Men with kidney failure and diabetes commonly have low testosterone (T) levels [1Gungor O. Kircelli F. Carrero J.J. Asci G. Toz H. Tatar E. et al.Endogenous testosterone and mortality in male hemodialysis patients: Is it the result of aging?.Clin J Am Soc Nephrol. 2010; 5: 2018-2023Crossref PubMed Scopus (0) Google Scholar], which have been associated with increased risk of all-cause mortality and cardiovascular (CV) events during hemodialysis (HD) [2van der Burgh A.C. Khan S.R. Neggers S.J.C.M.M. Hoorn E.J. Chaker L. The role of serum testosterone and dehydroepiandrosterone sulfate in kidney function and clinical outcomes in chronic kidney disease: A systematic review and meta-analysis.Endocrine Connections. 2022; 11e220061Crossref Scopus (1) Google Scholar]. Additionally, low T levels are associated with a higher risk of incident diabetes [3Yao Q.M. Wang B. An X.F. Zhang J.A. Ding L. Testosterone level and risk of type 2 diabetes in men: A systematic review and meta-analysis.Endocr Connect. 2018; 7: 220-231Crossref Scopus (39) Google Scholar] and lower health-related quality of life (HRQOL) [4Soleymanian T. Kokabeh Z. Ramaghi R. Mahjoub A. Argani H. Clinical outcomes and quality of life in hemodialysis diabetic patients versus non-diabetics.J Nephropathol. 2017; 6: 81-89Crossref Scopus (18) Google Scholar], which in turn is linked to an increased risk of CVD and premature mortality compared to otherwise similar men on HD with normal T levels [1Gungor O. Kircelli F. Carrero J.J. Asci G. Toz H. Tatar E. et al.Endogenous testosterone and mortality in male hemodialysis patients: Is it the result of aging?.Clin J Am Soc Nephrol. 2010; 5: 2018-2023Crossref PubMed Scopus (0) Google Scholar]. Despite known diabetes-mortality associations [4Soleymanian T. Kokabeh Z. Ramaghi R. Mahjoub A. Argani H. Clinical outcomes and quality of life in hemodialysis diabetic patients versus non-diabetics.J Nephropathol. 2017; 6: 81-89Crossref Scopus (18) Google Scholar], the effects of diabetes on the relationship between T, mortality rates, and HRQOL among HD patients is understudied. Investigating whether diabetes modifies the association between low T, mortality, and poor HRQOL in male HD patients could enable early risk stratification and improve patient care management, given the potential for worsened CV outcomes through the intersection of diabetes, abnormal hormonal milieu, and vascular endothelial dysfunction. Given the potential for increased mortality risk and adverse outcomes in men on HD with low T and diabetes, we extended our previous work [5Bello A.K. Stenvinkel P. Lin M. Hemmelgarn B. Thadhani R. Klarenbach S. et al.Serum testosterone levels and clinical outcomes in male hemodialysis patients.Am J Kidney Dis. 2014; 63: 268-275Abstract Full Text Full Text PDF PubMed Google Scholar] using data from a prospective, multi-centre study to evaluate whether diabetes or age would influence prospective associations between T levels, mortality, and adverse clinical outcomes in a cohort of male HD patients. The present study was reported according to Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Supplementary Material). Table 1 shows the baseline characteristics of 587 men analyzed in this study (Supplementary Methods and References), with 317 (54%) having low, 170 (29%) borderline, and 100 (17%) normal T levels. Of those with low, borderline, and normal T levels, 54.4%, 47.1%, and 39.9% had diabetes, respectively.Table 1Baseline characteristicsCharacteristicLow Testosterone (n = 318)Non-low Testosterone (n = 269)p-valueAge (years)64 (54, 73)61 (48, 72)0.009BMI (kg/m2)27 (23, 31)26 (23, 30)0.116Systolic BP (mmHg)135 (118, 151)141 (127, 152)0.002Ethnicity (%)White260 (81.8)201 (74.7)0.116Indigenous16 (5.0)18 (6.7)Other42 (13.2)50 (18.6)Smoker (%)56 (17.7)61 (22.8)0.125GN/autoimmune (%)54 (17.0)62 (23.0)0.066Comorbidities (%) aPrevalent CVD183 (57.5)115 (42.8)<0.001Mental health104 (32.7)91 (33.8)0.773Other serious illnesses107 (33.6)89 (33.1)0.885Health-related quality of lifeHUI30.71 (0.51, 0.91)0.84 (0.61, 0.93)<0.001KDQOL12-PCS31 (25, 37)35 (28, 42)<0.001KDQOL12-MCS48 (36, 55)47 (37, 55)0.600Albumin (g/L)33 (29, 37)35 (31, 38)0.040Sex hormone-binding globulin (nmol/L)33 (23, 46)47 (35, 62)<0.001Testosterone (nmol/L)4.8 (3.1, 6.4)11.1 (9.4, 13.7)<0.001Free testosterone (pmol/L)89 (56, 121)169 (137, 205)<0.001Measure of central tendencies are reported as medians with corresponding 25th and 75th percentiles.a Comorbidities include CVD (stroke, coronary artery disease, heart failure, peripheral vascular disease), mental health (substance misuse, psychiatric disorder), and other serious illnesses (cancer, chronic respiratory disease, chronic liver disease, dementia).Abbreviations: BMI: body mass index, BP: blood pressure, CVD: cardiovascular disease, GN: glomerulonephritis, KDQOL 12-MCS: Kidney Disease Quality of Life Mental Component Score, KDQOL 12-PCS: Kidney Disease Quality of Life Physical Component Score. Open table in a new tab Measure of central tendencies are reported as medians with corresponding 25th and 75th percentiles. a Comorbidities include CVD (stroke, coronary artery disease, heart failure, peripheral vascular disease), mental health (substance misuse, psychiatric disorder), and other serious illnesses (cancer, chronic respiratory disease, chronic liver disease, dementia). Abbreviations: BMI: body mass index, BP: blood pressure, CVD: cardiovascular disease, GN: glomerulonephritis, KDQOL 12-MCS: Kidney Disease Quality of Life Mental Component Score, KDQOL 12-PCS: Kidney Disease Quality of Life Physical Component Score. Over a median follow-up of 3.7 years (IQR: 1.4-8.0; range: 5 days–13.1 years), 318 (54%) participants died, and 118 (37%) of those deaths were attributed to CV-related causes. There were no significant associations between low T and all-cause mortality (HR = 1.24, 95% CI: 0.97, 1.59), or CV mortality (HR = 0.85, 95% CI: 0.57, 1.27) (Table S1). There was no significant association between low T and incident CVD (HR = 0.87, 95% CI: 0.53, 1.43) (Figure 1). In sensitivity analyses, continuous serum total T was associated with increased incident CVD risk (HR = 1.31, 95% CI: 1.04, 1.64). Low T was prospectively associated with lower HUI3 scores (MDHUI3=-0.06, 95% CI: -0.10, -0.02) and KDQOL12-PCS scores (MDKDQOL12-PCS=-1.87, 95% CI: -3.62, -0.11), but not with KDQOL12-MCS scores (MDKDQOL12-MCS=-0.01, 95% CI: -1.98, 1.97) (Table S2). The association between serum T levels and HUI3 varied between younger (<63 years) and older (≥63 years) men (interaction p=0.044) with serum T levels positively associated with better HUI3 scores only among older men (MDHUI3=0.04, 95% CI: 0.01, 0.06) (Table S2). Diabetes was not a significant effect modifier for any relationships (all interactions p ≥ 0.170). BIC values for linear models were lower for all-cause and CV mortality models, but for incident CVD, the BIC value for the non-linear model was lower than that of the linear model (BIC 858 vs. 859) (Figure S1C). These findings update a previous study conducted by Bello et al [5Bello A.K. Stenvinkel P. Lin M. Hemmelgarn B. Thadhani R. Klarenbach S. et al.Serum testosterone levels and clinical outcomes in male hemodialysis patients.Am J Kidney Dis. 2014; 63: 268-275Abstract Full Text Full Text PDF PubMed Google Scholar] using data collected over ≈7 years from a prospective cohort of men undergoing HD in Canada. In this study, over a period of ≈14 years, no significant associations were observed between low T, all-cause and CV mortality, and incident CV events. However, higher T levels may increase the risk of incident non-fatal CVD events. Furthermore, low T was significantly associated with lower HUI3 and KDQOL12-PCS scores, but not with KDQOL12-MCS scores. These associations were not modified by diabetes status or age, highlighting the potential of T as a modifiable risk factor across these HD patient subgroups. Although not statistically significant, our findings on the association between low T levels and all-cause and CV mortality align with recent reports [2van der Burgh A.C. Khan S.R. Neggers S.J.C.M.M. Hoorn E.J. Chaker L. The role of serum testosterone and dehydroepiandrosterone sulfate in kidney function and clinical outcomes in chronic kidney disease: A systematic review and meta-analysis.Endocrine Connections. 2022; 11e220061Crossref Scopus (1) Google Scholar, 6Nilsson E. Stenvinkel P. Liu S. Stedman M.R. Chertow G.M. Floege J. Serum testosterone concentrations and outcomes in hemodialysis patients enrolled in the evolve trial.Nephrology Dialysis Transplantation. 2022; : gfac278Google Scholar]. Men with lower T levels had a higher risk of all-cause mortality when T was examined as a continuous variable in a flexible model (Figure S1A). On another note, SHBG determines T and fT concentrations and higher SHBG levels has been associated with higher risk of death or CV event among men on HD [6Nilsson E. Stenvinkel P. Liu S. Stedman M.R. Chertow G.M. Floege J. Serum testosterone concentrations and outcomes in hemodialysis patients enrolled in the evolve trial.Nephrology Dialysis Transplantation. 2022; : gfac278Google Scholar]. Nilsson et al. [6Nilsson E. Stenvinkel P. Liu S. Stedman M.R. Chertow G.M. Floege J. Serum testosterone concentrations and outcomes in hemodialysis patients enrolled in the evolve trial.Nephrology Dialysis Transplantation. 2022; : gfac278Google Scholar] indicated that protein-energy wasting could explain the link between fT and outcomes during HD, but further investigation is warranted. Notably, the directionality of our estimates is inconsistent with prior observational reports of associations between low T and higher risk of incident CVD. Incident CVD in prior reports was defined as a composite outcome comprising fatal and non-fatal CV events [2van der Burgh A.C. Khan S.R. Neggers S.J.C.M.M. Hoorn E.J. Chaker L. The role of serum testosterone and dehydroepiandrosterone sulfate in kidney function and clinical outcomes in chronic kidney disease: A systematic review and meta-analysis.Endocrine Connections. 2022; 11e220061Crossref Scopus (1) Google Scholar], but we included only incident CVD in our definition. Therefore, different CV outcome definitions could explain the discrepancies. Also, our data suggest that for every 1 SD increase in serum T, the risk of incident CVD increases. Our observations align with results of a meta-analysis, showing that T supplementation increases the risk of a CV-related event for older men [7Xu L. Freeman G. Cowling B.J. Schooling C.M. Testosterone therapy and cardiovascular events among men: A systematic review and meta-analysis of placebo-controlled randomized trials.BMC Medicine. 2013; 11: 108Crossref PubMed Scopus (0) Google Scholar]. Thus, more studies are needed to replicate our findings. Diabetes did not modify relationships between low T, mortality, and adverse health outcomes. The concomitant presence of low T and diabetes may worsen the inflammatory burden of HD patients. However, the moderate decrease in T levels among patients with diabetes similar to the decrease observed in chronic illnesses such as kidney failure and suggest that low T could be a non-specific biomarker of ill health [8Corona G. Monami M. Rastrelli G. Aversa A. Sforza A. Lenzi A. et al.Type 2 diabetes mellitus and testosterone: A meta-analysis study.Int J Androl. 2011; 34: 528-540Crossref PubMed Scopus (276) Google Scholar]. The positive association between serum total T and HUI3 scores were seen only in older patients (≥ 63 years), possibly due to healthy user bias. The relationship between low HUI3 scores and low T levels in our data suggest clinically-important changes [9Drummond M. Introducing economic and quality of life measurements into clinical studies.Annals of Medicine. 2001; 33: 344-349Crossref PubMed Google Scholar], but contextualising minimal clinically-important scores (MCID) remains challenging in patients with kidney disease. Therefore, more research is needed in this area. Study strengths include data from a carefully-phenotyped HD population and a prospective (≈14 years), multi-center design. We were able to ascertain which participants were taking exogenous sex hormones, adjust for multiple confounders, such as malignancies and chronic respiratory disorders, and include data for HRQOL. Finally, serum T concentrations were measured using standardized methods in a well-established laboratory. Study limitations include a single T measurement at baseline, which did not allow monitoring of changes in T levels over time. Although the study tried to adjust for relevant confounders, residual confounding remains likely due to the nature of observational studies. We excluded participants with missing T measurements, but the findings were not likely impacted (Table S3). Our findings are not generalizable because data used were from HD patients. In conclusion, no significant associations were seen between low serum T levels and mortality (all-cause and CV). However, low T was associated with lower HUI3 and KDQOL12-PCS scores, and the positive association between serum total T and HUI3 scores were apparent only among older patients. Future work is needed to explore the impact of diabetes on serum T levels and relationship to adverse health outcomes in patients on HD. The authors of this report are grateful to the study coordinators (Sue Szigety, Nasreen Ahmad, Coralea Bignell, Edita Delic, Sharon Gulewich, Julie Leidecker, Lorena McCoshen, Lisa McFaull, Mary Morgen, Nancy Ruholl, Rafael Sibrian, Charlynn Ursu, Gwen Winter and Jessica Wagner), research assistants (Lois Hannam, Rosie Hernandez, Bill Liu, Sandra Mackay, Shezel Muneer, and Bev Vanderham), Ghenette Houston for administrative support, Len Hannam for database development, Sophanny Tiv for quality assurance, and Dawn Opgenorth for project management. The authors also thank the dialysis patients who participated in this research. Download .pdf (.64 MB) Help with pdf files