Flecainide or Sotalol: Antiarrhythmic Choice Without a Confirmed Diagnosis
Three converging cardiovascular risk factors, but no confirmed structural or ischemic heart disease — the case depends on how much testing is needed before a Class IC agent can safely be started.
D.K., a 56-year-old man who has spent the better part of three decades building up his own small landscaping business, first noticed his heart "flip and race" while loading mulch onto a trailer four months ago — an episode that lasted almost two hours before settling on its own. A nearly identical one, caught a week later on a pharmacy blood-pressure cuff at 156 beats per minute and irregular, sent him to the emergency department, where atrial fibrillation was confirmed; it converted on its own before any drug was given. He has had roughly four more episodes since, each lasting one to several hours, usually on the tail end of a physically demanding workday. He describes himself as otherwise in solid health for his age — he still does the heavy lifting the business requires most days — though he has carried a hypertension diagnosis for about eight years, well controlled on his current regimen, and was found to have type 2 diabetes two years ago on a routine physical, managed since with diet and a single oral medication, his last A1c 6.8%. His LDL has run in the 120s despite a statin, and his father had a heart attack at 51. He has never had chest pain, shortness of breath on a ladder, or any exertional symptom beyond the palpitations themselves.
The paroxysms are frequent and disruptive enough — roughly monthly, sometimes swallowing most of a workday — that he wants more than rate control, and his cardiology team agrees a trial of a rhythm-control drug is reasonable. The plan was flecainide, chosen first specifically because it avoids sotalol's monitoring burden — until the same three risk factors that brought him to a cardiologist in the first place surfaced as a real complication of that choice. Flecainide's mortality signal in the CAST trial was demonstrated in patients with a prior myocardial infarction and a reduced ejection fraction, not in anyone who looks like D.K. on paper — but "looks like him on paper" is exactly the problem: his echocardiogram is normal (LVEF 62%, no LVH, no wall-motion abnormality), and a normal echo was never designed to rule out early coronary disease. Hypertension, diabetes, and a premature paternal MI are three real, converging markers that some fraction of patients who look exactly like this actually carry underneath a clean study. The question the team is now working through isn't whether flecainide is contraindicated — nothing confirms that — it's how much reassurance a normal echo can honestly provide against a risk profile it was never built to screen.
Before starting a drug the guideline built a warning label around
I'd start flecainide today. The mortality signal that put Class IC agents under a warning label came from CAST — patients with a recent myocardial infarction and a depressed ejection fraction, given encainide or flecainide to suppress asymptomatic ventricular ectopy. D.K. has neither a prior infarction nor a depressed ejection fraction; his echo is normal in every dimension that trial actually measured. The guideline's own language withholds Class IC for structural or ischemic heart disease — not for hypertension, not for diabetes, not for a father's heart attack at 51. Those are real risk factors for developing disease someday. They are not evidence he has it now.
If his echo had shown reduced function or a wall-motion abnormality, I wouldn't be arguing this. This case is entirely about what a genuinely normal study is allowed to mean.
A normal echocardiogram tells you his ventricle is working. It was never built to tell you whether his coronary arteries are clean, and that's precisely the organ this decision actually depends on. He's carrying three independent, well-established markers of atherosclerotic risk at an age where covert coronary disease is not rare — and a premature paternal MI in particular tends to under-count, not over-count, real risk. A coronary CT angiogram directly answers the one question his echo cannot: whether there's disease there at all. It costs one appointment. Committing to a drug whose defining risk depends on that answer, without first getting the one test built to give it, isn't caution — it's skipping the test because the rest of the workup happened to come back clean.
I recognize CAST itself never enrolled a patient like him. That's not the same as saying the extrapolation doesn't apply — the guideline extrapolated it precisely because no one has run that trial in exactly this population, and I'd rather find out than assume the gap in the evidence resolves in our favor.
I don't think a clean CTA closes this the way it sounds like it would. Coronary CT is excellent at ruling out obstructive disease; it is not perfect at ruling out the diffuse, non-obstructive atherosclerosis that a risk profile like his would predict first, well before anything flow-limiting develops. If it comes back clean, we've narrowed the probability, not eliminated it. Given that the actual failure mode being screened for is a mortality signal, I'd rather choose the drug whose risks don't depend on that answer at all. Sotalol is not a free alternative — it needs baseline and follow-up QTc measurement, its dosing tracks renal function, and its FDA label requires at least three days of monitored inpatient initiation for exactly this indication. But every one of those costs is known, measurable, and manageable in a way "the CTA was probably enough" is not.
Agreed: a coronary CT angiogram before any antiarrhythmic is started, continuing metoprolol succinate for rate control through the paroxysms in the meantime.
Not agreed, and left as an explicit branch rather than a single expectation:
Flecainide starts as planned; the electrophysiologist and preventive cardiologist both accept this as sufficient reassurance.
The electrophysiologist would still favor flecainide, arguing CAST's population had scarred, poorly functioning myocardium, not incidental plaque. The clinical pharmacologist would move to sotalol regardless, arguing any confirmed coronary disease reopens the question CAST was built to answer. No default was set for this branch — it is revisited when the result is in hand.
All three agreed on one thing without qualification: whichever drug is started, it will be after the CTA, not instead of it.