LAUSR

In patients with chronic kidney disease, serum potassium level is a factor influencing sudden cardiac death (SCD). The aim of our analysis was to study the combined effect of serum potassium level and renal function on the onset of SCD in elderly hypertensive subjects. Data from the 3620 hypertensive patients aged over 70 years were extracted from three randomized clinical trials included in the INDANA database. During a mean follow up of 4.5 years, 81 patients (2.24%) died from SCD. Mean serum potassium levels and prevalence of chronic kidney disease were not different in patients who died from SCD. In addition to serum potassium and creatinine levels, 14 clinical and biological variables linked to cardiovascular diseases recorded at baseline were analyzed using a Bayesian network. The area under the receiver operating characteristic curve of the Bayesian model reached 0.91. Bayesian inference was used to simulate the combined effects of serum potassium and creatinine levels on SCD. Our analysis, using simulated data from Bayesian model, showed that the estimated probabilities of SCD was significantly increased in case of hyperkalemia (>5.0 mmol/l) and in case of hypokalemia (<3.5 mmol/l) and in case of chronic kidney disease. Combined effects of serum potassium level and renal function revealed that chronic kidney disease increased the probability of SCD whatever the serum potassium level. Our results using a Bayesian model confirm the deleterious effects of hypokalemia, hyperkalemia and chronic kidney disease on SCD in elderly hypertensive patients.