is a retrospective chart review. Charts of all adult (age >18 years) male patients attending the following clinics between January 2012 and December 2014 were reviewed at:- a) diabetes clinics at Texas Tech University Health Sciences Center-Permian Basin, Odessa, TX and at Diabetes-Endocrinology Center of Western New York, State University of New York, Buffalo, NY; b) nephrology clinic at the Permian Basin Kidney Center, Odessa, TX Patients who had the following information available were included in a password-locked database: - T (total and free) done by liquid chromatography tandem mass spectrometry (LC- MS/MS) followed by equilibrium dialysis, SHBG, LH, FSH and prolactin concentrations. Exclusion criteria were panhypopituitarism or congenital hypogonadotropic hypogonadism; severe depression or psychiatric illness; head trauma, hemochromatosis, cirrhosis, hepatitis C or HIV; treatment with testosterone, steroids, or opiates; treatment with medications that can cause hyperprolactinemia; patients with active infection or who had a recent surgery or hospitalization for any reason in the last 6 weeks. HIPAA forms were signed by all patients. The collected data was kept confidential by the investigators and patient identifiers were eliminated from the database once the master database had been created. Permission was obtained to review patient 6

7 Page 7 of charts to collect data from the Institutional Review Board of the Texas Tech University of Health Sciences Center. All patients had provided fasting blood samples between 8 and 10 a.m. to measure serum T, SHBG, LH, FSH and prolactin. The measurements were carried out by a commercial laboratory. T concentrations were measured by LC-MS/MS and free T was separated by tracer equilibrium dialysis method(14). The sensitivity of the assay, set at a coefficient of variation(cv) of 20%, was 0.3 ng/dl. The intra-assay CV ranged from % and inter-assay CV ranged from % at total T concentrations between nmol/l. Reference range for total T ( nmol/l) and free T ( nmol/l) was determined from 264 apparently healthy men by the laboratory. SHBG (normal range nmol/l), LH (normal range 2-10 IU/L), FSH (normal range 2-8 IU/L) and prolactin (normal range nmol/l) concentrations were measured by a solid-phase, immunochemiluminometric assays (ADVIA Centaur). The analytical sensitivity for SHBG by this assay was 2 nmol/l. The intra-assay CV was 6.2% at SHBG concentration of 4.14 nmol/ L and 3.2% at 66.9 nmol/l. The inter-assay CV was 4.1% at 4.14 nmol/l and 5.2% at 66.9 nmol/l. CKD stages were defined based on estimated glomerular filtrate rate (egfr) by the CKD-EPI (CKD Epidemiology Collaboration) Equation (15). 53 patients had egfr of ml/min/1.73m 2 (stage 3), 28 patients had egfr of ml/min/1.73m 2 (stage 4) and 30 patients had egfr <15 ml/min/1.73m 2 ) or were on dialysis (stage 5). Out of the thirty men in stage 5, eighteen were on hemodialysis and five were on peritoneal dialysis. One hundred and eight two patients had egfr 60 ml/min/1.73m 2 and were assumed not to have CKD. Data on age, height, weight, ethnicity, presence of hypertension, duration of diabetes, albuminuria, medication use and HbA1c were also collected. Hemoglobin data were collected in men who 7

8 Page 8 of were not on erythropoietin. 8 men with CKD stage 5 were on erythropoietin and their hemoglobin concentrations were not included in the database. No patient in CKD stages 3 or 4 was on erythropoietin. HH was defined as free T concentration less than nmol/l along with an inappropriately low LH of 9.4 IU/L. Hypergonadotropic hypogonadism was defined as free T concentration less than nmol/l along with LH >9.4 IU/L. Compensated hypogonadism was defined as free T concentration nmol/l along with LH >9.4 IU/L(16). Statistical Analysis: Group comparisons were performed by one-way ANOVA, t tests, Mann- Whitney rank sum tests, and χ 2 tests as appropriate. Adjustment for age and BMI in group comparisons was done with ANCOVA and generalized linear model analysis. Data that were not normally distributed (determined by Kolmogorov-Smirnov test as well as visual estimation of data) were log-transformed to perform the parametric statistical tests. BMI, LH, FSH, SHBG, A1c and prolactin concentrations were not normally distributed and were log-transformed. Pearson correlation and multiple linear regression analyses between variables were done using SPSS software(spss Inc, Chicago, Illinois). Since age and BMI affect T and SHBG concentrations (3), we adjusted T and SHBG concentrations as well as prevalence of low T for age and BMI differences when comparing groups. In addition, total T concentrations were adjusted for SHBG differences. Data are presented as means±sd for normally distributed data and median [25 th, 75 th percentile] for non-normal data. P<0.05 was considered significant RESULTS We screened charts of 451 men with type 2 diabetes for availability of relevant laboratory data. Forty-two men were excluded because they did not satisfy the study inclusion and exclusion 8

9 Page 9 of criteria. Commonest reasons for exclusion were hepatitis C, panhypopituitarism or treatment with testosterone. We excluded 116 men because their T concentrations were measured by inaccurate assays or because hormonal measurements were missing. The mean age and BMI of excluded patients (58±12 years and 34.3±7.3 kg/m 2 ) were similar to that of study patients (58±10 years and 33.9±7.3 kg/m 2 ; p=0.91 and 0.46 respectively). The median HbA1c were also similar (7.1 [6.4, 8.2] % and 7.2 [6.3, 8.2] %, p=0.93). Out of 293 men with type 2 diabetes that qualified for the study, 111 had CKD and 182 had normal egfr. Seven percent of men with CKD were African Americans, 61% were Caucasian, 25% were Hispanic and 7% identified themselves as other. Eight percent of men without CKD were African Americans, 63% were Caucasian, 21% were Hispanic and 8% identified themselves as other (P=0.97 for comparison with CKD group). Comparison of hormone concentrations in men with and without CKD: Men with CKD were older than men without CKD while their BMI tended to be lower (table 1). Men with CKD had diabetes for a longer duration than men without CKD. Higher proportion of CKD patients had insulin as part of their diabetes regimen while their use of metformin and sulfonylureas was lower than patient without CKD. Free T concentrations were related inversely to age (β= -0.35, p<0.001) and BMI (β= -0.23, p<0.001) but not to duration of diabetes (β= -0.05, p=0.39). Hence we adjusted T concentrations for differences in age and BMI among men with and without CKD (table 1). Total and free T concentrations were lower in men with CKD. The median gonadotropin concentrations were higher in men with CKD as compared to men with normal egfr. Consistent with this, men with CKD had a higher prevalence of hypergonadotropic hypogonadism but not of hypogonadotropic hypogonadism. 9

10 Page 10 of Table 2 depicts the hormone concentrations of men with CKD and hypogonadotropic, hypergonadotropic or compensated hypogonadism. Men with hypogonadotropic hypogonadism had higher BMI and lower SHBG concentrations than other groups. The free T concentrations of men with compensated hypogonadism were lower than eugonadal men but higher than those with hypogonadotropic or hypergonadotropic hypogonadism. As expected, men with CKD had lower hemoglobin concentrations than men without CKD (table 1). In men without CKD, the hemoglobin concentrations were lower in those with subnormal free T concentrations as compared to men with normal free T concentrations (139 ±13 vs. 144±10 g/l, P=0.02). Similarly, in men with CKD, the hemoglobin concentrations were lower in those with subnormal free T concentrations as compared to men with normal free T concentrations (119 ±19 vs. 128±19 g/l, P=0.04). Hemoglobin concentrations were positively related to egfr (r=0.61, P<0.001) and to free T concentrations (r=0.31, P<0.001), and inversely related to age(r= -0.17, P=0.01). In multivariate regression analyses, both egfr (β=0.60, P<0.001) and free T concentrations (β=0.21, P=0.001) were predictors of hemoglobin concentrations, but age was not (β=0.10, P=0.13) Comparison across CKD stages: We then compared the hormone concentrations of men in CKD stages 3, 4 or 5 with those of men without CKD (table 1). There was a progressive decline in T concentrations and an increase in the prevalence of subnormal free T concentrations with worsening renal insufficiency.. This was accompanied by an increase in LH, FSH and prolactin concentrations. While there was no change in the prevalence of hypogonadotropic hypogonadism, there was an increase in the prevalence of hypergonadotropic hypogonadism 10

11 Page 11 of with worsening CKD. Approximately half of all men with renal failure (CKD stage 5) had hypergonadotropic hypogonadism. There was an increase in prolactin concentrations with worsening CKD. The prevalence of supranormal prolactin concentrations (>0.78 nmol/l) was 3% in men with normal egfr, 3% in CKD stage 3, 11% in CKD stage 4 and 27% in CKD stage 5 (P=0.001 for comparison by chisquare across groups). The prevalence of hypogonadotropic hypogonadism was similar in men with and without hyperprolactinemia (25% vs 47%, P=0.24). Hyperprolactinemia >1.3 nmol/l was not detected in any patient with normal egfr and CKD stage 3. Prolactinemia above 1.3 nmol/l was present in one patient with stage 4 CKD. His prolactin concentration was 1.6 nmol/l and he had normal free T concentrations. Hyperprolactinemia was present in 4 patients with stage 5 CKD (prolactin concentrations were 1.4, 1.9, 2.3, 3.4 nmol/l). All 4 patients had low free T concentrations; 2 patients were hypogonadotropic and 2 were hypergonadotropic. All patients with prolactin concentrations >1.3 nmol/l were evaluated by an endocrinologist and were not found to have a prolactinoma. There was a progressive decrease in hemoglobin concentrations across progression of CKD stages (table 1). 13% of all patients with CKD had hemoglobin concentrations less than 100 g/l (the threshold for use of erythropoietin in clinical practice). Only one patient with hemoglobin <10 g/dl had normal free T concentrations. The rest of these patients had subnormal free T concentrations. We also compared the T concentrations of men without CKD but with or without albuminuria. 18% of men without CKD had micro-albuminuria and 3% had macro-albuminuria. The total and free T concentrations were similar in men with and without albuminuria 11

12 Page 12 of (13.16±7.01 and 12.53±5.14 nmol/l for total T, p=0.64; 0.21±0.09 and 0.19±0.18 nmol/l for free T, p=0.75) DISCUSSION We have shown that two-thirds of men with CKD and type 2 diabetes have low free T concentrations. Another 10% had compensated hypogonadism, a condition that is likely to convert into hypergonadotropic hypogonadism in future (16). Thus only 24% of men with CKD had normal T and LH concentrations. The situation was particularly dismal in stage 4 and stage 5 in which ~90% of men were either hypogonadal and or had compensated hypogonadism. Our study is also the first to comprehensively describe the prevalence of hypogonadotropic, hypergonadotropic and compensated hypogonadism in CKD. We found that the prevalence of hypogonadotropic hypogonadism in men with type 2 diabetes was similar across all stages of CKD (range: 32-46%). This is consistent with the prevalence described in type 2 diabetic men without CKD (2, 3, 6). In contrast, we found that the prevalence of hypergonadotropic hypogonadism increased with worsening CKD. The median gonadotropin concentrations of men with CKD stage 5 were 2-3 times higher than men with normal egfr. Studies that have measured gonadotropin concentrations in men with CKD stage 5 have also reported similarly high gonadotropin concentrations (17-19). Deconvolution analyses of frequently measured LH concentrations over 24 hours have shown that the half life of LH is doubled but the daily secretion rate of LH is lower by 50% in eugonadal men with CKD as compared to healthy men(20). This results in 58% higher mean LH concentrations over 24 hours in men with CKD. However, the T concentrations of men with CKD in our study were lower inspite of the higher circulating LH concentrations. Men with renal failure also have decreased spermatogenesis, 12

13 Page 13 of infertility and decreased concentrations of anti-mullerian hormone(8, 17). Thus, there is evidence of both leydig cell and sertoli cell dysfunction in men with renal failure. It is possible that CKD is associated with an increase in inactive LH fragments in the circulation. There are data to suggest that immunoreactive but biologically inactive LH fragments may be increased in renal failure (21, 22). We did not measure LH bioactivity in our patients. It is possible that a portion of LH concentrations in our patients with renal failure is not bioactive and that some men with hypergonadotropic hypogonadism may actually have hypogonadotropic hypogonadism(23). Similarly some men with compensated hypogonadism may be eugonadal. However, we found that the T concentrations of men with compensated hypogonadism were lower than those of eugonadal men (table 2). This suggests that LH concentrations in these men were appropriately elevated as a result of decreased T concentrations and not because they are inactive metabolites that have accumulated due to decreased clearance in CKD. Chronic renal failure is associated with an increase in prolactin concentrations (24, 25). Decreased clearance as well as increased production are responsible for the elevation of prolactin in CKD (26). Consistent with other studies, we found higher prolactin concentrations in patients with CKD(17). However, the mild elevations in prolactin had no impact on the prevalence of hypogonadotropic hypogonadism. Neither was there any difference in prolactin concentrations among men with hypogonadotropic or hypergonadotropic hypogonadism (table 2). The hyperprolactinemia in renal failure is not corrected by dialysis. Nor is it likely that hemodialysis is responsible for the low T concentrations in CKD because the clearance of T by hemodialysis is miniscule(27). 13

14 Page 14 of Prior studies have found that total T concentrations are frequently low in men with renal failure. Studies in hemodialysis patients have reported that 44-57% of the men have subnormal total T concentrations (12, 13, 28). In one study of 239 patients referred to a renal center, the prevalence of subnormal total T concentrations was 17%, 17%, 34%, 38% and 57% in CKD stages 1,2,3,4 and 5 respectively(10). In a large study of 2,419 patients with CKD stage 3 or 4, 53% were found to have subnormal total T concentrations(11). However, half of these patients had their blood sample drawn in the afternoon. Since T concentrations can decline by 30% as the day progresses, this would falsely increase the prevalence of subnormal T concentrations. Overall the literature suggests that approximately half of men with ESRD have low total T concentrations and approximately one-third of men with milder forms of renal failure have low total T. These studies also found that the total T concentrations were lower in men with type 2 diabetes and CKD as compared to men without diabetes and CKD(10-12, 29). These studies have used immunoasssays to measure total T concentrations in men with CKD. Immunoassays have poor accuracy at lower T concentrations and are not recommended for assessment of hypogonadism. The recommended assay to measure total T is LC-MS/MS. Only one study has reported total T concentrations measured by LC-MS/MS in men with CKD (29). However, there was no comparison group of men with normal egfr, nor was the prevalence of subnormal total T concentrations mentioned. Free T concentrations were not reported. Ours is the only study in men with CKD that has measured total and free T by the recommended gold-standard methodology : LC-MS/MS and equilibrium dialysis (5). We found that subnormal free T concentrations were associated with lower hemoglobin concentrations. This is consistent with the known erythropoetic effect of testosterone. We have previously shown that hemoglobin concentrations are lower in hypogonadal type 2 diabetic men 14

15 Page 15 of without CKD(30). Our current data indicate that this is true in type 2 diabetic men with CKD as well. One previous study has also found that total T concentrations are inversely related to hemoglobin concentrations in men with CKD(31). Furthermore, almost all men with hemoglobin concentrations <100 g/l had subnormal free T concentrations. It is well known that decreased erythropoietin concentrations are largely responsible for renal failure associated anemia. Since we excluded CKD patients who were on erythropoietin, we cannot comment on the relative contributions of erythropoietin and testosterone on hemoglobin concentrations in CKD. Neither is it known if T replacement therapy to normalize the T concentrations would decrease the need of erythropoietin therapy in anemic men with CKD(32, 33). It is well known that there is significant day-to-day variability in hormone concentrations, especially testosterone. As in most epidemiologic or cross-sectional studies, the hormone concentrations were measured only once in this study. In view of the variability in testosterone concentrations, this is a limitation. However, it is not likely that the prevalence of subnormal free testosterone concentrations would have altered following repeated measurements since the probability of testosterone concentrations rising or falling with repeated measurements is statistically equal. The issue of repeated measurements is important in the context of diagnosing hypogonadism clinically in individual patients. The fact that our study included a moderately large number of participants also helps to diminish the effect of hormonal variability on study results. Another limitation of our study is that we did not collect data on signs or symptoms of hypogonadism in patients. There is a high prevalence of symptoms of hypogonadism in men with renal failure(34-36). Many of the signs and symptoms of hypogonadism such as erectile dysfunction, fatigue, anemia, poor quality of life, sarcopenia and low bone mass are also 15

16 Page 16 of believed to be a consequence of renal failure(37). This limits the predictive value of screening for signs and symptoms of hypogonadism in men with CKD. This probably explains why very few CKD patients (12% in one study) have their T concentrations tested for clinical indications (11, 37). It is possible that T replacement therapy may improve sexual symptoms, fatigue, anemia, muscle mass and bone mass in hypogonadal men with CKD. The role of T in men with CKD needs to be thoroughly evaluated. This is imperative since low total T concentrations have been consistently associated with increased mortality in men with CKD(11, 29). Whether T is a marker or mediator of mortality in men with CKD can only be answered through trials of T replacement. In summary, we have demonstrated that two-thirds of men with type 2 diabetes and CKD have subnormal free T concentrations. As shown in prior studies by us and others, we found that approximately one-third of men with type 2 diabetes have hypogonadotropic hypogonadism. This was consistent across all stages of CKD. The presence of CKD was additionally associated with an increase in the prevalence of hypergonadotropic hypogonadism and compensated hypogonadism in men with type 2 diabetes. Subnormal free T concentrations were associated with lower hemoglobin concentrations in men with CKD Acknowledgements The study was supported in part by Clinical Research Institute of Texas Tech University Health Sciences Center. Paresh Dandona is supported by grants from NIH (R01-DK and R01- DK075877), Juvenile Diabetes Research Foundation ( ), and the American Diabetes Association (112CT20). Paresh Dandona also has support from Merck, Novo Nordisk, Boehringer Ingelheim, and AbbVie Pharmaceuticals. 16

17 Page 17 of Disclosure Statement SD and PD have received lecture fees from Abbvie. AR, JK, SB, AC, AM, KR and MB have nothing to disclose References Dandona, P., and Dhindsa, S Update: hypogonadotropic hypogonadism in type 2 diabetes and obesity. J Clin Endocrinol Metab 96: Dhindsa, S., Prabhakar, S., Sethi, M., Bandyopadhyay, A., Chaudhuri, A., and Dandona, P Frequent occurrence of hypogonadotropic hypogonadism in type 2 diabetes. J Clin Endocrinol Metab 89: Dhindsa, S., Miller, M.G., McWhirter, C.L., Mager, D.E., Ghanim, H., Chaudhuri, A., and Dandona, P Testosterone concentrations in diabetic and nondiabetic obese men. Diabetes Care 33: Chandel, A., Dhindsa, S., Topiwala, S., Chaudhuri, A., and Dandona, P Testosterone concentration in young patients with diabetes. Diabetes Care 31: Bhasin, S., Cunningham, G.R., Hayes, F.J., Matsumoto, A.M., Snyder, P.J., Swerdloff, R.S., and Montori, V.M Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 95: Kapoor, D., Aldred, H., Clark, S., Channer, K.S., and Jones, T.H Clinical and Biochemical Assessment of Hypogonadism in Men With Type 2 Diabetes: Correlations with bioavailable testosterone and visceral adiposity. Diabetes Care 30: United States Renal Data System Handelsman, D.J Hypothalamic-pituitary gonadal dysfunction in renal failure, dialysis and renal transplantation. Endocr Rev 6: de Vries, C.P., Gooren, L.J., and Oe, P.L Haemodialysis and testicular function. Int J Androl 7: Yilmaz, M.I., Sonmez, A., Qureshi, A.R., Saglam, M., Stenvinkel, P., Yaman, H., Eyileten, T., Caglar, K., Oguz, Y., Taslipinar, A., Vural, A., Gok, M., Unal, H.U., Yenicesu, M., Carrero, J.J Endogenous testosterone, endothelial dysfunction, and cardiovascular events in men with nondialysis chronic kidney disease. Clin J Am Soc Nephrol 6: Khurana, K.K., Navaneethan, S.D., Arrigain, S., Schold, J.D., Nally, J.V., Jr., and Shoskes, D.A Serum testosterone levels and mortality in men with CKD stages 3-4. Am J Kidney Dis 64: Bello, A.K., Stenvinkel, P., Lin, M., Hemmelgarn, B., Thadhani, R., Klarenbach, S., Chan, C., Zimmerman, D., Cembrowski, G., Strippoli, G., Carrero J.J., Tonelli, M Serum testosterone levels and clinical outcomes in male hemodialysis patients. Am J Kidney Dis 63: Carrero, J.J., Qureshi, A.R., Nakashima, A., Arver, S., Parini, P., Lindholm, B., Barany, P., Heimburger, O., and Stenvinkel, P Prevalence and clinical implications of testosterone deficiency in men with end-stage renal disease. Nephrol Dial Transplant 26:

18 Page 18 of Salameh, W.A., Redor-Goldman, M.M., Clarke, N.J., Reitz, R.E., and Caulfield, M.P Validation of a total testosterone assay using high-turbulence liquid chromatography tandem mass spectrometry: total and free testosterone reference ranges. Steroids 75: Levey, A.S., Stevens, L.A., Schmid, C.H., Zhang, Y.L., Castro, A.F., 3rd, Feldman, H.I., Kusek, J.W., Eggers, P., Van Lente, F., Greene, T., Coresh, J., CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) A new equation to estimate glomerular filtration rate. Ann Intern Med 150: Tajar, A., Forti, G., O'Neill, T.W., Lee, D.M., Silman, A.J., Finn, J.D., Bartfai, G., Boonen, S., Casanueva, F.F., Giwercman, A., Han, T.S., Kula, K., Labrie, F., Lean, M.E., Pendleton, N.,Punab, M., Vanderschueren, D., Huhtaniemi, I.T., Wu, F.C. EMAS Group Characteristics of secondary, primary, and compensated hypogonadism in aging men: evidence from the European Male Ageing Study. J Clin Endocrinol Metab 95: Eckersten, D., Giwercman, A., and Christensson, A Male patients with terminal renal failure exhibit low serum levels of antimullerian hormone. Asian J Androl 17: Molsted, S., Andersen, J.L., Eidemak, I., Harrison, A.P., and Jorgensen, N Resistance training and testosterone levels in male patients with chronic kidney disease undergoing dialysis. Biomed Res Int 2014: Albaaj, F., Sivalingham, M., Haynes, P., McKinnon, G., Foley, R.N., Waldek, S., O'Donoghue, D.J., and Kalra, P.A Prevalence of hypogonadism in male patients with renal failure. Postgrad Med J 82: Veldhuis, J.D., Wilkowski, M.J., Zwart, A.D., Urban, R.J., Lizarralde, G., Iranmanesh, A., and Bolton, W.K Evidence for attenuation of hypothalamic gonadotropin-releasing hormone (GnRH) impulse strength with preservation of GnRH pulse frequency in men with chronic renal failure. J Clin Endocrinol Metab 76: Schaefer, F., Veldhuis, J.D., Robertson, W.R., Dunger, D., and Scharer, K Immunoreactive and bioactive luteinizing hormone in pubertal patients with chronic renal failure. Cooperative Study Group on Pubertal Development in Chronic Renal Failure. Kidney Int 45: Mitchell, R., Bauerfeld, C., Schaefer, F., Scharer, K., and Robertson, W.R Less acidic forms of luteinizing hormone are associated with lower testosterone secretion in men on haemodialysis treatment. Clin Endocrinol (Oxf) 41: Handelsman, D.J., Spaliviero, J.A., and Turtle, J.R Bioactive luteinizing hormone in plasma of uraemic men and men with primary testicular damage. Clin Endocrinol (Oxf) 24: Katz, A.I., and Emmanouel, D.S Metabolism of polypeptide hormones by the normal kidney and in uremia. Nephron 22: Yavuz, D., Topcu, G., Ozener, C., Akalin, S., and Sirikci, O Macroprolactin does not contribute to elevated levels of prolactin in patients on renal replacement therapy. Clin Endocrinol (Oxf) 63: Veldhuis, J.D., Iranmanesh, A., Wilkowski, M.J., and Samojlik, E Neuroendocrine alterations in the somatotropic and lactotropic axes in uremic men. Eur J Endocrinol 131: Singh, A.B., Norris, K., Modi, N., Sinha-Hikim, I., Shen, R., Davidson, T., and Bhasin, S Pharmacokinetics of a transdermal testosterone system in men with end stage renal disease receiving maintenance hemodialysis and healthy hypogonadal men. J Clin Endocrinol Metab 86: Kyriazis, J., Tzanakis, I., Stylianou, K., Katsipi, I., Moisiadis, D., Papadaki, A., Mavroeidi, V., Kagia, S., Karkavitsas, N., and Daphnis, E Low serum testosterone, arterial stiffness and mortality in male haemodialysis patients. Nephrol Dial Transplant 26:

19 Page 19 of Grossmann, M., Hoermann, R., Ng Tang Fui, M., Zajac, J.D., Ierino, F.L., and Roberts, M.A Sex steroids levels in chronic kidney disease and kidney transplant recipients: associations with disease severity and prediction of mortality. Clin Endocrinol (Oxf). 30. Bhatia, V., Chaudhuri, A., Tomar, R., Dhindsa, S., Ghanim, H., and Dandona, P Low testosterone and high C-reactive protein concentrations predict low hematocrit in type 2 diabetes. Diabetes Care 29: Carrero, J.J., Barany, P., Yilmaz, M.I., Qureshi, A.R., Sonmez, A., Heimburger, O., Ozgurtas, T., Yenicesu, M., Lindholm, B., and Stenvinkel, P Testosterone deficiency is a cause of anaemia and reduced responsiveness to erythropoiesis-stimulating agents in men with chronic kidney disease. Nephrol Dial Transplant 27: Yang, Q., Abudou, M., Xie, X.S., and Wu, T Androgens for the anaemia of chronic kidney disease in adults. Cochrane Database Syst Rev 10:CD Brockenbrough, A.T., Dittrich, M.O., Page, S.T., Smith, T., Stivelman, J.C., and Bremner, W.J Transdermal androgen therapy to augment EPO in the treatment of anemia of chronic renal disease. Am J Kidney Dis 47: El-Assmy, A Erectile dysfunction in hemodialysis: A systematic review. World J Nephrol 1: Yavuz, D., Acar, F.N., Yavuz, R., Canoz, M.B., Altunoglu, A., Sezer, S., and Durukan, E Male sexual function in patients receiving different types of renal replacement therapy. Transplant Proc 45: Park, M.G., Koo, H.S., and Lee, B Characteristics of testosterone deficiency syndrome in men with chronic kidney disease and male renal transplant recipients: a cross-sectional study. Transplant Proc 45: Carrero, J.J Testosterone deficiency at the crossroads of cardiometabolic complications in CKD. Am J Kidney Dis 64:

20 Page 20 of 23 Table 1 shows the hormone concentrations and demographics of study participants. CKD stage Number of subjects CKD (egfr <60 ml/min/ 1.73m 2 ) Normal (egfr 60 ml/min/1.7 3m 2 ) P as compared to non- CKD Stage 3 (egfr ml/min/1.73 m 2 ) Stage 4 (egfr ml/min/1.73 m 2 ) Stage 5 (egfr <15 ml/min/ 1.73m 2 ) P across groups (normal, stages 3-5) egfr (ml/min/1.73m 2 ) 35±17 90±15 < ±10* # 23±6* # 11±4* <0.001 Age (years) 62±11 56±9 < ±10* # 67±10* # 58±11 <0.001 Weight (kg) 101±25 107± ±23 102±29 94±24* 0.06 BMI (kg/m 2 ) 31.9 [29.2, 35.4] 33.2 [29.2, 38.5] [29.2, 36.6] 32.0 [28.2, 38.2] 30.2 [26.7, 33.5] 0.39 Total T (nmol/l) 9.72± ±5.9 < ±5.73 # ±4.41* # 6.04±3.58* <0.001 Free T (nmol/l) 0.18± ± ±0.08 # 0.19±0.07 # 0.14±0.07* SHBG (nmol/l) 30±16 28± ±18 25±18 32± LH (IU/L) 7.9 [5.6, 12.3] 4.4 [3.0, 7.0] < [4.5, 8.2 [4.7, 12.3 [7.4, 10]* # 14.7]* # 16.1]* <0.001 FSH(IU/L) 8.2 [5.1, 12.3] 6.4 [3.8, 9.8] < [4.1, 9.8 [5.3, 11.8]* # 15.6] 12.3 [6.3, 17.9]* <0.001 Prolactin (nmol/l) 0.39 [0.32, 0.47] 0.28 [0.21, 0.37] < [0.27, 0.44 [0.37, 0.42]* # 0.62]* 0.47 [0.36, 0.83]* <0.001 A1c % 7.1 [6.5, 8.1] 7.1 [6.3, 8.4] [6.8, 8.3] 7.1 [6.5, 8.0] 6.7 [5.2, 8.9] 0.22 Duration of diabetes (years) 18±10 10±7 < ±10* 18±9* 22±6* 0.19 Hypertension 97% 86% % 100%* 100%* 0.40 Hemoglobin (g/l) 122±19 142±12 < ±23* # 116±15* 108±15* <

21 Page 21 of 23 % with subnormal free T Hypogonadotropi c Hypogonadism Hypergonadotropi c hypogonadism Compensated hypogonadism 66% 37% < %* 63%* 79%* < % 32% % 46% 25% % 5% < %* # 17%* # 54%* < % 7% % 25%* 11% 0.02 Medications Metformin 17% 82% < %* # 0%* 0%* <0.001 Sulfonylureas 23% 52% < % # 14%* # 0%* <0.001 Insulin 84% 60% < %* 93%* 87%* <0.001 GLP-1 agonists 14% 22% % # 11% 3%* 0.06 DPP-4 inhibitors 19% 18% % 14% 10% 0.15 Thiazolidinediones 26% 33% % # 39% # 0% Diet only 4% 8% % 7% 3% 0.01 ACE-I/ARB 57% 68% % # 33%* 33%* 0.04 Other hypertensives 83% 48% %* # 83%* 100%* Statin use 73% 67% % 67% 78% 0.89 Data are presented as means±sd for normally distributed data and median [25 th, 75 th percentile] for non-normal data. The 4 th and 8 th columns show the P values (by t test, χ 2 square, ANOVA or ANCOVA) for comparison across groups. SHBG, free T concentrations, prevalence of subnormal free T and prevalence of cardiovascular disease were adjusted for age and BMI differences between groups. Total T concentrations were adjusted for age, BMI and SHBG differences. To convert T into SI units (nmol/liter), divide by Data are presented as means±sd for normally distributed data and median [25 th, 75 th percentile] for non-normal data. SHBG concentrations were not normally distributed and are presented 2

22 Page 22 of 23 as reverse log-transformed values after adjustment for age and BMI. 5 men in CKD stage 5 were on erythropoietin; hence their hemoglobin concentrations were not included in the study. *P<0.05 as compared to normal egfr category # P<0.05 when comparing stages 3 or 4 to stage 5 ACE-I: Angiotensin converting enzyme-inhibitor; ARB: Angiotensin receptor blocker; GLP: Glucagon like peptide; DPP: Dipeptidyl peptidase 3

23 Page 23 of 23 Table 2: Comparison of men with CKD across gonadal categories Hypogonadotropic Hypogonadism Hypergonadotropic Hypogonadism Compensated Hypogonadism Eugonadal Number of subjects Age (years) 62±9 62±12 65±12 61±11 BMI (kg/m 2 ) 35.7±8.5 # 32.1± ± ±4.2 Total T (nmol/l) 10.1±4.6* # 7.9±4.5* 12.4±4.8* # 15.5±5.3 # Free T (nmol/l) 0.14±0.05* 0.14±0.06* 0.20±0.07* # 0.24±0.06 # SHBG (nmol/l) 27±15 37±12 27±8 33±14 LH (IU/L) 5.3 [4.2, 7.1] # 15.4 [12.3, 23.2]* 14.7 [12.0, 25.1]* 6.0 [4.5, 7.3] # FSH(IU/L) 6.5 [3.9, 10.8] # 13.4 [10.6, 22.5]* 23.0 [9.1, 42.9]* 7.2 [4.2, 9.9] # Prolactin (nmol/l) 0.38 [0.28, 0.48] 0.44 [0.33, 0.74] 0.46 [0.32, 0.68] 0.37 [0.27, 0.53] Total T concentrations were adjusted for age and SHBG differences among groups. Free T and SHBG concentrations were adjusted for age difference between groups. SHBG concentrations were not normally distributed and are presented as reverse log-transformed values after adjustment for age and BMI. *P<0.05 as compared to eugonadal men # P<0.05 as compared to men with hypergonadotropic hypogonadism