diuretics (for indications, see above). In the multivariate analysis, Chen et al. adjusted for estimated glomerular filtration rate, that is, kidney disease, but not for the other conditions, which had not been significantly related to mortality in univariate analyses. 3 Although hypertension was not related to mortality in our analysis, left ventricular dysfunction had a strong effect on mortality. 2 In conclusion, the data published by Chen et al. and the data available from our study do not sufficiently explain the mechanisms by which out-of-range potassium levels would affect mortality. 2,3 We are aware of the protective effect of surgery on survival. Therefore, all patients included in our study underwent surgery. 2 It is correct that cardiac dysfunction was already present at patient admission and did not change after surgery. 2 Since aortic regurgitation and pericardial effusion were surgically treated, we assumed that cardiac dysfunction was a chronic condition of these patients that persisted after surgery and cannot be surgically treated. Nevertheless, increased awareness of the high risk of these patients, especially during the first few postoperative months, should induce intensified surveillance. In addition, optimization of medical therapy according to current guidelines for treatment of cardiac failure might help to increase survival.