Let’s start with the number that ends the debate: 13.8%.
That’s the LAA thrombus rate found on TEE in the ACUTE trial — 1,222 patients, AF duration over two days, unselected. One in seven. Not a rare finding. Not an edge case. One in seven patients who present for cardioversion of AF lasting more than 48 hours has a clot that imaging will find.
If you cardiovert without looking, you don’t know which one in seven you have.
The 48-Hour Rule Is Not What You Think It Is
The 48-hour threshold was never designed to define a population where cardioversion is safe. It was a pragmatic estimate of how long organized thrombus takes to form. It’s a time frame, not a guarantee.
In practice, the stroke risk with cardioversion of AF presumed to be under 48 hours is 0.7 to 1.1%. That’s not zero. And it concentrates almost entirely in patients with thromboembolic risk factors.
“Presumed onset” is also doing a lot of work in that sentence. Plenty of patients presenting to the ED with “new AF” have had paroxysmal or silent episodes longer than they report. They don’t know. You don’t know. The 48-hour cutoff assumes accurate patient recall of a symptom that many patients don’t notice until it becomes persistent.
The 2023 ACC/AHA/ACCP/HRS Guideline put this plainly: thromboembolic risk in the under-48-hour window is “not homogeneous but, rather, varied by a patient’s risk profile.” That’s the guideline explicitly walking away from treating 48 hours as a bright line.
A 70-year-old with hypertension and diabetes presenting with “new AF” has a CHA₂DS₂-VASc score of at least 3 before sex category is factored in. The 48-hour rule doesn’t make that patient safe. It just makes the conversation shorter.
What the ACUTE Trial Found
The trial enrolled patients randomized to TEE-guided cardioversion versus conventional 3-week anticoagulation before cardioversion. The TEE group found LAA thrombus in 13.8% of patients. Those patients didn’t get cardioverted. They got anticoagulated until the clot resolved.
That’s the point. Imaging finds the patients you’d otherwise be embolizing.
For context, AF presumed under 48 hours has a thrombus rate of around 1 to 2%. Documented therapeutic anticoagulation for three or more weeks drops it to under 1%. The risk scales with duration and with risk factors. It doesn’t disappear below 48 hours. It just gets smaller.
Atrial Stunning: The Risk That Starts After Cardioversion
Here’s the part that gets underweighted in the ED setting.
Even when cardioversion succeeds and no pre-existing thrombus is present, the left atrium doesn’t snap back to normal mechanical function. Atrial stunning is the impairment of LAA contractility after cardioversion. Partial recovery takes 15 to 30 days. Full recovery can take 30 to 90 days.
During that window, the left atrial appendage is functionally impaired even though the patient is back in sinus rhythm. That’s when de novo thrombus forms. That’s when the embolism happens after a cardioversion that seemed clean.
This is why current guidelines require four weeks of therapeutic anticoagulation after cardioversion in patients with meaningful thromboembolic risk. Regardless of AF duration. Regardless of whether TEE was performed. Cardioverting without arranging post-procedure anticoagulation doesn’t end the risk. It starts a new phase of it.
What the 2023 Guidelines Require
For AF duration 48 hours or longer, or uncertain: - Therapeutic anticoagulation for at least three weeks before cardioversion, OR - TEE to exclude LAA thrombus, then cardioversion with heparin bridging, then anticoagulation for at least four weeks after
For AF duration under 48 hours: - Anticoagulation initiated before or at the time of cardioversion, continued for at least four weeks - TEE should be considered in patients with elevated thromboembolic risk, because the under-48-hour window doesn’t confer uniform safety
After cardioversion in all patients: - Long-term anticoagulation is determined by CHA₂DS₂-VASc score, independent of whether sinus rhythm is maintained
The guideline language on that last point is deliberate. Stroke risk doesn’t go away when you cardiovert. Many of these patients need anticoagulation regardless. The rhythm outcome doesn’t change the CHA₂DS₂-VASc calculation.
The Clinical Bottom Line
A patient cardioverted in the ED without TEE and without anticoagulation had roughly a one-in-seven chance of carrying an LAA thrombus if AF duration was over 48 hours or uncertain. A 0.7 to 1.1% stroke risk even if AF was truly under 48 hours and risk factors are present. And is now entering 30 to 90 days of impaired LAA mechanical function with nothing on board to prevent the clot that atrial stunning enables.
The argument for skipping TEE in a stable patient is speed. The argument against it is the possibility of embolizing a clot that a 20-minute exam would have caught. Those are not equivalent tradeoffs.
The only indication for immediate cardioversion without preceding TEE or anticoagulation is hemodynamic instability. A stable patient is not an emergency. Take the time to do it right.