Spherical adsorptive carbon use and risks of dialysis initiation and mortality in advanced chronic kidney disease: a real-world cohort study
Article information
Abstract
Background
The benefits of spherical adsorptive carbon (SAC) for individuals with chronic kidney disease (CKD) remain uncertain. This study investigated the association between SAC use, dialysis initiation, and all-cause mortality in patients with advanced CKD using large-scale, real-world data.
Methods
This multicenter, retrospective cohort study analyzed electronic health record data standardized through the Observational Medical Outcomes Partnership Common Data Model across three tertiary hospitals. Patients with CKD and serum creatinine levels ≥2.0 mg/dL who received SAC at ≥4 g/day for ≥90 days were included, along with non-users. Risks of dialysis initiation and all-cause mortality were estimated using inverse probability of treatment weighting (IPTW)-weighted Cox proportional hazards models adjusted for relevant clinical variables.
Results
Among 15,324 patients, 3,989 received SAC. In IPTW-adjusted analyses, non-use of SAC was associated with an increased risk of dialysis initiation (hazard ratio, 1.125; 95% confidence interval, 1.018–1.243; p = 0.02). This association was observed in selected subgroups, particularly among patients aged <65 years and those with preserved kidney function. In a sensitivity analysis using a landmark approach, the association with dialysis initiation was attenuated. SAC non-use was also associated with an increased risk of all-cause mortality (hazard ratio, 1.659; 95% confidence interval, 1.487–1.851; p < 0.001), and this finding remained consistent in sensitivity analyses.
Conclusion
The use of SAC was associated with a lower risk of all-cause mortality and showed a potential association with reduced risk of dialysis initiation in patients with advanced CKD, suggesting a possible role as an adjunctive therapy in this population.
Introduction
Chronic kidney disease (CKD) is a major cause of morbidity and mortality and is a well-established risk factor for cardiovascular disease, underscoring its substantial impact on global health [1,2]. Furthermore, CKD progression contributes to a growing economic burden, as patients who advance to end-stage kidney disease (ESKD) ultimately require renal replacement therapies such as dialysis or kidney transplantation [3,4]. Therefore, early detection and appropriate management of CKD are essential to improve clinical outcomes and reduce healthcare costs.
In patients with CKD, a wide range of uremic toxins accumulate in the circulation because of impaired renal clearance [5,6]. Among these toxins, protein-bound, gut microbiota-derived solutes such as indoxyl sulfate (IS) and p-cresyl sulfate (PCS) have received particular attention because they exhibit nephrotoxic and vasculotoxic properties and are associated with CKD progression and cardiovascular disease [7–9]. Both experimental and clinical studies have demonstrated that the accumulation of these solutes contributes not only to the decline in renal function but also to cardiovascular complications, thereby linking the gut–kidney axis to broader systemic disease burden [8,10,11].
Oral spherical adsorptive carbon (SAC), commercially available as AST-120 (Kremezin®, Renamezin®), is an oral carbonaceous adsorbent that removes intestinal precursors of uremic toxins, including indole and p-cresol, thereby lowering circulating IS and PCS concentrations. Several experimental and clinical studies have demonstrated that SAC may attenuate renal fibrosis, slow kidney function decline, and potentially reduce cardiovascular complications in CKD [12]. However, evidence regarding its renoprotective effects remains inconclusive. Although large randomized controlled trials, including the EPPIC studies, did not show significant benefits for composite renal outcomes [13], several smaller clinical studies have reported a slower decline in kidney function and reduced progression to ESKD in selected patient populations [14]. Despite these findings, real-world evidence regarding the long-term effectiveness and clinical impact of SAC in diverse CKD populations remains limited.
Given the limited and inconsistent evidence, we aimed to evaluate whether SAC use is associated with a reduced risk of progression to ESKD and all-cause mortality in patients with CKD, using large-scale real-world data standardized through a common data model (CDM).
Methods
Common data model database
This multicenter, retrospective, observational cohort study was conducted using data from the Observational Medical Outcomes Partnership Common Data Model (OMOP-CDM) databases of three tertiary hospitals: Korea University Anam, Ansan, and Guro Hospitals. The OMOP-CDM schema, provided through the Observational Health Data Sciences and Informatics (OHDSI) Collaboration, standardizes electronic health records into a unified structure within the CDM database (https://github.com/OHDSI/CommonDataModel/) [15]. Using the electronic health record systems of each participating hospital, data were extracted, transformed, and loaded into the OMOP-CDM, with all entries mapped to unique concept identifiers (IDs) according to internationally accepted coding standards, including the International Classification of Diseases, 10th Revision (ICD-10), for diagnostic information [16]. For each hospital, electronic health record data from February 1, 2002, to December 31, 2024, were converted to CDM version 5.3 of the OHDSI framework and stored separately on Microsoft structured query language (SQL) servers. Data of interest were then queried and extracted using SQL.
Study population
Patients with advanced CKD, defined as a serum creatinine level ≥2.0 mg/dL, who received SAC at a dosage of at least 4.0 g/day for a cumulative duration of 90 days or more between 2003 and 2022 were classified as SAC users. Patients with CKD with a serum creatinine level ≥2.0 mg/dL who did not receive SAC were included as the control cohort.
The study protocol was reviewed by the Institutional Review Board of the Korea University Guro Hospital (2024GR0012). As all data were anonymized prior to analysis, informed consent was not required. The algorithm for patient enrollment is shown in Fig. 1.
Study flowchart.
IPTW, inverse probability of treatment weighting; SAC, spherical adsorptive carbon; sCr, serum creatinine.
The primary clinical outcome, initiation of dialysis, was identified using dialysis-related procedure codes (Supplementary Table 1, available online). Mortality data were obtained from death certificate records integrated into the electronic health systems from 2002 to 2024. Comorbidity burden was assessed using the Charlson Comorbidity Index (CCI), with ICD-10 codes for each comorbidity listed in Supplementary Table 2 (available online). Detailed information regarding all concept IDs used in the analyses is provided in Supplementary Table 3 (available online). Covariates, including laboratory findings, were collected from the measurements obtained at the time closest to the index date, defined as the first SAC prescription date for the treatment group or the first recorded serum creatinine level ≥2.0 mg/dL for the control group. Baseline covariates included demographic characteristics (age, sex), clinical factors (body mass index, systolic blood pressure, comorbidities such as diabetes mellitus [DM], hypertension [HTN], and CCI), laboratory parameters (e.g., hemoglobin, estimated glomerular filtration rate [eGFR], blood urea nitrogen, total protein, albumin, uric acid, lipid profiles), and concomitant medications (e.g., angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, calcium channel blockers, beta blockers, diuretics, statins, and sodium-glucose cotransporter 2 inhibitors). The eGFR was calculated using the 2021 CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) creatinine equation without race, as proposed by Inker et al. [17].
Statistical analysis
Continuous variables are presented as means ± standard deviations, and categorical variables are summarized as counts and percentages. Group comparisons were made using the Student t test for continuous variables and the chi-square test for categorical variables. To evaluate progression to dialysis and all-cause mortality, outcomes in the treatment group (SAC users) were compared with those in the control group (SAC non-users).
To mitigate potential confounding effects and achieve balanced baseline characteristics between the two groups, inverse probability of treatment weighting (IPTW) was applied. Propensity scores were derived from a logistic regression model incorporating clinically relevant covariates, including age, sex, DM, HTN, CCI, and eGFR. The adequacy of covariate balance before and after the weighting process was confirmed using the standardized mean difference (SMD).
To estimate hazard ratio (HR) and 95% confidence interval (CI) for the outcomes, marginal structural Cox proportional hazards models incorporating the IPTW were utilized. To account for any residual confounding, these models were additionally adjusted for the same baseline covariates used in the propensity score estimation. To address potential immortal time bias related to the 90-day exposure definition, a sensitivity analysis using a landmark approach was performed. For SAC users, the index date was redefined as the time point at which the 90-day exposure requirement was fulfilled, and patients who experienced dialysis or death before this time point were excluded. All statistical procedures were performed using SAS software, version 9.4 (SAS Institute), and R software (R Foundation for Statistical Computing).
Results
Baseline characteristics
Among the 15,324 patients included in this study, 3,989 were SAC users. The overall cohort had a mean age of 63.1 ± 14.8 years and included 63.3% males. Baseline demographic and clinical characteristics are summarized in Table 1. In the unweighted cohort, SAC users were younger, had a higher proportion of males, and had better-preserved kidney function. However, the prevalence of DM and HTN was higher among SAC users than non-users. Given these baseline differences, IPTW was applied to minimize potential confounding.
Baseline characteristics of the study population
After applying IPTW, baseline characteristics were well balanced between SAC users and non-users, with most SMDs below 0.1. Demographic variables, comorbidities, and key laboratory parameters were comparable between the two groups after weighting. Although slight residual differences were observed in some nutritional markers, such as serum albumin and total cholesterol, the absolute differences were small. Differences in concomitant medication use persisted to some extent, reflecting real-world treatment patterns.
Medication adherence during follow-up was moderate, with a mean proportion of days covered of approximately 50%, and about 25% of patients demonstrating high adherence (≥80%).
Dialysis outcome
Follow-up completeness was assessed in the unweighted cohort and was comparable between SAC users and non-users, suggesting a limited risk of bias due to differential loss to follow-up (Supplementary Table 4, available online). Kaplan-Meier analysis demonstrated significantly higher dialysis-free survival among SAC users than among SAC non-users (Log-rank p = 0.007) (Fig. 2). In the IPTW-adjusted Cox regression analysis, SAC non-use was associated with an increased risk of dialysis initiation even after adjustment for age, sex, baseline eGFR, DM, HTN, and CCI (HR, 1.125; 95% CI, 1.018–1.243; p = 0.02) (Table 2). Reduced eGFR (eGFR ≥15 to <30 mL/min/1.73 m2: HR, 1.682 [95% CI, 1.476–1.917; p < 0.001]; eGFR <15 mL/min/1.73 m2: HR, 2.307 [95% CI, 1.978–2.692; p < 0.001]; reference: eGFR ≥30 mL/min/1.73 m2), DM (HR, 1.434; 95% CI, 1.262–1.629; p < 0.001), and higher CCI scores (CCI ≥1) were also significantly associated with a greater risk of dialysis initiation. Subgroup analyses demonstrated that SAC non-use was associated with an increased risk of dialysis initiation across clinically relevant subgroups (Table 3). In subgroup analyses stratified by age, older age was defined as ≥65 years. Notably, this association remained significant among patients aged <65 years (HR, 1.197; 95% CI, 1.045–1.372; p = 0.01), and those with preserved kidney function (eGFR ≥30 mL/min/1.73 m2: HR, 1.450; 95% CI, 1.187–1.771; p < 0.001). Although some variability was observed across subgroups, the overall direction of the association was consistent. In a sensitivity analysis using a landmark approach, the association between SAC use and dialysis initiation was attenuated, with a similar direction of effect observed (HR, 1.044; 95% CI, 0.945–1.153; p = 0.40).
IPTW-weighted Kaplan-Meier curves for dialysis-free survival and overall survival according to SAC use.
(A) Dialysis-free survival (end-stage kidney disease) according to SAC use. (B) Overall survival according to SAC use.
IPTW, inverse probability of treatment weighting; SAC, spherical adsorptive carbon.
IPTW-adjusted HR for dialysis initiation according to SAC use
Subgroup analyses of the association between SAC use and study outcomes
Mortality outcome
Kaplan-Meier analysis demonstrated higher overall survival among SAC users than among non-users (Fig. 2). In the IPTW-adjusted Cox regression analysis, SAC non-use was significantly associated with an increased risk of all-cause mortality (HR, 1.659; 95% CI, 1.487–1.851; p < 0.001) (Table 4). Older age (HR, 1.043; 95% CI, 1.038–1.048; p < 0.001), male sex (female vs. male: HR, 0.831; 95% CI, 0.735–0.940; p = 0.003), reduced eGFR (eGFR ≥15 to <30 mL/min/1.73 m2: HR, 1.293 [95% CI, 1.137–1.470; p < 0.001]; eGFR <15 mL/min/1.73 m2: HR, 1.399 [95% CI, 1.174–1.667; p < 0.001]; reference: eGFR ≥30 mL/min/1.73 m2), HTN (HR, 0.829; 95% CI, 0.744–0.924; p = 0.001), and higher CCI scores (CCI ≥1) were significantly associated with mortality risk. Subgroup analyses demonstrated that SAC non-use was associated with an increased risk of mortality across clinically relevant subgroups (Table 3). Although some variability was observed, the overall direction of the association remained consistent across age, sex, comorbidities, and kidney function categories. In a sensitivity analysis using a landmark approach, the association between SAC non-use and increased all-cause mortality remained statistically significant (HR, 1.529; 95% CI, 1.370–1.707; p < 0.001).
IPTW-adjusted HR for all-cause mortality according to SAC use
Discussion
In this study, SAC use was associated with a reduced risk of dialysis initiation and all-cause mortality among patients with advanced CKD. In sensitivity analyses using a landmark approach, the association with dialysis initiation was attenuated, although the overall direction of the association remained consistent. Subgroup analyses demonstrated that SAC non-users had a significantly higher risk of dialysis initiation than SAC users, particularly among patients aged <65 years and those with preserved kidney function (eGFR ≥30 mL/min/1.73 m2). These findings suggest that active SAC use may be especially beneficial in younger individuals and in those with relatively preserved renal function. In addition, SAC non-use was consistently associated with increased mortality across clinically relevant subgroups, and this finding remained robust in sensitivity analyses.
Previous studies have reported inconsistent findings regarding the renoprotective effects of SAC in patients with CKD. Large randomized controlled trials such as the EPPIC studies did not demonstrate significant benefits of SAC on hard renal outcomes, possibly due to suboptimal selection of patients with progressive CKD, regional variation in dialysis initiation practices, baseline imbalances, and issues with medication adherence [13]. Although the CAP-KD study did not show a significant improvement in the primary composite endpoint with AST-120 treatment, patients receiving AST-120 exhibited a slower decline in eGFR compared with those in the control arm [18]. However, that study was relatively small and had a short duration of only one year. Similarly, the Korean K-STAR randomized trial involving patients with advanced CKD found no significant effect of long-term AST-120 therapy on renal disease progression, proteinuria, mortality, or quality of life compared with that of standard therapy [19]. Nonetheless, post hoc analyses showed that patients with higher adherence to AST-120 had a lower risk of reaching the composite primary endpoint [20]. Among individuals with diabetic nephropathy, greater reductions in serum IS levels were linked to improved outcomes, and AST-120 demonstrated protective effects against major adverse cardiovascular events. The present study contributes to this evidence by demonstrating an association between SAC use and renal outcomes in a large, real-world cohort. While the magnitude of the association varied across analytic approaches, the findings suggest that the potential benefits of SAC may be more evident in routine clinical practice and among specific patient subgroups.
The potential benefits of SAC are thought to be mediated through reduction of circulating uremic toxins such as IS and PCS, which contribute to tubular injury, renal fibrosis, and vascular dysfunction [21]. In particular, IS has been shown to impair antioxidative pathways in renal tubular and glomerular mesangial cells and downregulate Klotho expression in proximal tubular cells, thereby promoting cellular senescence and progressive kidney injury [22]. By adsorbing intestinal precursors of these protein-bound solutes, SAC lowers serum and urinary levels of IS and advanced glycation end products, reducing oxidative stress, inflammation, and endothelial injury and potentially mitigating both renal and cardiovascular complications [14,23]. Furthermore, SAC can be used regardless of the underlying etiology of CKD, broadening its applicability across diverse patient populations. The more substantial benefits observed among patients with preserved kidney function suggest that early SAC initiation, before significant nephron loss occurs, may offer a greater opportunity to slow CKD progression. Clinically, our findings highlight the potential value of SAC as an adjunct therapy for patients with advanced CKD. The consistent association between SAC use and reduced mortality across all subgroups indicates that toxin removal may confer systemic benefits extending beyond renal protection. Given its oral administration and generally well-tolerated nature, SAC may represent a feasible long-term therapeutic option for patients at high risk of progression to ESKD or cardiovascular events.
This study has some limitations. First, as this was an observational study using real-world data, the possibility of residual confounding factors cannot be completely excluded, despite multivariable adjustments. In addition, dialysis initiation may be influenced not only by kidney disease progression but also by clinical decision-making, including physician judgment and patient preferences. Therefore, the observed association between SAC use and reduced risk of dialysis initiation may partly reflect differences in clinical practice patterns and warrants cautious interpretation. Second, advanced CKD was defined using a serum creatinine threshold (≥2.0 mg/dL) rather than eGFR, reflecting real-world reimbursement criteria, which may have led to some degree of misclassification. Importantly, this definition was applied consistently across both SAC users and non-users, with similar proportions of patients having eGFR ≥30 mL/min/1.73 m2 (28% vs. 27%), suggesting that any potential misclassification is likely to be non-differential. Third, detailed information regarding medication adherence was limited, and data on dietary intake and serum concentrations of uremic toxins were not available. Medication adherence during follow-up was moderate, and the potential for exposure misclassification due to varying treatment persistence cannot be excluded. Fourth, kidney transplantation events were not included as ESKD outcomes, which may have led to an underestimation of the overall incidence of renal replacement therapy. Fifth, the findings may not be generalizable to populations with different ethnic or clinical characteristics. Nonetheless, the strengths of this study include the use of a large-scale CDM database and long-term follow-up, which enhance the robustness of the findings and provide meaningful insights into the potential clinical impact of SAC in real-world settings.
In summary, this large, real-world analysis demonstrated that SAC use is associated with a reduced risk of dialysis initiation and all-cause mortality in patients with advanced CKD. These results support the potential role of SAC as an adjunctive therapeutic strategy to slow disease progression and improve clinical outcomes in CKD. Further prospective studies are warranted to validate its long-term benefits and identify patient populations most likely to benefit from therapy.
Supplementary Materials
Supplementary data are available at Kidney Research and Clinical Practice online (https://doi.org/10.23876/j.krcp.25.427).
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
G.J.K was supported by a grant of the Patient-Centered Clinical Research Coordinating Center (PACEN) funded by the Ministry of Health & Welfare, Republic of Korea (RS-2024-00444497).
Acknowledgments
The analysis of this work was supported by the Health-Data Platform at Korea University Guro Hospital.
Data sharing statement
The data presented in this study are available from the corresponding author upon reasonable request.
Authors’ contributions
Conceptualization, Supervision: GJK
Data curation: CMP, DRC, YSK, JJC, SKJ, SWO, MGK, GJK
Formal analysis, Software: CMP
Funding acquisition: GJK
Investigation: HJK, DRC, YSK, JJC, SKJ, SWO, MGK, GJK
Methodology: HJK, CMP, GJK
Writing–original draft: HJK
Writing–review & editing: HJK, GJK
All authors read and approved the final manuscript.