This week’s article by Cheng et al (JACC, 74 (18)) regarding the identification of predictors of early mortality post-atrial fibrillation (AF) ablation underscores key indicators affecting mortality rates for these patients. Defining early mortality as death during the origin admission or 30-day readmission, Cheng and colleagues evaluated 60,203 admissions for AF ablation in patients 18 years of age or older between 2010 and 2015. Early mortality following AF ablation affected nearly 1 in 200 patients, with procedure complications, congestive heart failure, and procedures conducted at hospitals with low AF ablation volume identified as predictors.
Prompt recognition of procedural complications and effective management is critical to prevent early death in these patients. Dr. Peter Noseworthy discusses complication management in his updated presentation “Atrial Fibrillation”.
IM Essentials learners can view the full lecture video:
Want To Try IM Essentials Content?
Faculty: Dr. Peter Noseworthy
The learning objective of this presentation is to design appropriate management strategies for patients with atrial fibrillation.
Assessing Stroke Risk in AF
Everybody in the [room] knows that atrial fibrillation increases the risk of stroke considerably. It's a five-fold increase over the general population, but the risk is not uniform.
So our job in the clinic is to look at that patient with atrial fibrillation and risk-stratify them and find the patients in whom anticoagulation is warranted in terms of the increased risk of bleeding. And we determine that not necessarily by the amount of atrial fibrillation or the pattern of atrial fibrillation or the burden, but rather by the underlying risk factors that it's tracking with. So they're patient factors rather than arrhythmia factors.
For a long time, we used the CHADS2 score.
That was hypertension, heart failure, age, diabetes, and stroke; getting 2 points for stroke. But all the more recent guidelines favor the CHA2DS2-VASc score over the CHADS2 score, and it's because it is a little bit more granular in the way it assesses stroke risk.
It acknowledges that nothing magic happens at 65. In fact, there's a graded increase of stroke risk with age and it gets at that by giving 2 points over 75 and 1 point for that intermediate rate. It acknowledges that women are at slightly higher risk of stroke particularly if they have other risk factors and it includes vascular disease.
If we add up these scores, we can make an assessment of what the overall stroke rate would be like without anticoagulation. And, if we compare the CHA2DS2-VASc score to the CHADS2 score, the real value is in the low end of the risk spectrum.
A patient who has none of those CHA2DS2-VASc risk factors really has a very low risk of stroke. Whereas a CHADS2 of zero, which may include people with peripheral vascular disease, women, people who are over 65 actually have some stroke risk in that range.
So, a CHA2DS2-VASc score of zero is a powerful number to recognize because those patients do not require anticoagulation. You only expose them to the risk of bleeding. Because the event rates are so low, there's really not a lot of benefit.
The current way we approach who to anticoagulate based on the most recent guideline updates just came out in January. So the low-risk group, we used to just say CHA2DS2-VASc of zero, but what we're realizing is that, if the only risk factor a patient has is that they're a woman and they have no other risk factors, they're still in that lowest risk group and they don't require anticoagulation.
If they have at least one risk factor—and then this should say two for women, not one, then they're in the intermediate risk. And there we have to sit down with the patient and figure out what makes sense for them based on their context and their preference and so forth. But anybody who has at least two CHA2DS2-VASc risk factors really should be anticoagulated.
The most recent guideline updates also mentioned is that we should not be using aspirin for stroke prevention.
And it's probably true that there's a very modest reduction in stroke with aspirin but it's on the order of 10% reduction and it's trivial and it comes with bleeding risks that are comparable to the safer NOACs that are on the market right now. We know from the AVERROES trial that apixaban, for instance, was associated with very similar bleeding risk to aspirin. So taken aspirin off because, in anybody who really has any significant stroke risk, you should be using an anticoagulant and aspirin kind of gave us all an out in those intermediate patients or patients who are worried about bleeding. You kind of feel like, "Well, I'll just give you an aspirin. At least we get some protection and we all feel good," but really that's probably just exposing patients to risk of bleeding with very little benefit so that's no longer suggested.
You'll see, when you see patients in clinic—I do when I see patients that I've been following for years that are on the low-risk, they're off on an aspirin because we used to always recommend that. So that's something you can kind of think about as you see patients in your office. Reassess in the sort of NOAC era which drugs to take.
So, which agent should we use? For five decades or more, there was only one option; that was warfarin. And you either anticoagulated with warfarin or you didn't, and it was cumbersome. But now we have a number of other options and increasingly these are the dominant agents for initiation of anticoagulation for stroke prevention in atrial fibrillation. So dabigatran, rivaroxaban, apixaban, and edoxaban. And right now apixaban and rivaroxaban make up the lion's share of new prescriptions. These drugs are at least as good and at least as safe as warfarin.
Looking at a metanalysis for stroke or systemic embolism, they're all either equivalent or better and as a class are better than warfarin for stroke prevention. For major bleeding, they're at least comparable and apixaban and edoxaban particularly at low dose are better for bleeding, overall slightly better. And the big difference is in the risk of intracranial bleeding which is the most feared complication, and anybody who's had a patient who's had intracranial bleed on warfarin, you know that this can be devastating. The risk with neuro agents is about half of that as what we would expect with warfarin. The caveat is that the event rates are very low. You'd have to switch 200 of your patients on warfarin and treat them for a year to prevent one intracranial bleed, at significant cost, and so forth. But, if you're starting a new agent, this does play into the calculation at least for me.
Choice of Anticoagulant
So warfarin is still the drug of choice for some patients. Those who have valvular atrial fibrillation...I use a relatively narrow definition of valvular atrial fibrillation to mean that they have a mechanical valve for which they really should be anticoagulated anyway or they have rheumatic mitral stenosis and those patients who are at very high risk of thromboembolism. Warfarin is probably used preferentially in patients with end-stage renal disease on dialysis. And for many patients, even though they've been on the market for a while, these other drugs are somewhat cost-prohibitive particularly for a daily medication that's going to be taken over somebody's entire life.
I think the NOACs are reasonable agents of choice if there's no contraindication, if they can be afforded or they're covered by the drug plans. There are a few minor caveats like we tend not to use dabigatran in very elderly patients due to increased risk of bleeding. Edoxaban is renally cleared. And, if you have a GFR over 95, you may actually clear it and you may be relatively underdosed, so we don't use that drug although we use edoxaban infrequently in the United States anyway. So we're in the category here of either choosing between rate control or rhythm control. Rate control, we either use beta-blockers, calcium channel blockers, occasionally digoxin for patients who are in chronic or long-standing persistent atrial fibrillation.
And an extreme version of rate control is AV nodal ablation and pacemaker implantation. That does nothing to affect stroke rate. It doesn't get rid of the atrial fibrillation but it allows you to have control over the heart rate by putting in a pacemaker, severing the atrium from the ventricle and rendering them dependent on their pacemaker, and then they no longer have the rapid rates and the irregularity that can drive symptoms. It's kind of sacrilegious for an EP to ablate the AV node. We try to avoid that but there are patients who benefit from that strategy.
If you want to try to maintain sinus rhythm, we have either antiarrhythmic drugs or catheter ablation and I'll go through this in a little bit more detail. Who, in the room, has used an antiarrhythmic drug in the past 3 months either prescribed flecainide or amiodarone or...? Yeah, so actually a fair number of you. It's one of those things that unless you do it frequently it doesn't feel comfortable, and you may want to be prescribing these in collaboration with somebody in cardiology or even EP. But if you're seeing enough of these patients, it's worthwhile being aware of how these drugs work and what the contraindications and implications of their use are.
There's basically two types of antiarrhythmic drugs. They're either the class I agents which are sodium channel blockers or the class III agents which are potassium channel blockers. The sodium channel blockers, the prototype of which are either flecainide or propafenone, they slow conduction. So the analogy I like to make is if you have a hot tub with all kinds of white water but then you turn down the jets by turning down the conduction velocity. It sort of calms all those eddy currents and it can help you restore and maintain sinus rhythm. Because they slow conduction, they result in a lengthening of the QRS complex because that's the depolarization wavefront as it goes through the myocardium. So if you're looking for toxicity of the sodium channel blocker, look at the EKG and look at the QRS duration. You probably have seen patients in the ICU with a tricyclic overdose which has some certain channel blocking effect and you probably remember from that that you look at the QRS duration. It's the same thing when you're monitoring the safety of a class IC agent. Look at the QRS duration, not the QT.
Now, because those drugs slow conduction, if there's any scar in the heart, then that slow wavefront can snake its way around the scar and result in a monomorphic ventricular tachycardia. So, those can be proarrhythmic because they can cause ventricular tachycardia. So we never use those drugs in people who have any scar in the heart for any reason, particularly myocardial infarction. And a few people in the room will remember the CAST trial during which sodium channel blockers were used post-infarct to suppress PVCs and it increased mortality and it caused sudden death related to ventricular tachycardia.
Now, class III agents worked totally differently. The increased tissue refractoriness. So they lengthen the QT interval. And because of that, the tissue is more refractory and there's less opportunity for that reentry. So it's a different mechanism but, instead of slowing conduction, the tissue is more refractory. The action potential duration is longer. As a result, if you're monitoring the effect of a class III agent, you look at the QT interval. So that's why we hospitalize patients to initiate sotalol or dofetilide, and it's why we measure the QT as we're titrating the effect and looking for proarrhythmia. So, the danger of a QT-prolonging medication is monomorphic VT due to reentry but torsade due to polymorphic ventricular tachycardia and QT prolongation.
Atrial Fibrillation: Choice of AAD
How do we choose an antiarrhythmic drug for a given patient? This is where the rubber hits the road. So, if they have no heart disease at all, then all the drugs are on the table. You can use whatever you like, but we tend to use flecainide or propafenone for atrial fibrillation first because it can be started as an outpatient. It's generally quite well-tolerated, and it has a better safety profile. Second-line agents would be dofetilide or amiodarone. Amiodarone has toxicity so we like to try to avoid... Now, if they have coronary artery disease, then we assume that there might be some myocardial scar and these patients cannot be on a class I agent so we have to treat them with dofetilide or sotalol. These patients are hospitalized for that and we monitor their QT interval. And if they have heart failure, we want to avoid the negative ionotropic properties of sotalol so we would use preferentially either amiodarone or dofetilide.
Complications of Catheter Ablation for AF
Now, it seems like an easy thing to do to do an ablation but there are considerable risks of the procedure and it's important for everybody in the room to be aware of these because they're often not immediate. We see things like growing complications, pericardial effusion, and we know that, while we're doing the procedure, but after we're done and several weeks later, some of these symptoms start to appear and it's often the person who's seeing the patient in follow-up who picks these up. So, there are things like the pulmonary veins can narrow if our ablation is too close to the pulmonary vein. That can cause sort of a regional pulmonary edema, shortness of breath, cough. The most devastating complication is an injury to the esophagus and that's because, as we heat the left atrium, if the esophagus is heated, it can develop a little bit of an ulcer. And then over time, if that ulcer forms a track, it can actually form a fistula to the left atrium. That can be a fatal complication. It's very rare. We tend to think it's about 1 in 5,000, but it occurs about 6 weeks after the ablation procedure.
And certainly for us at Mayo, if we're seeing patients and then they go back and are followed locally, this is somebody that needs to be picked up by the primary care physicians.
You would think it would cause bleeding if you have a fistula between the atrium and the esophagus but it tends not to. Instead, what it ends to cause is introduction of material from the esophagus and polymicrobial sepsis. So it looks like endocarditis where they can even get air emboli but it can be a very indolent and life threating condition. It would require surgery.
2014 AHA/ACC/HRS Guideline
Right now, catheter ablation is recommended as class I for symptomatic paroxysmal atrial fibrillation in patients who are refractory to antiarrhythmic drugs. And it's a class IIa as an initial strategy for paroxysmal atrial fibrillation or for persistent patients who are refractory to drugs. That's maybe a little in the weeds for the internal medicine boards. But in your practice, you can think about patients who either have highly symptomatic paroxysmal atrial fibrillation who don't want to take the drug. You could send them. Or, if they're failing other strategies, they could be considered for catheter ablation.
Weighing The Options
Antiarrhythmic drugs are less costly upfront. That avoids the procedure but there are side effects of these drugs. We talked about proarrhythmia. They're not terribly effective, about a 50% success rate at one year. Ablation is more cumbersome at least upfront. There's high upfront cost. There's a 2% to 3% complication risk from the procedure itself. But, in the long run, if it's effective, people can avoid daily suppressive medications and allows us to simplify their atrial fibrillation care.