Association of hyperkalemia with clinical outcomes in advanced chronic kidney disease

Abstract

IntroductionPatients with advanced chronic kidney disease (CKD) are at greatest risk of hyperkalemia (HK). The relationship between HK and negative outcomes (mortality or progression of renal insufficiency) in non-dialysis dependent CKD patients is controversial. AimsTo determine the incidence, prevalence, and factors related with HK in a cohort of CKD patients, and its relationship with mortality, hospitalization rate, CKD progression, and dialysis initiation. Material and methodsA retrospective, observational study in an incident cohort of adult patients with stage 4 or 5 CKD not on dialysis. Inclusion criteria were: having at least three consecutive estimated glomerular filtration rate (eGFR) measurements in a follow-up period >3 months. Decline in renal function was estimated as the slope of the individual linear regression line of eGFR over follow-up time. HK was defined as serum K levels ≥5.5meq/l. Associations of HK with outcomes were adjusted for major confounding variables in the multivariate analysis. ResultsThe study group consisted of 1079 patients (574 males, mean age: 65±14 years) with mean baseline eGFR 14.8±4.5ml/min/1.73m2. Mean follow-up time was 15 months with a median of 7 serum sample determinations per patient. HK was observed at baseline in 26% of patients; in at least one serum sample during the individual follow-up period in 68%; or chronically (>50% of samples) in 33% of patients. By multivariate logistic regression, the best determinants of chronic HK were: male sex (OR=1.529; 95% CI [1.154–2.025], p=.003), serum bicarbonate (OR=0.863 [0.829–0.900], p<.0001), diuretic treatment (OR=0.743 [0.556–0.992], p=.044), and angiotensin converting enzyme inhibitor and/or angiotensin receptor blockers (OR=4.412 [2.915–6.678], p<.0001). Patients whose serum K levels were in the upper quartile showed a significantly faster CKD progression (−4.05±5.22 vs. −2.69±5.61ml/min/1.73m2/year, p<.0001), and more frequent dialysis initiation (63% vs. 57%, p=.115), though lower mortality (9% vs. 17%, p=.003) and hospitalization rates (2.68±5.94 vs. 3.16±6.77 days per year, p=.301) than the other study patients. However, in the multivariate analysis, average serum K levels were not independently associated with the clinical outcomes investigated. ConclusionHK is a common biochemical finding in non-dialysis dependent CKD patients, mainly associated with prescribed medication. However, HK was not independently associated with major negative clinical outcomes.