Use of diuretics in patients with heart failure

INTRODUCTION — Evaluation and optimization of volume status is an essential component of treatment in patients with systolic or diastolic heart failure (HF) [1]. Removal of excess extracellular fluid with diuretics to treat peripheral and/or pulmonary edema is one of the mainstays of volume management. In contrast to other HF therapies, such as angiotensin inhibitors, beta blockers, and aldosterone antagonists, limited outcome data are available for diuretic therapy. However, for the majority of patients with HF, diuretics are essential for the control of volume status.

Issues related to the use of diuretics, as well as sodium and fluid restriction, in patients with HF will be reviewed here. The evaluation, including assessment of volume status, and overall management of HF are discussed separately. (See "Heart failure: Clinical manifestations and diagnosis in adults" and "Approach to diagnosis and evaluation of acute decompensated heart failure in adults" and "Treatment and prognosis of heart failure with preserved ejection fraction" and "Treatment of acute decompensated heart failure: General considerations".)

EFFICACY AND SAFETY — Diuretics are the cornerstone of treatment of volume overload in patients with HF. However, few clinical trials have studied the impact of diuretic therapy on clinical outcomes.

A meta-analysis of diuretic treatment in chronic HF demonstrated a beneficial effect on clinical outcomes, although only small trials were available [2]. Three trials with a total of 202 patients with chronic HF found a reduction in mortality with diuretic use compared with placebo (odds ratio [OR] 0.24, 95% CI 0.07-0.83). Admission for worsening HF was reduced by diuretics compared with placebo in two trials with a total of 169 patients (OR 0.07, 95% CI 0.01-0.52). These results are consistent with data from observational studies as well as diuretic trials demonstrating rapid improvement in dyspnea in patients treated largely with diuretic therapy [3,4].

However, observational data have raised safety concerns about diuretic treatment of HF. Several uncontrolled studies have found that higher diuretic doses are associated with worse outcomes, including mortality [4]. Use of nonpotassium sparing diuretics to treat HF has been associated with arrhythmic death [5]. However, these observations are confounded by the indications for high dose diuretic use (ie, the need for higher doses of loop diuretics is likely a marker of more severe HF with increased mortality risk).

Blood urea nitrogen and serum creatinine levels often rise during diuretic treatment of HF, and careful monitoring of renal function and electrolytes is recommended. These issues are discussed below. (See 'Effects on renal function' below.)

GENERAL PRINCIPLES — The major issues that should be considered about diuretic therapy in patients with HF include salt and water restriction, the choice and dose of diuretic, hemodynamic consequences of diuresis, improved survival with aldosterone antagonism, and variable effects on systemic vascular resistance.

Sodium and fluid restriction — The approach to sodium and fluid restriction and other aspects of self-management in patients with HF are discussed separately. (See "Heart failure self-management".)

Diuretic use — Most patients with HF and volume overload are initially treated with the combination of an oral loop diuretic (such as furosemide, torsemide, or bumetanide) and a low sodium diet. The three major manifestations of volume overload in HF are pulmonary congestion, peripheral edema, and elevated jugular venous pressure (see "Heart failure: Clinical manifestations and diagnosis in adults", section on 'Volume assessment'). Intravenous administration of loop diuretics (either as a bolus or a continuous infusion) is generally required for acute decompensation in order to achieve rapid diuresis or for severe disease, in which case there may be diuretic resistance. (See 'Refractory congestion' below.)

Loop diuretics are often required for fluid control and to relieve symptoms, but the risk of arrhythmic death may be increased if hypokalemia occurs [5]. However, the data supporting this observation may be confounded by the fact that the need for loop diuretics may be correlated with more severe HF and increased mortality risk.

Aldosterone antagonism with spironolactone or eplerenone can be added to loop diuretics to modestly enhance diuresis and minimize potassium loss. More importantly, these drugs improve survival in patients with systolic HF. The benefits of aldosterone antagonists in patients with HF are thought to be relatively independent of their diuretic effect and may be due to a beneficial influence on cardiovascular remodeling and other actions. Eplerenone is a more specific aldosterone antagonist that is associated with fewer side effects but is much more expensive than generic spironolactone.

Nonsteroidal antiinflammatory agents are among the drugs that should be avoided in patients with HF, in part because they interfere with the response to diuretics. (See "Drugs that should be avoided or used with caution in patients with heart failure", section on 'Nonsteroidal anti-inflammatory drugs'.)

Use of loop diuretics

Choice of loop diuretic — In patients with HF who require diuretic therapy, furosemide, torsemide, or bumetanide can be used as the initial therapy; all three are generally effective and well tolerated. We typically start oral diuretic therapy with furosemide. If the patient cannot be effectively diuresed with high-dose furosemide (eg, total daily dose is ≥200 mg), we switch from furosemide to torsemide or bumetanide.

This approach is based on the following: Furosemide is effective in most patients, its dosing is familiar to most health care providers, and it can be readily converted between its intravenous and oral forms if necessary. In patients who are refractory to furosemide, the pharmacologic properties of the loop diuretics favor the use of torsemide or bumetanide:

●Furosemide has a bioavailability of only about 50 percent and has substantial interpatient and intrapatient variability (range 10 to 90 percent) [6,7], whereas the bioavailability of torsemide and bumetanide is 80 to 100 percent. As a result, patients who are refractory to furosemide may have an improved diuretic response (ie, greater, more predictable) to torsemide or bumetanide [8-10].

●Torsemide has a longer half-life than furosemide and bumetanide, which may contribute to a more sustained diuretic response that reduces the need for a second daily diuretic dose.

One large trial that included comprehensive medical therapy for HF with reduced ejection fraction showed no advantage of torsemide when compared with furosemide, while the remainder of trials were smaller and older:

●In a trial that included 2869 patients who were randomly assigned to an equivalent dose of torsemide or furosemide prior to hospital discharge, both groups had similar mortality rates (26 percent) and rehospitalization rates (38 versus 40 percent) [11]. The trial was limited by its open-label treatment design and by immediate crossover from torsemide to furosemide (7 percent) and furosemide to torsemide (4 percent).

●A network meta-analysis aggregated data from trials that reported the relative efficacy of loop diuretics (ie, furosemide, torsemide, bumetanide, azosemide) on all-cause mortality, cardiovascular mortality, HF-related hospitalization, hypokalemia, and acute renal failure [12]. The study found no significant differences between the loop diuretics for the risk of all-cause mortality, cardiovascular mortality, or hypokalemia, while torsemide had the lowest risk of HF hospitalization and a trend towards less acute renal failure. Several features of this study limit the generalizability of its findings: most of the trials included in the meta-analysis were conducted prior to the year 2000, the trials had diverse methodologies (eg, diuretic dose, follow-up time, endpoint definitions), and torsemide was only superior to other diuretics for two of five outcomes examined.

●In one of the larger trials included in the network meta-analysis, 234 patients with chronic HF were treated with open-label furosemide (average daily dose 136 mg/day) or torsemide (average daily dose 72 mg/day) [8]. The primary endpoint of readmission for HF at one year was significantly lower with torsemide (17 versus 32 percent), but the rate of all-cause hospital admission was similar between the two groups (76 versus 71 percent). The limitations of this trial were its small size and the use of open-label treatments.

●Only three studies compared furosemide with bumetanide [10].

Loop diuretic dosing — Issues related to appropriate dosing of loop diuretics in patients with HF and the treatment of refractory edema are discussed in detail elsewhere. (See "Loop diuretics: Dosing and major side effects" and "Causes and treatment of refractory edema in adults".)

Summarized briefly:

●The diuretic response to loop diuretics is described by a threshold type dose-response curve linking the rate of diuretic excretion in the urine (reflecting the concentration of diuretic in the renal tubular lumen) and the degree of natriuresis [13]. There is virtually no increase in natriuresis until the threshold concentration of luminal diuretic is reached and, once the threshold is exceeded, the degree of natriuresis increases until a maximal ceiling is reached. Patients with HF have a lesser response to a given dose of diuretic than normal subjects for two reasons: decreased diuretic delivery to the kidney (and therefore less transport of diuretic into the tubular lumen) because renal blood flow is reduced; and increased sodium reabsorption at other sites due to hypoperfusion-induced activation of the renin-angiotensin-aldosterone and sympathetic nervous systems. Intestinal absorption of an oral loop diuretic may be delayed in patients with edema [13]. Slower absorption will generate a lower peak concentration despite total absorbed dose similar to that in healthy individuals [14], and this will reduce the diuretic effect if the threshold concentration is not achieved.

●An effective diuretic dose must be determined. The usual starting dose is 20 to 40 mg of furosemide or its equivalent. Subsequent dosing is determined by the diuretic response. Thus, if a patient does not respond to 20 mg of furosemide, the dose should be increased to 40 mg rather than giving the same dose twice a day. Maximum single doses are discussed below. (See 'Treatment of ADHF' below and 'Chronic therapy' below.)

●If the patient is already receiving the maximum loop diuretic dose and has partial but inadequate diuresis, the loop diuretic can be given twice or even three times a day.

Treatment of ADHF — Patients with acute decompensated HF (ADHF) and evidence of fluid overload should be promptly treated with intravenous diuretics as part of the initial therapeutic regimen. The expeditious initiation of an effective diuretic regimen is important in controlling dyspnea and other symptoms related to fluid overload and, in addition, may improve hospital outcomes. (See "Treatment of acute decompensated heart failure: Specific therapies", section on 'Diuretics'.)

Intravenous diuretics (either as a bolus or a continuous infusion) are more potent than equivalent oral doses, and elicit an earlier diuretic response, especially in the presence of significant interstitial edema in the gastrointestinal tract. The overall treatment of ADHF is discussed in detail elsewhere. (See "Treatment of acute decompensated heart failure: General considerations" and "Treatment of acute decompensated heart failure: Specific therapies".)

With intravenous therapy in patients with ADHF, the onset of diuresis is within 30 minutes, with peak diuresis usually at one to two hours. Most patients require continued diuretic therapy after the initial dose. This can be achieved with the administration of two or more doses per day or with a continuous intravenous infusion.

We switch from an effective intravenous dose to an oral regimen once the patient’s acute symptoms have been stabilized to help ensure that an effective outpatient dose is identified and prescribed.

The approach to diuretic therapy and concomitant monitoring in patients with ADHF and fluid overload is discussed further separately. (See "Treatment of acute decompensated heart failure: Specific therapies", section on 'Diuretic administration'.)

No single intravenous dosing regimen (bolus versus continuous infusion; high dose versus lower dose) has been shown to be superior to others, as illustrated by the following observations:

●A meta-analysis from the National Institute for Health and Clinical Excellence (NICE) included 10 randomized controlled trials comparing various intravenous furosemide regimens (mainly bolus versus continuous infusion) [15]. Comparisons between furosemide intravenous bolus and continuous infusion revealed no clear differences in outcomes such as weight loss, urine output, or change in renal function; ototoxicity was not assessed. An earlier meta-analysis found that tinnitus and hearing loss were less frequent with continuous infusion therapy, but these side effects were transient and infrequent [16].

●The best data come from the DOSE trial, which was included in the above NICE meta-analysis [3]. The trial randomly assigned 308 patients to receive furosemide administered intravenously via either a bolus every 12 hours or continuous infusion and at either a low dose (equivalent to the patient’s previous oral dose) or a high dose (2.5 times the previous oral dose). The efficacy endpoint was the global assessment of symptoms over the course of 72 hours, and the safety endpoint was change in serum creatinine level from baseline to 72 hours. Worsening renal function was defined as an increase in the serum creatinine level >0.3 mg/dL (>26.5 micromol/L) at any time during the 72 hours after randomization.

The following findings were noted:

•There was no significant difference in efficacy or safety endpoints for bolus versus continuous infusion. Patients assigned to intravenous bolus therapy were more likely to require a dose increase at 48 hours; however, the total dose of furosemide over 72 hours in the bolus group was not significantly different from that in the continuous infusion group (592 versus 480 mg, p = 0.06).

•High-dose furosemide, compared with low-dose furosemide, produced greater net fluid loss, weight loss, and relief from dyspnea, but also more frequent transient worsening of renal function (23 versus 14 percent). There was no significant difference in global assessment of symptoms in the high-dose group (p = 0.06); the mean change in the serum creatinine level was less than 0.1 mg/dL (9 micromol/L) in both groups.

In summary, the available data suggest that intravenous continuous infusion and bolus loop diuretic therapy have similar efficacy in patients with ADHF. There have been no trials comparing the efficacy of high-dose oral furosemide with intravenous regimens.

Chronic therapy — Patients with fluid overload and chronic HF (and those who have been stabilized following acute decompensation) are generally treated with oral loop diuretics.

●For patients who have not received prior loop diuretic therapy:

•The usual starting oral dose is 20 to 40 mg of furosemide or its equivalent once or twice a day [1]. If a patient does not respond to initial dosing, the dose should be increased rather than giving the same dose twice a day. For patients with normal glomerular filtration rate (GFR), maximum single oral doses of furosemide are 40 to 80 mg. In patients with renal insufficiency, a higher maximum dose of 160 to 200 mg of furosemide can be given (maximum daily dose of 600 mg).

•The usual initial oral dose of torsemide is 10 to 20 mg with maximum individual dose of 100 mg (maximum daily dose of 200 mg) [1].

•The usual initial oral dose of bumetanide is 0.5 to 1.0 mg with maximum individual dose of 5 mg (maximum daily dose of 10 mg) [1].

●Dosing in patients who have received prior loop diuretic therapy is based upon the response to prior therapy. The oral dose of furosemide is approximately twice the intravenous dose, but the bioavailability of an oral dose varies widely (mean approximately 50 percent, range 10 to 90 percent) [17] and other factors such as rate of absorption impact efficacy [18], so individualized titration of dose is required.

●Once the patient is clinically euvolemic (eg, has reached a "dry" body weight), the dose of diuretic should be adjusted to the minimum dose required to maintain euvolemia.

Intravenous therapy is warranted in patients who do not respond to maximum oral therapy. The regimen is the same as that described above for acute decompensated HF. (See 'Treatment of ADHF' above.)

Effects on renal function — Blood urea nitrogen (BUN) and serum creatinine levels often rise during diuretic treatment of HF, and careful monitoring is recommended. In the absence of other causes for an elevated BUN level, a disproportionate rise in BUN level relative to serum creatinine (BUN/serum creatinine ratio >20:1) suggests a prerenal state with increased passive reabsorption of urea. An initial rise in BUN level may be accompanied by a stable serum creatinine level, reflecting preserved GFR. Further elevations in BUN level along with a rise in serum creatinine level are likely if diuresis is continued in such patients. (See "Etiology and diagnosis of prerenal disease and acute tubular necrosis in acute kidney injury in adults", section on 'Blood urea nitrogen/serum creatinine ratio'.)

An otherwise unexplained rise in serum creatinine level, which reflects a reduction in GFR, may be a marker of reduced perfusion to the kidney and other organs. Patients in whom this occurs before euvolemic status is achieved have a worse prognosis. Nevertheless, fluid removal may still be required to treat signs and symptoms of congestion, particularly pulmonary edema. On the other hand, a stable serum creatinine suggests that perfusion to the kidneys (and therefore to other organs) is being well maintained and that diuresis can be continued if the patient is still edematous. (See "Cardiorenal syndrome: Definition, prevalence, diagnosis, and pathophysiology", section on 'Reduced renal perfusion' and "Cardiorenal syndrome: Prognosis and treatment", section on 'Change in glomerular filtration rate during therapy for heart failure'.)

Changes in cardiac output and the consequent changes in renal perfusion are not the only determinant of changes in GFR in patients with HF. Among patients with an elevated central venous pressure, the associated increase in renal venous pressure can reduce the GFR, while lowering venous pressure with diuretics and other therapies might therefore increase the GFR. (See "Cardiorenal syndrome: Definition, prevalence, diagnosis, and pathophysiology".)

Guidelines for management of patients with HF with elevated or rising BUN and/or serum creatinine levels include the following [19]:

●Other potential causes of kidney injury (eg, use of nephrotoxic medications, urinary obstruction) should be evaluated and addressed.

●Patients with severe symptoms or signs of congestion, particularly pulmonary edema, require continued fluid removal independent of changes in GFR. In the presence of elevated central venous pressure, renal function may improve with diuresis. (See "Cardiorenal syndrome: Definition, prevalence, diagnosis, and pathophysiology", section on 'Increased renal venous pressure'.)

●If the BUN level rises and the serum creatinine level is stable or increases minimally, and the patient is still fluid overloaded, diuresis can be continued to achieve the goal of eliminating clinical evidence of fluid retention with careful monitoring of renal function. (See 'Goals of therapy' below.)

●If increases in serum creatinine level appear to reflect intravascular volume depletion, then reduction in or temporary discontinuation of diuretic and/or angiotensin converting enzyme (ACE) inhibitor/angiotensin II receptor blocker (ARB) therapy should be considered. Adjunctive inotropic therapy may be required. (See "Treatment of acute decompensated heart failure: Specific therapies", section on 'Inotropic agents'.)

If substantial congestion persists and adequate diuresis cannot be achieved, then ultrafiltration or dialysis should be considered. (See "Kidney replacement therapy (dialysis) in acute kidney injury in adults: Indications, timing, and dialysis dose", section on 'Urgent indications' and "Management of refractory heart failure with reduced ejection fraction", section on 'Ultrafiltration'.)

Goals of therapy — The goal of diuretic therapy is to reduce the signs and symptoms of volume overload. This goal should be pursued while adverse effects are monitored (see "Treatment of acute decompensated heart failure: Specific therapies", section on 'Monitoring'):

●Electrolyte imbalances (particularly hypokalemia, hypomagnesemia, and metabolic alkalosis) that develop during diuresis should be promptly treated while diuresis is continued.

●If hypotension or worsening renal function develops before the goals of treatment are achieved, diuresis may be slowed. Diuresis should be maintained until fluid retention is eliminated even if this results in asymptomatic mild to moderate decreases in blood pressure or renal function. Excessive concern about hypotension and azotemia can lead to underutilization of diuretics and persistent volume overload. Persistent volume overload contributes to continued symptoms, may reduce the efficacy of drug therapy for HF, and, as suggested in the following study, may be associated with increased mortality.

The potential importance of aggressive fluid removal was suggested by an observational study of 336 patients hospitalized with HF [20]. Patients developing significant hemoconcentration (defined as elevations in at least two of the following three parameters: hematocrit, serum albumin, and serum total protein) during treatment received higher doses of loop diuretics, lost more weight, and had greater reductions in intracardiac filling pressures. Although hemoconcentration was strongly associated with worsening renal function (odds ratio 5.3), it was also associated with a large and significant reduction in 180-day mortality (adjusted hazard ratio 0.16, 95% CI 0.02-0.44).

Time course of diuresis and associated complications — A potential fall in cardiac output induced by loop diuretic therapy (see 'Filling pressures and cardiac output' below) has important implications for the time course of diuresis. In both normal individuals and those with HF, the ensuing activation of sodium-retaining mechanisms (angiotensin II, aldosterone, and norepinephrine) as well as flow-dependent hypertrophy in the distal tubule limits the response to continued diuretic therapy [21-23].

In stable patients with HF, negative balances for sodium and potassium are greatest during the first few days of loop diuretic therapy (with most sodium losses typically occurring after the first dose). The time required to achieve a new steady state (when sodium and potassium excretion equals intake) depends on the amount of fluid overload at the start of therapy, but it averages two weeks (figure 1). Once a new steady state is reached, continued diuretic therapy will maintain the fluid loss that has been attained but will not induce further fluid or potassium loss unless the dose is increased, a second type of diuretic is added, or there is a change in HF severity or sodium intake. Thus, monitoring of serum electrolyte levels is warranted if the diuretic regimen is augmented or changes in HF severity or sodium intake are suspected. (See "Time course of loop and thiazide diuretic-induced electrolyte complications".)

The risk of hypokalemia is reduced by concurrent treatment with an aldosterone antagonist. When used for diuresis or potassium sparing effects, higher doses of spironolactone may be needed.

An aldosterone antagonist (spironolactone or eplerenone) is preferred to a sodium channel blocker (amiloride or triamterene) because of evidence that aldosterone antagonist therapy improves patient outcomes and high levels of aldosterone contribute to cardiovascular disease in patients with HF. In addition, hyperaldosteronism may contribute to diuretic resistance, an effect that can be minimized by aldosterone antagonist therapy [24].

Hyperkalemia is a potential complication of aldosterone antagonist therapy, particularly in patients who are also treated with an angiotensin inhibitor (ACE inhibitor, ARB, or angiotensin receptor-neprilysin inhibitor), who have impaired renal function, or who develop worsening cardiac function.

Duration of therapy — Once begun, diuretic therapy is generally continued indefinitely for fluid control unless cardiac function improves. However, the diuretic dose should be reconsidered after the patient is placed on other HF medications. Daily assessment of patient weight may be the most effective method for documenting effective diuresis [1]. For accurate comparisons, daily measurements should use the same scale and should be performed at the same time each day, usually in the morning, prior to eating, and after voiding. (See "Heart failure self-management".)

Long-term therapy can be facilitated by having patients record their weight each day and allowing them to make prespecified changes in dose if the weight increases or decreases beyond a specified range. Increases in body weight of more than 0.9 kg have been associated with an increased risk of HF hospitalization; these weight gains typically began at least one week before admission [25,26]. During long-term follow-up, a patient’s target weight may change over time as dry body mass changes with fluctuations in nutritional status.

There is a subpopulation of stable patients with less severe HF who can be withdrawn from diuretics. This issue was directly addressed in a 12-week clinical trial of 41 patients with stable HF who had been treated with diuretics but not an ACE inhibitor [27]. Diuretics were discontinued and the patients randomly assigned to lisinopril or placebo. The probability of remaining off diuretics at six weeks was 71 percent in patients who had no history of hypertension, had a left ventricular ejection fraction above 27 percent, and had been controlled on a furosemide dose (or its equivalent) of 40 mg/day or less.

HEMODYNAMIC EFFECTS

Filling pressures and cardiac output — Diuretic-induced fluid losses decrease intravascular volume and capillary pressure, permitting mobilization of edema fluid from the interstitium. Refilling of the vasculature with interstitial fluid derived from all capillary beds takes place with a half-life of approximately 20 minutes. The net effect is that systemic hemodynamics are maintained unless the rate of diuresis is extremely rapid.

Despite often marked symptomatic improvement and increased exercise tolerance, the decrease in intracardiac filling pressure induced by diuresis can, by reducing ventricular preload, lower the cardiac output (by as much as 20 percent) [28] and increase neuroendocrine activation (eg, higher plasma levels of renin and norepinephrine) [29,30]. However, some patients initially have little or no reduction in cardiac output because they are on the flat part of the Frank-Starling curve in which changes in left ventricular end-diastolic pressure have little or no effect on cardiac performance (figure 2) [31].

An otherwise unexplained rise in serum creatinine, which reflects a reduction in glomerular filtration rate (GFR), is a marker of reduced tissue perfusion.

A rise in serum creatinine level during therapy for HF may occur before euvolemic status is achieved; such patients have a worse prognosis. A rise in creatinine level may also reflect a rate of diuresis that exceeds the rate at which interstitial fluid can be mobilized, in which case a slower rate of diuresis may be tolerated without a rise in creatinine level. On the other hand, a stable serum creatinine level suggests that perfusion to the kidneys (and therefore to other organs) is being well maintained and that diuresis can be continued if the patient is still edematous. (See "Cardiorenal syndrome: Definition, prevalence, diagnosis, and pathophysiology", section on 'Reduced renal perfusion' and "Cardiorenal syndrome: Prognosis and treatment", section on 'Change in glomerular filtration rate during therapy for heart failure'.)

However, changes in renal perfusion are not the only determinant of changes in GFR in patients with advanced HF. Severe congestion of splanchnic veins results in increased intraabdominal pressure in 60 percent of patients admitted with advanced HF. The associated increase in renal venous pressure can reduce the GFR. Lowering initially elevated intraabdominal pressure with diuretics and other therapies has been associated with an increase in GFR, a relationship not seen with other hemodynamic variables (figure 3) [32]. (See "Cardiorenal syndrome: Definition, prevalence, diagnosis, and pathophysiology", section on 'Increased renal venous pressure'.)

Venodilatory effect in acute pulmonary edema — When administered to patients with pulmonary edema due to an acute myocardial infarction, intravenous furosemide often induces transient venodilation similar to morphine; the net effect is a fall in cardiac filling pressures and decreased pulmonary congestion prior to the onset of diuresis [33]. Studies in animals and humans have shown that loop diuretics increase renal and perhaps local production of vasodilator prostaglandins [34-37], and that the venodilator response can be prevented by nephrectomy or the administration of aspirin or other nonsteroidal antiinflammatory drugs [34-38].

Vasoconstriction in chronic HF — A deleterious acute hemodynamic response may be seen when a loop diuretic is given intravenously to patients with advanced chronic HF [39]. In this setting, there may be an acute increase in plasma renin and norepinephrine levels, leading to arteriolar vasoconstriction and a rise in systemic blood pressure; this increase in afterload then induces a reduction in cardiac output and a transient rise in pulmonary capillary wedge pressure that can last for an hour, with possible worsening of dyspnea. Fortunately, by four hours, pulmonary capillary wedge pressures fall below prediuretic values due both to diuresis and a decrease in vasoconstrictors [39]. A similar initial vasoconstrictor response has been demonstrated in patients with cirrhosis [40].

It is uncertain how to reconcile these different hemodynamic effects of diuretic therapy. One possibility is that patients with chronic HF (or cirrhosis) have persistent elevations in renin and angiotensin II release (induced by decreased renal perfusion) that are associated with hyperplasia of the renin secretory apparatus. As a result, administration of a loop diuretic may lead to an exaggerated rise in renin release and systemic vasoconstriction. By comparison, the vasodilator prostaglandin response may predominate in patients with an acute myocardial infarction who do not have chronic hypersecretion of renin.

REFRACTORY CONGESTION

Causes — Patients with HF who are not responsive to moderate oral doses of a loop diuretic (eg, furosemide 60 to 80 mg daily or its equivalent) should be evaluated for potential causes of diuretic refractoriness, including:

●The presence of advanced HF (eg, low cardiac output) or advanced kidney disease. (See "Clinical manifestations and diagnosis of advanced heart failure" and "Clinical manifestations and diagnosis of advanced heart failure", section on 'Clinical manifestations'.)

●Hypoalbuminemia, high sodium or fluid intake. (See "Causes and treatment of refractory edema in adults", section on 'Causes of refractory edema'.)

●Possible hypovolemia.

●Lack of adherence to the diuretic regimen.

●Inability to absorb oral diuretics.

Management options — In patients with an inadequate response to diuretics, possible interventions include:

●Change the type of diuretic – Oral furosemide is associated with substantial interpatient and intrapatient bioavailability. Thus, patients with an inadequate diuretic response to oral furosemide at the maximum recommended dose may respond to higher doses of furosemide or to replacement of furosemide with either oral torsemide or bumetanide, which are more predictably absorbed. (See 'Choice of loop diuretic' above.)

●Switch to intravenous diuresis – Patients who fail oral therapy may respond to more reliable diuretic delivery via intravenous administration by either intermittent bolus therapy or continuous intravenous infusion. Both methods of delivery appear to have similar efficacy and safety in patients with HF. The efficacy of a continuous intravenous infusion has not been evaluated in patients with refractory edema. (See 'Treatment of ADHF' above.)

●Add another diuretic to the regimen – Use of a loop diuretic with another diuretic agent ("combination therapy") augments diuresis by blocking sodium reabsorption at additional sites in the nephron (figure 4). In patients with signs and symptoms of congestion who are refractory to a total daily oral dose of furosemide of at least 200 mg (or its equivalent) or an intravenous dose of furosemide of at least 200 mg daily (or its equivalent), our contributors typically attempt combination diuretic therapy in patients who can be adequately monitored (see below). We suggest adding metolazone to the existing loop diuretic regimen rather than other agents (table 1). However, other thiazide diuretics, mineralocorticoid receptor antagonists (MRA), or carbonic anhydrase inhibitors are reasonable choices for combination therapy.

In patients with a high suspicion for decreased absorption of oral diuretics, use of an intravenous rather than oral agent to augment diuresis is reasonable. In patients with persistent hypokalemia, combination therapy with a potassium-sparing diuretic (eg, amiloride, MRA) may be preferable to combination therapy with other agents.

•Requirements and dosing – The requirements of combination therapy include daily monitoring of the diuretic response and electrolytes, time-limited dosing, and supervision by a provider experienced with this approach to diuresis. Due to these requirements, our experts rarely use combination therapy in outpatients for acute or chronic management of volume status.

•Dosing – The initial and maximum doses of diuretics used in combination therapy are listed in the table (table 1).

Some contributors to this topic typically administer a thiazide diuretic (or other agent) 30 to 60 minutes prior to administration of a bolus dose of loop diuretic to maximize the effect of combination therapy. Other contributors do not time the administration of thiazide diuretics with the administration of the loop diuretic due to the long half-lives of these agents, which lead to sustained sodium blockade. Our contributors agree that timing of the first dose is most important and that, in patients on a continuous loop diuretic infusion, a thiazide or other agent can be added at any time during the infusion.

In patients who respond to combination therapy, subsequent doses of a nonloop diuretic are typically given daily until adequate volume reduction has been achieved or the patient has an adequate diuretic response to monotherapy with a loop diuretic.

If after six to eight hours following the initial dose of combination diuretic therapy the rate of diuresis has not sufficiently increased, we typically double the dose of the nonloop diuretic or use an intravenous nonloop diuretic to augment diuresis. If the diuretic response remains inadequate, management is individualized and is based on the severity of congestion.

•Rationale and evidence – The use of combination diuretic therapy is based on our experience. There are no high-quality trials to suggest the superiority of one agent over another or that established the safety of combination therapy in patients who are truly refractory to loop diuretics. Among the options for combination therapy, hydrochlorothiazide and acetazolamide are the only agents that have undergone evaluation in trials. However, these trials have methodologic issues that limit their application to patients with HF refractory to loop diuretics:

-In a trial that included 320 patients with acutely decompensated HF who were treated with a loop diuretic, patients randomly assigned to receive hydrochlorothiazide or placebo had similar changes in dyspnea after 72 hours of treatment (mean area under the curve of visual analog scale dyspnea rating 960 versus 720; p-value 0.50), but patients assigned to hydrochlorothiazide had more weight loss (-2.3 kg versus -1.5 kg in the placebo group; adjusted 95% CI -1.8 to -0.42 kg) [41]. Worsening kidney function and hypokalemia were more common in the hydrochlorothiazide group.

The trial did not report the diuretic dose at the time of enrollment and was stopped early due to slow enrollment.

-In a randomized controlled trial of 519 patients hospitalized with acute decompensated HF, treatment with an intravenous loop diuretic alone (at twice the outpatient oral maintenance dose) was compared with treatment with a loop diuretic combined with 500 mg intravenous acetazolamide daily [42]. Patients receiving acetazolamide had higher cumulative urine output and natriuresis, a greater incidence of successful decongestion (42.2 versus 30.5 percent), and a shorter length of stay (geometric mean 9 versus 10 days; adjusted difference 0.9 days, 95% CI 0.81-0.98). There were similar rates of death or readmission after three months (30 versus 28 percent in the control group; hazard ratio 1.1, 95% CI 0.8-1.5). The overall results were similar for patients with HF with reduced ejection fraction and HF with preserved ejection fraction, and the rate of adverse kidney events was similar between the groups.

The trial had several limitations including its sample size, use of time to decongestion as the primary outcome, and inability of the investigators to change diuretic doses in the control group. In addition, the patients studied were not truly diuretic resistant; the inclusion criteria required an outpatient diuretic dose of only 40 mg of furosemide or the equivalent dose bumetanide or torsemide.

●Addition of hypertonic saline – In diuretic-resistant patients, administration of hypertonic saline is an alternative to hemodialysis or hemofiltration that should only be used after the interventions above have been attempted and only by providers experienced with the management of patients with HF and congestion refractory to diuretics. In patients with HF who are diuretic resistant, the use of hypertonic saline could increase total body sodium content and, accordingly, total body water.

The rationale for using hypertonic saline solution includes an osmotic effect that may promote refilling of the intravascular compartment and a possible increase in renal blood flow that might promote diuretic action. One proposed mechanism of action is increased tumor necrosis factor-alpha production, which may inhibit Na-K-2Cl cotransporter activity and angiotensinogen expression [43].

High-quality trials of hypertonic saline are lacking, and, therefore, the safety and effectiveness of this approach are uncertain. Several studies have suggested greater diuretic efficiency and improved outcomes from combined intravenous hypertonic saline solution plus intravenous furosemide compared with intravenous furosemide alone in treating acute decompensated HF [44,45]:

•A meta-analysis included nine randomized controlled trials that compared intravenous hypertonic saline solution plus intravenous furosemide with intravenous furosemide alone [46]. All-cause mortality was reported in five trials and was lower with combined hypertonic saline solution plus furosemide compared with furosemide alone (risk ratio 0.57, 95% CI 0.44-0.74). However, there was substantial heterogeneity among the studies and no significant benefit remained if two of the component trials were excluded [47,48].

•A meta-analysis that included four observational studies and ten randomized trials found that the combination of hypertonic saline solution plus furosemide was associated with more weight loss (mean difference [MD] 0.99 kg), reduced hospital length of stay (MD -2.72 days), readmissions (relative risk [RR] 0.63), and mortality (RR 0.55) [44]. However, the findings of this analysis are limited by significant risk of bias (eg, selection bias among the studies, absence of blinding in the trials, selective outcome reporting).

●Hemodialysis and ultrafiltration – In patients who cannot be managed with diuretics and appropriate use of inotropic agents, vasopressors, and mechanical support, hemodialysis may be indicated. Issues related to the use of hemodialysis in patients with HF and the role of ultrafiltration are discussed separately (See "Management of heart failure in patients on dialysis" and "Management of refractory heart failure with reduced ejection fraction", section on 'Ultrafiltration'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Heart failure in adults".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Beyond the Basics topic (see "Patient education: Heart failure (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Goal of therapy – The goal of diuretic therapy is to eliminate clinical evidence of fluid retention, such as elevated jugular venous pressure and peripheral edema. This goal should be pursued while adverse effects are monitored. Persistent volume overload contributes to continued symptoms, may reduce the efficacy of drug therapy for heart failure (HF), and may be associated with increased mortality. (See 'Goals of therapy' above.)

●Therapy with diuretics – Patients with HF and volume overload require diuretic therapy.

•Type of agent – In patients with HF who require diuretic therapy, furosemide, torsemide, or bumetanide can be used as the initial therapy; all three are generally effective and well tolerated. We typically start oral diuretic therapy with furosemide. If the patient cannot be effectively diuresed with high-dose furosemide (eg, total daily dose is ≥200 mg), we switch from furosemide to torsemide or bumetanide. (See 'Choice of loop diuretic' above.)

Patients with renal insufficiency require higher maximum bolus doses of up to 160 to 200 mg of furosemide, 100 to 200 mg of torsemide, or 4 to 8 mg of bumetanide . (See "Causes and treatment of refractory edema in adults".)

•Route of administration – Intravenous diuretics (either as a bolus or a continuous infusion) are more potent than equivalent oral doses and are sometimes required for unstable or severe disease. (See 'Treatment of ADHF' above.)

•Dosing

-Oral – For patients with more chronic HF, the usual starting oral dose is 20 to 40 mg of furosemide. If a patient does not respond to initial dosing, the dose should be increased rather than giving the same dose twice a day (see 'Chronic therapy' above). If there is a good but short response, more frequent dosing may be needed.

Maximum single oral doses of furosemide are 40 to 80 mg for patients with a normal glomerular filtration rate (GFR). In patients with renal insufficiency, a higher maximum dose of 160 to 200 mg of furosemide can be given (maximum daily dose of 600 mg).

Torsemide and bumetanide are more predictably absorbed than furosemide. The usual initial oral dose of torsemide is 5 to 10 mg with maximum individual dose of 100 mg (maximum daily dose of 200 mg). The usual initial oral dose of bumetanide is 0.5 to 1.0 mg with maximum individual dose of 5 mg (maximum daily dose of 10 mg).

-Intravenous – The usual initial intravenous bolus dose of furosemide is 20 to 40 mg or up to 2.5 times the previously ineffective oral dose; if no response is obtained, the dose may be repeated every two hours with doubling of the dose as needed up to maximum doses. In patients with a normal GFR (typically estimated from the serum creatinine concentration), the maximum intravenous doses are usually 40 to 80 mg of furosemide, 20 to 40 mg of torsemide, or 1 to 2 mg of bumetanide. However, at times, higher doses are needed.

•Monitoring – In stable patients on a stable diuretic regimen, the fluid and electrolyte complications of diuretic therapy (eg, volume depletion, hypokalemia with loop diuretics, and hyperkalemia with spironolactone) are complete at two to three weeks. (See 'Time course of diuresis and associated complications' above.)

●Refractory congestion – Patients with HF who are not responsive to moderate oral doses of a loop diuretic (eg, furosemide 60 to 80 mg daily or its equivalent) should be evaluated for potential causes of diuretic refractoriness that include advanced HF, advanced kidney disease, hypoalbuminemia, high fluid or sodium intake, possible hypovolemia, lack of adherence to the diuretic regimen, and inability to absorb oral diuretics. (See 'Causes' above.)

In patients with refractory edema despite therapy with a high-dose loop diuretic, options for the management of volume overload include (see 'Management options' above):

•Switch to a more reliably absorbed oral loop diuretic (ie, torsemide or bumetanide in place of furosemide).

•Use intravenous diuretics.

•In patients who meet monitoring criteria, add a thiazide or other nonloop diuretic to augment diuresis (table 1). We suggest adding metolazone to the existing loop diuretic regimen (Grade 2C).

•Perform hemofiltration if other measures are unsuccessful.