American College of Cardiology

STEVENSON AND KORMOS, ET AL., MECHANICAL CARDIAC SUPPORT 2000
JACC Vol. 37, No. 1, January 2001:340-70

II. Evolution of Therapies for Heart Failure

A. Medical Therapies for Heart Failure

The evolution of therapy for heart failure presently includes many strategies never tested by properly controlled clinical trials (Table 2). Many treatments have been abandoned without formal testing after unrewarding anecdotal experience. Over two millennia ago, treatment for what was once termed “dropsy” was aimed at restoring a balance of fundamental elements and complementary humors.^15,16 A historical overview of more modern therapies ¹⁷ reveals that in 1683, Thomas Sydenham recommended bleeding, purges, blistering, garlic and wine. A century later, William Withering provided a precise description of the benefits of foxglove in the Shropshire maid’s cure for dropsy. Catharsis and venesection continued through the nineteenth century, with amyl nitrate, mercurial diuretics and digitalis glycosides becoming available in the early part of the twentieth century.

While the laboratory experience was developing that allowed human cardiac transplantation to proceed, medical therapy for heart failure included only digitalis, thiazide diuretics (introduced in 1962) and furosemide (introduced in 1965). Controlled trials of withdrawing or administering digoxin did not take place until 1993^18,19 and 1997,²⁰ and there were no trials of diuretics except as substudies of two trials testing other drugs.^21,22 The quest to establish a basis of evidence from which to prescribe effective therapies for specified populations has been relatively recent.²³ The concept of vasodilators for heart failure was introduced by the acute use of nitroprusside in 1974, followed by hydralazine in 1977. The first large randomized clinical trial in heart failure with mortality end points was not completed until 1986,²⁴ demonstrating improved survival with the hydralazine-isosorbide dinitrate combination. With the release of captopril in 1980 and enalapril in 1984, multiple large, randomized, placebo-controlled trials established angiotensin-converting enzyme inhibitors as the cornerstone of therapy, with extensive unforeseen benefits for this drug class occurring beyond that expected only from vasodilation.^25–28

Trials have also demonstrated the lack of sustained clinical benefit from many therapies with sound theoretical rationale. Although acute hemodynamic improvements in heart failure patients were readily demonstrated with dopamine in 1972 and dobutamine in 1974, inotropic agents have not been associated with sustained hemodynamic benefit or mortality reduction during chronic therapy. In fact, mortality is increased in these patients, as suggested by early experiences and confirmed in larger trials.²⁹ Although excess myocyte calcium concentrations have been implicated in progression and death, calcium channel blockers have worsened heart failure and survival in retrospective analyses and prospective trials. Many anti-arrhythmic agents that suppress ventricular arrhythmias were shown in large trials to increase death in patients with heart failure. Amiodarone, the only currently available anti-arrhythmic agent that does not increase mortality in heart failure, may in fact have more benefit for heart failure end points than for sudden death. Beta-adrenergic blocking agents worsen hemodynamics initially but, when tolerated, lead eventually to improved hemodynamics and survival in recent large trials of mild-to-moderate heart failure.

Reviewing the history of introduction, adoption and, in some cases, abandonment of therapies for heart failure, reveals the contribution of large controlled trials in defining the additive impact of our interventions. In the process of establishing a basis of evidence to guide current medical therapy for heart failure, a template has been created for the rigorous testing of medications that can be administered in parallel with placebo therapy. End points of survival, clinical status, cardiovascular function and cost-effectiveness can be evaluated using this template without either patient or physician knowing who has received the new therapy being tested.

However, the randomized placebo-controlled trials have, in general, not included patients desperate for relief from severe heart failure symptoms or hoping to be rescued from imminent death. For example, the rapid impact of intravenous diuretics in treating dyspnea from pulmonary edema in heart failure, and the rapid benefit of inotropic therapy to improve perfusion acutely in CS have not been put to the test of placebo-controlled, randomized trials. The immediate cause-effect response typically observed renders a physician unlikely to substitute placebo therapy in these situations. Even in a less compromised group of hospitalized patients, placebo-controlled trials have either excluded patients with urgent indications for intravenous therapy or limited placebo therapy to a short period with early crossover to active treatment.

B. Surgical Therapies for Heart Failure

Early surgical procedures for heart failure included thyroidectomy, pericardiectomy and valve replacement. Subsequent procedures, such as intra-aortic balloon counterpulsation for CS,³⁰ proposed in 1961, and LV aneurysmectomy introduced for chronic HF in 1962,³¹ were more systemically studied and reported but without specific control groups against which to compare benefit. As soon as orthotopic cardiac transplantation was performed in humans, it was tried in many centers with poor initial results. In large part through the perseverance of the Stanford team, outcomes steadily improved. Approval by Medicare of heart transplant as standard therapy was based on careful description of outcomes for a cohort of patients assumed to have over 50% six-month mortality without transplantation (estimates based on early waiting list deaths, but not on any control groups). Increasing waiting times for transplantation have led to expanding use of mechanical circulatory support as bridging devices for cardiac transplantation. Comparisons with patient cohorts without bridging devices suggested better survival to transplantation and discharge, but no randomized trials were done before the widespread acceptance of bridging strategies.

With a limited supply of donor hearts, research continued into other surgical options for heart failure. Coronary revascularization and valvular heart surgery, once thought to be contraindicated in the presence of a low ejection fraction, were extended into the heart failure population, where their roles are not yet defined. A variety of cardiac remodeling procedures (aneurysmorrhaphy/aneurysmectomy, infarct exclusion, application of cardiac restraining and ventricular splinting devices) have recently been introduced and reported in small numbers. More systematic evaluations have been recommended.³¹ In fact, there are developing plans for a national randomized trial in ischemic heart failure to compare medical therapy with surgical therapy, with further randomization of the surgical arm with or without ventricular reconstruction.

Despite the obstacles, large randomized clinical trials have been performed with surgical therapies of advanced cardiac disease. Three landmark trials of coronary artery bypass surgery clarified its role in ameliorating morbidity and mortality from coronary heart disease.^32–34 The smaller analyses of patients with three-vessel disease with decreased LVEF demonstrated particular benefit but included few patients with typical heart failure. With enthusiasm generated by uncontrolled experiences of cardiomyoplasty, the Cardiomyoplasty-Skeletal Muscle Assist Randomized Trial (C-SMART) was an ambitious trial³⁵ that included a non-blinded, control arm of patients without cardiomyoplasty. Due to early problems with patient recruitment and withdrawal to receive active therapy, the protocol was changed to allow crossover to active treatment after one year. After recruitment of only 100 patients over five years because of ongoing problems with both patient recruitment and reimbursement, the trial was terminated, despite a trend for improved outcomes in the surgical group.

Revascularization is commonly employed as standard therapy for CS due to an acute ischemic event. The Should We Emergently Revascularize Occluded Coronaries for Cardiogenic Shock? (SHOCK) trial³⁶ of revascularization for acute coronary syndromes causing CS was completed in 302 patients only after five years. Survival benefit from revascularization was not apparent at one month but shown by the six-month evaluation for patients under 75 years. At the same time, the Swiss Multicenter Angioplasty for Shock Trial was terminated because of inadequate enrollment.³⁷

The ongoing REMATCH trial¹³ faces the double challenge posed by both a surgical trial and study of a more compromised heart failure population than was ever enrolled in a controlled trial. Candidate criteria were originally designed to include patients with an expected 25% two-year survival. Considering previous information from cohort experiences suggesting a large benefit from “bridging“ in transplant candidates, concern was raised that this trial was unethical because it denied patients a life-saving therapy. In fact, the 21-patient pilot trial prior to REMATCH demonstrated a three-month mortality of almost 30% without apparent difference between the medical and surgical arms. Attempting to find a population with intermediate risk, the inclusion criteria for REMATCH were subsequently expanded to require 60 rather than 90 days of severe symptoms and either dependence on intravenous inotropic agents or a peak oxygen consumption <14 ml/kg/min, compared with the previous limit of 12 ml/kg/min. Enrollment in the trial has been limited by issues of reimbursement for the surgical procedures, difficulty in regional recruitment at designated centers and reluctance of patients and families as well as physicians to accept randomization in the setting of a life-threatening illness for which a new therapy might be life-saving. Still, it is anticipated that the completion of this trial in 2001 will provide new benchmarks for both the medical and device arms of future trials.

C. Downshifting of Risk for New Surgical Therapies

The recognized success of new surgical procedures for advanced disease may be followed in some cases by a cycle of improving results and expanding population definition. The evolution of such therapy contrasts with the development of pharmacologic and exercise interventions, which have usually been initiated in patients with mild disease, validated in trials of moderate disease and ultimately extended to patients with severe disease who would have been excluded from the landmark trials.³⁸ Surgical therapies for heart failure carry front-loaded risk that is easier to absorb for patients expecting high early mortality. As survival and improved function are realized by these desperate patients, the procedure is then sought by patients at earlier stages of the disease. These patients are more likely than the initial subjects to obtain good results from the procedure. With the downshifting of risk, however, the actual benefit, calculated as the difference between outcome with the procedure and outcome without the procedure, may become less significant. An appropriate example of “downshifting” the risk is the evolution of cardiac transplantation.^39–41 Candidates were originally expected to have “less than six months to live,” at which time survival with transplantation was 60% to 70% at one year. The current one-year survival rate after heart transplant is 80% to 85%, with a 10-year survival rate of about 50%. For ambulatory heart failure patients not requiring intravenous inotropic agents, the survival without transplantation has also improved to 60% to 70% without death or urgent transplantation at one year in many studies, leaving a smaller margin of early benefit. The positive impact of heart transplant remains striking, however, for patients in critical status or dependent on inotropic infusion. After initial experiences, risk can shift up as well, as has happened for candidates developing organ failure while awaiting transplantation, such that procedures may be extended to patients who are more severely ill than their predecessors. As new surgical therapies for heart failure are introduced and accepted into broadening populations, it remains crucial to monitor the target populations and ensure that the benefits expected from earlier experience are being derived.