Competing-risk analysis of death and end stage kidney disease by hyperkalaemia status in non-dialysis chronic kidney disease patients receiving stable nephrology care

Abstract

Hyperkalaemia burden in non-dialysis chronic kidney disease (CKD) under nephrology care is undefined. We prospectively followed 2443 patients with two visits (referral and control with 12-month interval) in 46 nephrology clinics. Patients were stratified in four categories of hyperkalaemia (serum potassium, sK ≥ 5.0 mEq/L) by sK at visit 1 and 2: Absent (no-no), Resolving (yes-no), New Onset (no-yes), Persistent (yes-yes). We assessed competing risks of end stage kidney disease (ESKD) and death after visit 2. Age was 65 ± 15 years, eGFR 35 ± 17 mL/min/1.73 m2, proteinuria 0.40 (0.14–1.21) g/24 h. In the two visits sK was 4.8 ± 0.6 and levels ≥6 mEq/L were observed in 4%. Hyperkalaemia was absent in 46%, resolving 17%, new onset 15% and persistent 22%. Renin-angiotensin-system inhibitors (RASI) were prescribed in 79% patients. During 3.6-year follow-up, 567 patients reached ESKD and 349 died. Multivariable competing risk analysis (sub-hazard ratio-sHR, 95% Confidence Interval-CI) evidenced that new onset (sHR 1.34, 95% CI 1.05–1.72) and persistent (sHR 1.27, 95% CI 1.02–1.58) hyperkalaemia predicted higher ESKD risk versus absent, independently from main determinants of outcome including eGFR change. Conversely, no effect on mortality was observed. Results were confirmed by testing sK as continuous variable. Therefore, in CKD under nephrology care, mild-to-moderate hyperkalaemia status is common (37%) and predicts per se higher ESKD risk but not mortality.