We’ve been all over the road with beta-blockers and COPD. For a long time, there was a belief that beta-blockers were contraindicated in COPD. Eventually that concept fell out of favor. Subsequently, some correlative data suggested that beta-blockers might be beneficial in COPD. This led to the current BLOCK-COPD trial which (spoiler alert) shows that metoprolol isn’t beneficial for COPD. Indeed, this study raises concerns about the safety of metoprolol in COPD, which actually puts us back to where we were initially! We’ve come full circle on the beta-blocker roller coaster.
BLOCK-COPD trial: design
This study was designed to test the concept that beta-blockers could reduce the incidence of COPD exacerbation. Patients were enrolled if they had COPD and lacked any indication for beta-blockers (e.g., prior myocardial infarction or systolic heart failure). The primary endpoint was median time until a COPD exacerbation.
The study was stopped prematurely based on a combination of futility (very low likelihood that the trial could possibly show benefit from metoprolol) and some concerns regarding safety. Premature termination increases the likelihood of spurious findings due to transient statistical anomalies (which would have regressed towards the mean, had the study been completed).
At baseline, patients in the metoprolol group may have been a bit sicker. They had a higher rate of COPD exacerbation within the year prior to study enrollment (63% vs. 50%, p=0.005).
The primary endpoint was time to first COPD exacerbation. This was exactly the same between groups:
In the study protocol (published at clinicaltrials.gov) the following endpoints are listed:
- Number of COPD exacerbations in a year
- Number of ED visits for COPD over a year
- Number of hospital admissions for COPD over a year
- Number of hospital days due to COPD exacerbations over a year
- Major adverse coronary events (MACE) over a year
- All-cause mortality
- Incidence of presumed metoprolol side-effects
- Short form health survey 36
- Modified Medical Research Council dyspnea scale
- Forced expiratory volume in one second (FEV1)
- Exercise capacity in six minutes (six-minute walk distance)
- Cardiac biomarkers
- Brain natriuretic peptide (BNP)
- High-sensitivity CRP
- Saint George’s Respiratory Questionnaire
- COPD assessment test
- San Diego Shortness of Breath Questionnaire
- Combined rate of acute exacerbations of COPD and major adverse coronary events (MACE)
- “Serum biomarkers”
So, there are a lot of secondary endpoints. The vast majority of these secondary endpoints were negative. Many of them disappeared from the manuscript entirely. There were no changes in mortality or all-cause hospitalization. Objective measurements of lung function were the same (e.g. spirometry and six-minute walk distance).
The authors report that metoprolol caused an increase in dyspnea based on two subjective dyspnea scales (San Diego Shortness of Breath Score and the COPD Assessment Test). However, looking at the totality of the subjective data, this may just represent statistical noise:
The other difference was that patients in the metoprolol group had a higher rate of COPD-related hospitalization (defined below as a “severe” or “very severe” exacerbation of COPD):
interpretation of a positive secondary endpoint when the primary endpoint is negative?
This is always a bit complicated. A common interpretation is that any study with a negative primary endpoint is negative, so the secondary endpoints are solely for hypothesis generation. According to this logic, Christopher Columbus’s voyage was negative, because he failed to reach China (his pre-specified endpoint).
My preference is to use some judgement in these studies, based on numerous factors (e.g. pre-test probability that the hypothesis is valid and the overall constellation of data findings). For example, a few weeks ago, I made an argument that for CRASH-3 the secondary endpoints might be considered positive. The opposite argument may be most appropriate here.
There are several reasons why this secondary endpoint shouldn’t be taken too seriously:
- The study was stopped prematurely, due largely to futility. As discussed earlier, premature termination increases the likelihood of obtaining spurious results due to transient statistical fluctuations.
- The results are not statistically robust (especially considering the myriad of secondary endpoints). If correction for multiple comparisons was performed, the results would probably be deemed statistically insignificant.
- There is no coherent pattern of harm across the numerous secondary endpoints. For example, there are trends toward fewer mild COPD exacerbations and fewer moderate COPD exacerbations in the metoprolol group.
- The possibility that beta-blockers may increase COPD exacerbation isn’t consistent with prior clinical data (in fact, the authors initially postulated the exact opposite!).
- Baseline imbalances exist between the two patient groups. Patients in the metoprolol group may have been sicker (with a substantially higher rate of COPD exacerbation in the year prior to study enrollment).
publication in NEJM and associated hype
The authors presented these results in a rather dark light:
The trend in mortality is mentioned here, which seems to imply that metoprolol might be causing an increased mortality. However, this mortality difference doesn’t come anywhere close to statistical significance (p = 0.2). As such, focusing on this trend within the abstract seems a bit irresponsible.
This study will almost certainly be misinterpreted to mean that beta-blockers are unsafe to give in patients with COPD. For example, the following punchline has already appeared on MedPage Today:
- The use of beta-blockers in COPD has been subject to repeated reversals over the past few decades. Unfortunately, this pattern shows no signs of abating today.
- The BLOCK-COPD trial tests the hypothesis that metoprolol could be used to prevent COPD exacerbations. This is a strange hypothesis, so it comes as little surprise that the primary endpoint was neutral (metoprolol had no effect on the time delay before a COPD exacerbation).
- Among numerous secondary endpoints, there was an increase in the rate of severe exacerbations within the metoprolol group. This isolated secondary endpoint is of dubious significance for several reasons – it’s statistically weak, the study was stopped prematurely, baseline imbalances exist between the groups, and there isn’t a coherent signal of harm across multiple secondary endpoints.
- This study should not change practice. Cardioselective beta-blockers are generally safe among patients with COPD.
- Beta-blockers aren’t completely benign medications: they do have some side-effects. Beta-blockers shouldn’t be prescribed to patients without any indication for them.
- BLOCK-COPD trial
- Pubmed link here.
- Beta-blockers best avoided in COPD patients without cardiovascular disease (MedPage today)
- Prior material on beta-blockers in COPD
- Beta-blockers are safe for most patients with asthma and COPD? PulmCCM 2014
- Beta-blocker myths (CHEST physician, Douglas Paauw)
- COPD and Beta-blockers: another myth dispensed… (John Mandrola)