Catheter ablation for atrial fibrillation is recommended for symptom relief, and the evidence behind that recommendation has always compared it with drugs or with nothing — never with a convincing placebo. PVI-SHAM-AF supplied one. At nine sites in Germany and Poland, 262 patients with symptomatic paroxysmal or persistent atrial fibrillation were randomised 2:1 to catheter ablation or to a sham procedure, with both patients and assessors blinded. Median age was 67 and the split between men and women was even. The endpoint was the change at six months in the Atrial Fibrillation Effect on Quality-of-life questionnaire, a validated patient-reported score.
Both groups got much better. The score rose from 61.3 to 81.1 after ablation and from 59.2 to 74.9 after sham. The difference between those improvements was 2.6 points (95% CI -2.7 to 8.0, p=0.36) — not significant, and smaller than the threshold most people consider clinically meaningful. One death occurred in each arm, neither procedure-related, and procedure-related serious adverse events occurred in six ablation and four sham patients.
The finding to take away is not that ablation does nothing. It is that a large part of what patients report after it is not attributable to the ablation. Undergoing a procedure, being watched closely, expecting to improve, and simply regressing from the bad spell that prompted referral all produce real improvement in a symptom score. Sham-controlled trials in interventional cardiology have repeatedly delivered this lesson and this is another instance of it.
For a clinician outside cardiology, the practical use is in the conversation before referral. A patient asking whether ablation will make them feel better deserves the honest answer that most people do feel better afterwards, and that a blinded trial could not show the procedure itself was responsible. Ablation still has arguments in its favour — rhythm control, reduced anticoagulation burden in some, outcomes in heart failure — but symptom relief alone is now a weaker one. Note the trial's own limitation: 1,199 patients were approached and only 262 agreed, so those who consented to possibly receiving a sham may not resemble everyone offered ablation.
- Tell patients that improvement after ablation is real but is not all caused by the ablation.
- Do not use expected symptom relief as the sole justification for referral.
- The trial measured quality of life at six months; it did not test stroke, mortality or heart failure outcomes.
- Only about one in five patients approached agreed to be randomised, which limits how far the result generalises.
- Apply the same scepticism to other unblinded procedural evidence where the endpoint is how the patient feels.
The statistics, in plain English
The confidence interval of -2.7 to 8.0 around a 2.6-point difference is the whole story. It includes zero, so no difference is compatible with the data; its upper end of 8 points means a modest real benefit has not been excluded either. What makes this trial persuasive is the size of the improvement in the sham arm — nearly 16 points — which shows how much a patient-reported score can move without any effective treatment. That is exactly why unblinded trials of procedures with symptom endpoints tend to look positive.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for top clinical updates, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free