Usefulness of urinary potassium to creatinine ratio to predict diuretic response in patients with acute heart failure and preserved ejection fraction

Abstract Background Patients with acute heart failure (AHF) require intensification in the diuretic strategy. However, the optimal diuretic strategy remains unclear. In this work, we aimed to evaluate the role of urinary potassium to creatinine ratio (K/Cr) to predict diuretic and natriuretic response to thiazide or mineralocorticoid receptor antagonists (MRAs) in a cohort of patients with AHF and preserved ejection fraction (AHF‐pEF). Hypothesis Patients with a high urinary K/Cr ratio will have a better diuretic and natriuretic response with spironolactone versus chlorthalidone. Methods This is a study of 44 patients with AHF‐pEF with suboptimal loop diuretic response. The primary endpoint was the baseline K/Cr associated with natriuretic and diuretic effect of chlorthalidone versus spironolactone at 24 and 72 h. Mixed linear regression models were used to analyze the endpoints. Estimates were reported as least squares mean with their respective 95% confidence interval (CIs). Results The median age of the study population was 85 years (82.5−88.5), and 30 (68.2%) were women. The inferential multivariate analysis suggested a greater natriuretic and diuretic effect of chlorthalidone across K/Cr levels. In the upper category, chlorthalidone translated into a statistically increase in natriuresis at 24 and 72 h. Chlorthalidone versus spironolactone showed ∆uNa of 25.7 mmol/L at 24 h (95% CI = −3.7 to 55.4, p = .098) and ∆uNa of 24.8 mmol/L at 72 h (95% CI = −4 to 53.6, p = .0106). The omnibus p value is .027. Multivariate analyses revealed a significant increase in 72 h cumulative diuresis irrespective of K/Cr status in those on chlorthalidone. Conclusions In patients with AHF‐pEF and suboptimal diuretic response, diuresis and natriuresis are higher with the administration of chlorthalidone over spironolactone. These data don't support the hypothesis that the K/Cr ratio can help guide the choice of thiazide diuretic versus MRA in AHF‐pEF patients on loop diuretic.

help guide the choice of thiazide diuretic versus MRA in AHF-pEF patients on loop diuretic.

K E Y W O R D S
acute heart failure, diuretic response, preserved ejection fraction, urinary potassium creatinine ratio 1 | INTRODUCTION Fluid overload or fluid redistribution is the main reason for hospital admission in acute heart failure (AHF), followed by increased cardiac filling pressures. 1 The mainstay of treatment for AHF consists of the administration of intravenous (i.v.) loop diuretics, with or without the addition of other diuretics. 2,3 In most patients, administration of i.v. loop diuretics is sufficient to obtain clinical relief and mitigate signs of fluid overload. However, in some patients a sequential nephron blockage with additional diuretic is necessary. 4,5 The evidence supporting the efficacy and safety of the diuretic sequence in patients with suboptimal loop diuretic response is scarce. Indeed, the benefit of adding thiazides to loop diuretics in patients with AHF is based on small observational studies and not on randomized clinical trials. 2,6,7 In addition, the diuretic effect of mineralocorticoid receptor antagonist (MRA) in patients with diuretic resistance is controversial. [8][9][10][11] We recently published that in patients with AHF treated with i.v. furosemide, the addition of chlorthalidone was associated with a short-term greater natriuresis and urine output compared to addition of spironolactone. 12 However, given that hyperaldosteronism is a key feature of diuretic resistance, and it translates into an increase in urine potassium levels, we hypothesize that high urine potassium will identify a subset of patients that benefits from MRA administration. Conversely, low urine potassium will identify those in which thiazides will provide a stronger natriuretic response.
In this work, we aimed to evaluate the role of urinary potassium to creatinine ratio (K/Cr) to predict diuretic and natriuretic response to thiazide or MRA in a cohort of patients with AHF with preserved ejection fraction (AHF-pEF) and suboptimal loop diuretic response.

| Study population and protocol
This is a substudy of ICARPo, a prospective observational study of hospitalized patients with AHF-pEF, treated with i.v. loop diuretics in which an additional diuretic (chlorthalidone or spironolactone) was prescribed. 12 The study population included patients with AHF-pEF admitted for AHF to the Internal Medicine Department of a third-level teaching hospital center between June 2020 and March 2021.
Patients meeting all of the following criteria were included: (1) older than 18 years; (2)

| Endpoint
The endpoint of interest was between treatment differences (chlorthalidone vs. spironolactone) in 24 and 72 h natriuresis and 72 h diuresis across K/Cr.

| Statistical analysis
Continuous variables were presented as median and interquartile range (IQR), expressed as (Q 1 −Q 3 ). Discrete data were expressed as frequency and percentages. The χ 2 test or Wilcoxon rank sum test was used as appropriate to compare baseline characteristics between the two treatment groups. K/Cr was categorized by the median of the distribution for exploratory purposes.

| Loop diuretic treatment
The group treated with spironolactone showed a statistical trend to receive greater doses of i.v. furosemide (Table 1).

| Length of stay
The median length of stay was 7 (p25 5 to 9 p75) days. Patients    Table 2. increased intake. 15 We speculate that patients with an elevated urine K/Cr ratio might respond better to thiazides due to better renal flow.
Considering the determinants of potassium excretion in urine, potassium intake was similar throughout the whole cohort since we did not educate our patients on a restriction in potassium intake. In our cohort, we did not measure plasma aldosterone levels, since 2/3 of the patients received drugs that were inhibitors of the renin

| Study limitations
Several limitations in this study should be noted. First, this is a small single-center observational study that only included patients with AHF and LVEF ≥50%. Second, it should be noted that urine potassium values were relatively low in the entire sample, probably due to the advanced disease and high resistance to diuretics. Third, in the current work we did not measure aldosterone levels or other surrogates of aldosterone activity.
Thus, we could not examine the role of urinary potassium as a proxy of aldosterone activity. Finally, it is a small prospective single-center observational study, in which the risk of significant residual confounding cannot be ruled out.
T A B L E 2 Estimates (β-coefficients) of the covariates included in the final model for natriuresis and diuresis at 72 h. Abbreviations: eGFR, estimated glomerular filtrate rate; K/Cr, urinary potassium to creatinine ratio.
In patients with an episode of AHF-pEF and suboptimal loop diuretic response, both diuresis and natriuresis are higher with the administration of chlorthalidone compared to spironolactone, irrespective of K/Cr status.