[The prevalence, influencing factors and prognosis of abnormal serum potassium levels in patients with chronic kidney disease].

To estimate the prevalence of hyperkalemia and hypokalemia in patients with chronic kidney disease (CKD), analyze the influencing factors and explore the impact on disease prognosis. A total of 3 190 patients with CKD stage 1-4 from 39 tertiary clinical centers in China between November 2011 and December 2016 were recruited. The baseline characteristics of the patients were collected through face-to-face questionnaire investigation, physical examination and laboratory test. Meanwhile, the data of patient's end-stage renal disease, cardiovascular disease events and deaths were obtained up to December 2017 through active monitoring. The patients were categorized into three groups based on their baseline level of serum potassium (hypokalemia:<3.5 mmol/L, normal range: 3.5-<5.0 mmol/L, hyperkalemia: ≥5 mmol/L). Multi-nominal logistic regression was employed to evaluate the association between clinical characteristics and the presence of hyperkalemia or hypokalemia. The competing risk-based subdistribution Cox proportional hazards regression was used to assess the association between baseline level of serum potassium and various outcomes. The mean age of the patients was (50±14) years, with a male rate of 57.6% (1 839/3 190) and a majority of glomerulonephritis (59.7%, 1 668/2 792). Patients with CKD stage 3-4 accounted for 70.8% (2 260/3 190), and the mean level of serum potassium was (4.4±0.7) mmol/L. The prevalence of hypokalemia and hyperkalemia was 3.7% (=118) and 17.6% (=561), respectively. In the multivariable adjusted analysis, presence of history of cardiovascular disease (=0.33, 95%: 0.13-0.83, =0.019) and estimated glomerular filtration rate (=0.95, 95%: 0.91-0.98, =0.001) were inversely associated with hypokalemia, while use of thiazide or loop diuretic (=2.06, 95%: 1.51-2.81, <0.001) and estimated glomerular filtration rate (=1.13, 95%: 1.12-1.16, <0.001) were positively associated with hyperkalemia. After adjusting for relevant cardiovascular and renal risk factors, the result only showed a significant association between hypokalemia and risk of all-cause mortality (=2.12, 95%: 1.06-4.24, =0.034). Hypokalemia and hyperkalemia were not rare in patients with CKD in China, with the latter more prevalent. Hypokalemia was independently associated with the risk of death.