Is There a Two-Way Risk between Decreased Testosterone Levels and the Progression and Prognosis of Chronic Kidney Disease? A Cohort Study Based on the National Health and Nutrition Examination Survey Database

Abstract

Purpose

The causal relationship between the incidence and prognosis of chronic kidney disease (CKD) and serum testosterone levels in patients is not yet fully understood. This study aims to use the National Health and Nutrition Examination Survey (NHANES), a large-scale nationally representative sample, to investigate the relationship between CKD and testosterone.

Materials and Methods

This study included six NHANES cycles for linear regression analysis, verified by multiple imputation methods. Stratified analysis and subgroup analysis were used to demonstrate the stability of CKD’s effect on testosterone. Furthermore, we used Kaplan-Meier plots and log-rank tests to evaluate differences in survival rates between CKD male patients with low and normal levels of testosterone.

Results

From a total of 71,163 subjects, the cohort selected 28,663 eligible participants. Results showed that CKD patients had testosterone levels 28.423 ng/mL (24.762, 32.083) lower than non-CKD patients. The results of multiple imputations (β=27.700, 95% confidence interval: 23.427, 31.974) were consistent with those of linear regression analysis, and the numerical match was good. Stratified regression analysis, and subgroup analysis results showed that CKD had a significant impact on testosterone at different dimensions. Kaplan-Meier plots showed significantly reduced survival rates in low testosterone CKD male patients (p<0.0001).

Conclusions

The results of this big data analysis suggest that there may be a two-way risk between low levels of testosterone and CKD. The testosterone levels of CKD patients were significantly lower than those of the non-CKD population, and CKD patients with low testosterone levels had poorer prognoses. These results suggest that correcting testosterone levels in a timely manner can have preventive and therapeutic effects on the progression of CKD.

INTRODUCTION

Chronic kidney disease (CKD) is a chronic, progressive disease in which there is a gradual decline in glomerular function. CKD has emerged as a major global public health concern, affecting more than 10% of the world’s population and the number of cases continues to rise [1]. Hemodialysis and kidney transplantation have significantly increased the life expectancy of CKD patients. Doctors and patients are increasingly focusing on the impact of complications related to CKD on overall health.

Previous research has shown that CKD patients often experience hypogonadism, which is a multifactorial complex of signs and symptoms in the areas of body composition, sexual and psychological health, as well as metabolic disorders, including lipid and carbohydrate abnormalities. In males, this manifests as erectile dysfunction, while in females it leads to menstrual irregularities and decreased fertility due to steroid hormone secretion disruption, especially testosterone deficiency [2]. Testosterone replacement therapy (TRT) has been shown to alleviate hypogonadism in CKD patients [3], and males with low testosterone levels are at a significantly greater risk for developing CKD than those with normal testosterone levels [4]. Testosterone plays an important role in improving alertness and physical performance, enhancing libido, increasing muscle strength, and boosting the immune system. The accumulation of toxins such as creatinine (Cr) and urea nitrogen in the bodies of CKD patients can disrupt the hypothalamic-pituitary-gonadal axis [5]. Shoskes et al [6] found that low testosterone levels at the time of kidney transplantation are a risk factor for patient mortality and graft failure. These studies suggest that CKD may lead to a decrease in testosterone levels and testosterone deficiency may be one of the risk factors for CKD, with testosterone playing a protective role in kidney function [7]. However, it is contradictory that some studies have shown a faster decline in renal function in male CKD patients than in female patients, which may be related to estrogen’s protective effects or the detrimental effects of testosterone [8].

The mechanism by which testosterone regulates kidney function is still not fully understood, and previous studies examining the correlation between CKD and testosterone have been mostly limited to small-scale surveys, resulting in no definitive conclusions. This study intends to capitalize on the extensive size and national representation of the National Health and Nutrition Examination Survey (NHANES) database to explore differences in testosterone levels between individuals without CKD and those with CKD and examine risk factors that influence testosterone secretion. The objective is to provide valuable insights for preventing and treating complications associated with testosterone deficiency in CKD patients.

MATERIALS AND METHODS

1. Study participants

All data in this study are from NHANES, a representative cross-sectional survey. It is designed to obtain information about the health and nutritional status of civilians in the United States, using multi-stage stratified samples. Trained interviewers utilize a computer-assisted personal interview system at Mobile Examination Centers (MECs) to conduct interviews, while participants undergo comprehensive physical examinations and provide blood and urine samples. Additional information about this program can be obtained through the NHANES website [9]. The study analyzed data from six NHANES cycles. Since testosterone was not measured during the years 2005–2010 and 2017–2018, we included specific periods of 1999–2000, 2001–2002, 2003–2004, 2011–2012, 2013–2014, and 2015–2016 for our analysis. All participants in the cohort underwent either Cr measurement in blood and urine or completed the KIQ025 questionnaire (‘received dialysis in past 12 months’). The survey protocol received approval from the NCHS Ethics Review Board (https://www.cdc.gov/nchs/nhanes/irba98.htm), and all participants provided written informed consent.

In this study, the covariates were categorized as demographics, medical conditions, and lifestyle factors. Demographic factors comprised sex, age, race/ethnicity, education status (merged data from codes ‘DMDEDUC2’ and ‘DMDEDUC3’), and marital status. Medical conditions included high blood pressure (HBP; defined as a history of physician-diagnosed hypertension, an average systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or current use of antihypertensive medication) and diabetes mellitus (DM; defined as a history of physician-diagnosed diabetes, hemoglobin A1c level ≥6.5%, or current use of diabetes medication). We also included body mass index (BMI; kg/m²), waist circumference (cm), hemoglobin (mg/dL), total cholesterol (TC; mg/dL), high-density lipoprotein cholesterol (HDL; mg/dL), anemia (hemoglobin level less than 12.7 mg/dL is defined as anemia [10]) and low-density lipoprotein cholesterol (LDL; mg/dL) in this analysis. Lifestyle factors comprised smoking status (never, ever, current) and alcohol consumption (yes, no). Public-use Linked Mortality Files for six NHANES cycles are available, providing mortality data from the survey enrollment date through December 31, 2019. The data were generated on April 28, 2022. We used survival time data in terms of the number of Person-Months of follow-ups from the NHANES interview date. The mortality source and cause of death were determined using death certificates. The research process is outlined in Fig. 1.

Fig. 1. A flowchart to study the relationship between CKD and testosterone. NHANES: National Health and Nutrition Examination Survey, CKD: chronic kidney disease, BMI: body mass index, TC: total cholesterol, HDL: high-density lipoprotein cholesterol, LDL: low-density lipoprotein cholesterol, DM: diabetes mellitus, HBP: high blood pressure.

2. Definition

Serum and urine samples were collected in MEC. Serum Cr (Scr) was measured using the Jaffé rate method, with calibration performed for subjects from 1999–2000 and 2001–2002 [11].

3. Assessment of CKD

Scr was calibrated for participants in all six NHANES cycles. The Cr level of each participant was measured at least four times and the average value was used in the formula below. The six NHANES survey cycles’ Cr levels were utilized to estimate glomerular filtration rate (eGFR) by applying the CKD Epidemiology Collaboration equation:

• eGFR=141×min (Scr/κ,1)^α×max (Scr/κ,1)^-1.209×0.993^Age×1.018 (if the participant is female)×1.159 (if the participant is African American)

• (κ is 0.7 for females and 0.9 for males, α is -0.329 for females and -0.411 for males. min indicates the minimum of Scr/κ or 1, and max indicates the maximum of Scr/κ or 1) [12].

CKD was defined as participants with an eGFR less than 60 mL/min/1.73 m² or a urinary albumin-to-creatinine ratio greater than or equal to 30 mg/g [13].

Total testosterone measurements in serum were collected from the NHANES database using isotope dilution liquid chromatography-tandem mass spectrometry (ID-LC-MS/MS) [14]. Analyses were conducted following the Strengthening the Reporting of Observational Studies in Epidemiology statement guidelines; females were excluded [15]. Low testosterone concentrations were determined as total testosterone <350 ng/dL, according to the recommendations of the International Society for the Study of the Aging Male (ISSAM) [16] and the International Society for Sexual Medicine (ISSM) [17]; otherwise, they were classified as normal total testosterone.

All-cause mortality: Using a unique study identifier, we determined mortality status for all-cause mortality by probabilistic matching with the National Death Index until December 31, 2019. Additional information about the matching method can be obtained from the National Center for Health Statistics [18].

4. Statistical analysis

In our study, we followed the official guidelines provided by NHANES for selecting weights, which involve identifying the variable that includes the smallest population and selecting the appropriate weights. Our data included MEC examination data, so we selected the sub-weights corresponding to MEC following the recommended weight selection guidelines. To compute the new sampling weights for the combined survey cycles as per the analysis guidelines of NHANES, we used the following formula: 2/6בwtint4yr’ for the 1999–2002 cycle comprising two periods, and 4/6בwtint4yr’ for the 2011–2016 cycle comprising four periods [19].

To explore the independent association, our study constructed three regression models: unadjusted, adjusted for age and race, and adjusted for all covariates. To avoid the bias that may arise from simply deleting data or using mean imputation to replace missing values, we used multiple imputations (MI) to increase the amount of covariate data and assess whether the error resulting from missing covariate data would affect our study outcomes. We utilized MI with five replications and the chained equation approach method in the R MI procedure to handle missing data. Moreover, we used stratified multivariate linear regression to conduct subgroup analyses to find acceptable groups. We then used Kaplan-Meier (K-M) curves and log-rank tests to assess the relationship between testosterone levels and survival among individuals with low CKD. The data were recorded using categorical, dichotomous, and continuous variables, expressed as counts/proportions or mean standard deviation as appropriate. Differences in clinical characteristics between groups were evaluated using weighted chi-square tests (categorical variables), weighted one-way analysis of variance (ANOVA) (continuous variables with normal distribution), or weighted Kruskal-Wallis's test (continuous variables with skewed distribution). Statistical analyses were performed using R 3.5.3 (http://www.r-project.org/) and Empower Stats software (http://www.empowerstats.com), with a p-value of less than 0.05 considered statistically significant.

RESULTS

Following a rigorous screening process, a total of 28,663 participants were deemed eligible for the final analysis. Among these individuals, 1,584 (5.53%) members of the general population were diagnosed with CKD. The sociodemographic and baseline clinical characteristics of both the investigated populations in the two groups are summarized in Table 1. In comparison to non-CKD patients, CKD patients were older, male, married, and white. They also had higher BMI, thicker waist circumference, and TC levels. While, serum HDL levels are lower than non-CKD groups. The CKD and non-CKD groups also displayed significant differences in terms of educational status and race. Moreover, a greater proportion of individuals with CKD who did not smoke or drink alcohol and had no DM or HBP were observed.

Table 1. Weighted descriptive statistics for participants whose demographic basic status and clinical database.

Characteristic		Overall (n=28,663)	CKD status		p-value
			Yes (n=1,584)	No (n=27,079)
Age (y)		38.201±22.365	69.364±12.079	39.710±19.582	<0.00001
BMI (kg/m2)		26.979±7.409	29.639±6.745	27.707±7.220	<0.00001
Waist (cm)		92.043±19.249	103.276±15.082	94.757±18.344	<0.00001
Hemoglobin (mg/dL)		13.867±1.496	13.444±1.683	14.130±1.423	<0.00001
Total cholesterol (mg/dL)		174.64±52.756	189.488±121.947	185.932±107.447	0.0035
	Missing	24.06
	HDL-cholesterol (mg/dL)	52.475±14.576	51.376±14.652	52.306±14.722	<0.00001
	Missing	46.69
	LDL-cholesterol (mg/dL)	82.844±55.688	89.659±114.714	95.249±122.974	0.0001
	Missing	65.21
Sex					0.0016
	Male	52.01	52.017	50.28
	Female	47.99	47.983	49.72
Race/ethnicity					<0.00001
	Mexican American	16.86	4.042	10.416
	Other Hispanic	10.42	3.449	6.536
	Non-Hispanic White	35.39	78.769	63.333
	Non-Hispanic Black	22.54	9.032	11.328
	Other race (including multi-racial)	14.79	4.708	8.388
Education status					<0.00001
	Less than high school	26.82	9.209	16.261
	High school	31.65	37.444	29.823
	More than high school	41.53	53.347	53.916
Marital status					<0.00001
	Married	41.42	76.123	48.464
	Divorced	7.3	11.425	8.078
	Single	23.03	12.026	24.854
	Missing	28.17	0.426	18.603
DM					<0.00001
	Yes	12.93	37.601	10.99
	No	72.52	53.814	76.41
	Missing	14.54	8.585	12.6
Smoking status					<0.00001
	Never	25.04	32.256	28.67
	Current	6.91	5.129	7.929
	Former	1.86	1.133	2.083
	Missing	65.09	61.482	61.319
HBP status					<0.00001
	Yes	25.04	67.618	23.745
	No	65.36	29.933	69.975
	Missing	9.6	2.45	6.279
Alcohol status					<0.00001
	Yes	14.93	22.445	17.136
	No	85.07	77.555	82.864
Anemia					<0.00001
	Yes	19.94	31.691	13.884
	No	80.06	68.309	86.116

Values are presented as mean±standard deviation (for age, BMI, waist, TC, HDL, hemoglobin, LDL. p-value was calculated by the weighted linear regression model) or percent (%) (for sex, race/ethnicity, education status, marital status, DM, Smoking status, HBP status, anemia, and alcohol status. p-value was calculated by the weighted chi-square test).

CKD: chronic kidney disease, BMI: body mass index, HDL: high-density lipoprotein, LDL: low-density lipoprotein, DM: diabetes mellitus, HBP: high blood pressure.

We conducted multivariate linear regression analysis. In an unadjusted model, Trends in testosterone were associated with CKD (β=19.819, 95% confidence interval [CI]: 14.897, 24.741). Following that, after adjusting for age, race, and BMI, we developed a second model, the pattern becomes more prominent (β=59.020, 95% CI: 55.522, 62.518). After adjusting for all variables, the results are consistent with the direction indicated by the previous two models and are also significant (β=28.423, 95% CI: 24.762, 32.083). To ensure the reliability of our results, we conducted regression analyses after using the MI method to fill in missing covariate data five times. The results of this analysis are displayed in Table 2. As sex, age, and BMI had no missing values, they were not adjusted for in the MI process. Both the Unadjusted and Adjusted I analyze in Table 3 were consistent with the original regression results. The MI results primarily appear in the Adjusted II analysis. The final result, obtained by combining the data from these five simulations, was 27.700 (23.427, 31.974). Overall, the five MI analyses indicated relatively stable results that trended similarly to those obtained from the original data's multiple linear regression analysis (β=28.423, 95% CI: 24.762, 32.083), thereby strengthening our conclusion that the ability to promote testosterone secretion in the no-CKD population is 28.423 ng/mL greater than in individuals with CKD.

Table 2. Presents the results of linear regression analysis examining the relationship between CKD and testosterone.

		β (95% CI)
		Non-adjusted	Adjust I	Adjust II
CKD
	Yes	Reference	Reference	Reference
	No	19.819 (14.897, 24.741)	59.020 (55.522, 62.518)	28.423 (24.762, 32.083)
	p-value	<0.00001	<0.00001	<0.00001

CKD: chronic kidney disease.

p-value was calculated by the weighted chi-square test.

Table 3. Displays the results of five multiple imputations performed to address the missing data in regression analysis of CKD and testosterone.

			Exposure
			Non-adjusted	Adjust I	Adjust II
Model 1
	CKD
		Yes	Reference	Reference	Reference
		No	19.819 (7.762, 31.876)	59.020 (50.451, 67.589)	27.839 (19.469, 36.209)
		p-value	0.00128	<0.00001	<0.00001
Mode 2
	CKD
		Yes	Reference	Reference	Reference
		No	19.819 (7.762, 31.876)	59.020 (50.451, 67.589)	27.602 (19.233, 35.972)
		p-value	0.00128	<0.00001	<0.00001
Mode 3
	CKD
		Yes	Reference	Reference	Reference
		No	19.819 (7.762, 31.876)	59.020 (50.451, 67.589)	27.889 (19.519, 36.260)
		p-value	0.00128	<0.00001	<0.00001
Mode 4
	CKD
		Yes	Reference	Reference	Reference
		No	19.819 (7.762, 31.876)	59.020 (50.451, 67.589)	27.466 (19.101, 35.832)
		p-value	0.00128	<0.00001	<0.00001
Mode 5
	CKD
		Yes	Reference	Reference	Reference
		No	19.819 (7.762, 31.876)	59.020 (50.451, 67.589)	27.707 (19.334, 36.080)
		p-value	0.00128	<0.00001	<0.00001
Summary			Pool estimates from multiple imputed data=27.700 (23.427, 31.974)

Values are presented as β (95% confidence interval).

Non-adjusted model adjust for: none; Adjust I model to adjust for: sex, age, BMI; Adjust II model adjust for: sex, age, BMI, total cholesterol, HDL, LDL, DM, waist, hemoglobin, race/ethnicity, education status, marital status, smoking status, HBP status, anemia, alcohol status.

CKD: chronic kidney disease, BMI: body mass index, HDL: high-density lipoprotein cholesterol, LDL: low-density lipoprotein cholesterol, DM: diabetes mellitus, HBP: high blood pressure.

After receiving positive feedback on the results of MI, we further conducted hierarchical regression analyses on all covariates in the CKD population to explore the impact of exposure variables on testosterone secretion levels in various covariates. As shown in Fig. 2, it can be observed that the exposure variable CKD has a significant impact on testosterone at different stages or levels of many covariates. Specifically, in the male (β=16.941, 95% CI: 3.605, 30.277), severe LDL (β=90.500, 95% CI: 47.686, 133.314), and obese (BMI ≥24.9 kg/m²) (β=43.110, 95% CI: 15.855, 70.365) populations, there was a considerable effect on testosterone secretion. Smoking, drinking, and hypertension also affect testosterone secretion levels, while CKD patients with DM do not have significantly different testosterone levels than those without CKD (β=12.869, 95% CI: -5.524, 31.263). Furthermore, to demonstrate that CKD significantly reduces testosterone secretion in different subgroups, we conducted subgroup analyses on age, sex, BMI, DM, anemia, and HBP. These analyses were divided into the no-adjusted, adjusted I, and adjusted II groups for statistical analysis, with specific results presented in Table 4. It can be observed that the negative impact of CKD on testosterone is consistently present.

Fig. 2. (A, B) Forest plot: to explore of the effect of CKD on testosterone in different subgroups. The two plots focus on CKD with different specific factors. Plot (A) examines the impact of race, education, smoking, HBP, marital status, waist circumference, and HDL cholesterol. On the other hand, Plot (B) focuses on sex, age, BMI, total cholesterol, LDL cholesterol, DM, and alcohol status. ^*The population and β values (95% CI) of the summary are the sum of plots (A) and (B). N: quantity, CKD: chronic kidney disease, CI: confidence interval, HDL: high-density lipoprotein, HBP: high blood pressure, BMI: body mass index, LDL: low-density lipoprotein, DM: diabetes mellitus.

Table 4. Subgroup analysis of testosterone levels according to CKD status.

Characteristic		Non-adjusted	Adjust I	Adjust II
Age (y)
	6–22	121.618 (17.447, 225.789)	131.519 (44.954, 218.084)	-
	23–49	44.957 (-10.484, 100.398)	114.007 (80.196, 147.817)	115.160 (21.702, 208.617)
	50–85	23.649 (10.055, 37.243)	33.496 (25.116, 41.877)	45.002 (19.434, 70.571)
Sex
	Male	46.79 (31.10, 62.47)	65.63 (50.44, 80.82)	89.38 (41.80, 136.96)
	Female	2.76 (0.94, 4.57)	0.77 (-1.04, 2.58) 0.4054	0.27 (-4.78, 5.32) 0.9168
BMI (kg/cm2)
	≤18.5	-21.37 (-95.93, 53.19)	60.22 (-9.97, 130.42)	63.99 (-195.29, 67.30)
	>18.5, <24.9	52.51 (21.92, 83.11)	46.75 (26.44, 67.07)	73.90 (18.68, 129.12)
	≥24.9	19.74 (6.49, 32.99)	30.01 (21.55, 38.46)	37.50 (9.71, 65.29)
HBP status
	Yes	37.410 (22.457, 52.363)	29.889 (19.937, 39.841)	39.152 (13.439, 64.866)
	No	-5.547 (-28.250, 17.156)	63.010 (47.621, 78.399)	73.265 (9.507, 137.024)
DM status
	Yes	12.87 (-5.52, 31.26)	22.39 (11.08, 33.70)	41.40 (6.68, 76.12)
	No	23.17 (6.21, 40.13)	27.16 (16.18, 38.14)	36.30 (4.35, 68.25)
Anemia
	Yes	-43.091 (-46.874, -39.308)	23.803 (20.350, 27.255)	6.718 (3.061, 10.375)
	No	4.821 (-1.458, 11.101)	57.604 (53.152, 62.057)	33.503 (28.887, 38.118)

Values are presented as β (95% confidence interval).

Non-adjusted model adjust for: None; Adjust I model to adjust for: sex, age, BMI; Adjust II model adjust for: sex, age, BMI, total cholesterol, HDL, LDL, DM, waist, hemoglobin, race/ethnicity, education status, marital status, smoking status, HBP status, anemia, alcohol status.

CKD: chronic kidney disease, BMI: body mass index, HDL: high-density lipoprotein cholesterol, LDL: low-density lipoprotein cholesterol, DM: diabetes mellitus, HBP: high blood pressure.

After verifying that the exposure factor CKD is a risk factor for testosterone secretion in the human body, we further investigated whether testosterone shortens the life expectancy of CKD male patients. CKD male subjects were divided into low and normal testosterone groups based on testosterone values, with 544 individuals in the low testosterone group and 319 in the normal testosterone group. The average follow-up time was 57.8±27.8 months for the low group and 61.8±26.6 months for the normal group. As shown in the K-M curve in Fig. 3, the survival rate of CKD males with normal testosterone levels was significantly better than that of those with low testosterone levels within 125 months of follow-up (log-rank tests p<0.0001).

Fig. 3. Kaplan–Meier plot of the effect of testosterone levels on prognostic survival in males with chronic kidney disease.

DISCUSSION

As the global prevalence of CKD continues to rise, an increasing number of patients are undergoing dialysis or kidney transplant surgery. These patients are at a high risk of developing endocrinological disorders, and other comorbidities [20]. One of the endocrine disorders associated with CKD is hypogonadism, which has a prevalence of 27% to 66% in this patient population [21]. Hypogonadism can disrupt the balance of the body's internal environment, leading to obesity, metabolic syndrome, anemia, and decreased erectile function [22,23]. Compounding matters, many of these comorbidities frequently occur concomitantly in CKD patients with hypogonadism [24], potentially contributing to decreased life expectancy. In this study, we aimed to explore the relationship between CKD and testosterone using the rich data on population health and nutrition provided by NHANES. Our results consistently demonstrated that CKD was a significant contributor to reduced testosterone secretion in the body. Furthermore, our K-M plot showed that low testosterone levels significantly impacted the average lifespan of male CKD patients.

There is a significant difference in the prevalence of low testosterone between individuals without CKD and CKD patients. Our results (Fig. 2B) show that healthy males have significantly higher blood testosterone levels than male CKD patients (β=46.786, 95% CI: 31.101, 62.471). This suggests that low testosterone levels may be associated with the onset or progression of CKD. Previous studies have also reported differences in testosterone levels among CKD patients compared to the general population, as well as across different stages of CKD. For example, Hylander and Lehtihet [25] found that total testosterone levels were lower in CKD patients than in normal individuals, and among 101 CKD patients in different stages, higher stages of CKD (lower kidney function) showed a significant negative linear trend for both total testosterone levels (p<0.01) and free testosterone levels (p<0.01). The possible reasons for the decrease in serum testosterone levels among CKD populations may include reduced production, increased metabolic or dialytic clearance, or alterations in testosterone-binding activity or capacity [26].

Following the discovery of decreased testosterone levels in CKD patients, literature has confirmed that TRT has favorable anabolic effects on malnourished CKD patients of both sex who were not hypogonadal [26]. However, can TRT improve the prognosis of CKD? In our large-scale study, we found a strong correlation between testosterone levels and prognosis in CKD patients. Our K-M plot showed that as the follow-up time increased, the all-cause mortality rate significantly increased in the low testosterone group, indicating that low testosterone is a risk factor for the survival prognosis of CKD. In recent years, some single-center research institutions have attempted to use TRT to observe whether it can improve the prognosis of CKD and reduce the incidence of complications. For example, according to Skiba et al [3], TRT administration in CKD patients undergoing dialysis or predialysis significantly improved clinical symptoms of hypogonadism and erectile dysfunction. Yeo et al [27] found that grip strength significantly improved and quality of life improved markedly after TRT in a retrospective study on CKD patients. Kurita et al [28] discovered a significant association between low testosterone levels and poorer kidney function in a cohort study involving 848 participants. Although the sample sizes of previous studies were limited, the results have led us to pay increasing attention to and be curious about whether TRT treatment could benefit CKD patients with imbalanced testosterone levels.

It is well known that males have much higher levels of testosterone secretion than females. In a 15-year follow-up study, Amiri et al [4] found that males with low testosterone levels had a significantly higher risk of developing CKD than those with normal testosterone levels (hazard ratio=1.38; 95% CI: 1.05, 1.80). Furthermore, Carrero et al [29] found that the clinical state of testosterone deficiency was independently associated with cardiovascular complications (OR=2.51; 95% CI: 1.32, 4.76) and death (OR=2.00; 95% CI: 1.01–3.97). However, Khurana et al [30] compared the efficacy of the low or TRT group with the normal testosterone group in a total of 2,419 patients over a median follow-up duration of 2.3 years and found no significant difference in survival rate between the two groups. Previous reports on the association between serum testosterone and CKD events are scarce, and this study adds to this area of research. Although the debate on whether TRT can improve the prognosis of CKD patients has been ongoing, as a potentially treatable pathological phenomenon, active exploration of treatment based on multidisciplinary clinical literature is always encouraging. Since the baseline testosterone levels in females are influenced by various factors such as menstrual cycle, pregnancy, age, etc., females were excluded from the survival analysis of CKD patients with testosterone in this study.

One limitation of this study is that most of the data have a cross-sectional design, which limits our ability to make temporal statements about the relationship between testosterone levels and renal function damage. Additionally, the current study still lacks randomized clinical trials to confirm the conclusions we have drawn. In future research, this task will be included on the agenda. Lastly, part of the data came from questionnaires, including the counts of exposure variables. Although we screened out some abnormal data, the accuracy of the data still needs to be improved.

CONCLUSIONS

In conclusion, our study found a two-way interaction between CKD and testosterone levels, where CKD progression may lead to decreased testosterone levels and low testosterone may be a risk factor for CKD. Timely correction of testosterone levels may have preventive and therapeutic effects on the occurrence and progression of CKD. However, to confirm this conclusion, further support from clinical and basic experimental evidence will be necessary in the future.

Data Sharing Statement

The datasets presented in this study can be found in online repository. The name of the repository and accession number can be found below: https://www.cdc.gov/nchs/nhanes/.