A large cohort study found that among patients admitted with heart failure, the adjusted risk of death in the following 12 months was similar among patients taking atenolol and carvedilol, but slightly higher in those taking metoprolol tartrate.
This new data supports the NICE recommendations on the choice of beta-blocker in heart failure. People who develop heart failure due to left ventricular systolic dysfunction and who are not previously taking a beta-blocker should normally be offered a beta-blocker licensed for heart failure (bisoprolol, carvedilol or nevibolol). This should be offered after diuretic and ACE inhibitor therapy (regardless of whether or not symptoms persist). However, people who are already on treatment with another beta-blocker for a concomitant condition (for example, angina or hypertension) should continue with a beta-blocker – either their current beta-blocker or an alternative licensed for heart failure treatment.
Prescribers should continue to follow NICE guidance on chronic heart failure.
What is the background to this?
Some beta-blockers (bisoprolol, carvedilol, nevibolol and extended-release metoprolol succinate [not licensed in the UK]) have been shown to be beneficial in people with heart failure compared to placebo. However, atenolol is widely used in the UK for other cardiovascular indications but is not licensed for use in heart failure
What does this study claim?
This was a cohort study of 11,326 patients who survived hospitalisation for heart failure, of whom 7,976 received beta-blockers. After adjustment for confounders and differences in patient characteristics, the risk of death compared with atenolol was higher for metoprolol tartrate (adjusted hazard ratio [HR] 1.16; 95% confidence interval [CI] 1.01 to 1.34) and no beta-blockers (HR 1.63; 95% CI, 1.44 to 1.84) but was not significantly different for carvedilol (HR 1.16; 95% CI 0.92 to 1.44).
How does this study fit with other evidence?
Compared with placebo, bisoprolol, carvedilol and extended-release metoprolol succinate have been shown to reduce mortality in people with heart failure (the CIBIS II, COPERNICUS and MERIT-HF trials, respectively). In the SENIORS trial, nevibolol reduced the risk of a composite of death or hospitalisation from cardiovascular causes compared with placebo. The only comparative RCT of beta-blockers was the COMET trial, in which carvedilol was shown to be superior to standard-release metoprolol tartrate (HR 0.83, 95%CI 0.74 to 0.93, P=0.002).
This study provides some reassurance that in patients who are already taking atenolol, and who develop heart failure, continuing with the atenolol is a reasonable option, in line with NICE guidance on chronic heart failure.
The study does not provide data on the relative merits of switching from a beta-blocker that is not licensed for treatment of heart failure to one that is. It indicates only that mortality in patients taking atenolol (but perhaps not metoprolol tartrate) was similar to that in those taking carvedilol. However, this study looked only at the first year after diagnosis, and the CIs around the HR were wide.
Given the limitations of observational data, it does not provide grounds to change NICE guidance, i.e. that, all other things being equal, a beta-blocker licensed for use in heart failure is the preferred option in people who develop heart failure and have not previously taken a beta-blocker.
Well conducted RCTs would be necessary to answer definitively the questions which remain: is atenolol really as effective as carvedilol in reducing mortality in people with heart failure; are patients who were taking atenolol or other beta-blockers not licensed for use in heart failure before developing it better off or worse off if they change to a licensed beta-blocker, or does it make no difference?
Another cohort study with broadly similar conclusions was published in the same edition of the journal. However, patients had to survive for at least 30 days after discharge to be included, and also their cohort group (different type of beta-blocker, or no beta-blocker) was determined by them filling a prescription in this first thirty days after discharge. Although this study provides additional support for the conclusions of the first study, we suspect it may have been subject to immortal time bias: we have blogged about this bias in the past.
Patients: 11,326 adults patients from Kaiser Permanente Northern California and Harvard Pilgrim Healthcare (operating in New England) who were hospitalized between January 1, 2001, and December 31, 2003, with a primary discharge diagnosis of heart failure for whom data was available for 12 months before admission and for at least 12 months after admission (or to death).
Intervention and comparison: Using pharmacy data, patient exposure was classified as one of the following atenolol, metoprolol tartrate, carvedilol, other beta-blockers, and no beta-blocker. The time exposed to each beta-blocker was calculated for each individual: 1191 (14.9%) of patients received more than one type of beta-blocker during the follow-uppatient characteristics including demographic data and medication history. Death was identified from administrative, state mortality, and Social Security Administration databases
Outcomes: The crude rate (per 100 person-years) of death was lowest while receiving carvedilol (17.7), followed by atenolol (20.1), other beta-blockers (21.9), metoprolol tartrate (22.8), and no beta-blocker (37.0). Compared with atenolol, observed differences were significant only for metoprolol tartrate (P=0.04) and for periods of not using beta-blockers (P<0.001). After adjustment for confounders and the propensity to receive carvedilol, the risk of death compared with atenolol was higher for metoprolol tartrate (adjusted HR 1.16; 95% CI, 1.01 to 1.34) and no beta-blockers (HR1.63; 95% CI 1.44 to 1.84) but was not significantly different for carvedilol (HR 1.16; 95% CI 0.92 to 1.44).
Sponsorship: This project was funded under contract HHSA290-2005-0033-I from the Agency for Healthcare Research and Quality, US Department of Health and Human Services as part of the Developing Evidence to Inform Decisions about Effectiveness (DEcIDE) program.