This systematic review and meta-analysis was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines13 and the Cochrane Handbook for Systematic Reviews of Interventions.14 The study was registered in the PROSPERO database under the registration number CRD42024603403. A completed PRISMA checklist is provided as supplementary material. Statement of human and animal rights was deemed unnecessary.

Studies were eligible for inclusion if they: (1) were randomized or nonrandomized clinical trials (including post-hoc analyses); (2) compared SZC with placebo, other potassium binders, or a potassium binder-free control arm; (3) enrolled patients receiving RAASi with prior hyperkalemia or at elevated risk for hyperkalemia; and (4) reported quantifiable outcomes on SZC's efficacy in maintaining RAASi therapy and preventing its discontinuation; (5) with a minimum follow-up of 3 months.

We excluded (1) studies without control groups; (2) overlapping patient populations; (3) case reports or case series; (4) review articles, comment articles, editorials, and letters to the editor.

A comprehensive and systematic search was conducted in the following electronic databases from their inception: Cochrane Central Register of Controlled Trials (CENTRAL), PubMed/MEDLINE, Embase, and Web of Science (WOS) which were consulted from inception until August 2025. The search strategy was developed in consultation with a medical librarian and utilized a combination of controlled vocabulary (MeSH terms) and free-text keywords. The complete electronic search strategy is described in the supplementary file. The references from all included studies, previous systematic reviews, and meta-analysis were also searched manually for any additional studies. Three authors independently extracted the data following predefined search criteria and quality assessment.

All records were imported into Zotero for reference management and deduplicated via Rayyan.ai, which also facilitated collaborative screening. Three reviewers independently screened titles/abstracts and full texts. Four authors then performed double-blinded, independent data extraction from primary and supplementary materials. Disagreements were resolved by consensus among the authors.

The risk of bias and quality assessment of the RCTs were conducted using the version 2 of the Cochrane Risk of Bias assessment tool.15 Nonrandomized clinical trials were assessed with the Risk of Bias in Non-randomized Studies of interventions tool (ROBINS-I).16 The risk of bias evaluation was independently performed by two authors. While disagreements were resolved through consensus after discussing the reasons for discrepancy. Publication bias was investigated using the Funnel-plot analysis and Egger's regression test.

Endpoints were analyzed using risk ratio (RR) with 95% confidence intervals (CIs) to compare treatment effects for categorical endpoints. For the primary analysis, all eligible comparator arms (placebo, other potassium binders and a potassium binder-free control arm) were pooled together to estimate the overall association between SZC use and RAASi-related outcomes. A random-effects model was applied to account for variability in effect sizes across studies. We assessed heterogeneity using the I2 statistic and Cochran Q test; p-values <0.10 and I2>25% were considered significant. In the meta-regression the significance of the regression slope was tested using the Q-test (reported as a two-sided p-value). Review Manager 5.4 (Cochrane Center, The Cochrane Collaboration, Denmark) and R version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analysis.

The primary outcomes included: (1) proportion of patients remaining on any dose of RAASi therapy at the end of follow-up (RAASi maintenance); (2) proportion of patients who up-titrated their RAASi dose during follow-up; (3) proportion of patients who down-titrated or discontinued RAASi therapy. Given potential clinical heterogeneity introduced by differing comparator types, prespecified subgroup analyses were performed, using (4) studies reporting outcomes specifically in patients receiving mineralocorticoid receptor antagonist (MRA) therapy; (5) patients with baseline serum potassium >5.0mEq/L who remained on any RAASi; and (6) using RCTs only. To explore the potential influence of comparator type, additional subgroup analyses were performed according to control group strategy (placebo or active potassium binder). These analyses were conducted when sufficient studies were available.Given the limited number of head-to-head trials and anticipated heterogeneity, we did not perform network meta-analysis or indirect comparisons between potassium binders. To explore potential effect modification, we conducted an exploratory random-effects univariate meta-regression and examined whether differences in net serum potassium change (ΔK, mmol/L) had any significant impact on the results and explained the heterogeneity in the pooled estimates. ΔK was calculated as mean end follow-up serum potassium minus baseline serum potassium (in mmol/L); thus, negative values indicate potassium reduction. Effects were modelled on the log(RR) scale and reported as RRs per 1.0mmol/L decrease in potassium (i.e. ΔK=−1.0).

The primary meta-analysis, including six studies, evaluated the association between SZC therapy and maintenance of any RAASi treatment. As described in the forest plot of Fig. 2A, the pooled analysis showed a significant 26% increase in the likelihood of RAASi maintenance in the SZC group compared to the control group (RR, 1.26; 95% CI, 1.14–1.40; p<0.0001). Significant heterogeneity was observed across the included studies (I2=77%).

Fig. 2.

Forest plots of primary endpoints: (A) RAASi maintenance: pooled risk ratio (RR) for remaining on any dose of RAASi at end of follow-up. (B) RAASi up-titration: RR for dose escalation during follow-up. (C) RAASi discontinuation or down-titration: RR for therapy reduction or withdrawal. Results are shown as risk ratios with 95% confidence intervals. The size of each square is proportional to the study's weight in the meta-analysis; horizontal lines represent 95% CI. The diamond represents the pooled estimate using a random-effects model.

The benefit of SZC was further demonstrated in therapy optimization. Patients treated with SZC were 47% more likely to undergo up-titration of their RAASi therapy compared to controls (RR, 1.47; 95% CI, 1.20–1.80; p=0.0002), as shown in Fig. 2B, with moderate heterogeneity (I2=62%, p=0.05). Conversely, the risk of down-titration or complete withdrawal of RAASi therapy was significantly reduced by 44% in the SZC group (RR, 0.56; 95% CI, 0.46–0.68; p<0.00001, Fig. 2C), with low heterogeneity (I2=18%, p=0.30).

Absolute risk differences corresponded to numbers needed to treat (Table S4 of the supplementary file) ranging from 3 to 12 for maintenance of RAAS inhibitor therapy, indicating a clinically meaningful improvement across studies. For prevention of RAASi withdrawal, NNT values ranged from 5 to 10.

To investigate the sources of heterogeneity and test the robustness of the primary findings, we performed several subgroup analyses. In the subgroup of patients who remained on MRA therapy, the point estimate favored SZC, but the result was not statistically significant (RR, 1.38; 95% CI, 0.98–1.93), with high heterogeneity (I2=90%, Fig. 3A). A sensitivity analysis including only patients with serum potassium levels >5.0mEq/L at baseline showed a consistent and significant 26% higher rate of RAASi maintenance with SZC (RR, 1.26; 95% CI, 1.06–1.49; Fig. 3B), with moderate heterogeneity (I2=51%, p=0.11). Furthermore, an analysis restricted to the three RCTs demonstrated a significant 31% benefit for RAASi continuation with SZC (RR, 1.31; 95% CI, 1.04–1.66; Fig. 3C), although, with higher heterogeneity (I2=64%, p=0.06). Subgroup analyses stratified by comparator type were performed to explore potential differences according to control strategy (supplementary file, Fig. S1). Effect estimates in placebo-controlled studies were consistent with the primary analysis.

Fig. 3.

Subgroup analyses of RAASi persistence with SZC. (A) Effect in patients receiving mineralocorticoid receptor antagonists (MRAs). (B) Effect in patients with baseline serum potassium >5.0mmol/L. (C) Analysis restricted to randomized controlled trials (RCTs) only. Data are presented as risk ratios (RR) with 95% confidence intervals. Heterogeneity was assessed using the I2 statistic.

An exploratory meta-regression of four studies was performed to assess whether the magnitude of serum potassium reduction (ΔK, in mmol/L) explained the treatment effect of SZC on RAASi maintenance. As shown in Fig. 4, the meta-regression slope was −0.39 (SE=0.126, p=0.002), i.e. for every 1.0mmol/L increase in ΔK the ln(RR) decreases by 0.39. Because ΔK is defined as (follow-up−baseline) potassium, a 1.0mmol/L decrease corresponds to ΔK=−1.0, which increases the log(RR) by +0.39. Equivalently, for every 1.0mmol/L decrease in serum potassium the RR for RAASi persistence is multiplied by exp(0.39)=1.48 (RR, 1.48; 95% CI, 1.16–1.90), corresponding to an estimated 48% higher relative likelihood of remaining on RAASi therapy. The model explained a large proportion of between-study variance (τ2≈0, R2=100%), suggesting a dose-dependent relationship (supplementary file, Tables S1–S3). However, given the limited number of studies and use of study-level data, these findings should be interpreted as exploratory.

Fig. 4.

Meta-regression of potassium reduction and RAASi persistence. Scatter plot showing the association between mean change in serum potassium (ΔK, mmol/L; defined as follow-up minus baseline) and the log risk ratio for RAASi persistence across four studies. Each circle represents a study, positioned at its mean potassium change (ΔK, mmol/L) and log risk ratio for RAASi persistence. Bubble area is proportional to study precision (inverse variance). The regression line (blue) shows the association between potassium reduction and RAASi maintenance, with dashed lines=95% CI. A 1mmol/L decrease in potassium was associated with a 48% higher likelihood of maintaining RAASi (RR, 1.48; 95% CI, 1.16–1.90).

RCTs showed low risk of bias, with only one of them showing some concerns, particularly in reporting and randomization. Observational studies demonstrated high risks of bias, especially for confounding, deviations from intended interventions, and selection of participants (supplementary file). Publication bias of the primary endpoint was investigated using the Funnel-plot analysis showing a symmetric distribution of studies with similar weights (supplementary file). Egger's regression test for funnel plot asymmetry was non-significant (z=0.28, p=0.78), suggesting no evidence of small-study effects or publication bias.

In this systematic review and meta-analysis of RCTs and observational studies including nearly 80,000 patients, SZC was associated with a 26% higher likelihood of remaining on RAASi therapy, 47% greater probability of up-titration, and 44% lower risk of dose reduction or discontinuation. Subgroup analyses, including restriction to randomized trials and stratification by comparator type, yielded results consistent with the primary findings. However, heterogeneity across studies was substantial, reflecting differences in study design, populations, and clinical settings. Reporting absolute risks with NNT values alongside relative risks improves interpretation of the magnitude and clinical relevance of treatment effects. However, NNT values should be interpreted in the context of study design heterogeneity.

The meta-regression analysis provides exploratory insights into the relationship between potassium reduction and RAASi maintenance. Our findings suggest a statistically significant ecological association between the magnitude of potassium reduction and RAASi maintenance across studies (coefficient=−0.39, p=0.002; R2=100%). While the result is based on only four studies and reflects group-level trends, it suggests that better potassium control may facilitate continued RAASi therapy and thus support cardiorenal protective strategies. Leave-one-out analysis (supplementary file, Table S3) revealed that the association remained statistically significant when any single study was omitted, except when the Kosiborod et al. trial was excluded (RR, 1.42, p=0.34), suggesting the overall result is sensitive to this trial, which reported the largest potassium reduction (ΔK=−1.0mmol/L). Although causal relationships cannot be established, the observed dose–response pattern indicates a link between potassium reduction and RAASi persistence, that align with the clinical importance of achieving and maintaining potassium levels within the target range (4.0–4.5mmol/L).23,1 These findings should be considered hypothesis-generating and require confirmation in patient-level analyses and prospective studies.

Hyperkalemia is common in CKD and HF, affecting up to half of patients with CKD, and is associated with its recurrence and adverse prognosis.3,6,7,24 SZC is a non-systemically absorbed potassium binder with generally favorable gastrointestinal tolerability in clinical and real-world settings,25 supporting its use in cardiorenal patients requiring chronic potassium control. Despite international guidelines discouraging RAASi withdrawal due to hyperkalemia and supporting use of newer potassium binders,5 registry data indicate that a substantial proportion of cardiorenal patients receive suboptimal RAASi therapy, with complete discontinuation in 14–16% of cases,4,9 particularly among those treated with MRAs.26

Our findings suggest that SZC may help mitigate hyperkalemia-related barriers to RAASi optimization, hence, facilitating the achievement of optimal quadruple therapy in HF and maximal renoprotective dosing in CKD, both linked to improved survival and fewer hospitalizations.5,27,28 However, RAASi persistence represents a process measure and surrogate outcome; this meta-analysis did not evaluate mortality, hospitalization, or renal endpoints. Therefore, the results should be interpreted as reflecting treatment optimization rather than demonstrated clinical benefit.

The inclusion of the ZORA real-world multicenter observational study19 and the REALIZE-K double-blinded randomized trial,21 enhances the clinical relevance and novelty of our findings. Notably, the effect of SZC was also observed in patients with baseline potassium >5.0mEq/L, supporting the early initiation (proactive approach) of the novel potassium binders in patients with higher potassium levels.29

Several limitations should be acknowledged. First, the evidence base is largely derived from observational studies, introducing potential residual confounding despite adjustment methods. Although analysis restricted to RCTs showed consistent results, their number remains limited. Second, the substantial heterogeneity in the primary analysis (I2=77%), though reduced by subgroup and sensitivity analyses, highlights variability in study design, populations, and clinical practice.

Third, evidence directly comparing SZC with other potassium binders was limited to only two studes.17,18 In the Onogi et al. study,18 the CPS arm was selected for analysis to avoid double counting and ensure methodological consistency. Although pharmacologic differences between SZC and patiromer have been described,25 this study was not designed to evaluate comparative effectiveness of these two novel potassium binders, highlighting the need for direct comparative trials.

Although dosing regimens were not specifically analyzed, most included studies used maintenance doses of 5–10g/day targeting serum potassium 4.0–4.5mmol/L, consistent with KDIGO 2024 guidance.5

Finally, while the meta-regression demonstrates a statistically significant ecological association between the magnitude of potassium reduction and RAASi persistence, this relationship is derived from a small number of studies using aggregate data, not individual patient effects. Therefore, the findings are exploratory and should not be interpreted as evidence of a causal or mechanistic relationship.