Background <p>Serum potassium levels are regulated by renal function; however, the extent to which the association with mortality is affected by renal function remains unknown. We aimed to explore the association of admission serum potassium levels with mortality in patients with coronary artery disease (CAD), stratified by renal function.</p> Methods and results <p>Renal function was categorized into normal (≥ 90&#xa0;mL/min/1.73&#xa0;m²), mildly decreased (60–89&#xa0;mL/min/1.73&#xa0;m²), and impaired (&lt; 60&#xa0;mL/min/1.73&#xa0;m²) according to estimated glomerular filtration rate (eGFR). The outcome was all-cause mortality within 3&#xa0;years after hospitalization. 7739 patients from a prospective multicenter registry were included. All-cause mortality occurred in 366 patients (4.7%). In normal renal function, restricted cubic spline analysis showed a U-shaped mortality curve. However, higher or lower potassium levels were not significantly associated with mortality risk compared with 4.0–4.5&#xa0;mmol/L. In mildly decreased renal function, the slope of the U-shaped curve on the right side was steeper, and potassium levels of 4.6–5.0&#xa0;mmol/L (hazard ratio (HR) 2.34, 95% confidence interval (CI) 1.48–3.70) and &gt; 5.0&#xa0;mmol/L (HR 5.76, 95% CI 2.54–13.12) were associated with higher mortality risk. In impaired renal function, the U-shaped curve flattened, and potassium levels &gt; 5.0&#xa0;mmol/L (HR 2.02, 95% CI 1.02–3.98) were linked to increased mortality risk. A significant interaction was observed between potassium levels and eGFR.</p> Conclusions <p>Our findings highlighted the independent association of elevated potassium levels with increased mortality in the mildly decreased renal function, indicating that renal function should be considered when evaluating mortality risk related to serum potassium in CAD.</p> Clinical Trial Registration information <p> <a href="https://clinicaltrials.gov/">https://clinicaltrials.gov/</a>, NCT04044066.</p> Graphical Abstract <p></p>