β-blockers in heart failure. From No! to Maybe? to Yes!
For decades, the use of β-blockers in heart failure was contraindicated. Then, Finn Waagstein made a bold move, successfully treating a single patient. Thus began a miraculous 180.
Early in medical school, I memorized that β-blockers (e.g., propranolol, metoprolol, carvedilol) are “negative inotropes” and therefore decrease the force of heart contractions. Later, in residency, I was indoctrinated that they were part of the backbone of heart failure treatment, alongside ace inhibitors and aldosterone antagonists.1
At first blush, these observations are a contradiction, particularly for patients with heart failure with reduced ejection fraction. Why would we possibly want to give a drug that might decrease the contractile power of the heart? Wouldn’t that be a bad thing? And so the thinking went, from the first introduction of β-blockers in the 1960s, through the initial observations of their benefit in the late 1970s, all the way through the mid-1990s, when multiple randomized trials demonstrated that not only were β-blockers not harmful to the heart failure patient, they actually save lives.
The story of how we got there is fascinating. It involves a young Swedish cardiologist willing to test a hypothesis on a single patient and persevering for decades while the medical establishment called his proposals perverse and negligent. In the end, the science prevailed.
1964-1975: β-blockers are contraindicated in HF
Propranolol, the first β-blockers, was introduced to the world in 1964 by James Black. Initially used as a treatment for hypertension and angina, β-blockers were quickly cast as potentially harmful in heart failure. In 1965, Epstein, Robinson, Kahler, and Braunwald found that propranolol caused a significant reduction in cardiac output. Two years later, Epstein and Braunwald wrote, “In patients with serious heart disease, these actions may impair cardiac function to the point of seriously intensifying heart failure” and that β-blockers are “potent agents and must be used with considerable care.” Early (and some later) trials of β-blockers in the treatment of hypertension and angina would often exclude patients with heart failure, fearing they would worsen this condition. They were, after all, negative inotropes.
Though their use was initially focused on hypertension and chronic stable angina, it did not take long for some to expand the use of β-blockers to acute myocardial infarction. Although there was no consensus on the mechanism of benefit, some wondered whether blunting an overly aggressive sympathetic response might mitigate infarct size and/or arrhythmia.
In the early 1970s, Finn Waagstein was training in cardiology at Sahlgrenska University Hospital in Sweden. Though contraindicated at the time, he administered cardioselective β-blockers to several patients with large transmural infarcts, ongoing pain, and signs of heart failure. Amazingly, he observed no worsening of congestion despite the use of these negative inotropes. Recounting his experience decades later, he said, “I was banned from the unit because I had risked the life of the patients by using an unauthorized treatment. The nurses supported me because they had seen the patients improve.”
1975-1980: β-blockers ain’t so bad?
Emboldened by his experience, Waagstein took another leap. In 1974, he was caring for a 59-year-old woman with many years of heart failure secondary to idiopathic dilated cardiomyopathy. In the middle of the night, the patient arrived with acute pulmonary edema. At the time, the standard of care included oxygen, morphine, furosemide, and rotating tourniquets on the arms and legs (to reduce preload). After observing that the patient also had sinus tachycardia, he administered a β-blocker. Miraculously, there was almost immediate relief from the pulmonary edema, and the patient survived the night. They were later maintained on alprenolol and then metoprolol. In 1998, it was reported that she remained well, at 82 years old, on a combination of metoprolol, enalapril, diuretics, and digitalis.
In 1975, the British Medical Journal reported the experience with this patient and six others. The opening of the letter Waagstein received from the BMJ editor is telling.
Waagstein expanded his study of β-blockers in heart failure and published a report in 1979 in The Lancet suggesting prolonged survival with these drugs. Alongside first author Karl Swedberg, he studied 37 patients with dilated cardiomyopathy and heart failure. Twenty-four received β-blockers, and 13 did not. As shown in the figure below, survival at one, two, and three years was significantly higher in those who had received β-blockers (83% versus 49% at one year, 66% versus 19% at two years, and 52% versus 10% at three years).
Over the next couple of years, this Swedish team published additional observational studies suggesting that β-blockers were not harmful but instead helpful in patients with heart failure. In one report, they withdrew β-blockers therapy in 15 patients who had shown improvement on the drugs. Six patients worsened clinically, and in all others, there was a significant decrease in ejection fraction. When they were placed back on β-blockers, their status improved.
Waagstein and his colleagues’ work did not receive a warm response. A 1981 editorial in The Lancet began by writing, “To give a beta blocker to a patient in heart failure is a perverse form of treatment. Yet a group in Sweden has been doing just this for the past 5 years with apparent success.”
1980-1995: β-blockers get lost in the background
The 1980s saw the emergence of vasodilator therapy in heart failure. First, hydralazine and isosorbide, followed by ACE inhibitors. β-blockers remained scarcely used and largely contraindicated in heart failure. The 1987 CONSENSUS study provides a good example of how uncommon they were prescribed in this period.2 This landmark study demonstrated an enormous reduction in mortality at one year (52% versus 36%) for those who had been randomized to receive captopril. Table 1 offers an interesting lens to the standard of care for heart failure when this trial was conducted (1985-1986).
Just 2-4% of patients in CONSENSUS were receiving β-blockers. Not much would change over the next ten years. It isn’t as though β-blockers were no longer being studied. They were. However, the results weren’t consistent, and the studies were often small and remained observational. Some even suggested harm. For many, the belief was that patients were improving despite β-blockers use, not as a result of their use.
A 1986 review offers a sense of where things stood. Entitled “Beta-Blockers in Congestive Cardiomyopathy Conceptual Advance or Contraindication?” it includes a table summarizing the available data. The sections are separated with the following headers: “Swedish Optimism,” “American Support,” “British Empire Pessimism,” and “Miscellaneous Mishaps.”3 The burden of proof remained on those who wanted to use β-blockers in heart failure.
In 1989, Waagstein published another study looking at what happens when you remove β-blockers from patients whose heart failure had improved after their administration. Once again, he showed that they got worse and that reintroduction restored the benefit. Much like a version of Koch’s postulates, these data were making a strong case that β-blockers weren’t just along for the ride but were actually part of a causal chain. And still, rightly, the call was for more and larger data. This table from a 1993 review shows just how small the trials were at the time.
Unsurprisingly, reviews and guidelines continued to show caution. In a 1994 editorial published in The Lancet, David Barnett offered that “Beta adrenergic blocking drugs are well known for their negative inotropic properties. For most medical practitioners their use in heart failure is not only strongly contraindicated but also possibly negligent.” That same year, a JAMA review concluded that “β-Blockers and first-generation calcium channel blockers must be used with caution because of their negative inotropic effect.” The ACC/AHA guideline was similarly cautious, writing, “At the present time, the use of β-blockers for the treatment of chronic heart failure remains investigational, but the official status of β-blockers may change as recent data are reviewed. Hence, physicians might consider the use of a β-blocker in selected patients with chronic heart failure. However, this approach should be taken only with considerable caution because the initiation of β-blockade may exacerbate heart failure in some patients.”
Things were about to change.
1995-Today: β-blockers are great? Yes, they are!
First came the Metoprolol for Dilated Cardiomyopathy (MDC) study. Published in 1993 in The Lancet, MDC randomized 383 patients with heart failure from idiopathic dilated cardiomyopathy (ejection fraction <40%) to either metoprolol or placebo. Though they found no difference in mortality at one year, fewer patients who received metoprolol required transplantation. The metoprolol group also showed greater improvement in ejection fraction and exercise time.
Next came the Cardiac Insufficiency Bisoprolol Study (CIBIS) trial, published in Circulation in 1994. Another 641 patients with heart failure with reduced ejection fraction (<40%) were randomized to bisoprolol or placebo. After a mean follow-up of 1.9 years, those randomized to bisoprolol showed more improvement in functional status and fewer hospitalizations for heart failure. But, again, mortality was no different.
And this was the main lament. Though many were coming around to the idea that β-blockers were indeed safe in heart failure, the cry for “show us a mortality benefit!” needed to be satiated.
Finally, in 1996, The New England Journal of Medicine published the US Carvedilol trial. Larger than the previous randomized trials, it included 1094 patients with heart failure, allocating them to either carvedilol or placebo. Mortality was cut by almost two-thirds (7.8% versus 3.2%). Reductions in cardiovascular hospitalizations and worsening heart failure were also seen in those randomized to carvedilol.
What followed was a deluge of additional studies showing benefit:
MERIT-HF (1999): 7% mortality for metoprolol versus 11% for placebo (mean follow-up 2 years)
CIBIS-2 (1999): 12% mortality for bisoprolol versus 17% for placebo (mean follow-up 1.3 years)
COPERNICUS (2001): 11% mortality for carvedilol versus 19% for placebo (mean follow-up 1.0 year)
BEST (2001): 30% mortality for bucindolol versus 33% for placebo - no difference (mean follow-up 2 years)
Guidelines quickly followed, with the 2001 ACA/AHA task force writing, “Most patients with symptomatic left ventricular dysfunction should be routinely managed with a combination of 4 types of drugs: a diuretic, an ACE inhibitor, a beta-adrenergic blocker, and (usually) digitalis.” Later, they offer the following Class I recommendation: “Beta-adrenergic blockade in all stable patients unless contraindicated (Level of Evidence: A).”
β-blockers had completed their 180, going from contraindicated between 1964 and 1996 to a must-give unless a contraindication could be identified.
In reading this literature, I was struck by Finn Waagstein's boldness. He gave β-blockers to patients despite a consensus belief that they were harmful. He was proved right, but one might imagine the opposite scenario playing out. Fortunately for patients with heart failure, Waagstein’s hypothesis was correct, and his cavalier treatment has become one of the core features of goal-directed medical therapy in heart failure.
Could you imagine such a scenario taking place in 2024?
I was a resident from 2006-2009, well before we had neprilysin inhibitors and SGLT2 inhibitors.
It’s also worth noting that many of the Swedish investigators who studied β-blockers in the late 1970s were the same ones who started studying ACE inhibitors. Karl Swedberg was the lead author of CONSENSUS.
In reading this review, I am reminded of the current state of sodium administration in acute heart failure. It seems that we’re in a “The Italian Group” versus the world stand-off.