Diltiazem vs metoprolol in patients with AF on direct acting oral anticoagulants. Is there a difference? Is it correlation or causation? Does it matter?
Great read. Just goes to show something I’ve been harping on for a while: namely afib is a cardio-cerebral disease, not a problem that can be cleanly partitioned between cardiology and neurology. I think most would agree with that framing, however, once neurologists accept responsibility for stroke prevention, they also need enough command of the surrounding cardiovascular literature to recognize when ‘non-neurologic’ management decisions materially alter neurologic outcomes.
Great article, I think the important point you brought up about this being only a 1 year study and that risk would increase over 5-10 years is an important one. When I'm worried about my patient on AC having a potential bleeding risk/event, I'm thinking in terms of years, even decades. A low absolute risk after 1 year is good to see but might not be representative of the whole story, especially if the KM curves keep diverging as the years progress.
As a PhD Pharmacologist and MD Emergency Physician, I am not surprised by the results of this study. And I am not surprised to see that too many physicians understand pharmacology/pharmacotherapuetics so poorly that this issue wasn't confronted in the prescribing recommendations or clinical practice guidelines (as this column seems to suggest). During my active clinical career, the common lack of knowledge/understanding of phamcology by many of my colleagues always impressed me - it demonstrated how inadequate was medical school phamcology education.
At what increase in absolute risk would you set a trigger in an electronic medical record to alert you to such a possible interaction for any two drugs. From my experience these minor increases in absolute risk become almost a black box risk in the eyes of the pharmacist.
Yes that happens all the time in propensity matching, which makes it non-reproducible, arbitrary, sometimes data-order-dependent, and inefficient. Patients who would normally be considered excellent matches are discarded because the patient they would match was already matched to an earlier patient. m:n matching with variable m and n would fix that but nobody goes to that trouble. Regression adjustment using ALL patients would have been much preferred.
I’m guessing that since verapamil has the same method of action as diltiazem (L-type calcium channel blocker), it would have the same effect on the CYP3A4 metabolism.
Were the genetic variants of CYP3A4 (and possibly CYP3A5) taken into account in this study, especially to ensure that they were not responsible for the observed differences?
I bet all of the authors of the study are receiving stipends from Bristol-Myers Squibb. I can’t document it because they hide the funding so well, but someone younger and smarter than me can probably pull up the numbers.
It’s a very good article. It brings up some good principles. But the biggest principal is mentioned by some of the commenters in that the study is more about promoting one medicine over the other than it is about good clinical care for patients. Beta blockers didn’t slow my a fib, diltiazem did. That’s what counts.
Is also a recent article comparing the two an outpatient treatment of pulmonary embolism. Ella crisp was also found to be superior, but they were comparing apples to oranges. At the initiation of treatment they gave the Eliquis for a week, but they gave the xarelto for three weeks. That’s when all the bleeding occurred. The editorialist picked up on that concept, but then he caved to Bristol Meyers. There’s a lot of balder dash in pharmaceutical company driven medicine.
The bleeding rates observed here are much higher than the 2018 cancelling of Aspirin with the 100 mg daily dose out of UK and Australia published in NEJM 10-2018. Now apixaban is $6 billion a year in revenue!! The same group that published the Aspree has now again found that aspirin doesn't have the cancer benefits originally ascribed. Yet, I cannot cancel aspirin after 40 years of data!!
From https://www.nejm.org/doi/full/10.1056/NEJMoa1805819 Major hemorrhage: 8.6 events per 1,000 person-years in the aspirin group vs. 6.2 events per 1,000 person-years in the placebo group (p < 0.001). One conclusion is The use of low-dose aspirin as a primary prevention strategy in older adults resulted in a significantly higher risk of major hemorrhage and did not result in a significantly lower risk of cardiovascular disease than placebo. Yet, the absolute risk of bleeding was not much different! A cooridinated strike on Aspirin!!!
So now most physicians do not recommend low-dose aspirin and place people on expensive blood thinners that have more bleeding than aspirin! Something is rotten in Denmark!!
I agree with your assessment of this study, John. This is the type of observational study that has a high likelihood of providing causal inference, mainly because of your fundamental observation that IN PRACTICE, the decision between these 2 different drugs is unlikely to be based on important clinical differences (both measured and unmeasured) between the 2 groups of patients. And as I have explained previously, this type of insight is only possible for clinicians who understand what types of considerations go into these medical decisions. In short, this study is a natural experiment much more than a propensity-matched study.
The other methodologic strength of this study was the ability of the investigators to take advantage of a distinct event (initiation of a DOAC) as time zero, which is a huge problem for many observational comparative effectiveness studies.
Having worked with the authors in Reference 5 on the apixaban clinical pharmacology program from first-in-human until I retired 3 years ago, I can attest to the goodness of this study. It’s spot on and, again, distinguishes apixaban as the DOAC of choice for AF and other indications.
Pharmacology was the most intense course I have ever taken. One year long . Every exam was cumulative meaning anything covered test on week one was fair game to be repeated week 40. No Multiple Choice!! They would give you a drug and you would list 5 side effects. Same for metabolism. Given the instant access to information we have now I doubt this it is still taught this way. It did however make my life much easier practicing medicine- as i had instant access as well!
Great read. Just goes to show something I’ve been harping on for a while: namely afib is a cardio-cerebral disease, not a problem that can be cleanly partitioned between cardiology and neurology. I think most would agree with that framing, however, once neurologists accept responsibility for stroke prevention, they also need enough command of the surrounding cardiovascular literature to recognize when ‘non-neurologic’ management decisions materially alter neurologic outcomes.
Since I'm on both Xarelto and metoprolol I thank Dr Mandrola for his sage analysis. :)
Great article, I think the important point you brought up about this being only a 1 year study and that risk would increase over 5-10 years is an important one. When I'm worried about my patient on AC having a potential bleeding risk/event, I'm thinking in terms of years, even decades. A low absolute risk after 1 year is good to see but might not be representative of the whole story, especially if the KM curves keep diverging as the years progress.
Thanks for sharing Dr. Mandrola
Given the dose-response relationship demonstrated in the study, would you anticipate absolute risk increasing over time? Just curious, thanks!
As a PhD Pharmacologist and MD Emergency Physician, I am not surprised by the results of this study. And I am not surprised to see that too many physicians understand pharmacology/pharmacotherapuetics so poorly that this issue wasn't confronted in the prescribing recommendations or clinical practice guidelines (as this column seems to suggest). During my active clinical career, the common lack of knowledge/understanding of phamcology by many of my colleagues always impressed me - it demonstrated how inadequate was medical school phamcology education.
34 and 29 per 1000 pt-years is 3.4 and 2.9 per 100 pt-years
At what increase in absolute risk would you set a trigger in an electronic medical record to alert you to such a possible interaction for any two drugs. From my experience these minor increases in absolute risk become almost a black box risk in the eyes of the pharmacist.
The authors started with 85,000 Metoprolol patients and only analyzed 23,000 of them. Is that scientific?
propensity matching, no? Does that happen regularly with uneven groups? I am asking.
Yes that happens all the time in propensity matching, which makes it non-reproducible, arbitrary, sometimes data-order-dependent, and inefficient. Patients who would normally be considered excellent matches are discarded because the patient they would match was already matched to an earlier patient. m:n matching with variable m and n would fix that but nobody goes to that trouble. Regression adjustment using ALL patients would have been much preferred.
I’m guessing that since verapamil has the same method of action as diltiazem (L-type calcium channel blocker), it would have the same effect on the CYP3A4 metabolism.
Were the genetic variants of CYP3A4 (and possibly CYP3A5) taken into account in this study, especially to ensure that they were not responsible for the observed differences?
I bet all of the authors of the study are receiving stipends from Bristol-Myers Squibb. I can’t document it because they hide the funding so well, but someone younger and smarter than me can probably pull up the numbers.
It’s a very good article. It brings up some good principles. But the biggest principal is mentioned by some of the commenters in that the study is more about promoting one medicine over the other than it is about good clinical care for patients. Beta blockers didn’t slow my a fib, diltiazem did. That’s what counts.
Is also a recent article comparing the two an outpatient treatment of pulmonary embolism. Ella crisp was also found to be superior, but they were comparing apples to oranges. At the initiation of treatment they gave the Eliquis for a week, but they gave the xarelto for three weeks. That’s when all the bleeding occurred. The editorialist picked up on that concept, but then he caved to Bristol Meyers. There’s a lot of balder dash in pharmaceutical company driven medicine.
The bleeding rates observed here are much higher than the 2018 cancelling of Aspirin with the 100 mg daily dose out of UK and Australia published in NEJM 10-2018. Now apixaban is $6 billion a year in revenue!! The same group that published the Aspree has now again found that aspirin doesn't have the cancer benefits originally ascribed. Yet, I cannot cancel aspirin after 40 years of data!!
From https://www.nejm.org/doi/full/10.1056/NEJMoa1805819 Major hemorrhage: 8.6 events per 1,000 person-years in the aspirin group vs. 6.2 events per 1,000 person-years in the placebo group (p < 0.001). One conclusion is The use of low-dose aspirin as a primary prevention strategy in older adults resulted in a significantly higher risk of major hemorrhage and did not result in a significantly lower risk of cardiovascular disease than placebo. Yet, the absolute risk of bleeding was not much different! A cooridinated strike on Aspirin!!!
So now most physicians do not recommend low-dose aspirin and place people on expensive blood thinners that have more bleeding than aspirin! Something is rotten in Denmark!!
An excellent study. Thanks for analyzing. Kudos to Vanderbilt IM section for doing the study.
I agree with your assessment of this study, John. This is the type of observational study that has a high likelihood of providing causal inference, mainly because of your fundamental observation that IN PRACTICE, the decision between these 2 different drugs is unlikely to be based on important clinical differences (both measured and unmeasured) between the 2 groups of patients. And as I have explained previously, this type of insight is only possible for clinicians who understand what types of considerations go into these medical decisions. In short, this study is a natural experiment much more than a propensity-matched study.
The other methodologic strength of this study was the ability of the investigators to take advantage of a distinct event (initiation of a DOAC) as time zero, which is a huge problem for many observational comparative effectiveness studies.
Having worked with the authors in Reference 5 on the apixaban clinical pharmacology program from first-in-human until I retired 3 years ago, I can attest to the goodness of this study. It’s spot on and, again, distinguishes apixaban as the DOAC of choice for AF and other indications.
Pharmacology was the most intense course I have ever taken. One year long . Every exam was cumulative meaning anything covered test on week one was fair game to be repeated week 40. No Multiple Choice!! They would give you a drug and you would list 5 side effects. Same for metabolism. Given the instant access to information we have now I doubt this it is still taught this way. It did however make my life much easier practicing medicine- as i had instant access as well!
You are correct, I think, in the change in teaching methods and I suspect the brains of those taught now will not be as helpfully changed.
Thank you. Very helpful.