The 30 seconds was determined by consensus to be a reasonable threshold by which most would consider calling that an episode of Afib.  However, it is still unclear how little is still significant.  We have pacemaker patient data saying that both several minutes of Afib per month is enough for increased events and that strokes occur in sinus rhythm in patents with a history of Afib (see ASSERT and IMPACT trials).  The 2014 AHA/ACC/HRS Atrial Fibrillation Guidelines took out the 30 sec comment and left it up to the reader to conclude how long of an episode should count.

As noted by Dr. Chung, the 30s limit was an arbitrary consensus. With respect to stroke and the role of AF, it is still a point of active research whether the AF is a reflection of underlying endothelial dysfunction, or required to promote thrombosis. Even brief episodes (6 minutes of subclinical AF) have been associated with increased risk of stroke (PMIDs:22236222, 25460864).

I view as suspect all "consensus" estimates of pseudo-science.  The probability of a stroke is probably related to minutes of AF (thrombus formation), that becomes one of the variables in the equation.  The frequency of the episodes, in addition to the time of individual episodes, becomes a factor in cardiac physiology separate from stroke risk.

Indeed, the 30 seconds duration of AF was determined by consensus. It is of notice though that the 30second duration within trials is often based on technical failings of detection of algorithms rather on scientific grounds. These algorithms used by automated detection mechanisms demand a minimum number of R-R intervals to assess variability with enough confidence to reach a safe diagnosis of atrial fibrillation. Less duration of AF and stroke burden might be a good idea to be studied.

Gentlemen, would you consider a three second burst of afib to quality as an official afib episode? I ask having in mind to correlate your responses with professional practices in Greece.