β-Blocker withdrawal: The song of Orpheus☆☆☆★★★

See related article on page 456.

Calliope, the muse of heroic poetry, had a mortal son named Orpheus who sang almost as beautifully as the muses themselves. He brought the joy of music to earth, and his voice rang so pure and true that warriors put down their swords and savage beasts were spellbound. Orpheus was in love with Eurydice, who, on the day of their wedding, stepped on a snake and fell dead from its poisonous bite. Orpheus ceased singing and wandered the earth grieving. At the end of the world, he found the entrance to the underworld and went into the realm of the dead where he sang a song for Hades, the god of the dead. His music had the power to melt the cold heart of Hades, and he begged Hades to release Eurydice from the underworld. Hades gave his consent, but made one condition: Orpheus must not look back at Eurydice before reaching the land of the living. Overcome with joy, Orpheus started up the dark path toward the light, hearing footsteps behind him. As he walked, doubt began to creep into his mind. Had Hades deceived him? He had almost reached the upper world when he could bear his doubts no longer and turned for an instant to see if Eurydice was really behind him. As he did so, Hermes appeared to lead her back into the land of the dead and Orpheus heard her whisper farewell. Because of his lack of faith, he had lost her forever. Orpheus never again found joy on earth and wandered on singing songs that made all who heard them cry.

The withdrawal of β-blockade from patients with heart failure reminds me of the song of Orpheus. In this issue of the Journal (see page •••), Morimoto et al1 studied 13 patients with dilated cardiomyopathy who were receiving long-term β-blockade; the investigators withdrew the therapy to see if the patients would continue to do well. Seven of the 13 patients deteriorated, including 4 who died either suddenly or of progressive pump dysfunction. Here were patients who were generally doing relatively well on β-blockade; after withdrawal, heart rate increased, and many had worsening symptoms or died. Like Orpheus, if we turn to see if β-blocker therapy is really still necessary, if we withdraw therapy, our patients, like Eurydice, will find their way to the underworld, and we will sing only sad songs. We must have more faith than Orpheus and continue the therapy.

β-Blocker withdrawal was first tested in patients with heart failure by the same Swedish group who first presented evidence that these agents were beneficial in patients with heart failure.2 In that early case series, 15 patients with congestive cardiomyopathy who had improved with β-blockade were withdrawn from this medication. Six patients had a pronounced deterioration and all the remaining patients had a decrease in ejection fraction with an increase in third heart sound. Readministration of β-blockers resulted in subsequent improvement in these patients. A subsequent hemodynamic study from this same group3 examining 24 patients who improved while receiving metoprolol therapy demonstrated that metoprolol withdrawal produced deterioration in 16 of 24 patients and death in 4 patients (3 died suddenly). Again, readministration resulted in improvement in the survivors.

In addition to the deleterious effects β-blocker withdrawal might have in patients with heart failure, we have known since 1973 that β-blocker withdrawal might result in a “rebound” phenomenon in patients with coronary artery disease, producing ischemia, myocardial infarction, and death.4, 5 Miller et al6 noted a 50% incidence of withdrawal syndrome (including ventricular tachycardia, fatal myocardial infarction, or sudden death in 3 of 20 patients) after crossover to placebo during a double-blind trial of propranolol in the treatment of angina pectoris. Although a 50% incidence is probably higher than that normally found in the general population withdrawing from β-blockade,4 this study serves as a warning. Alderman et al7 had a similar experience in 6 patients with stable coronary disease who all had unstable angina develop after propranolol cessation. Three (50%) of these patients had a myocardial infarction and 1 died suddenly.

Why might β-blocker withdrawal result in deterioration in patients with heart failure and ischemia in patients with coronary disease? To understand this, the pathophysiologic characteristics of heart failure and why β-blockers work must be understood. Our paradigms for heart failure progression have dramatically changed in the last 30 years. Current concepts of why left ventricular function deteriorates over time are based on progressive biologic damage resulting in pathologic remodeling.8 Left ventricular dysfunction results in activation of compensatory neurohormonal systems (renin-angiotensin system, adrenergic nervous system, vasopressin, cytokines, endothelins) mediated by increased wall stress (stretch) and reduced stroke volume (resetting baroreceptors). Neurohormonal activation results in short-term hemodynamic effects (vasoconstriction, positive inotropy, positive chronotropy, and salt and water retention) but has long-term deleterious biologic effects (pathologic growth and remodeling, cell death, phenotypical alterations). These more chronic deleterious effects on growth in the heart lead to more left ventricular dysfunction, more neurohormonal activation, and a vicious cycle is established. β-Blockade inhibits and reverses the deleterious effects of norepinephrine on the heart (including phenotypic alterations in contractile proteins,9, 10 myocyte death,11 and pathologic remodeling12). β-Blockers slow heart rate, improve contractility, function, and energetics through a biological effect, not a pharmacologic mechanism.8, 12, 13 If β-blockers are withdrawn, the downhill biologic vicious cycle of cell death and phenotype alterations in contractile proteins from sympathetically mediated norepinephrine occur repetitively, leading to more pathologic remodeling and further neurohormonal activation. Thus heart failure returns because of the deleterious growth and death effects of norepinephrine on the myocytes.

Once myocytes die and are replaced by fibrosis, we are no longer able to create new myocytes because they are terminally differentiated and cannot enter the cell cycle.14 Progressive cell loss, fibrosis, and remodeling lead to downhill left ventricular dysfunction and patient death. Thus we have forever lost Eurydice by not trusting the therapy.

The reason for the β-blocker withdrawal syndrome in patients with coronary disease is caused by a sudden increase in myocardial oxygen consumption from elevated heart rate, blood pressure, contractility, and free fatty acid use (an inefficient fuel),13, 15 and at times a concomitant reduction in myocardial oxygen delivery. Heart rate may especially increase in patients who have received a β-blocking agent that upregulates β-1 receptors such as metoprolol or propranolol.3, 16 In this “hypersensitive” state17, 18 there are more receptors to transduce the signal intracellularly, and heart rate can “rebound” upwards. Even in patients who have received a β-blocker that does not upregulate β-receptors, such as carvedilol or bucindolol, the unmasking of and exposure of β-receptors can produce a marked increase in heart rate. On the side of oxygen delivery, platelet aggregation is produced by adrenergic stimulation and ischemia.19, 20 As heart rate and coronary flow rate increase when β-blockers are withdrawn, sheer stress within the coronary vasculature increases and ischemia increases (from increased metabolic demand). With increased sheer stress, ischemia, and pro-aggregatory stimuli (norepinephrine, epinephrine) present, the likelihood of platelet aggregation or disruption of a plaque increases, creating a milieu for myocardial infarction.

Finally, adrenergic stimuli create a milieu for sudden death. Surges of norepinephrine can mediate hypokalemia through the β-2 receptor21 and can cause increased risk of severe ventricular arrhythmias.17 As the heart pathologically remodels and becomes hypertrophic, and as fibrosis is formed, the dispersion of refractoriness increases, predisposing to arrhythmias. Nonselective β-blockade reduces the risk of sudden death in the patient after myocardial infarction22 and β1 –selective agents reduce sudden death in the patient with chronic heart failure.23, 24 Long-term β-blockade in patients with heart failure (both selective and nonselective) reduces QTc dispersion, a marker of the inhomogeneity of refractoriness.25 Thus sudden exposure of β-receptors after withdrawal of β-blockade will predispose the patient to sudden death. In the study by Morimoto et al,1 2 of the 4 patients who died did so suddenly, and in the study by Waagstein et al,3 3 of 4 died suddenly.

Occasionally, β-blocking agents must be withdrawn in patients for good reasons. Patients in cardiogenic shock or incipient shock need to have their β-blockade stopped immediately.22 Situations may arise in which withdrawal is necessary in a more elective fashion. If so, guidelines for withdrawing this therapy and its reinstitution have been previously published.22 β-blocking agents should not be routinely stopped for fluid overload in the absence of cardiogenic shock. In addition, they should not be stopped “cold turkey” if at all possible. They should be tapered over a period of several days to weeks, cutting the dose in half at each downtitration. If for any reason β-blocker therapy is discontinued, reinstitution of β-blockade for patients who are again well compensated should be performed carefully with the following suggested guidelines22: (1) For patients who have been off this medication <72 hours and who have not been in incipient cardiogenic shock, restart the β-blocker at the dose he or she was taking just before discontinuation. (2) For patients who have been off this medication >72 hours but less than 7 days and who have not been in incipient cardiogenic shock, restart the β-blocker at half the dose he or she was taking just before discontinuation. (3) For patients who have been off this medication ≥7 days or have recently been in incipient cardiogenic shock, restart the study medication at the lowest dose (3.125 mg of carvedilol twice daily, 3 to 6 mg of metoprolol or labetolol twice daily [which can be given orally by mixing 3 to 6 mg of the intravenous preparation in juice], 1.25 mg of bisoprolol orally each day) and retitrate on a weekly basis as tolerated.

Guidelines for a full uptitration have been previously published.22 Because β-blocking agents may protect the heart from arrhythmia and progressive pump dysfunction, they should be reinitiated if at all possible, even in the patient who has stabilized from decompensation.

Because β-blocking agents appear to protect the heart from sudden death and progressive pump dysfunction, they are invaluable agents in the treatment of heart failure. The large β-blocker mortality trials (BEST, CIBIS-2, COPERNICUS, and MERIT-HF) will help us determine the ultimate benefit of these agents in reducing mortality rate and understand which patients should be treated with this therapy. The CIBIS-1 trial suggested that some patients will have a deterioration in pump function with β-blockade and will have a concomitant increase in mortality rate compared with placebo.24 Thus much has yet to be learned about this therapy and who the appropriate patients are for its use. However, when patients are doing well taking β-blockers, we should have faith in the therapy and indefinitely maintain its protective effects. Patients who are doing well taking β-blockers should thus not suffer the fate of Eurydice, and their physicians should not look back.