Comparison of Oral or Intravenous Thiazides vs Tolvaptan in Diuretic Resistant Decompensated Heart Failure

Background: The investigators aim to evaluate the optimal regimen for restoring diuretic efficacy in patients with decompensated heart failure demonstrating loop diuretic resistance, for which guideline-based recommendations are weak secondary to a lack of evidence. By comparing the efficacy, cost, and adverse effects of currently recommended therapies and testing a novel diuretic combination, the investigators will augment the dearth of data that exists regarding this clinical challenge. Current heart failure guidelines recommend addition of a thiazide diuretic, listing either oral metolazone or intravenous chlorothiazide, to loop diuretic therapy as strategy to overcome loop diuretic resistance. At equipotent doses, these two therapies differ 250 fold in cost. To date, no prospective trial has compared the efficacy of these two commonly utilized therapies. Tolvaptan, an oral vasopressin 2 receptor antagonist, could restore diuretic efficacy when used in combination with loop diuretics. While the safety of this combination has been established in the EVEREST trials, tolvaptan has been formally studied in a limited capacity as combination therapy to restore loop diuretic resistance. Hypokalemia is a common adverse effect of combining a thiazide and loop diuretic, increasing the risk of atrial and ventricular arrhythmias in a population who is already at high risk. Hypokalemia as not been reported with the combination of tolvaptan and loop diuretics, likely due to tolvaptan's distinctive mechanism of action. This potential benefit could provide tolvaptan a unique advantage for combination diuretic therapy in environments when electrolyte monitoring cannot be routinely performed or in patients with frequent arrhythmic events. Methods: All patients will provide informed consent prior to enrollment. All patients will be randomized in a 1:1:1 fashion using an electronic randomization tool embedded in REDCAP. All patients will be started on a 2L/day fluid restriction and a 2g/day sodium restriction. Decisions regarding the initiation, titration, or discontinuation of standard heart failure medications (Angiotensin Converting Enzyme Inhibitors, Angiotensin Receptor Blockers, Aldosterone Antagonists, Beta Blockers, digoxin, hydralazine, nitrates) are left to the discretion of the treating physicians. Patients will be randomized to either intravenous chlorothiazide 500mg IV Q12H + an oral placebo capsule Q12H or intravenous placebo infusion Q12H + a capsule containing either oral metolazone 5mg PO Q12H or oral tolvaptan 30mg once daily and placebo capsule in the evening dose. (Relative potency: Metolazone 100 fold more potent than chlorothiazide) All electrolyte repletion, loop diuretic dose titration, and concomitant therapies to enhance diuresis if needed will be utilized at the provider's discretion. To prevent confounding heterogeneity in the diuretic treatment approach, a stepped care algorithm similar to the CARRESS-HF trial will be utilized for loop diuretics, both initial doses and subsequent dose changes, and for concomitant inotropes and vasodilators. A minimum furosemide equivalent dose of 580mg/24hrs (100mg IV bolus + 20mg/hr infusion rate) must be ordered at enrollment. Outcomes The primary outcome will be 48-hour standing scale weight change (kg) from enrollment among the metolazone, intravenous chlorothiazide, and tolvaptan arms, using metolazone group as the comparator group for all other groups. Secondary outcomes, using metolazone as the comparator group for each, will be: * 48 hour net urine output (mls) * mean change in serum creatinine,, blood urea nitrogen, and eGFR at 24 hours, 48hours, and at hospital discharge * mean change in diuretic efficiency at 24 and 48 hours from baseline value at enrollment * mean change in serum potassium at 24 and 48 hours from baseline value at enrollment * mean change in serum sodium at 24 hours, 48hrs, and at discharge from baseline value at enrollment * cumulative dose of potassium (mEq) and magnesium (g) supplementation administered at 24 and 48 hours * incidence of severe hypokalemia * need for escalation in study-directed loop diuretic therapy at 24 and 48 hours * addition of vasoactive or inotropic medication at 24 and 48 hours * Treatment failure (definition below) * Patient-scored congestion visual analog scale score at baseline, 24 and 48 hours * new cardiac arrhythmias (atrial and ventricular) during the study period * receipt of inotropic therapy, dopamine, or nitroglycerin; requirement of ultrafiltration or hemodialysis during index hospitalization * in-hospital mortality * pharmacoeconomic analysis of the direct costs in each arm including the cost of: study medication, additional non-trial protocol laboratory analysis cost related to monitoring of electrolytes, treatment of study medication related adverse effects (arrhythmias, hypotension, electrolyte repletion), escalation of loop diuretic therapy doses, addition of additional therapies for suboptimal diuresis (inotropic therapy, vasodilators), and new initiation of renal replacement therapies (hemodialysis or ultrafiltration). Study Definitions * Urine output: Total urine volume (ml) from time of study enrollment to 48 hours * Hypokalemia: Serum potassium value \< 3.5mEq/L * Severe Hypokalemia: Serum potassium value \< 3.0mEq/L * Hyponatremia: Serum sodium value \< 135mEq/L * Severe Hyponatremia: Serum sodium value \< 130mEq/L and a decrease of 5mEq/L or more from enrollment serum sodium * Overcorrection of serum sodium: increase in serum sodium from baseline by \>12mEq/L in 24 hours, increase in \>8mEq/L in 12 hours, or receipt of intravenous fluids because of symptoms of overcorrection of serum sodium regardless of the numerical rise * Hypomagnesaemia : Serum magnesium value \< 2mEq/L * Diuretic efficiency = 24hr urine output/ 24hr Lasix equivalents in milligrams * Weight: Standing weight on the same scale as used for baseline weight measurement * New Atrial Arrhythmia: A "new" diagnosis of atrial arrhythmia (includes atrial fibrillation, atrial flutter, ectopic atrial tachycardia) lasting \> 30 seconds OR any atrial arrhythmia which causes hemodynamic instability (MAP \< 60 and requiring intervention) * New Ventricular Arrhythmia: Ventricular tachycardia lasting longer than 30 seconds, or frequent non-sustained VT causing hemodynamic instability with MAP \< 60 mmHg requiring intervention or \> 1 intra-cardiac defibrillation or external cardiac defibrillation shock or ventricular fibrillation requiring defibrillation * Hypotension: SBP \< 85 for 2 repeated measurements within 30 minutes or lasting at least 30 minutes or symptomatic hypotension necessitating clinical intervention (defined as vasopressor support, intravenous fluid boluses, or initiation of inotropes) * Treatment failure: Patients requiring additional non-study diuretic (spironolactone doses \>75mg/day, eplerenone \> 75mg/day, non-study thiazides (at a dose of metolazone 2.5mg or greater equivalence) or loop diuretics, or systemic acetazolamide (for diuretic indication), triamterene, or amiloride therapy) at any time during the 48-hour randomization period. These patients will be considered treatment failures for the purpose of analysis of the primary endpoint and all secondary endpoints. * Patients whose cardiologist adds inotropic or vasodilator medications will not be considered treatment failures. Patients who require an increase in the loop diuretic regimen will also not be considered treatment failures. * Medication costs will be defined as the Redbook average wholesale price at the time of the trial to reduce inter-institutional price differences and improve external validity of the analysis. Statistical Analysis The investigators have collaborated with Department of Biostatistics at Vanderbilt University Medical Center to employ the best statistical methods that allow ther study to be realistic and achievable. Power calculations are difficult because of the lack of prospective trials comparing combination diuretic therapy and the numerous flaws in the methods of these previous studies. The investigators will utilize change in weight as the primary outcome because weight change has been utilized as a primary efficacy outcome in landmark heart failure diuretic trials (CARRESS-HF) and has less standard deviation than net urine output. In previous studies standard deviation of weight loss changes between groups varied with an approximate value of 1.6kg. If the minimum clinically meaningful difference in the experimental and control means is 1.5kg, the investigators will be able to reject the null hypothesis that the population means of the experimental and control groups are equal with 82.3% power. The Type I error probability associated with this test of this null hypothesis is 0.05. The investigators will utilize an intention-to-treat univariate Wilcoxon rank sum analysis for the independent continuous primary outcome variable using metolazone as the comparison group for both intravenous chlorothiazide and oral tolvaptan. The investigators will also perform a multivariate linear model adjusted analysis of the primary outcome to correct for baseline weight and loop diuretic regimen.