I will be discussing revisiting the use of diuretics in heart failure.

Acutely decompensated heart failure affects about 1 million individuals presenting in the hospital each year in the United States. The in-hospital mortality is approximately 4%. The mean length of stay in these individuals is about 6.5 days. The break-even point for a Medicare diagnosis-related group (DRG) is about 4.8 to 5 days, so many hospitals lose money on these acutely decompensated heart failure patients.

If we look at how well we are doing in treating these patients and at what happens after hospital discharge, we see that there is a very high rate of readmission after a patient has been hospitalized with acutely decompensated heart failure. There is also a very high mortality rate, After hospitalization for acutely decompensated heart failure, a patient is at substantially higher risk for mortality and rehospitalization than a chronic, stable outpatient with heart failure -- even one with the same degree of clinical symptoms and ventricular dysfunction. This raises the possibility that some of what we have been doing in the hospital to treat and stabilize these patients and reduce symptoms may actually have been contributing to these very high morbidity and mortality rates. It is hard to look at these data and say that our current treatment strategies are more than adequate for these patients.

I suggest that our therapeutic goals in acutely decompensated heart failure should be to reverse the acute hemodynamic abnormalities, rapidly relieve symptoms, initiate treatments that will slow heart failure disease progression as well as improve long-term clinical outcome, and try to apply these therapies in the most cost-effective fashion. We can think of a variety of end points to help us meet these therapeutic goals.

Intravenous (IV) diuretics have been a mainstay of therapy applied to patients with acutely decompensated heart failure. One of our major ways of removing the congestion in these patients is to diurese them. It is certain that IV loop diuretic therapy results in diuresis and natriuresis, and we believe this would be favorable. However, IV loop diuretic therapy, administered as monotherapy without a vasodilator or natriuretic peptide, also results in a variety of potentially deleterious effects that have been largely overlooked. We get further activation to the renin-angiotensin-aldosterone system, further activation of the sympathetic nervous system, reflex vasoconstriction, and an increase in afterload. There are also reductions, not improvements, in stroke volumes, a reduction in cardiac output, and a substantial and early reduction in glomerular filtration rates (GFRs).

If we look for clinical trial evidence that has supported the use of IV loop diuretics, we recognize that there have been no randomized clinical outcome studies of loop diuretics for the treatment of acutely decompensated heart failure. Thus, the true effects on the morbidity and mortality of this condition are largely unknown. As we look at observational data, it becomes even more of a concern.

Is there really activation of the neurohormonal systems when IV loop diuretics are administered? A variety of studies demonstrate this. Here we see the before and after results of IV diuretic for patients with decompensated heart failure. There is a marked activation in renin activity as well as in plasma aldosterone levels.

From this study, which looked at individuals 20 minutes after a relatively low dose of IV furosemide (40 mg), we see a modest reduction in wedge pressure. However, there is already evidence for vasoconstriction, with an increase in systemic vascular resistance (SVR), a reduction in stroke volume index, a marked activation (doubling) in plasma-renin levels, and a statistically significant increase in plasma norepinephrine levels. Thus, there is unquestionable activation of the renin-angiotensin-aldosterone and sympathetic nervous systems.

The impact of IV loop diuretics on vasoconstriction was part of the reason, in early studies, why there was believed to be a Frank-Starling curve with regard to decompensated heart failure. These patients actually had to have their wedge pressure elevated. If we lowered it too much, they would have a substantial drop in stroke volume and cardiac output. We saw this when the IV loop diuretic was used alone. As we got down to wedge pressures of even 24, there was a substantial decrease in stroke volume due to the excess vasoconstriction. However, when the IV diuretic was combined with an IV vasodilator, we could normalize wedge pressure with an improvement in stroke volume due to the reduction in mitral regurgitation and the reduction in afterload. Again, this shows us the adverse hemodynamic consequences when IV loop diuretics were used alone.

This study looked at the change in glomerular filtration rate within 8 hours of giving 80 mg of IV furosemide. There was a net diuresis of about 1500 cc, but the GFR, within 8 hours after administering the 80 mg of IV furosemide, fell by 20%. The deleterious effects on renal function have clearly been defined, but not necessarily recognized clinically, as playing a role.

Look at the real world experience. This comes from the Acute Decompensated Heart Failure National Registry (ADHERE) data involving over 100,000 patients hospitalized with heart failure in 270 United States hospitals. Among the IV therapies used for patients with acutely decompensated heart failure, the medication most commonly administered (88%) was an IV loop diuretic. For 64% of patients, this was the only IV therapy they received during hospitalization for acutely decompensated heart failure.

What do we observe with this practice pattern with regard to clinical outcomes? The majority of patients hospitalized with acutely decompensated heart failure -- looking at a change from admission to discharge creatinine -- actually have evidence of worsening renal function; 70% have a rise in creatinine of 0.1 mg/dL or more. About 50% have an increase of 0.3 mg/dL or more, and about 22% of patients experience a creatinine rise of greater than 0.5 mg/dL. We see frequent evidence of worsening renal function in patients hospitalized with heart failure when the major IV therapy they have received is IV loop diuretics.

Looking at the independent predictors of worsening renal function, a variety of clinical and treatment factors, we see in this analysis that loop diuretic dose (in 20-mg increments) was an independent predictor of 4% increased risk of worsening renal function. We also see other risk predictors. Note that angiotensin-converting enzyme (ACE) inhibitors, adrenergic-receptor binders (ARBs), and aldosterone antagonists were not associated with worsened renal function during the hospitalization, in contrast with what is a frequent clinical perspective among clinicians. Also, note that higher doses of loop diuretics were not associated with either increased diuresis or more effective weight change in these patients. We are seeing a lack of efficacy and an independent increased risk of worsening renal function.

We recognize that renal dysfunction is frequent in patients with heart failure who are hospitalized. Looking at this in great detail in the ADHERE registry, we see that 22% of hospitalized patients have creatinine greater than 2 mg/dL. These individuals are no older than those without renal insufficiency, are less likely to be female, and are more likely to be diabetic. Ventricular function, however, between those with and without significant renal insufficiency is similar. We do see that the B-type natriuretic peptide (BNP) levels are high in both, but a little higher in those with renal insufficiency.

If we look at clinical outcomes during hospitalization in individuals with renal insufficiency, we see a significantly prolonged length of stay, looking at mean length of stay, and about double the in-hospital mortality rate. It has been frequently recognized that renal insufficiency is a major predictor of adverse outcome in patients hospitalized with heart failure.

This analysis will be covered in greater detail in the presentation by Dr. Abraham, but of all the predictors of in-hospital mortality, the single best predictor was admission blood urea nitrogen (BUN). This highlights how critically important renal function is and the potentially adverse consequences of deterioration of renal function during the hospitalization.

Looking at the potential relationship between the use of IV loop diuretics and adverse outcomes, we see that although we do not have randomized trials, we do have a variety of observational data. They certainly seem to confirm that some of these deleterious mechanisms demonstrated in patients may actually have adverse clinical consequences. These data are taken from ADHERE -- patients who received IV loop diuretics during their hospitalization for decompensated heart failure vs those who did not. There are differences in baseline characteristics, so this study tries to adjust for 33 clinically important covariates as well as propensity adjustment. Those getting IV loop diuretics, controlling for the baseline differences, had a 49% increased risk for a prolonged length of stay and a 29% independent increased risk of mortality.

The level of effectiveness in terms of relieving congestion and getting excellent diuresis in patients may be something we need to put up with. However, when looking at the change from admission to discharge weight, we see that a large number of patients had little relief of congestion, as evidenced by a net change in weight. This slide shows that 20% had no change in weight or an increase in weight during their hospitalization; in about 50% there was either very mild diuresis or a net gain.

Many patients being discharged from hospitalization for acutely decompensated heart failure are going home with signs and symptoms of congestion and heart failure. Even when applying IV loop diuretics, many patients are leaving the hospital with signs and symptoms of congestion.

If we move beyond the hospital and look at the long-term consequences of loop diuretics, how do they look in observational data? There is also evidence for concern here. This slide is looking at higher vs lower doses of loop diuretics in patients with severe heart failure in the Prospective Randomized Amlodipine Survival Evaluation (PRAISE) database. In those patients, higher doses of loop diuretics were associated with increased risk of total mortality, sudden cardiac death, and progressive heart failure death. Does neurohormonal activation play a role in this potential deleterious effect? Higher doses of ACE inhibitors can partially overcome this increased risk, supporting that hypothesis.

An analysis of the Studies of Left Ventricular Dysfunction (SOLVD) subdatabase, likewise when adjusting for baseline differences, also showed that those receiving a loop diuretic were more likely to have an increase in all-cause mortality. This was not observed with aldosterone antagonists. Thus, we see in this observational data that there is potential increased risk.

In a situation of acute renal failure reported in The Journal of the American Medical Association (JAMA) the IV loop diuretic (administered within 24 hours of the patient presenting with acute renal failure) was also associated with excess risk of in-hospital mortality and nonrecovery of renal function.

Supporting the observations that these deleterious mechanisms may not play out well, we could begin to recognize, as we talk about the cardiorenal syndrome of heart failure, that one of the major precipitants of this syndrome and potentially the morbidity/mortality of this condition may be diuretic therapy itself. It is a neurohormonal activation leading to vasoconstriction, diminished renal blood flow, decreased perfusion, impaired renal function, further diuretic resistance, further neurohormonal activation, and pathologic ventricular remodeling; and all of these contribute to the morbidity and mortality we observe here.

There is 1 small randomized study that tried to spare the dose of IV loop diuretics by using low-dose diuretic combined with an IV vasodilator compared with patients randomized to high doses of furosemide. In this trial, which enrolled 110 patients, there was a substantially lower risk for the adverse outcome -- a need for mechanical ventilation -- when patients were started off with lower doses of loop diuretics. The overall adverse clinical outcomes, including mortality, were lowered by sparing the dose of loop diuretic.

You may say, "But wait, we've got to get these patients relief from congestion; our hemodynamic goals are to reduce wedge. How can we approach these patients?" We are certainly not going to use any loop diuretics. But if we combine this with natriuretic peptides, we add the potential ability to get the same or greater degree of natriuretic diuresis, and we do it in a more physiologic fashion, exploiting the human BNP for its hemodynamic effects, favorable neurohormonal effects, favorable renal effects and direct potential cardiac effects.

What are the renal effects of natriuretic peptides? In renal blood flow, they are the dilation of the preglomerular (afferent) vessels, constriction of the postglomerular (efferent) vessels, reduction of the venous renal filling pressure, and an increase in glomerular hydrostatic pressure. There is relaxation of mesangial cells, increasing the filtration surface area; inhibition of the proximal tubular sodium reabsorption; and inhibition of the inner medullar collecting duct's sodium reabsorption. These facts can be unrelated to the underlying pathologic state; we can see this in normal individuals as well as in patients with heart failure.

There are multiple mechanisms within the kidney where the natriuretic peptides can promote diuresis and natriuresis and sodium chloride reabsorption: the inner medullary collecting duct, via the cyclic guanosine monophosphate (GMP) method, as well as the sodium channels and sodium potassium adenosine triphosphatase (ATPase) inhibition shown here.

The mechanisms by which the natriuretic peptides may increase GFR are the increases in hydrostatic pressure and filtration area as well as the favorable effects in vasodilation that increase renal blood flow.

You may say, "But wait, we've got to get these patients relief from congestion; our hemodynamic goals are to reduce wedge. How can we approach these patients?" We are certainly not going to use any loop diuretics. But if we combine this with natriuretic peptides, we add the potential ability to get the same or greater degree of natriuretic diuresis, and we do it in a more physiologic fashion, exploiting the human BNP for its hemodynamic effects, favorable neurohormonal effects, favorable renal effects and direct potential cardiac effects.

What are the renal effects of natriuretic peptides? In renal blood flow, they are the dilation of the preglomerular (afferent) vessels, constriction of the postglomerular (efferent) vessels, reduction of the venous renal filling pressure, and an increase in glomerular hydrostatic pressure. There is relaxation of mesangial cells, increasing the filtration surface area; inhibition of the proximal tubular sodium reabsorption; and inhibition of the inner medullar collecting duct's sodium reabsorption. These facts can be unrelated to the underlying pathologic state; we can see this in normal individuals as well as in patients with heart failure.

There are multiple mechanisms within the kidney where the natriuretic peptides can promote diuresis and natriuresis and sodium chloride reabsorption: the inner medullary collecting duct, via the cyclic guanosine monophosphate (GMP) method, as well as the sodium channels and sodium potassium adenosine triphosphatase (ATPase) inhibition shown here.

The mechanisms by which the natriuretic peptides may increase GFR are the increases in hydrostatic pressure and filtration area as well as the favorable effects in vasodilation that increase renal blood flow.

What about patients with renal insufficiency; how did they respond? The response to loop diuretics is somewhat attenuated in these patients; they are more likely to have diuretic resistance. Looking at patients in the Vasodilatation in the Management of Acute Congestive Heart Failure (VMAC) study with baseline creatinines greater than 2 mg/dL, we actually see similar reductions in wedge pressure. These patients have a similar clinical and hemodynamic response.

If we look at the effects of nesiritide on the symptoms of acutely decompensated heart failure patients with and without renal insufficiency, we see that both groups have substantial improvement compared with baseline on their shortness-of-breath score and their global clinical status. The patients with renal insufficiency respond very well to the natriuretic peptides.

Tolerability was excellent. This slide is comparing, in patients with renal insufficiency, IV nitroglycerin and nesiritide; overall adverse effects were slightly lower with nesiritide. The risk of symptomatic hypotension is low and not dissimilar in the 2 groups.

Is there evidence to support, even compared with another IV vasodilator, that we have a more potent natriuretic effect and are less likely to require subsequent dosings of the loop diuretics? There are some observations from the VMAC trial that support this. In patients with creatinine above 2 mg/dL at baseline and comparing nesiritide with nitroglycerin, a greater number do not require subsequent IV loop diuretic dosing over the first 24 hours when they receive nesiritide relative to nitroglycerin.

There will be further data regarding the outpatient use of nesiritide, but I wanted to highlight that the outpatient administration of nesiritide, compared with standard care, was associated with a decrease in serious adverse renal effects, which included a rise in BUN or creatinine, developing acute renal failure, or oliguria. Thus, we appear to see in these patients, even in the outpatient setting, that nesiritide may work favorably and counteract some of the deleterious effects of the loop diuretics.

A study in Circulation looked at the impact on GFR of nesiritide in 15 patients. It did not find an improvement, but it did not find a deterioration, either. This study selected for patients who had not really responded to nesiritide; if their net weight had changed by 1 kg over 24 hours, they were excluded. These patients were enrolled fairly far into their hospitalizations; thus, it was not clear that they were still congested and decompensated.

If we look at the balance of the evidence that we have so far, it does suggest that the natriuretic peptides could restore diuretic responsiveness and perhaps allow us to limit the renal toxicity and adverse neurohormonal effects of the loop diuretics. We know a variety of the effects caused by the underlying heart failure condition and by the use of loop diuretics. As we think about the natriuretic peptides in each case, they help us potentially counteract and reverse these effects.

Some data make a compelling case that nesiritide may be a preferential approach in those patients with renal insufficiency and heart failure. It allows us to maintain GFR and renal blood flow during natriuresis, limits the need for repeat dosing of the loop diuretic, is safe and effective in patients with moderate/severe renal insufficiency, rapidly lowers ventricular filling pressures, and brings about symptomatic, as well as hemodynamic, improvement. From the data, we see no proarrhythmic or tachyphylaxic effects. It can be safely administered in the emergency department; intensive care unit (ICU) monitoring is not needed; and patients with renal insufficiency may have improved outcomes.

We could also make a case for nesiritide as initial treatment even in patients with acutely decompensated heart failure. Why wait for these patients to develop worsening renal failure? In the absence of baseline renal insufficiency, we recognize that we get greater efficacy of sodium and water removal compared with loop diuretics alone, rapid and sustained reduction in wedge pressure compared with even IV nitroglycerin, and the ability to relieve congestion obtained via dry weight compared with the relative inefficiency of diuretic monotherapy-based strategies. We are also able to avoid early worsening in renal function, which we know is associated with adverse outcomes, and reduce the deleterious neurohormonal activation that occurs when we use IV loop diuretics and monotherapy strategies.

Thus, the use of the IV loop diuretic and nesiritide as initial therapy upon presentation supports this concept of initial treatment for the heart failure patient who is volume overloaded and short of breath but has preserved systemic perfusion.

What other strategies can we utilize to try to overcome the potential adverse effects of loop diuretics? Other therapies undergoing evaluation include vasopressin antagonists and ultrafiltration, a mechanical method that I will highlight briefly.

The vasopressin antagonists are highlighted in the Acute and Chronic Therapeutic Impact of a Vasopressin Antagonist (ACTIV) in Congestive Heart Failure (CHF) trial published in JAMA. We do see a net change in weight at 24 hours with the different doses. Tolvaptan was also sustained at discharge in 2 of these doses.

Looking at clinical outcome at 60 days, we see a trend for a favorable effect on mortality. Clearly, this needs to be confirmed, but it did appear to reach significance in those with higher BUN and in those with congestion, making vasopressin a potentially promising agent that deserves further testing in clinical trials.

The other method undergoing evaluation is ultrafiltration, the concept of removing volume mechanically. A simplified peripheral ultrafiltration system has been developed that uses a miniaturized disposable circuit and allows patients without a central line to undergo ultrafiltration.

There have been observational studies that looked at this, and further randomized control studies are under way. Data from one such study were published in the Journal of Cardiac Failure. After undergoing the ultrafiltration, there was a net change of 2.6 kg -- highly statistically significant -- and improvement, compared with baseline, of the signs and symptoms of heart failure. Some potentially promising data show that mechanical methods may allow us to get rid of excess volume without necessarily exposing the patient to the loop diuretics to the same degree and to some of the adverse effects.

What strategies can we put into place to protect heart failure patients from these potentially deleterious effects of the IV loop diuretics? In the hospital, we can initiate and continue to optimize our ACE inhibitor dosing, protecting the patient from the deleterious consequences of activation of renin-angiotensin system. We can get beta-blockers started as soon as possible in these patients, once they are compensated prior to hospital discharge; initiate or continue the aldosterone antagonists in patients without contraindications; make use of IV nesiritide, as we have covered, and continue evaluation of ultrafiltration vasopressin.

On the outpatient basis, we can try to use the minimal diuretic dose that will maintain the euvolemic state and optimize our neurohormonal antagonists, and we can also avoid nonsteroidal anti-inflammatory agents. You will hear about the Follow-Up Serial Infusions of Nesiritide (FUSION) data evaluating nesiritide on that outpatient basis. At least one of the potential mechanisms, a benefit here, may be limiting the adverse consequences of chronic use of the loop diuretics.

Intravenous loop diuretics are the most frequently used treatment for decompensated heart failure. When used as monotherapy for acute heart failure, IV loop diuretic therapy produces a variety of potentially deleterious effects that have been underappreciated: neurohormonal activation, vasoconstriction, decreased perfusion, and decreased GFR. These may contribute to the adverse clinical outcomes that we have seen.

There are no loop diuretic mortality trials; observational studies do not look good. They suggest that loop diuretics may contribute to the high risk of rehospitalization mortality. In the era of evidence-based medicine, we certainly need to revisit this. The use of nesiritide, as well as other potential strategies, may allow us to decrease the amount of loop diuretics required for diuresis. This will likely allow us to translate this change in strategy into improved clinical outcomes.