LDs are first line therapy in the treatment of patients with ADHF. Of all diuretic classes, LDs have the most immediate onset of action when given intravenously and the greatest impact on sodium and water excretion. Furosemide, bumetanide, and torsemide are sulfonamide loop diuretics that reversibly bind to and reversibly inhibit the Na⁺: K⁺:2Cl⁻ co-transporter on the apical membrane of epithelial cells in the thick ascending limb of the loop of Henle. LDs inhibit sodium transport at this site in the nephron. Loop diuretics can increase sodium excretion by as much as 20–25 % of filtered sodium and augment excretion of both sodium and water. Inhibition of the co-transporter also inhibits reabsorption of calcium and magnesium. The increased sodium delivered to the distal convoluted tubule significantly increases the excretion of urinary potassium via the sodium potassium co-transporter. This effect is amplified by elevated levels of aldosterone typically seen in patients with ADHF. Loop diuretics indirectly decrease the reabsorption of water in the collecting ducts by decreasing the concentration of sodium in the medullary interstitium resulting in a decrease in the driving force for water reabsorption in the collecting duct. Urine produced in response to LDs is mildly hypotonic when compared with plasma [19–26].

Ethacrynic acid is a non-sulfonamide loop diuretic that inhibits reabsorption of sodium and chloride in the ascending loop of Henle and distal renal tubule by interfering with the chloride-binding co-transport system, causing increased excretion of water, sodium, chloride, magnesium, and calcium. Ethacrynic is a less effective diuretic than the other three loop diuretics, is probably more ototoxic at high doses, and is cumbersome to administer intravenously because of its relative insolubility. Its use is limited to patients who have an allergy to sulfonamide LDs [27].

An adequate intraluminal concentration of diuretic is needed for LDs to be effective. Because LDs are extensively bound to plasma proteins, delivery to the loop of Henley by glomerular filtration is limited. LDs are secreted into the intraluminal space of the nephron by the organic acid transport system in the proximal tubule. Intraluminal concentration is dependent on dose, bioavailability, adequate renal blood flow and adequate proximal tubule secretory function. Secretion into the proximal tubule may be impaired in heart failure by a decrease in renal plasma flow. In addition, in renal insufficiency, the accumulation of organic anions (e.g., blood urea nitrogen) competes with LDs for the receptor sites of the organic anion transporter. Higher doses are required to overcome this competitive inhibition and to obtain therapeutic urinary concentrations in patients with heart failure and renal impairment [20, 25].

Loop diuretics have an “S” shaped dose-response curve characterized by a threshold concentration below which no diuresis occurs (the minimal effective concentration), a steep increase in dose response, and a concentration ceiling above which no additional diuretic effect is seen. Diuretic effect is dependent on achieving an adequate or “threshold” intraluminal concentration of diuretic in the thick ascending limb of the loop of Henle. In heart failure, the dose response curve is shifted downward and to the right so that a higher concentration of LD is needed to induce a diuresis and the peak response is decreased (see Fig. 11.2) [23].

The dose response curve has a number of clinical implications. First, there is a threshold concentration of LD that needs to be achieved at the active site to elicit any diuretic response. Diuretic doses that achieve intraluminal concentrations below threshold are ineffective so that it is important to demonstrate a response to a specific dose of LD rather than give an inadequate dose more frequently. Because of individual differences in diuretic sensitivity and pharmacokinetics, the dose that achieves threshold concentrations differs among patients. Second, the diuretic ceiling limits the maximal effect of increasing diuretic dose. A ceiling dose of diuretic can be identified in individual patients (the lowest dose of LD that elicits a maximal response). If additional diuresis is needed, the dosing frequency should be increased rather than increasing the LD dose above the ceiling dose [19, 25].

When a bolus of LD is administered, there is generally a diuretic response within 30 minutes that peaks in 1 hour provided that the intraluminal concentration of diuretic is above the threshold concentration in the loop of Henle. When the intraluminal concentration of LD declines below the threshold concentration, urinary excretion of sodium stops and compensatory sodium retention occurs (post-diuretic salt retention or “rebound”) until another diuretic dose is administered and threshold concentration is again achieved. It takes four half-lives for LDs to reach steady state so that administration at a frequency of longer than 4 half-lives will allow for a period of “post-diuretic sodium retention” [23]. Diuretic rebound is directly related to sodium and water excretion and can be prevented by replacement of excreted sodium with normal saline. If dietary intake is not limited, diuretic-induced net sodium excretion may be nullified by post-diuretic salt retention. This phenomenon suggests that in patients hospitalized with ADHF, sodium intake should be restricted and LDs should be given at least 2–3 times/day or administered as a continuous infusion [19, 21, 26, 28].

Two forms of diuretic tolerance have been described. Acute tolerance develops within the first several days of LD therapy and refers to a progressive time-dependent decline in sodium excretion in response to the same dose of LD. This form of early tolerance has been referred to as the “breaking phenomenon”. A number of studies have demonstrated decreases in mean blood pressure and eGFR and increases in renin, angiotensin II, aldosterone and plasma norepinephrine levels after a single dose of IV furosemide. The pathophysiology of early tolerance is likely multi-factorial with contributions from: a decrease in renal blood flow mediated by hypotension, increased angiotensin II and sympathetic activation; increases in renin, angiotensin II and aldosterone levels; an increase in sodium reabsorption in the proximal tubule mediated by angiotensin II; and an increase in distal sodium absorption mediated by aldosterone [25, 26, 29]. The release of adenosine by the macula densa/juxtaglomerular apparatus in response to a LD-induced increase in sodium chloride concentration in the distal loop of Henley may contribute to early tolerance by reducing renal blood flow and GFR [30–32]. In addition, vasopressin-induced upregulation of the Na⁺: K⁺:2Cl⁻ co-transporter may also contribute to tolerance [33].

A second type of LD tolerance occurs with chronic administration of LD. With LD administration, the nephron distal to the loop of Henle is flooded with solute. This causes diuretic-induced hypertrophy of the distal convoluted tubule and functional changes in the distal nephron that result in increased sodium reabsorption in the distal nephron. This attenuates loop diuretic-induced sodium and water excretion [20, 23].

Furosemide, bumetanide, and torsemide all act by reversibly inhibiting the Na⁺:K⁺: 2Cl⁻ co-transporter in the ascending limb of the loop of Henle. They differ primarily in oral bioavailability, dose, metabolism, relative potency and duration of action. Furosemide is the most commonly used diuretic in patients with ADHF. In the ADHERE Registry, 84 % of patients received IV furosemide, 7 % received IV bumetanide and 2 % received IV torsemide [34]. All of the loop diuretics (including ethacrynic acid) are extensively bound to plasma proteins, have limited glomerular filtration and are secreted into the intraluminal space of the nephron by the organic acid transport system in the proximal tubule. Furosemide has the most variable bioavailability ranging from 40–70 % of the oral dose. The bioavailability of bumetanide and torsemide are 80–100 % of the oral dose. Furosemide is metabolized and excreted by the kidney. Bumetanide and torsemide are primarily metabolized by the liver. The onset of diuretic effect when given intravenously is 5 min for furosemide, 2–3 min for bumetanide and 10 min for torsemide. The onset of diuretic effect when given orally is 30–60 min for all three LDs. Furosemide has a half-life of 1.5–2 h, bumetanide 0.8 (range 0.3–1.5) h and torsemide 3.5 h. The duration of action of furosemide is 6–8 h; bumetanide 4–6 h and torsemide 12 h (range 6–16). Furosemide and bumetanide are available for oral and intravenous administration while torsemide is currently available only for oral administration in the United States. The relative potency of LDs given intravenously is furosemide 40 mg: bumetanide 1 mg: torsemide 20 mg. IV to PO conversion is 1:2 for furosemide and 1:1 for bumetanide and torsemide [35–38]. Table 11.1 summarizes the pharmacokinetics of the sulfonamide loop diuretics [24, 39]

Ethacrynic acid has a bioavailability of 100 % of the oral dose, is significantly protein bound and is secreted into the proximal tubule via the organic acid transporter. It has a half-life of approximately 1 h, duration of action of 4–6 h, and potency relative to IV furosemide of 0.7 [24].

Loop diuretics should be given intravenously in ADHF as the bioavailability of furosemide is highly variable and the absorption of all loop diuretics may be impeded by bowel edema or intestinal hypoperfusion [40]. Loop diuretics have a number of benefits in volume-overloaded patients with ADHF. Intravenous administration of an effective dose of loop diuretic generally results in a diuretic effect that peaks within an hour of administration [19, 23, 24, 38]. Diuretic induced excretion of sodium and water affects a reduction in right- and left-sided filling pressures with a decrease in pulmonary and systemic venous congestion and a decrease in left ventricular dilation. In patients with LV systolic dysfunction and volume overload, the left ventricle is operating on the flat part of the LV performance curve where stroke volume is relatively independent of LV filling pressures. Diuretic induced reductions in LV filling pressure generally do not cause hypotension or a reduction in stroke volume or cardiac index [41]. Diuretic-related reduction in left and right sided ventricular filling pressures is associated with an improvement in stroke volume and cardiac output related to: a decrease in functional mitral and tricuspid regurgitation; a decrease in right ventricular volume with relief of ventricular-interdependent LV compression; improved endocardial blood flow; and a reduction in LV wall tension resulting in a decrease in myocardial oxygen consumption. The reduction in wall tension may be particularly important in patients with coronary artery disease. The associated reduction in secondary mitral regurgitation and left ventricular wall tension results in an improvement in cardiac output and overall myocardial performance [42–45].

Loop diuretics also have hemodynamic effects independent of their diuretic effects. Several studies have suggested that administration of an intravenous loop diuretic results in an early reduction in pulmonary capillary wedge pressure that may be independent of diuretic effect [46–48]. This effect is likely due to dose-dependent direct venodilation mediated by the release of vasodilatory prostaglandins [49, 50]. These findings may explain why LDs can produce clinically significant reductions in left- and right-sided filling pressures and improvement in symptoms in as little as 15 min after administration [47].

In contrast, in some patients, IV loop diuretics may cause an early increase in systemic vascular resistance and systolic blood pressure. This effect is not mediated by a direct vascular effect but rather, by neurohormonal activation of the sympathetic nervous system and RAAS [42, 43, 50]. In a study of the hemodynamic and neurohormonal responses to IV furosemide in 15 patients with severe chronic heart failure, mean arterial pressure, heart rate, systemic vascular resistance, and left ventricular filling pressure increased and stroke volume index decreased 20 min after the administration of intravenous furosemide. These changes were associated with increases in plasma norepinephrine, plasma renin activity and plasma arginine vasopressin. At 2 h, patients had diuresed and had a reduction in filling pressures to below baseline. Neurohormone levels returned toward baseline [51]. In another study, neurohormonal activation was followed in 34 patients admitted with acute decompensation of chronic HFrEF. Patients were treated with hemodynamically guided therapy with diuretics and sodium nitroprusside titrated to reduce filling pressures and lower systemic vascular resistance. Patients were then transitioned to oral therapy. Neurohormone levels were obtained at baseline, during parenteral treatment (mean 1.4 days) and after transition to oral therapy (mean 3.4 days). Filling pressures (PCWP 31 to 18 mmHg; RA 15 to 8 mmHg) and cardiac index (1.7 to 2.6 L/min/m²) improved during treatment with parenteral medications. Plasma norepinephrine levels did not change during parenteral therapy but decreased after transition to oral medication. Plasma aldosterone and plasma renin activity increased during parenteral therapy. Aldosterone levels returned to baseline and plasma renin activity remained elevated after transition to oral medication. Plasma endothelin levels decreased during parenteral therapy and remained lower after transition to oral medication [52]. These observations support the early use of a vasodilator added to a diuretic to attenuate the effects of LD-induced neurohormonal activation.

There has been little prospective evidence to guide diuretic use in ADHF. It has been hypothesized that the continuous infusion of a loop diuretic should avoid the periods of ineffective diuretic concentration in the loop of Henle seen with intermittent bolus therapy and should result in greater diuresis. It has also been suggested that continuous infusion should be associated with fewer electrolyte abnormalities, preservation of renal function and a shorter length of stay. A Cochran review published in 2005 compared intermittent bolus administration with continuous intravenous infusion of loop diuretics in eight studies of a total of 254 patients. The studies were small (8–107 patients) and heterogeneous. There were significant differences among the studies with respect to diuretic dose, method of administration, follow-up period and clinical outcomes reported [53]. The authors concluded that the data were insufficient to assess the merits of the two methods of LD administration.

The Diuretic Optimization Strategies in Acute Heart Failure (DOSE) trial was a multicenter, randomized, controlled, double-blind trial that assessed the effect of diuretic dose and mode of administration in patients admitted with ADHF [54]. 308 patients hospitalized for ADHF with a history of chronic heart failure treated with an oral loop diuretic at a daily dose of furosemide of 80–240 mg/day or equivalent for at least one month prior to hospitalization were enrolled in the study. Using a 2 × 2 factorial design, patients were randomized in a 1:1:1:1 ratio to receive furosemide administered intravenously by means of either a bolus every 12 h or continuous infusion and with either a low dose (equivalent to the patient’s previous oral dose) or a high dose (2.5 times the previous oral dose). Dose adjustments could be made at 48 h. There were two co-primary end-points: patients’ global assessment of symptoms, quantified as the area under the curve on a visual-analogue scale over the course of 72 h (the efficacy end-point); and the change in the serum creatinine level from baseline to 72 h (the safety end-point). There was no significant difference in the patients’ global assessment of symptoms or mean change in creatinine between the bolus and continuous infusion groups. Patients in the bolus therapy group were more likely to require an increase in furosemide dose at 48 h. There was a non-significant trend toward greater improvement in symptoms assessed by the visual analogue scale in the high dose group compared to the low dose group (P = 0.06). The high dose strategy was associated with greater relief of dyspnea (P = 0.04), greater fluid (p = 0.001) and weight (p = 0.01) loss and fewer adverse events (p = 0.033). Patients in the high dose group were less likely to require an increase in furosemide dose and more likely to be changed to oral therapy at the 48-h reevaluation. There were more cases of ventricular tachycardia and myocardial infarction with bolus than with continuous infusion and with the low-dose strategy than with the high-dose strategy. A higher proportion of patients in the high dose group met the pre-specified secondary safety end-point of an increase in serum creatinine of more than 0.3 mg per deciliter at any time during the 72 h after randomization. Although prior studies have suggested that an increase in creatinine during hospitalization for ADHF is associated with worse long-term outcome [55, 56], there was no evidence of worse clinical outcome in the high dose group at 60 days. This is consistent with other reports that found that a transient increase in creatinine during hospitalization may not be associated with poorer outcomes after discharge [57, 58].

Two recent meta-analyses that included the results from the DOSE Trial compared the efficacy of continuous infusion versus IV bolus of loop diuretics in patients hospitalized with ADHF. One meta-analysis included ten randomized controlled trials with a total of 518 patients in the analysis. Continuous infusion of diuretics was associated with a significantly greater weight loss compared with bolus injection but no significant differences in urine output, the incidence of electrolyte abnormality, change in creatinine level, hospital length of stay, the incidence of ototoxicity, cardiac mortality, or all-cause mortality [59]. The second meta-analysis included 7 cross-over and 8 parallel-arm randomized trials in adults with a total of 844 patients in the analysis. 8/15 studies included patients with ADHF, 3 included ICU patients, 2 included cardiac surgery patients and 2 included patients with chronic kidney disease. A non-significant net increase in daily urine output was seen with continuous infusion diuretic therapy. When the 8 studies that included an initial loading dose were analyzed, continuous loop diuretic infusion was associated with a significant net increase in daily urine output of 294 ml/day and a significant net negative weight loss of 0.78 kg when compared with intermittent infusion [60].

The optimal dose of LD is uncertain and needs to be individualized based on age, prior diuretic use, severity of heart failure, the presence of impaired systemic perfusion, creatinine and estimated glomerular filtration rate calculated using either the Cockroft-Gault (CG) or the Modified Diet in Renal Disease (MDRD) equations. Diuretics should be given at a dose and frequency sufficient to relieve symptoms and signs of congestion and normalize volume status without causing excessively rapid reduction in intravascular volume or clinically significant electrolyte abnormalities [8].

The ACCF/ACC guidelines recommend that in patients already receiving a loop diuretic, the initial diuretic dose should equal or exceed their chronic oral daily dose and be given either as intermittent intravenous boluses or continuous infusion [28]. The ESC guidelines recommend that an initial dose of furosemide 20–40 mg IV (or 0.5–1.0 mg bumetanide or 10–20 mg of torsemide) be given on admission; in patients with evidence of volume overload, the dose of parenteral diuretic may be higher based on renal function and a history of chronic oral diuretic use [18]. Continuous infusion may also be considered after an initial starting bolus dose. The ESC recommends that the total furosemide dose should remain <100 mg in the first 6 h and <240 mg during the first 24 h. Diuretic dosing guidelines from the ASCEND-HF trial based on creatinine clearance and the chronicity of heart failure are summarized in Table 11.2 [61]. The recommended doses of LDs given by continuous infusion are summarized in Table 11.3 [23, 62].

There is conflicting data concerning the rapidity of response to parenteral diuretics in patients with ADHF. There is a general understanding that most patients have a rapid and significant improvement in dyspnea after administration of a parenteral diuretic although not all studies confirm this [8, 63]. In the Value of Endothelin Receptor Inhibition With Tezosentan in Acute Heart Failure (VERITAS) trial, patients admitted to the hospital with acute heart failure, dyspnea, tachypnea and evidence of volume overload or LV systolic dysfunction were treated with standard heart failure therapy which consisted mostly of parenteral diuretics and randomized to an infusion of tezosentan, an intravenous short-acting endothelin receptor antagonist, or placebo. The primary end-point was the change in dyspnea over the first 24 h measured at 3, 6 and 24 h using a visual analog scale. There were rapid and significant improvements in dyspnea in the placebo and tezosentan groups at 3, 6, and 24 h. Interestingly, these changes were associated with small changes in hemodynamic parameters. In the placebo group, baseline PCWP was 25.6 mmHg and decreased by 1.5, 1.9, and 2.9 mmHg at 3, 6, and 24 h, respectively. Baseline RAP was 15.9 mmHg and changed by 0.8, −0.2 and 0.7 mmHg at 3, 6, and 24 h respectively. Baseline cardiac index was 2.01 L/min/m² and increased by 0.18, 0.18 and 0.15 L/min/m² at 3, 6, and 24 h, respectively. Baseline systemic vascular resistance was 1813 dyne-sec/cm⁵ and changed by −157, −54 and 136 dyne-sec/cm⁵ at 3, 6, and 24 h [64].

In the Efficacy of Vasopressin Antagonism in Heart Failure Outcome Study with Tolvaptan (EVEREST) trial, patients with chronic heart failure and systolic dysfunction who were hospitalized for worsening HF received standard therapy for acute heart failure and were randomized to receive tolvaptan, an oral non-peptide selective V₂-receptor antagonist or placebo. Approximately 2/3 of the placebo patients reported mild, moderate or marked improvement in dyspnea but only one third of patients reported moderate or marked improvement in dyspnea at day 1 of the study. Physician assessment of signs and symptoms of heart failure on day 1 noted improvement in dyspnea in 47.1 % of patients, orthopnea in 59.2 %; JVD in 43.8 %; rales in 43.7 % and edema in 52.6 % [65].

The rate of dyspnea relief was reviewed in a post hoc analysis of data from the Placebo-controlled Randomized study of the selective A1 adenosine receptor antagonist rolofylline for patients hospitalized with acute heart failure and volume Overload to assess Treatment Effect on Congestion and renal funcTion (PROTECT) clinical trial pilot study. In PROTECT, patients with acute heart failure and mild to moderate renal impairment defined as an eGFR of 20–80 mL/min using the Cockcroft-Gault equation with dyspnea at rest or with minimal activity, signs and symptoms of volume overload and an elevated BNP or NT-pro-BNP were randomized within 24 h of presentation to treatment with placebo or one of three doses of rolofylline, a selective A1 adenosine receptor antagonist. Dyspnea relief was defined as a moderate to marked improvement in dyspnea using a seven-point Likert scale and occurred in the placebo group in 49 %, 64 %, 78 %, and 75 % of patients at 24 h, 48 h, day 7 and day 14, respectively. Improvement at 24 and 48 h was associated with greater improvement in patient-reported general wellbeing, and greater decreases from baseline in edema, rales, jugular venous distention and orthopnea. Patients with relief of dyspnea did not have a greater decrease in weight compared to patients without relief of dyspnea [66].

The frequency of early dyspnea relief was assessed in the Preliminary study of RELAX in Acute Heart Failure (Pre-RELAX-AHF) study. Patients in Pre-RELAX-HF had acute heart failure, evidence of volume overload, a systolic blood pressure of >125 mmHg, and impaired renal function with an eGFR of 30–75 ml/min/1.73 m² calculated using the simplified Modification of Diet in Renal Disease (sMDRD) equation. Patients were randomized to placebo or one of four doses of relaxin, a naturally occurring peptide vasodilator. Early dyspnea relief was defined as moderate or marked improvement in dyspnea assessed by the Likert scale at 6, 12, and 24 h after the start of study medication. Moderate or marked improvement in dyspnea at 6, 12 and 24 h was observed in only 23 % of patients in the placebo group. 70 % of all patients had moderate or marked improvement in dyspnea at 14 days. Improvement in dyspnea was correlated with improvements in general well-being measured by a visual analogue scale, orthopnea, dyspnea on exertion, edema, and rales but not with improvements in JVP or weight. Early dyspnea relief was predicted by a higher initial systolic blood pressure and respiratory rate [67].

The Ularitide Global Evaluation in Acute Decompensated Heart Failure (URGENT) Dyspnoea study was an international, multi-center, observational cohort study of 524 acute heart failure patients managed conventionally and enrolled within 1 h of first hospital medical evaluation [68]. The primary outcome was patient assessed dyspnea at 6 h measured by a 5-point Likert scale. Dyspnea improvement was reported in 76 % of patients after 6 h of standard therapy.

The zinc metalloenzyme carbonic anhydrase plays a key role in NaHCO₃ ⁻ resorption and acid secretion in the proximal tubule and, to a lesser extent, in the collecting duct. Acetazolamide is the prototype of a group of carbonic acid inhibitors and the only one used in clinical practice. Acetazolamide inhibits the absorption of NaHCO₃ ⁻ in the proximal tubule. In the absence of another diuretic that inhibits sodium reabsorption in the distal nephron, treatment with acetazolamide does not cause a significant diuresis. It does however, cause significant excretion of HCO3 ⁻, an increase in urinary pH, significant excretion of potassium and metabolic acidosis [80]. Acetazolamide has been shown to cause a marked natriuresis when added to either hydrochlorothiazide or furosemide in patients resistant to either diuretic and who have a low fractional excretion of sodium [81]. However, its use is limited by the development of metabolic acidosis. In clinical practice, acetazolamide is only used in ADHF in the setting of marked diuretic-induced normokalemic, hypochloremic metabolic acidosis [82].

Benzothiadiazides were the first drugs to inhibit the Na⁺-Cl⁻ cotransporter in the distal convoluted tubule. This class of medication came to be known as thiazide diuretics (hydrochlorothiazide, chlorothiazide). Subsequently, drugs that were pharmacologically similar to thiazides and inhibited the Na⁺-Cl⁻cotransporter were termed thiazide – like diuretics (metolazone, chlorthalidone) [27]. Thiazide diuretics inhibit the Na⁺-Cl⁻ cotransporter in the distal convoluted tubule and promote Na⁺ and Cl⁻ excretion. Thiazides are only moderately effective diuretics – they block the reabsorption of 5–10 % of filtered sodium compared with LDs which block 25 % of filtered sodium. Thiazides are ineffective as monotherapy in the treatment of more than mild heart failure and are not used as single agents in the treatment of ADHF. However, thiazide diuretics are very useful when added to LDs in treating patients with ADHF who remain volume overloaded despite appropriate treatment with adequately dosed LD (see below). Side effects from thiazide diuretics include hypokalemia, hyponatremia (unrelated to arginine vasopressin release), metabolic alkalosis, hypomagnesemia, hypercalcemia and hyperuricemia [23, 25, 83].

Triamterene and amiloride have the same mechanism of action and are the only drugs in this class. Spironolactone and eplerenone also decrease the excretion of potassium but are more appropriately classified as mineralocorticoid receptor antagonists (see below). Triamterene and amiloride block epithelial Na⁺ channels (amiloride-sensitive Na+ channels or ENaC) in the luminal membrane of principal cells in the late distal tubule and collecting duct. The effect of ENaC blockade on lumen negative voltage decreases potassium (and H⁺, Ca²⁺ and Mg²⁺) excretion and increases serum potassium concentration. Na⁺ excretion is minimally increased (2 % of filtered load of sodium). The major side effect of this class of medication is the risk of hyperkalemia. As a result, these medications are contraindicated in in patients with hyperkalemia or patients with conditions that put them at risk of hyperkalemia including renal insufficiency, treatment with ACEI or ARB, treatment with potassium supplements, or treatment with MRAs. In light of their modest natriuretic effects and risk of hyperkalemia, triamterene and amiloride have little use in the treatment of patients with ADHF [25, 27].

Aldosterone is a steroid hormone with mineralocorticoid effects that is primarily produced in the renal cortex. Aldosterone plays a major role in the control of sodium and potassium homeostasis by increasing sodium and water resorption and increasing K⁺ and H⁺ excretion. Aldosterone binds to cytosolic mineralocorticoid receptors (MRs) with high aldosterone affinity in epithelial cells in the late distal convoluted tubule and collecting duct. MRs are members of the superfamily of receptors for steroid hormones, thyroid hormones, vitamin D and retinoids. When aldosterone binds to MR, the MR-aldosterone complex translocates to the nucleus and binds to specific DNA sequences that regulate the expression of multiple gene products. Transepithelial NaCl transport is enhanced. The resulting increase in lumen-negative transepithelial voltage increases the driving force for K⁺ and H⁺ secretion into the tubular lumen. The genomic effects of aldosterone take several hours to begin to take effect. Aldosterone also has non-genomic effects that are fast acting (occur within minutes) and are probably mediated by binding to plasma membrane MRs. These effects have not been well characterized but probably include an increase in blood pressure that is independent of sodium retention [27, 84, 85].

The MRAs spironolactone and eplerenone competitively inhibit the binding of aldosterone to the MR. The effects on urinary excretion of sodium are similar to those of the renal epithelial Na⁺ channel inhibitors. However, unlike the effects of Na⁺ channel inhibitors, the clinical effects of MRAs are dependent on endogenous aldosterone levels. Aldosterone levels are increased in heart failure despite treatment with ACEI or ARB and are increased further after administration of loop diuretics [86]. The higher the aldosterone level, the greater the impact of MRAs on urinary excretion. In the Randomized Aldactone Evaluation Study (RALES), spironolactone 25 mg daily did not increase urinary sodium excretion [87]. However, there is literature suggesting that higher doses (50–100 mg daily) of spironolactone significantly increase urinary excretion of sodium when added to standard therapy in patients with ADHF [88, 89].

MR antagonists are the only diuretics that do not require access to the tubular lumen to induce a diuresis. Spironolactone is absorbed partially (~65 %), is metabolized extensively by the liver, undergoes enterohepatic recirculation, is highly protein bound, and has a short half-life of ~1.6 h. The half-life is prolonged to 9 h in patients with cirrhosis. Eplerenone has good oral availability, is eliminated primarily by metabolism by CYP3A4 to inactive metabolites, and has a half-life of ~5 h.

MRAs are generally not used as diuretics in ADHF. They may be used to attenuate potassium losses from LD therapy. They should be initiated or continued during ADHF hospitalization as guideline directed medical therapy for their long-term benefit in patients with HFrEF (see below) [90].

The major side effect of MRAs is life-threatening hyperkalemia. Because spironolactone has affinity for other steroid receptors (progesterone and androgen receptors), it may cause gynecomastia, impotence, decreased libido, hirsutism, and menstrual irregularities. Eplerenone has very low affinity for androgen and progesterone receptors and generally does not cause these side effects.

Several studies have suggested that hemoconcentration (an increase in low hemoglobin or a relative increase in the cellular elements in blood) is a surrogate marker for intravascular volume status and may be a useful indicator of adequate decongestion in patients hospitalized for ADHF [99, 100]. A post-hoc analysis from the ESCAPE trial assessed the impact of hemoconcentration on outcome [58]. Three hundred thirty-six of 433 randomized patients who had paired baseline and pre-discharge hematocrit, albumin or total protein values were included in the analysis. Baseline to discharge differences in the laboratory values that fell within the top tertile of the group were defined as indicators of hemoconcentration. Patients with ≥2 paired laboratory values in the top tertile were considered to have evidence of hemoconcentration. The group of patients with evidence of HC received higher doses of diuretics, lost more weight and fluid and had greater reductions in filling pressures. HC was strongly associated with worsening renal function whereas changes in RA and PCWP were not associated with worsening renal function. HC was strongly associated with a lower 180-day mortality rate that persisted after adjustment for baseline differences in risk (HR 0.16; P = 0.001).

A post hoc analysis of 1969 patients enrolled in the PROTECT study assessed the relationship between the change in hematocrit during heart failure hospitalization and outcome [101]. Anemia at baseline was defined, according to the World Health Organization (WHO) criteria, as a baseline hemoglobin level of <13 g/dl for men and <12 g/dl for women and was present in 50.3 % of patients. Hemoconcentration was defined as an increase in hemoglobin levels between baseline and day 7 and was seen in 69.1 % of patients. HC was associated with better renal function at baseline, more weight loss and greater deterioration in renal function. The total dose of diuretics was lower in the patients with HC. Greater weight loss and better baseline renal function were associated with a more rapid increase in hemoglobin concentration. The absolute change in hemoglobin was an independent predictor of outcome. There was a 34 % reduction in all-cause mortality at 180 days for each gram/dl increase in hemoglobin between baseline and day 7.

A retrospective analysis of 1684 patients assigned to the placebo arm in the EVEREST trial found that 26 % of patients had evidence of hemoconcentration defined as a ≥ 3 % absolute increase in hematocrit between baseline and discharge or day 7 [102]. Patients with greater increases in hematocrit tended to have better baseline renal function. HC was associated with a greater risk of in-hospital worsening of renal function, which generally returned to baseline at 4 weeks after discharge. Patients with HC were less likely to have clinical congestion at discharge, and experienced greater in-hospital decreases in body weight and natriuretic peptide levels. After adjustment for baseline clinical risk factors, every 5 % absolute increase in in-hospital hematocrit change was associated with a 19 % reduction in all-cause mortality following discharge over an average follow up 9.9 months. HCT change was also associated with a significantly decreased risk of cardiovascular mortality or HF hospitalization at ≤100 days following randomization (HR 0.73). In the Korean Heart Failure (KorHF) Registry, HC (defined as an increase in hemoglobin levels between admission and discharge) occurred in 43 % of patients and was found to be an independent negative predictor of the combined primary end-point of all-cause mortality and HF hospitalization after adjusting for other HF risk factors (HR = 0.671; P < 0.001) [103].

There may be a difference in the prognostic value of hemoconcentration based on when it occurs during the HF hospitalization. In a single center study of 845 patients hospitalized for ADHF, hemoconcentration (defined as an increase in both hemoglobin and hematocrit) occurred in 422 patients. HC was defined as “early” or “late” based on whether the maximal increase in hemoglobin and hematocrit occurred in the first or second half of the hospitalization. Early and late patients had similar baseline characteristics, cumulative in-hospital diuretic administration, and degree of worsening renal function. Late patients had higher average daily loop diuretic doses, greater weight loss, later transition to oral diuretics, and shorter length of stay. Late HC was associated with a significant survival advantage over a median follow-up of 3.4 years compared with early HC (HR: 0.73, p = 0.026) and no HC (HR: 0.74; p = 0.0090). Early HC was not associated with a survival advantage [104].

This data suggests that more complete decongestion is associated with a better post-discharge prognosis and that HC may be a reasonable surrogate for assessing the adequacy of decongestion during an admission for decompensated heart failure. However, the data supporting the use of HC to guide therapy have significant limitations. A prospective randomized study comparing HC with usual clinical care as a therapeutic strategy to guide treatment has not been performed. It is possible that patients with HC are healthier and more diuretic responsive. Most of the data comes from studies of patients with HFrEF. Because HC has been variably defined, it is unclear what degree of change in hemoglobin, hematocrit or both is clinically significant [99]. Also, in some patients, the absence of hemoconcentration may not reflect residual volume overload but instead, occult blood loss, poor nutritional status, medications, or the effects of serial phlebotomy. Therefore, the data do not support the routine intensification of diuretic therapy in patients who do not demonstrate an increase in hemoglobin or hematocrit [105].

Invasive hemodynamic monitoring (IHM) with a pulmonary artery catheter (PAC) has been used in small non-randomized studies to guide diuretic and vasodilator therapy to achieve pre-specified near normal filling pressures in patients with severe refractory heart failure. This approach resulted in sustained improvements in hemodynamics, severity of mitral regurgitation, exercise tolerance and symptoms [45, 112, 113].

The Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization (ESCAPE) trial evaluated whether the routine use of a PAC was safe and improved clinical outcomes in patients with ADHF [114]. Four hundred forty-three patients with severe symptomatic heart failure at 26 sites were randomized to receive therapy guided by clinical assessment and PAC or clinical assessment alone. The target patient was sufficiently ill with advanced heart failure to make the use of the PAC reasonable, but also sufficiently stable to make crossover to PAC for urgent management unlikely. Patients who were felt to require a PAC for management of heart failure were not included in the study but were included in a PAC registry. The goal of treatment in both groups was resolution of clinical signs and symptoms of congestion, particularly jugular venous pressure elevation, edema, and orthopnea. Additional goals in the PAC group included achieving a pulmonary capillary wedge pressure of 15 mm Hg and a right atrial pressure of 8 mm Hg. Medications were not specified, but inotrope use was explicitly discouraged. The patient population was a particularly ill group as reflected by an average LVEF 19 %, urea nitrogen 35 mg/dL, creatinine1.5 mg/dL, RAP 14 mmHg, PCWP 25 mmHg, and cardiac index 1.9 l/min/m².

Treatment in both groups led to substantial reductions in symptoms, JVP and edema. The use of a PAC did not significantly affect the primary end-point of number of days alive and out of the hospital during the first six months. There was no difference between groups with respect to number of days hospitalized, mortality at 180 days, or in-hospital plus 30-day mortality. Exercise and quality of life end- points improved in both groups with a trend toward greater improvement in the PAC group. The Minnesota Living with Heart Failure questionnaire improved in both groups with greater improvement in the PAC group at 1 but not 6 months. The time trade-off tool showed great improvement for the PAC group compared with the clinical assessment group at all time points suggesting a greater improvement in quality of life. There was a consistent trend of greater improvement in clinical status in patients in whom therapy had been adjusted using PAC. There were no deaths related to PAC use. Adverse events specifically attributable to PACs occurred in 9/215 (4.2 %) patients and included: PAC-related infection (4 patients), bleeding (2 patients), catheter knotting (2 patients), pulmonary infarction/hemorrhage (2 patients) and ventricular tachycardia (1 patient).

The neutral findings from ESCAPE are consistent with recent meta-analysis data concerning the use of PACs in more diverse populations of patients in the critical care setting [115, 116]. These findings cannot be extrapolated to all patients in the critical care setting as patients enrolled in ESCAPE and other randomized studies were not felt to require invasive hemodynamic monitoring – that is, physicians had clinical equipoise about the need for a PAC.

Complications associated with the use of PACs include hematoma, pneumothorax, infection, pulmonary infarction, pulmonary hemorrhage, hemothorax, arterial puncture, catheter knotting, heart block in the setting of LBBB, and ventricular tachycardia. In the UK multicenter randomized Pulmonary Artery Catheters in Management of Patients in Intensive Care (PAC-Man) trial, the incidence of complications directly attributable to PAC insertion was significantly higher (10 %) than in the ESCAPE trail (4.2 %). However, the patients enrolled in the PAC-Man trial were significantly sicker than those in ESCAPE: 65 % had multi-system organ failure; only 11 % had decompensated heart failure; and the in-hospital mortality rate was 66 % in the control group [117].

The HFSA, ACCF/AHA, and ESC guidelines do not recommend the routine use of invasive hemodynamic monitoring in ADHF but do outline indications for PAC in carefully selected patients [8, 28, 118]. IHM should be considered in patients whose volume status and filling pressures are uncertain. In addition, placement of a PA catheter should be strongly considered in patients who are hypotensive, in cardiogenic shock, are unresponsive to initial medical therapy, or who develop clinically significant hypotension and/or significant worsening of renal function during initial treatment for ADHF. IHM will help identify patients who have hypotension because of low filling pressures, can guide the use of higher dose diuretic therapy and/or parenteral vasodilator therapy in patients with marginal blood pressure and persistent pulmonary congestion, and can indicate the need for inotropic therapy in patients with end-organ dysfunction, hypotension or worsening renal failure. IHM is indicated in patients with persistent severe symptoms despite adjustment of recommended therapies. In addition, invasive measurement of hemodynamics is indicated in patients being considered for heart transplantation to determine PA pressures, pulmonary vascular resistance, and transpulmonary gradient and in patients being considered for a long-term left ventricular assist device to better assess right ventricular performance [8, 119].

There is increasing evidence that elevated central venous pressure and elevated intra-abdominal pressure (IAP) contribute to diuretic resistance and worsening renal function in patients with ADHF. This has been appreciated since 1950 [124]. A retrospective review of 2557 patients who underwent right heart catheterization at a single academic medical center found that elevated CVP was the only hemodynamic parameter that was associated with low eGFR on multivariate analysis [125]. In another study, 51 patients with heart failure underwent pulmonary artery catheterization. GFR and renal blood flow were assessed by (125)I-Iothalamate and (131)I-Hippuran clearances, respectively. High right atrial pressure had little impact on GFR in patients with normal or high renal blood flow but was associated with a significant reduction in GFR in patients with low renal blood flow [126]. In a study of 145 consecutive patients admitted with ADHF treated with intensive medical therapy guided by hemodynamic monitoring using a pulmonary artery catheter, elevated CVP on admission (and after intensive medical therapy) was the only hemodynamic measure that was associated with worsening renal function defined as increase in serum creatinine of ≥0.3 mg/dL. There was an incremental risk of worsening renal function with increasing CVP; 75 % of patients with baseline CVP > 24 mmHg developed worsening renal function during hospitalization [127].

Mullens measured intraabdominal pressure (IAP) serially using a transvesicle technique [128, 129] in 40 patients admitted for treatment of ADHF. 60 % of patients had increased intraabdominal pressure (>7 mmHg) and 10 % had intra-abdominal hypertension (≥12 mmHg). Increased IAP was associated with worse renal function at baseline. Intensive medical therapy resulted in improved hemodynamics and lower IAP. There was a strong correlation noted between reduction in IAP and improvement in renal function in patients with elevated IAP at baseline [130]. In a small group of patients hospitalized for ADHF who had increased IAP and ascites who developed progressive elevation in serum creatinine in response to intravenous loop diuretics, mechanical fluid removal by paracentesis decreased IAP and improved renal function [131].

These data suggest that high right atrial pressure and elevated intraabdominal pressure contribute to renal dysfunction and diuretic resistance in patients with ADHF. Elevated IAP may compress the renal veins, the ureters or the renal parenchyma and probably impacts intra-glomerular hemodynamics. Glomerular filtration pressure can be calculated by subtracting proximal tubular pressure (which can be estimated by IAP) from mean blood pressure. The renal filtration gradient, which strongly correlates with GFR, is calculated by subtracting proximal tubular pressure from glomerular filtration pressure (mean blood pressure – 2 IAP). As IAP increases, especially in patients with relatively low mean arterial pressure, GFR significantly decreases [130].

Combined disorders of cardiac and renal function are classified as cardiorenal syndromes. CRS Type 1 is defined by the development of acute kidney injury occurring in the setting of an acute cardiac illness, most commonly ADHF. It occurs in approximately 25–30 % of patients hospitalized for ADHF. The most commonly used criteria to identify patients with CRS Type 1 are an increase in serum creatinine >0.3 mg/dL or a reduction in eGFR of 20–25 %. Risk factors for developing CRS include prior heart failure, male gender, diabetes, admission creatinine ≥1.5 mg/dL, and hypertension [132, 133]. The clinical syndrome is characterized by a rise in serum creatinine (generally within the first 3 days of hospitalization), oliguria and diuretic resistance with or without worsening HF symptoms. While alterations in renal hemodynamics including renal hypoperfusion, low systemic blood pressure and elevated venous congestion play a role in diuretic resistance and renal insufficiency in ADHF, changes in renal hemodynamics alone do not account for this syndrome. In an analysis from the ESCAPE trial, no correlation was found between any baseline hemodynamic parameter or change in hemodynamic parameter and the development of worsening renal function during heart failure hospitalization [134]. In an analysis of hemodynamic data from patients enrolled in the Vasodilation in Management of Acute Congestive Heart Failure (VMAC) trial who had undergone right heart catheterization, no correlation was found between worsening renal function and baseline right atrial pressure (RAP) or change in RAP. Smaller net fluid loss in the first 24 h was strongly associated with an increased risk of developing worsening renal function [135]. Emerging evidence suggests that the pathophysiology of CRS Type I is multifactorial, involves complex heart and renal crosstalk and is mediated by hemodynamic changes, neurohormonal activation, inflammation, immune cell signaling, hypothalamic-pituitary stress reaction, systemic endotoxin exposure from the abdominal viscera, and oxidative stress resulting in bidirectional organ injury. This pathophysiology is incompletely understood and definitive approaches to treatment have not been established [32, 122, 136–138].

The addition of an oral or intravenous thiazide diuretic has been shown to be helpful in patients with diuretic resistance. Thiazide diuretics are weak natriuretic agents that block only 5–10 % of filtered sodium when used as monotherapy. They are ineffective as single agents in the treatment of patients with moderate to severe heart failure. However, in the setting of LD therapy, sodium delivery to the distal tubule increases significantly. Hypertrophy and hyperfunction of the distal nephron contribute to diuretic resistance. Blocking sodium uptake using TD has been shown to increase diuresis significantly. The combination of diuretics that act at different sites in the nephron has been termed “sequential nephron blockade”. Generally, it is appropriate to consider adding a TD when adequate diuresis has not been achieved despite an intravenous dose of furosemide of 180–360 mg daily (or other LD equivalent). There is little evidence to suggest that one TD is superior to another when used in combination with an intravenous LD. It has been suggested that metolazone is superior to other TDs in patients with renal insufficiency but other TDs have been shown to be effective in this patient population. Traditionally, TDs have been given 30 min before LD administration. However, this dosing regimen has not been studied – most studies of CDT looked at the effect of giving an LD and TD at the same time [20, 83, 141, 142].

The most serious complication of CDT is hypokalemia which can be profound. Hyponatremia is common. Hypomagnesemia can occur and may make hypokalemia worse. Occasionally, CDT results in a marked diuresis which may result in hypotension. Electrolytes, renal function, urine output and vital signs need to be followed closely.