Purpose <p>To evaluate the efficacy and safety of oral sodium zirconium cyclosilicate (SZC) administered on non-dialysis days for the long-term management of chronic hyperkalemia in patients undergoing maintenance hemodialysis (MHD).</p> Methods <p>This single-center, prospective, open-label, parallel-group randomized controlled trial enrolled 116 MHD patients with chronic hyperkalemia and randomly assigned them 1:1 to an SZC group or a control group (58 per group). Both groups received routine hemodialysis and standardized dietary counseling. The SZC group additionally received oral SZC on non-dialysis days, with dose adjustment according to pre-dialysis serum potassium after the long inter-dialytic interval (LIDI). The total observation period was 9 months, including 6 months of treatment and 3 months of post-withdrawal follow-up after discontinuation of study SZC. The primary endpoint was the between-group difference in pre-dialysis serum potassium after LIDI at month 6. Repeated potassium measurements were analyzed using a linear mixed-effects model, with complete-case, multiple-imputation, and conservative missing-not-at-random sensitivity analyses.</p> Results <p>Baseline characteristics were generally balanced. The group-by-time interaction was significant (<i>P</i> &lt; 0.001), indicating different potassium trajectories over time. At month 6, the SZC group had a lower pre-dialysis serum potassium level after LIDI than the control group, with a model-estimated mean difference of − 0.363 mmol/L (95% CI − 0.542 to − 0.184). Sensitivity analysis yielded consistent results. During the 6-month treatment period, the frequency of target-range potassium measurements (4.0–5.0 mmol/L) was higher in the SZC group (IRR 1.88, 95% CI 1.476–2.392). Exploratory secondary clinical outcomes, including hyperkalemia-related unplanned hemodialysis, RAASi discontinuation burden, and cardiovascular hospitalizations, were lower in the SZC group, but the small event counts and open-label design preclude definitive clinical conclusions. No obvious within-group changes over time were observed in calcium, magnesium, albumin, hemoglobin, dry weight, or inter-dialytic weight gain-related indices. Five patients (8.6%) in the SZC group experienced mild hypokalemia (3.0–3.5 mmol/L), and four patients (6.9%) experienced constipation; no severe hypokalemia (&lt; 3.0 mmol/L) was observed.</p> Conclusion <p>Under the conditions of this study, oral SZC on non-dialysis days produced a modest but statistically significant reduction in pre-dialysis serum potassium after LIDI and was generally tolerated. The exploratory clinical-event findings are hypothesis-generating and require confirmation in larger blinded multicenter trials.</p> Trial registration <p>Chinese Clinical Trial Registry, ChiCTR2600119050. Registered February 14, 2026, retrospectively registered. <a href="https://www.chictr.org.cn">https://www.chictr.org.cn</a>. </p>