Association between Serum Klotho and All-Cause Mortality in Chronic Kidney Disease: Evidence from a Prospective Cohort Study

Chronic kidney disease (CKD) is a progressive condition characterized by the gradual decline of renal function, imparting a significantly increased risk of all-cause mortality worldwide [1]. CKD is also associated with various comorbidities, such as cardiovascular disease, infection, and osteodystrophy [2]. Given the annual increase in global CKD incidence, early intervention and targeted treatment are crucial for improving patient survival rates [3]. Despite the emergence of several therapeutic strategies that have effectively slowed disease progression and improved comorbidities in recent years, the mortality rate in CKD patients remains relatively high [4]. The global all-age mortality rate attributable to CKD increased by 41.5% between 1990 and 2017 according to the Global Burden of Disease, Injuries, and Risk Factors Study [5]. Therefore, it is essential to identify biomarkers associated with CKD mortality to target high-risk populations for early intervention.

α-Klotho (referred here as Klotho) is a protein product of the Klotho gene, initially linked to aging [6]. Two forms of Klotho protein exist in the human body: membranous and soluble types, of which soluble Klotho exerts its biological functions through circulation [7]. Klotho has been found to play various protective roles in kidney disease, such as enhancing kidney resistance to injury, promoting kidney recovery, suppressing renal fibrosis, improving mineral metabolism, and reducing blood pressure [8]. Unfortunately, circulating soluble Klotho levels decline as CKD progresses, as the kidneys are the primary source of soluble Klotho, and serum Klotho indeed serves as a reliable indicator of renal Klotho status [9]. Previous studies have suggested that low circulating soluble Klotho concentrations are associated with adverse kidney disease outcomes and could serve as a prognostic marker in CKD population [10]. A recent meta-analysis, encompassing six cohort studies and 655 patients, found that lower circulating soluble Klotho level was associated with an increased risk of all-cause mortality among patients with CKD [11]. However, the results from existing researches were somewhat inconsistent [12]. Additionally, potential biases may affect the findings of existing evidence, due to small sample sizes and a lack of adjustment for potential confounders. Furthermore, it remains unclear whether there is a linear or nonlinear association between serum Klotho levels and all-cause mortality in CKD patients. Given these uncertainties, this study was designed to elucidate the association between circulating soluble Klotho and all-cause mortality among CKD patients through a prospective cohort study based on a nationally representative sample of adults in the USA.

The data were sourced from the National Health and Nutrition Examination Survey (NHANES), a comprehensive cross-sectional survey administered by the National Center for Health Statistics. NHANES was designed to gather information on the representative nutritional and health status of American civilians through structured interviews, questionnaires, physical examinations, and laboratory tests. NHANES was approved by the National Center for Health Statistics Ethics Review Board, and all participants have provided the informed consent forms. Serum Klotho concentrations were tested in participants aged 40–79 years from NHANES cycles spanning from 2007 to 2016 and were included for analysis in this study.

CKD was defined as an estimated glomerular filtration rate (eGFR) of less than 60 mL/min/1.73 m² calculated using the CKD-EPI creatinine equation or a urinary albumin-to-creatinine ratio (ACR) equal to or exceeding 30 mg/g [13]. To accurately identify individuals with CKD and explore the association between Klotho and all-cause mortality in CKD, we excluded the following participants: (1) those with incomplete data on age, sex, race, serum creatinine, ACR, and serum Klotho (n = 37,000); (2) individuals who were currently pregnant (n = 9); (3) patients with cancer or malignancy (n = 1,573); and (4) those without CKD (n = 9,550). Consequently, 2,456 patients with CKD were ultimately included for analysis among the 50,588 participants from NHANES 2007 to 2016. The selection process is illustrated in Figure 1.

Pristine serum samples were stored at −80°C until assay analysis. Soluble Klotho was detected using an ELISA kit (IBL International, Tokyo, Japan), with the implementation of several quality control measures. Prior to the study commencement, an extensive validation of the ELISA method for measuring Klotho concentration in human samples was conducted. All sample analyses were carried out in duplicate, and the final Klotho concentration was calculated as the average of the two values. Any samples with duplicate results differing by more than 10% were flagged for subsequent reanalysis. The quality control measures for the Klotho assay encompassed the following aspects: (1) consistency of the assay standard curves and the relative signals of the calibrator concentrations with the manufacturer’s specified criteria; (2) the lower limit of detection was 4.33 pg/mL; (3) evaluation of assay linearity through the assessment of two samples with notably high Klotho concentrations, analyzed at different dilutions; the results indicated exceptional linearity within the assay’s measurement range, as evidenced by R² values of 0.998 and 0.997; (4) evaluation of intra-assay precision, resulting in coefficient of variation values of 3.2% and 3.9% for recombinant Klotho samples and 2.3% and 3.3% for human samples; (5) assessment of inter-assay precision, with samples analyzed in duplicate over four separate days, yielding coefficient of variation values of 2.8% and 3.5% for recombinant samples and 3.8% and 3.4% for human samples; (6) determination of reference ranges based on 114 samples obtained from apparently healthy donors, revealing Klotho levels ranging from 285.8 to 1,638.6 pg/mL, with an average concentration of 698.0 pg/mL.

The mortality status of participants from NHANES 2007 to 2016 was determined by cross-referencing their records with the National Death Index. The public-use mortality files contained mortality follow-up data from the date of survey participation to December 31, 2019. Cardiovascular mortality data, used for sensitivity analysis, was categorized according to the International Classification of Diseases, 10th Revision (ICD-10) codes (I00–I09, I11, I13, I20–I51).

Several confounding factors that have previously shown potential associations with serum soluble Klotho and all-cause mortality in CKD were identified as covariates, including demographic variables [13], body mass index (BMI) [14], smoking habits [15], hypertension [16], diabetes [17], cardiovascular disease (CVD) [18], dyslipidemia [19], albuminuria [20], and eGFR [9]. Given Klotho’s role in regulating calcium and phosphate homeostasis [21], we included serum phosphorus, total calcium, and 25‐hydroxyvitamin D as additional covariates. Smoking behavior was classified into three categories: never smoker, current smoker, and former smoker. Specifically, individuals who had smoked fewer than 100 cigarettes in their lifetime were categorized as never smokers, while those who had smoked more than 100 cigarettes and were currently smoking were designated as current smokers. Participants who had smoked more than 100 cigarettes but had quit smoking were classified as former smokers. Hypertension was defined as a self-reported medical diagnosis of hypertension or systolic blood pressure equal to or exceeding 140 mm Hg, or diastolic blood pressure equal to or exceeding 90 mm Hg during the physical examination. Diabetes was defined as a self-reported medical diagnosis of diabetes or a glycated hemoglobin A1c level equal to or exceeding 7.0%. CVD was defined as heart failure, coronary heart disease, angina pectoris, or a history of heart attack diagnosed by a medical professional. Serum cholesterol, triglycerides, calcium, and phosphate levels were obtained from the standard biochemistry profile. The quantification of 25‐hydroxyvitamin D was performed using ultra-high-performance liquid chromatography-tandem mass spectrometry.

Given NHANES’ complex, multistage, probability sampling design, we meticulously integrated the survey design details, encompassing sample weights, strata, and primary sampling units, into all the statistical models used in this study. This method aligns with NHANES survey methods and analytical guidelines provided by the US Centers for Disease Control and Prevention [22]. Descriptive statistics were presented as numbers (percentages) for categorical variables and as medians (interquartile ranges [IQRs]) for continuous variables. Baseline characteristics across different Klotho quartiles were compared using appropriate statistical tests, including one-way ANOVA tests, Kruskal-Wallis tests, and χ² tests.

Multivariate Cox proportional hazards models incorporating complex sampling design parameters were employed to estimate hazard ratios (HRs) and their corresponding 95% confidence intervals (CIs) for all-cause mortality among CKD patients with different serum Klotho concentrations. Three distinct Cox regression models were performed. Model 1 did not adjust for any covariate. Model 2 was adjusted for age (continuous), sex (male or female), and race/ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, or others). Model 3 was further adjusted for additional covariates, including BMI (<25, 25 ∼ 29.9, or 30 kg/m²), smoking status (former, current, or never smoker), hypertension (yes or no), diabetes (yes or no), CVD (yes or no), ACR (≥30 or <30 mg/g), CKD stages, 25‐hydroxyvitamin D (continuous), total calcium (continuous), phosphorus (continuous), cholesterol (continuous), triglycerides (continuous), and NHANES cycles (2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016). Restricted cubic spline (RCS) models were conducted to investigate the nonlinear correlation between serum Klotho levels and all-cause mortality in CKD patients, with four knots placed at the 5th, 35th, 65th, and 95th percentiles of the Klotho distribution.

Subgroup analysis was performed to explore the association of interest among varying subgroups. The stratified factors considered were gender (male or female), age (≥60 or <60 years), race/ethnicity (white or others), eGFR (≥60 or <60 mL/min/1.73 m²), ACR (≥30 or <30 mg/g), diabetes (yes or no), hypertension (yes or no), BMI (≥25 or <25 kg/m²), and 25‐hydroxyvitamin D (≥50 or <50 nmol/L).

Sensitivity analysis was performed by excluding participants who died within the first and second years of follow-up to minimize potential reverse causation bias. Given the pivotal role of cardiovascular death in CKD mortality, the correlation between Klotho and cardiovascular death in CKD patients was analyzed as part of additional sensitivity analysis.

All analyses were conducted using R Studio (version 2022.07.2 Build 576, open-source edition). A two-sided p value less than 0.05 was considered statistically significant.

A total of 2,456 participants with CKD were included in this study. Baseline characteristics stratified by quartiles of serum Klotho levels are detailed in Table 1. The median age of the participants was 62 years, with 51.55% being women. The median serum Klotho level was 760 pg/mL (IQR: 625, 958). Participants with higher Klotho concentrations tended to be younger (Q4 vs. Q1, median age 58 [IQR: 49, 68] vs. 64 [IQR: 56, 72] years). They were also less likely to be non-Hispanic white (Q4 vs. Q1, 29.25% vs. 38.38%). Individuals with increased serum Klotho levels had higher ACR (Q4 vs. Q1, 46.12 [IQR: 26.71, 103.76] vs. 38.85 [IQR: 8.88, 87.60] mg/g) and elevated eGFR (Q4 vs. Q1, 79.86 [IQR: 56.95, 99.73] vs. 57.32 [IQR: 48.89, 81.47] mL/min/1.73 m²). Additionally, participants with higher Klotho concentrations had lower serum 25‐hydroxyvitamin D levels (Q4 vs. Q1, 56.70 [IQR: 38.42, 76.59] vs. 65.60 [IQR: 43.41, 86.69] nmol/L). They also exhibited a higher prevalence of diabetes (Q4 vs. Q1, 44.34% vs. 38.53%), but conversely, a lower likelihood of having hypertension (Q4 vs. Q1, 73.43% vs. 81.12%). No significant differences were observed among the quartiles of serum Klotho levels in terms of gender, BMI, smoking status, serum cholesterol, triglycerides, phosphorus, and total calcium.

During a median follow-up period of 82 months, 550 deaths (22.40%) were recorded. Klotho was found to be inversely correlated with all-cause mortality in CKD (shown in Table 2). Cox proportional hazard regression models indicated that the risk of all-cause death in CKD decreased by 4–7% for every 100 pg/mL increase in serum Klotho concentration (model 1, HR = 0.93, 95% CI, 0.89, 0.96; model 2, HR = 0.96, 95% CI, 0.92, 0.99; model 3, HR = 0.96, 95% CI, 0.92, 0.99). CKD patients with higher quartiles of serum Klotho exhibited a significantly reduced risk of death compared to those in the lower quartile of Klotho in all three regression models (Q4 vs. Q1: model 1, HR = 0.58, 95% CI, 0.43, 0.78; model 2, HR = 0.71, 95% CI, 0.54, 0.94; model 3, HR = 0.73, 95% CI, 0.56, 0.96).

Using RCS models, we observed an “L”-shaped association between serum Klotho levels and all-cause mortality in participants with CKD. As shown in Figure 2, the risk of all-cause mortality initially demonstrated a significant decline with increasing serum Klotho levels, followed by a plateau phase. This pattern was most pronounced in model 3, with mortality leveling off after reaching a serum Klotho concentration of 760 pg/mL. When serum Klotho levels were below this inflection point, the risk of all-cause mortality decreased by 14% in model 1 (HR = 0.86, 95% CI, 0.78, 0.94) (shown in Fig. 2a), by 13% in model 2 (HR = 0.87, 95% CI, 0.79, 0.95) (shown in Fig. 2b), and by 14% in model 3 for every 100 pg/mL increase in Klotho (HR = 0.86, 95% CI, 0.78, 0.95) (shown in Fig. 2c).

We conducted subgroup analyses based on sex, age, race, 25‐hydroxyvitamin D, diabetes, hypertension, BMI, 25‐hydroxyvitamin D, ACR, and eGFR using model 3. While subgroup analyses suggested that the correlation between Klotho and CKD mortality risk (Q4 vs. Q1) might be more relevant in specific stratified populations, none of the interaction tests showed a statistically significant difference (all p values >0.05) (shown in Fig. 3). To further investigate the association between serum Klotho and mortality risk within different stages of CKD, we conducted regression analyses based on CKD stages, as detailed in Table 3. We observed that an increased serum Klotho concentration was only associated with a reduced risk of all-cause mortality in patients with CKD stage 3a (HR per 100 pg/mL increase of Klotho = 0.90, 95% CI, 0.84–0.97).

To eliminate the possibility of reverse causality, we excluded patients who died within 1 and 2 years after the initial survey. The results remained robust (shown in Fig. 4a, b). We observed a negative correlation between serum Klotho concentration and cardiovascular mortality in CKD patients; however, statistical significance was not achieved (model 3: HR per 100 pg/mL increase in Klotho = 0.96, 95% CI, 0.88, 1.04, P nonlinear = 0.310, P overall = 0.044) (shown in Fig. 4c).

This cohort study documented that CKD patients with high serum Klotho concentrations experience lower mortality rates compared to those with lower Klotho levels. The RCS analysis revealed a nonlinear “L”-shaped association between serum Klotho level and all-cause mortality among CKD patients. Specifically, we observed a significant negative correlation until a plateau was reached at a Klotho concentration of 760 pg/mL. Beyond this point, the risk of mortality did not significantly change with increasing serum Klotho levels.

Our findings were consistent with those of Yang et al. [23], who discovered an independent association between lower serum Klotho levels and increased all-cause mortality and cardiovascular events in nondiabetic predialysis CKD patients. These results were further corroborated by a recent meta-analysis, which demonstrated that CKD patients with lower circulatory Klotho concentrations faced a significantly elevated risk of all-cause mortality (risk ratio: 1.88; 95% CI: 1.29–2.74) [11]. However, conflicting findings have also emerged. Seiler et al. [24] reported that plasma Klotho did not predict adverse outcomes, defined as a combination of initiating renal replacement therapy or mortality, in 312 patients with stage 2–4 CKD, even after adjusting for age, eGFR, and ACR. These disparities may be attributed to a shorter follow-up duration (average, 2.2 years), variations in covariates, and differences in the definition of outcomes.

CKD progression is characterized by a progressive loss of functional nephrons, compensating with a rise of fibroblast growth factor 23 (FGF-23), followed by a decrease in active vitamin D and an increase in parathyroid hormone (PTH). These changes can suppress the expression of Klotho, leading to the loss of its protective properties [25]. Klotho has been extensively documented to exhibit multiple renoprotective effects, including anti-oxidation, anti-senescence, anti-apoptosis, inhibition of fibrogenesis, and preservation of stem cells [26, 27]. It also plays a vital role in managing mineral and bone disorders in CKD by regulating phosphate homeostasis, a factor associated with higher mortality in both dialysis and non-dialysis CKD patients [28]. Furthermore, emerging data indicated that dysregulation of Klotho and FGF-23 has various effects on the cardiovascular system, significantly contributing to increased cardiovascular morbidity and mortality rates among CKD patients [29]. In progressive CKD, the decreasing kidney expression of Klotho and the rising circulatory FGF-23 promoted adverse outcomes such as left ventricular hypertrophy, heart failure, atrial fibrillation, arterial calcification, atherosclerotic disease, and death [30]. Klotho was therefore considered as a promising therapeutic target to delay CKD progression, mitigate its complications, and show positive results in both in vivo and in vitro experiments [31]. Our study further confirmed the negative correlation between serum Klotho levels and long-term mortality in CKD, suggesting the potential of Klotho as a therapeutic target to reduce mortality in patients with CKD.

Our study uncovered a potential protective effect of Klotho in patients with high levels of 25-hydroxyvitamin D. This finding aligned with the results of Chen et al. [32], who identified interactive effects between Klotho and 25-hydroxyvitamin D for all-cause and cardiovascular mortality in American adults. Notably, Klotho expression is positively regulated by activated vitamin D, and low plasma vitamin D levels in CKD patients can reduce Klotho expression, rendering these organs resistant to FGF-23 [33]. However, our study did not reveal a significant interaction effect between Klotho and 25-hydroxyvitamin D. Therefore, further validation may be warranted with a larger cohort. It is worth noting that the prevalence of diabetes mellitus among the participants included in this study was unusually high. This could be attributed to our inclusion criteria, which involved the presence of albuminuria, condition that diabetic patients are more likely to experience. Despite this, our subgroup analysis did not reveal a statistically significant correlation between Klotho and the risk of death among CKD patients with diabetes. This lack of significance might be due to the relatively higher eGFR observed in these patients, particularly among those with diabetic kidney disease. Furthermore, among participants with eGFR levels below 60 mL/min/1.73 m², the protective effect of Klotho in reducing the risk of death was specifically observed in CKD 3a patients, rather than in those with CKD 3b, CKD 4, or CKD 5. However, it is important to note that the number of participants with an eGFR of less than 45 mL/min/1.73 m² was quite small. Confirming this finding may require a larger sample size for further data analysis.

The advantage of this study is that we identified a relatively large sample size of CKD patients who were rigorously detected for serum Klotho concentration, from a nationally representative population. Furthermore, we conducted a comprehensive examination of the association between serum Klotho and all-cause mortality in CKD patients. This was achieved by utilizing a combination of RCS analysis and multiple regression models, with meticulous incorporation of appropriate weights and covariates within the statistical models. These rigorous methodological approaches significantly enhance the credibility of the results presented. Nevertheless, it is important to acknowledge certain limitations of our study. First, Klotho was measured only once, preventing us from observing the association between dynamic changes in Klotho levels and the mortality of CKD. Additionally, the lack of data on FGF-23 and PTH in the NHANES 2007–2016 is a notable limitation. FGF-23 is closely associated with Klotho and has been found to be correlated with the risk of death in patients with CKD [34], as well as PTH [28]. The inability to adjust for FGF-23 and PTH as covariates may introduce potential bias into our findings. Future studies should seek to address these limitations for a more comprehensive understanding of the relationship between Klotho and CKD outcomes.

Utilizing nationally representative NHANES data, we identified an “L”-shaped association between serum Klotho concentration and all-cause mortality in CKD patients. Further research is warranted to validate our findings.

NHANES protocol was approved by the National Center for Health Statistics Ethics Review Board. Detail is available at https://www.cdc.gov/nchs/nhanes/irba98.htm. Informed consents were obtained from all the participants included in NHANES 2007–2016. Additional ethical approval and consents were not required, because this study relied on publicly available de-identified data.

The authors have no conflicts of interest to declare.

This work was supported by the National Natural Science Foundation of China (Grant No. 82274391), National Key Research and Development Program of China (Grant No. 2019YFC1709401), and the Project from Science and Technology Commission of Shanghai Municipality (Grant No. 20Y21902100).

Li Shang: conceptualization. Shisheng Han and Xiaolu Zhang: data analysis. Xiaojun Wang: writing original draft. Yi Wang and Yanqiu Xu: project administration and editing manuscript. All authors contributed to and approved the final manuscript.

Shisheng Han, Xiaolu Zhang, and Xiaojun Wang contributed equally to this work.

Publicly available datasets were used in this study. These can be found in NHANES website at https://www.cdc.gov/nchs/nhanes/. Further inquiries can be directed to the corresponding author.