Epidemiology of hyperkalemia in CKD patients under nephrological care: a longitudinal study.

Hyperkalemia is a potential life-threatening condition among chronic kidney disease (CKD) patients. Available estimates of the burden of this alteration in CKD are mainly derived from large administrative databases. Since K measurements in patients in these databases are often dictated by clinical reasons, longitudinal studies including pre-planned measurements of potassium independently of clinical complication/symptoms may produce more reliable estimates of the frequency and the risk factors underlying hyperkalemia in CKD patients. We estimated the prevalence and the incidence of hyperkalemia in a longitudinal study in 752 stages 2-5 CKD patients lasting 3 years and including up to seven pre-planned assessment of key biochemical measurements including K. At baseline, 203 out of 752 patients (27%) had serum K > 5.0 mM/L and 33% had acidosis (HCO3 ≤ 22 mmol/L). Among those without hyperkalemia at baseline (n = 549), 284 patients developed this alteration across the 3-year follow-up. The point prevalence of hyperkalemia rose from 27% (baseline) to 30% (last visit) (P = 0.001). In a multivariate model, hyperkalemia at baseline [odds ratio (OR):7.29, 95% CI 5.65-9.41, P < 0.001], venous bicarbonate levels [OR (1 mmol/l): 0.92, 0.89-0.96, P < 0.001], eGFR [OR (1 ml/min/1.73m2): 0.98, 0.97-0.99, P < 0.001], use of ACE inhibitors (OR: 1.68, 1.28-2.19, P < 0.001) and angiotensin II antagonists (OR: 1.30, 1.01-1.68, P = 0.045) were related to hyperkalemia over time. Of note, venous bicarbonate levels emerged as an independent risk factor of hyperkalemia over time also in a separate analysis of patients with and without hyperkalemia at baseline. In a cohort of CKD patients including pre-planned measurements of K, 27% of patients had hyperkalemia. Metabolic acidosis and the use of drugs interfering with renin-angiotensin system were the strongest modifiable risk factors for this potentially life-threatening alteration in CKD in longitudinal analyses in the whole study cohort and in patients developing de novo hyperkalemia over time.